Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2007-10-15

00:00:27 Quit ompaul (Client Quit)
00:00:31 Quit petur ("Zzzz")
00:04:30Mouser_XJust to make sure I'm doing this right: I've done "svn revert -R ." then I updated my SVN, then added 7331, and then 5241. Then I removed my "build" directory. I'm about to "make" a new Build. Is this correct?
00:04:38Mouser_XOr should I do "make clean" ?
00:05:27bluebrotherrun "make clean" first.
00:05:45Mouser_XWill that begin the build process?
00:07:18Mouser_XDoesn't appear so.
00:07:18 Join kugel|afk [0] (i=kugel@unaffiliated/kugel)
00:07:31n1sno, it will just remove old intermediate files
00:08:01pixelmaMouser_X: I had a problem with the mod patch and just doing an svn revert because of the new files which were added - those couldn't be reversed, svn doesn't know about them
00:08:29Mouser_XSVN compiled just fine with Rockbox + 5241
00:08:43 Quit BigBambi_ ("Please insert girder")
00:09:01Mouser_XIt also compiled just fine with Rockbox + 7331. However, when I did Rockbox + 7331 + 5241, I got problems.
00:09:19pixelmaok, just wanted to share my experience :)
00:09:55 Quit kugel (Nick collision from services.)
00:10:03 Nick kugel|afk is now known as kugel (i=kugel@unaffiliated/kugel)
00:10:06Mouser_X(I've been doing all kinds of builds for the last 2-3 days. So far, the only one that doesn't compile is when I do *both* 7331 and 5241.)
00:11:34Mouser_X(Just to clarify: I've been using 7331, 5241, and 2911. However, 2911 would just be "icing on the cake." While nice, I would be just as happy without it, as I would be with it.)
00:12:26 Quit Rob2222 (Read error: 104 (Connection reset by peer))
00:14:01 Quit ptw419 (Read error: 110 (Connection timed out))
00:14:18 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
00:15:16 Quit Robin0800 (" Try HydraIRC -> <-")
00:15:31 Quit linuxstb (Nick collision from services.)
00:15:33 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
00:18:52 Quit hannesd ("Client suicide")
00:19:39 Quit Llorean (Read error: 110 (Connection timed out))
00:21:23 Quit bertrik ("bye")
00:21:44 Join Llorean [0] (
00:22:04ashesis rockbox supposed to support utf8 characters in tag info, with ogg (or flac)
00:22:20ashesmy characters are being displayed badly
00:22:26ashesnot file names
00:22:56bluebrotherashes: are you using a font that has the characters in question?
00:23:04bluebrothervorbis comments are always utf-8.
00:23:14asheswould i get better performance if i compile rockbox myself specifically for my cpu model?
00:23:20ashesbluebrother: maybe not
00:23:33n1sashes: it _is_ compiled for your cpu
00:23:33bluebrotherRockbox _is_ build for the specific cpu.
00:23:37 Quit iamben (Read error: 104 (Connection reset by peer))
00:24:04bluebrotherand you might want to try unifont. It's the most complete font in terms of characters present afaik.
00:24:14 Quit Zagor ("Client exiting")
00:24:55n1swell, it's technically not compiled for the correct coldfire model but gcc doesn't support the one we have before 4.3
00:24:58LloreanIf the tags are being displayed poorly, but not the filenames, and they contain the same text, isn't that usually a codepage issue?
00:25:30ashesmy theme font is utf8, i changed it to something else which isn't utf8. it's okay now
00:26:19 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
00:26:21 Quit linuxstb (Nick collision from services.)
00:26:25 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
00:26:27ashesbuilding a snapshot version of rockbox with a snapshot version of gcc probably isn't a great idea
00:27:46bluebrotherLlorean: yes, usually. But AFAIK vorbis comments are always utf-8 per definition.
00:28:09bluebrotherit's an issue with id3 tags as those don't hold information about the used codepage
00:28:50 Join iamben [0] (
00:29:22 Join Rob2222 [0] (
00:29:29 Quit barrywardell_ ()
00:30:28pixelmaID3 does but only since version 2.4
00:30:58LloreanWhat was it about 2.4 that tends to cause people to prefer 2.3?
00:31:43linuxstbI think it's just less widely supported
00:32:30amiconnInteresting... HD66773R and HD66789R have a very similar command set, but there are some details where they behave different
00:32:55amiconnE.g. on HD66789R, setting the update window resets the address pointer. On HD66773R it does not.
00:33:27Mouser_Xbluebrother: Should I paste the entire log to pastebin, or just the "interesting" stuff?
00:33:39bluebrotherah, id3v2 also allows utf-16, even earlier than 2.4
00:33:52Mouser_X(Specifically, though I doubt it, there might be some stuff that I didn't think was important that I might leave out.)
00:34:15bluebrotherMouser_X: if you're sure you know what's the relevant parts you can trim it down. Otherwise just use the complete output
00:35:06 Join karashata [0] (
00:35:21 Quit Rob2222 (Read error: 104 (Connection reset by peer))
00:35:30 Join webguest23 [0] (i=4a469ecd@gateway/web/cgi-irc/
00:35:44Mouser_Xn1s: ^ If you can, that might be useful to you (if you're willing to help of course)
00:36:19Mouser_X(I patched in 7331 followed by 5241, otherwise there would have been more stuff there.)
00:37:05Mouser_XI'm building for the Gigabeat, if that's of importance.
00:37:27webguest23i installed rockbox on my sansa e200r with the e200rsansapatcher but i tried to update it. After it was done updating, i tried to install it back on. When the writing on the screen comes up, it say "unknown bootloader" then i have to shut it down. Anyone know what this means?
00:37:51webguest23when i updates the sansa, rockbox didn't boot
00:37:56pixelmaMouser_X: same error that I got, take a look at the metadata/mod.c file - mine was 3 times as long as it should be, because after the svn revert without the file being removed every try to patch added the file "inside" the one that was still there again
00:38:05 Join BHSPitLappy [0] (n=steve-o@unaffiliated/bhspitmonkey)
00:38:08n1swebguest23: that your bootloader isn't supported
00:38:20webguest23what does that mean
00:38:39n1sthat e200rpatcher doesn't know how to patch it
00:38:58webguest23do i have to find another bootloader
00:39:06Mouser_Xpixelma: Hmmm. Possible. Though I know it *does* compile just fine. I guess I could delete my SVN and checkout again.
00:39:18n1sI would suggest using an earlier version of the firmware
00:40:02LloreanMouser_X: What she said means "the patch may compile just fine, but what you've done combined with how the patch is made could cause this error by having applied, removed, then re-applied the patch"
00:40:08n1sMouser_X: no need to check out again just revert and delete any files added by the patches
00:40:15bluebrotherMouser_X: you could just rm mod.c
00:40:27 Join Rob2222 [0] (
00:40:38amiconnah, no, I just overlooked sth
00:41:27Mouser_XWhere would I find "mod.c" to remove it?
00:41:54*bluebrother is too slow
00:42:05amiconnThere are 2 different mod.c
00:42:39n1samiconn: did you ever find out what license that modplayer is under?
00:42:43pixelmayes - probably the other one in apps/codecs is affected too
00:43:38amiconnn1s: No. *someone* should drop 'myth' a mail...
00:44:30n1swell apparently that tumu person is working on porting dumb so I guess we can see how that turns out...
00:44:33webguest23do you know where i can download earlier firmware for my e200r
00:45:51jhMikeSamiconn: which is more imporant on SH? icode or idata?
00:46:23amiconnDepends on what the code does. What do you want to do?
00:46:54jhMikeSkeep iram use to only what's nescessary. I don't think switch_thread need to be icode for anything.
00:47:13amiconnI'd think so too
00:47:51amiconnThere isn't much iram, and switch_thread() does neither contain excessive loops, nor is it called extremely often
00:48:04 Quit Mouser_X (Read error: 104 (Connection reset by peer))
00:48:43jhMikeSthe thread data gets far more use than the code in general
00:48:44amiconnOn SH1, dram is as fast as iram for data _within a fast page access sequence_, i.e. no wait state
00:48:58amiconnOpening the page costs 2 cycles
00:49:28amiconnPutting code in iram helps to avoid contention of instruction fetches with external memory accesses, _if coded carefully_
00:49:39 Join Mouser_X [0] (
00:50:00Mouser_XSorry, my laptop locked up.
00:50:14jhMikeSso having data in iram and code in ram should help avoid contention?
00:50:30amiconnIt's not that simple
00:50:57amiconnSH instructions are 16 bit. DRAM is 16 bit too, IRAM is 32 bit
00:51:13amiconnWhen code is running from IRAM, the SH fetches them in pairs
00:52:12amiconnThat means if you put memory access instruction in the correct slot, the memory access doesn't contend with instruction fetch, no matter whether the data is in iram or dram
00:53:07amiconnIf either code or data is in iram and the other in dram, it is slower than having both in iram, but not much, as the cpu can make efficient use of fast page mode
00:53:39amiconnBoth code and data in dram still hass that advantage, as long as the data is located close enough (within the same page)
00:53:39 Quit webguest23 ("CGI:IRC (EOF)")
00:54:13amiconnIf they're further apart, that's the worst situation, as each and every fetch (code and data) switches pages
00:54:55 Quit obo ("bye")
00:55:02jhMikeSI should get the datasheet out here but what's the page size then?
00:55:22 Quit Nico_P (Remote closed the connection)
00:56:11jhMikeSso keeping blocking functions that call switch_thread near switch_thread itself could help
00:56:29ashes in case any of you want
00:56:35amiconn1KB iirc
00:56:50amiconnThat's a ram feature, not cpu
00:56:56amiconn(fast page mode)
00:58:10 Quit ender` (" I think computer viruses should count as life. I think it says something about human nature that the only form of life we ha)
00:58:37bluebrotherashes: why not install the cross compiler to /usr/local? That will get rid of possible conflicts with the host cc
00:59:50asheslfs doesn't install anything to /usr/local, and i've fixed any conflicts. installing libiberty.a to /usr/local/lib isn't a good idea either
01:00:11ashesit's up to you
01:00:47bluebrothersure. But I also don't use LFS ;-)
01:00:59bluebrotheranyway, time for sleep for now.
01:01:10ashesthat hint should work with any linux
01:01:44ashesor bsd
01:01:46bluebrotherof course.
01:01:57 Quit bluebrother ("sleep!")
01:03:09Lloreanashes: Is there some great flaw in the instructions and methods on our site for setting up a rockbox dev environment?
01:03:13jhMikeSgcc is playing size tricks on /me again...this time with SH and there's hardly any additional code for that
01:03:41 Join webguest68 [0] (i=4a469ecd@gateway/web/cgi-irc/
01:03:45ashesLlorean: i found it overly complicated
01:03:59Llorean"Run and add the appropriate line to your path"?
01:04:14ashesLlorean: maybe we're not talking about the same wiki/doc
01:04:42webguest68if i deleted my bootloader on my e200r, how do i get it back?
01:04:47LloreanVery possibly.
01:05:08Lloreanashes: ?
01:05:52ashesit has much more information than i need
01:06:13LloreanDid you see the "automatic build" section of it?
01:06:42 Quit mirak (Remote closed the connection)
01:06:57ashesexcept for that
01:07:00superkaybeei was testing out some themes on my e260 and the latest one i've selected doesn'
01:07:10superkaybeedoesn't show text in the menus
01:07:19superkaybeehow do i reset the theme when i can't see what i'm doing?
01:07:21linuxstbwebguest68: Which bootloader? What exactly did you do?
01:07:46Lloreanashes: Not only that, but following our processes insures you're using both the right versions of binutils and gcc, and have installed any necessary patches so that compiling doesn't fail
01:08:24LloreanWe don't really accept bug-reports on versions of Rockbox not compiled with the expected environment, because of how picky some of these processors can be (and how strangely buggy gcc has shown itself to be cross-compiling in some cases)
01:08:35webguest68well, i deleted the file OF.mi4
01:08:42webguest68that was in the system menu
01:09:08Lloreansuperkaybee: Connect to your computer and delete config.cfg in the .rockbox folder
01:09:20superkaybeeLlorean: thanks!
01:09:37linuxstbThat's the Sansa original firmware. You'll need to enter recovery mode and reinstall the Sansa firmware to get USB mode again, then install Rockbox again.
01:10:07 Join JdGordon [0] (
01:10:33webguest68my computer doesn't detect it in recovery mode
01:10:51***Saving seen data "./dancer.seen"
01:17:08 Quit n1s ()
01:20:19jhMikeSamiconn: according to the instruction reference, loading memory byte=>register or storing reg=>memory byte should be efficient on SH, right?
01:20:43amiconnIf you want sign extension, then yes
01:21:32JdGordonlinuxstb: hey, still around? doesn't really matter if it is extended or not. I only care about bits 0-7 anyway.
01:21:50amiconnOn H300 and X5 it might in fact pay off to write the data in lcd_write_yuv420_lines() into an IRAM buffer instead of the LCD controller, and then set up a DMA transfer and calculate the next line pair in the background...
01:22:19jhMikeSoh, so you will try that trick :)
01:22:48amiconnDrawback is that it needs 2 additional buffers in iram, each sized 2*LCD_WIDTH
01:22:49linuxstbJdGordon: Just about
01:23:17JdGordondo i use bootloader/bootloader.bin with sansapatcher to update the ppbl? or something else?
01:23:20amiconnThat is 880 iram bytes on H300, and still 640 bytes on X5
01:23:23jhMikeSthat's my misgiving about it too
01:23:52jhMikeSunless bursts are used then it's not much slower
01:23:54linuxstbJdGordon: According to barrywardell's commit message, it needs a raw binary, so yes, bootloader/bootloader.bin should be it.
01:24:44 Quit Toxicity999 (Read error: 104 (Connection reset by peer))
01:24:46amiconnThe intermediate Y, Cb and Cr buffers are already on stack, that's 660 bytes (H300) resp. 480 bytes (X5)
01:25:21amiconnPutting the extra buffers on stack too might overflow the stack...
01:26:36amiconnErm, and the 2 additional buffers on X5 would be 1280 bytes, not 640 (32 bits per pixel in order to use dma)
01:28:34 Quit Thundercloud (Remote closed the connection)
01:28:45*jhMikeS curses freescale for no data cache on cf
01:28:48*amiconn wonders whether an intermediate iram buffer + dma would even speed up lcd_update() on X5
01:29:58 Quit webguest68 ("CGI:IRC (EOF)")
01:30:05 Quit animeloe ("This computer has gone to sleep")
01:30:06JdGordonwell this sucks...
01:30:07amiconnIt's just having the LCD controller waitstates handled asynchronously (by the dma controller) instead of synchronously
01:30:20JdGordonmy sansa wont connect to usb, so i cant tst this out :(
01:30:23 Quit BHSPitLappy (Read error: 110 (Connection timed out))
01:31:31 Nick fxb__ is now known as fxb (
01:32:07 Quit jhulst ("Konversation terminated!")
01:34:15superkaybeethanks for all the help getting me started today −− things are working great
01:36:15 Quit lee-qid ("aufwiederbyebientotsayonara")
01:38:48 Quit spiorf (Remote closed the connection)
01:39:16 Quit moos ("Time to sleep a bit !")
01:39:50 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
01:42:41amiconnjhMikeS: Oh, plus letting the dma controller do the longword split (on X5)
01:44:03jhMikeSlongword split?
01:44:32jhMikeSsplits into 2x16-bits?
01:46:21 Quit bb (Nick collision from services.)
01:46:28 Join bb_ [0] (n=bb@unaffiliated/bb)
01:48:07zichohow do i access sleep timer?
01:48:39scorchethe manual should say if it is supported by your device
01:49:11zichoi know it is, but how do i access it?
01:49:32pixelmawhat does the manual say?
01:50:53 Join Logomachist [0] (
01:51:13pixelmascorche: the sleep timer is there on non rtc targets too
01:51:20zichofound it.
01:51:52LogomachistQ: What's the best video player that's Rockbox compatible?
01:52:11LloreanLogomachist: Do you mean "What's the best Rockbox target for playing video?"
01:54:45 Join bb [0] (n=bb@unaffiliated/bb)
01:54:46 Part pixelma
01:54:54LogomachistSort of... let me rephrase...
01:55:42 Join Asteriskk_ [0] (
01:55:49LogomachistWhat video player best supports Rockbox? If there are multiple video players that are all equally compatible with Rockbox, is there one that clearly gives you the best value, taking into account things like battery life and price?
01:56:08LloreanRockbox runs on music players
01:56:34Asteriskk_can somebody send me the lame3.97 source code? I can't seem to get sourceforge to respond and I need it for cygwin
01:56:37scorcheLogomachist: rockbox doesnt run on any "video players" runs on digital audio players, which can play video with rockbox
01:56:39LloreanMost of the devices weren't designed for video, and have it as a more or less "added feature" to the music playback
01:56:49LloreanLogomachist: As for which player runs video best in Rockbox, it's the Gigabeat.
01:57:22LogomachistI'm confused. Rockbox plays video on devices that aren't made to play video, but on devices that do play video, it doesn't work?
01:57:24jhMikeSVideo is hardly an addon to a 300 or 500MHz chip (when the latter is ready)
01:57:48LloreanLogomachist: There are no devices that Rockbox runs on that were designed as "video players"
01:58:01LloreanWith the possible exception of the iPod Video, which is an odd case.
01:58:08 Join webguest72 [0] (i=4a469ecd@gateway/web/cgi-irc/
01:58:24Asteriskk_if you want video whynot get one of those little portable dvd players?
01:58:30LloreanBut if you want to watch video on Rockbox, hands down your best choice is the Gigabeat.
01:58:49webguest72when i am in recovery mode with the usb plugged in, it son't connect to the computer. Why is it doing this.
01:59:12LloreanAt anywhere from US $90 to $130 depending on luck and persistence for a 40gb device, and a 320x240 screen that's quite nice, and a fast enough processor to run fullscreen at 30fps, it's a no brainer.
01:59:41webguest72rockbox is booting but the original firmware is not. I'm using a sansa e200r
01:59:48LogomachistI want the ability to watch video, but I don't want to buy a seperate video player, and I don't want dvd player because it would not be convinient at the gym..
02:00:25Mouser_XLogomachist: Rockbox can play MPEG-1 and MPEG-2 vidoes.
02:00:48LloreanLogomachist: Well, there's really only one choice with Rockbox. Well, two if you count the Gigabeat F and Gigabeat X as separate players, but they're more or less separate looks for the same player.
02:00:58Mouser_XIn other words, buy a Gigabeat, and convert your videos to MPEG-2, and play them in Rockbox.
02:01:21advcomp2019webguest72, did you add OF.mi4 when doing the step
02:01:24Asteriskk_to add to that, you can use virtualdub to do that
02:01:37Asteriskk_incase you didnt know
02:02:03markunI watched video on a Archos 605 today, looked pretty good
02:02:05LogomachistMaybe I will.
02:02:18LogomachistIs there anything like Rockbox designed for video players?
02:02:18scorchemarkun: back home yet?
02:02:26Mouser_XCheck out MpegPlayerPlugin in the Wiki (I think that's the name.)
02:02:29markunscorche: yes, arrived today
02:02:38Mouser_XIt will tell you how to convert videos for use in Rockbox.
02:02:41LloreanMouser_X: PluginMpegplayer
02:02:42scorcheMouser_X: PluginMpegplayer
02:02:49LloreanMouser_X: All plugin pages start with Plugin, or should
02:02:51webguest72if i could get into recovery mode thn i could fix it but i don't know how to get it to connect
02:03:24markunscorche: I should go to sleep now to get back into a normal rithm..
02:03:41krazykitLogomachist, no, not really.
02:03:43scorchemarkun: good luck with that...let me know when you get the pictures up ;)
02:03:54scorche...not like many were taken, but meh
02:04:12markunscorche: I don't have the pictures with the 3 of us
02:04:20markunwas karl's camera used for that?
02:04:23Mouser_XLogomachist: If you're wondering if there's any devices that *natively* support video playback, and run Rockbox, then your only option is the iPod video.
02:04:24scorcheoh...was that karl?
02:05:13markunI hope so
02:05:34Mouser_XLogomachist: However, I would strongly suggest that you look at the Gigabeat first. It's a great player, and video works great on it as well.
02:05:45LogomachistYeah... If Gigabeat plays video better than the iPod running Rockbox, there's no reason for me to prefer the iPod.
02:06:10webguest72`it says that the usb cable is onnected on the screen but i still get no results on the computer
02:06:27*amiconn wonders why we don't use buffered writes by default
02:06:39LogomachistBut is the Gigabeat really going to compare, in its video playback, with a device designed from the ground up to play video? For example, the Zune?
02:07:02markunLogomachist: depends how you want to compare them
02:07:05 Quit bb_ (Read error: 110 (Connection timed out))
02:07:06LloreanLogomachist: Rockbox doesn't run on the Zune anyway
02:07:12Mouser_XThe Zune uses hardware that is very similar to the Gigabeat S (though, the Gigabeat S doesn't yet run Rockkbox).
02:07:18advcomp2019webguest72, you do not see 16m-format in my computer
02:07:40krazykitLogomachist, there are other things to consider, of course. The Zune is crippled to Windows, Rockbox players are not, among other things.
02:08:11LogomachistI just used the Zune as an example, I wasn't seriously considering getting it.
02:08:43LloreanLogomachist: The only real advantages other players might have, is that if some of them have fast enough processors, filesizes at a given quality may be a little bit smaller.
02:08:44Mouser_XLogomachist: The Gigabeat F (which runs Rockbox, and Rockbox can play videos) has a 300 mhz CPU. The Gigabeat S (which has nearly identical hardware to the Zune, if I'm told correctly) has a 533 mhz CPU.
02:09:17webguest72do i have to install a certain driver
02:09:33Mouser_XThere's a port for the Gigabeat S in the making, but it hasn't gotten very far yet.
02:09:34markunMouser_X: I also believe the i.mx31 has some logic used for video playback (maybe decoding or just scaling, I don't know)
02:09:42Asteriskk_any chance of getting rockbox to play content?
02:09:54LloreanAsteriskk_: No. It's DRM restricted.
02:11:07 Quit bb (Nick collision from services.)
02:11:14 Join bb_ [0] (n=bb@unaffiliated/bb)
02:11:35jhMikeSamiconn: what's not using buffered writes?
02:11:46amiconnColdfire (except for dram)
02:12:23amiconnI enabled buffered writes as default in crt0.S and system-target.h, testing on H300 - and guess what?
02:12:35Asteriskk_it's drm restricted but they have players that will work with their content listed on their site
02:12:37amiconn38% speedup in lcd_yuv_blit() when boosted!
02:12:45jhMikeSnot bad :)
02:13:11amiconnI'll test this for a few days, whether something unusual happens
02:14:07amiconnI might have to recalibrate the remote lcd timing, as it may become too fast with buffered writes
02:14:14amiconnPerhaps some other timings too
02:14:22jhMikeSwhat's the mod?
02:14:41Mouser_XLogomachist: I am not the one to ask (since I've never used the Zune), but I would think that the Zune would have the advantage in 2 areas, over the Gigabeat. The Zune has a larger screen (I think. I don't know this), and a faster CPU. Because it has a faster CPU, the Zune might support seeking while watching a video (something that Rockbox can't do. There's not enough extra CPU available to watch the video, and fast forward/rewind).
02:14:59amiconnfirmware/target/coldfire/crt0.S, line 222: change 0x80000000 to 0x80000100
02:15:08advcomp2019webguest72, there is no drivers needed for recovery mode
02:15:22 Quit Toxicity999 (Read error: 113 (No route to host))
02:15:24amiconnSame in firmware/target/coldfire/system-target.h, line 158
02:15:45amiconn(the value loaded into %cacr in the next instruction)
02:16:01advcomp2019webguest72, what operating system are you running
02:16:11webguest72don't laugh
02:16:21*Asteriskk_ laffs
02:16:24Asteriskk_sorry can't help it
02:16:39amiconnYea, X5 needs a few more LCD waitstates when boosted
02:16:48webguest72i was kind of expecting it
02:17:16amiconnBut that has to wait... bed time
02:17:28Mouser_XBlast! Thus far, it appears that pixlma was right. My MOD.c file was corrupted (I'm compiling Rockbox + 7331 + 5241, and it's gotten past the point that normally causes problems)
02:17:39advcomp2019webguest72, i do not know if ME has the drivers to see recovery mode
02:18:09webguest72Oh, but is there any way i can find those drivers
02:18:30 Join bb [0] (n=bb@unaffiliated/bb)
02:18:44webguest72any other way...
02:19:05Asteriskk_what player is it again?
02:19:10amiconnLooks like the coldfire video capabilities will improve soon :)
02:19:22keanuAsteriskk_, e200 IIRC
02:19:33Asteriskk_have you looked on the sansa site itself?
02:19:37Mouser_XI think it's the "r:
02:19:41Asteriskk_they have legacy drivers tehre
02:19:54webguest72i'll check it out now
02:21:34*keanu confused...
02:21:47keanudoes calendar on the e200 data abort for anyone else?
02:22:03 Nick fxb is now known as fxb__ (
02:22:38jhMikeSamiconn: It even helps the SPC codec a bit even though it's mostly IRAM there
02:22:38 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
02:23:31Mouser_XjhMikeS: Nice!
02:23:36*Mouser_X likes SPCs.
02:23:55 Join bb__ [0] (
02:24:14 Quit webguest72 ("CGI:IRC (EOF)")
02:24:18jhMikeSnot too much since cf is already fast as hell with all the emac DSP stuff in there (1%-2% boost reduction)
02:24:44amiconnjhMikeS: Well, that's probably due to the offload in other parts of rockbox
02:25:31amiconnBuffered writes don't apply to iram, as it's internal. For sdram, buffered writes were already enable (via %acr0)
02:26:08amiconnBut all the register writes (MBAR +MBAR2 area), as well as lcd, were not
02:26:45amiconnBut now, really sleep
02:27:41Asteriskk_cya amiconn
02:29:14 Quit Rob2222 (Read error: 110 (Connection timed out))
02:29:48 Quit bb_ (Read error: 110 (Connection timed out))
02:30:33 Join NIHIL [0] (n=NIHIL@
02:30:41NIHILI need some help
02:31:03NIHILI think I bricked my Gigabeat f40 when I reinstalled rockbox
02:31:20NIHILwhen I reinstalled it I got codec failure
02:31:37krazykitplug in usb, then turn it on. you can install a different build like that
02:31:43NIHILSo I deleted the extra rockbox.gigabeat file in my root
02:31:45 Quit jhMikeS (Nick collision from services.)
02:31:48 Join Rob2222 [0] (
02:31:51 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
02:31:52NIHILnow I get rockbox error: -2
02:32:28NIHILI tried using the usb, but it doesnt contact the pc
02:32:53krazykitin bootloader usb mode?
02:33:07 Quit superkaybee ()
02:34:00krazykitif you can't access the gigabeat via Rockbox or the bootloader USB mode, you'll have to go through the recovery procedure:
02:34:26NIHILwait I got it via bootloader
02:34:52NIHILgreat might be system errors
02:34:52 Quit ramon8 (Read error: 104 (Connection reset by peer))
02:34:55NIHILthanks anyway
02:34:58krazykitthe fact that you had rockbox.gigabeat in the root of your drive suggests that your bootloader is out of date too, nihil
02:34:59LogomachistGuys, I'm looking up portable video players... do you know anything about the Archos 705?
02:35:13NIHILI'll reinstall then
02:35:13krazykitLogomachist, please stay on topic
02:35:23scorcheLogomachist: this is #rockbox...a channel for discussion...rockbox
02:36:25 Quit bb (Connection timed out)
02:36:28NIHILwhat is rockbox error: -2?
02:39:05krazykitmeans there's a problem with the checksum. try updating the bootloader with the one in the wiki (from april) and install the latest build of rockbox while you're at it
02:40:11 Quit Logomachist ("ChatZilla 0.9.61 [Mozilla rv:1.7.12/20050915]")
02:45:56 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
02:56:28 Quit kfazz ("Leaving")
03:06:07 Quit Asteriskk_ ("I ran out of quit messages.")
03:10:53***Saving seen data "./dancer.seen"
03:11:00 Join Davide-NYC [0] (
03:16:35 Join psycho_maniac [0] (
03:25:52 Join Aevum [0] (
03:27:23Aevumjust wanted to know if since the toshiba players and the kenwood players share plataform, could i use to toshiba rockbox build on a kenwood HD20GA7
03:27:28Aevumthank you in advnace
03:27:49TMMAevum: I'd recommend against just trying it
03:28:07TMMAevum: I'm no export on those targets though, perhaps someone else can give a better answer
03:28:40Aevumi do understand that using a firmware that isnt costum for that player can lean to bricking, thats why im asking first
03:29:22TMMAevum: I'm just saying :)
03:29:49Aevumi know, and sorry if my response has been hostile, you're just reminding me not to do stuff that might brick my player
03:30:32TMMAevum: I have experience in bricking players :)
03:32:21 Quit Toxicity999 (Remote closed the connection)
03:33:51LloreanAevum: Unless all the hardware is exactly identical, it will need at least some modification to work.
03:34:00 Quit NIHIL (""Help! I've been g:lined from my mIRC!!" Bersirc 2.2: less n00bs [ - Open Source IRC ]")
03:34:01 Quit miepchen^schlaf ("Verlassend")
03:34:23Aevumthe hardware isnt identical, thats why the toshiba sounds ok and the kenwood sounds like it does
03:34:42Aevumif the hardware was identical and it was a rebadge, i wouldnt ask
03:35:42Aevumim also thinking about replacing the hardrive, but thats not rockbox related
03:36:00Aevumneed to know if its 5mm or 9mm toshiba first
03:36:02 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
03:36:07 Join webguest62 [0] (i=456afa47@gateway/web/cgi-irc/
03:43:33 Join ToHellWithGA [0] (
03:47:30Soapis there objective tests somewhere showing the Kenwood is a more accurate audio reproducer?
03:48:13TMMperhaps the mp3 codec in the OF just sucks? :)
03:48:16*scorche pushes Soap off to somewhere else like #anythingbutipod, where those people are...
03:48:19SoapYou know what? I take that question back. That question and the response are not Rockbox related.
03:54:27webguest62I'm writing a paper for my college Operating Systems class on Rockbox. I've found a lot of great resources on, but there's a nice architecture summary on Wikipedia I want to use that I can't find on It's here: Would you say it is an acurate summary?
04:03:46 Quit kugel ("ChatZilla [Firefox]")
04:08:49 Join Brian10161 [0] (
04:09:54Brian10161grettings comrads
04:10:05 Quit TNTales ("I-n-v-i-s-i-o-n 2.0 Build 3515")
04:13:46 Quit Brian10161 (Client Quit)
04:15:12 Quit ashes (Remote closed the connection)
04:15:17 Join ashes [0] (n=ashes@2001:5c0:8fff:ffff:0:0:0:4d)
04:17:20 Join Rob222241 [0] (
04:18:26 Quit Davide-NYC ("ChatZilla [Firefox]")
04:19:01 Quit Rob2222 (Read error: 110 (Connection timed out))
04:23:29 Quit Soap ()
04:44:05 Join JdGordon|uni [0] (i=82c20d6a@gateway/web/cgi-irc/
04:44:20JdGordon|unianyone with a sansa or ipod around?
04:44:46JdGordon|unikeanu: fell like testing a patch quickly?
04:45:00scorcheJdGordon|uni: nano work?
04:45:04jhMikeSJdGordon: what sort of patch?
04:45:14JdGordon|uninano should work
04:45:21JdGordon|uni7952 and 7956
04:45:26JdGordon|uniui fixes
04:45:39JdGordon|unii forgot my microsd adaptor so i cant test till i get home :(
04:46:14jhMikeSmicroSD adaptor on a sansa?
04:46:35JdGordon|"> if you can be bothered checking
04:46:49JdGordon|unijhMikeS: i need the adaptor coz these comps dont have a microsd slot
04:47:58jhMikeSfirst I have to see if the bug happens for me
04:48:49psycho_maniacJdGordon: you mean a microsd to usb adapter?
04:49:11scorchejhMikeS: you on it, then?
04:49:23JdGordon|uniusually trhrough a msd -> sd -> usb adaptor though
04:50:50JdGordon|unihmm.. actually 7956 is confusing me
04:50:56 Quit jhMikeS (Nick collision from services.)
04:51:02 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
04:52:10jhMikeSugh, had to reboot the router
04:52:25JdGordon|unidont worry about it, ill be home in 90 min, ill fiddle then
04:52:40jhMikeSoh, I was gonna be on it but ok
04:52:40 Quit JdGordon|uni ("CGI:IRC (EOF)")
04:54:44 Join hidden212 [0] (i=d15ed6ca@gateway/web/cgi-irc/
04:55:23hidden212is there a new version to the sansapatcher
04:56:37hidden212good day to all
04:58:13Aevumi really dislike the fact that sundisk abandoned mini USB in favor of a propietary connector
04:59:56Aevumthat should be a nice project
05:00:10Aevumthe mini usb port soldering guide for different players
05:00:49 Quit Mouser_X (Read error: 104 (Connection reset by peer))
05:01:43 Join Mouser_X [0] (
05:01:53 Quit hidden212 ("CGI:IRC (Ping timeout)")
05:10:54***Saving seen data "./dancer.seen"
05:13:20 Join Bam2550 [0] (
05:14:16Bam2550When i install Cygwin it allways stops at 98 percent... (stays like that for 30 minutes) Also my computer is going very slow after installing Cygwin, is this normal?
05:15:16Mouser_XProbably not.
05:15:33 Quit Aevum ("Saliendo")
05:15:38Mouser_XI'm running cygwin (right now even) on my 200 mhz laptop, and overall usage is about normal.
05:15:49 Join bb [0] (n=bb@unaffiliated/bb)
05:16:27Mouser_XBam2550: When installing, maybe try a different server?
05:17:06alienbiker99Mouser_X why don't you just throw linux on that haha
05:17:18Mouser_Xalienbiker99: I probably should.
05:17:27Mouser_XI'm low on disk space though.
05:17:42Mouser_X(Meaning, I'm not sure if I have enough to re-partition it.)
05:17:48alienbiker99ah ok.
05:20:05Bam2550it gets stuck at /ect/postinstall/
05:22:17Mouser_Xalienbiker99: Here's a thought. The answer is probably yes, but I'm not sure (I've never done a dual-boot system before). I'm running Win98 (therefore, it's a FAT32 drive). Can I install Linux on the same drive as Windows? If yes, how much space, roughly, would you say it would take (with a minimalistic install)?
05:23:19 Join stevenm [0] (n=stevenm@
05:23:39alienbiker99hm, i'm pretty sure you can repartition it for linux, but i don't know how much space you would need, although i think it depends on the distro.
05:26:06 Quit Bam2550 ()
05:28:00 Quit bb__ (Read error: 110 (Connection timed out))
05:37:32stevenmAnyone seen n1s lately ?
05:40:23 Quit Mouser_X (Read error: 104 (Connection reset by peer))
05:42:25 Join Mouser_X [0] (
05:46:21 Quit BHSPitMonkey (Read error: 104 (Connection reset by peer))
05:46:54Mouser_XYay! Rockbox + 7331 + 5241 has successfully compiled. It's now chugging through the plugins.
05:47:21Mouser_X(I've been trying to get that to work all weekend. At least now I know what the problem was.)
05:51:19 Quit stevenm ("Connection reset by beer")
06:01:19psycho_maniacwhat patches are those? sound codecs?
06:03:22Mouser_XGBS and MOD
06:05:05 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
06:17:57 Quit webguest62 ("CGI:IRC")
06:25:53Mouser_XWhen I update Rockbox, how will that effect the database? Should I export it first?
06:26:12LloreanNot normally, unless the database format changes.
06:26:15Mouser_X(It's not important, as I don't use it much, but it'd be nice to be able to keep the info it has in it right now.)
06:31:02 Quit atsea-34 (Remote closed the connection)
06:34:02 Quit Mouser_X (Nick collision from services.)
06:34:19 Join Mouser_X [0] (
06:38:21 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
06:38:59 Join reverendnathan [0] (
06:39:16reverendnathanHay guyz, need some halp here
06:39:43scorcheplease read the guidelines linked in the topic first ;)
06:39:51reverendnathanSo I moved .rockbox and all it's contents to the "/" directory of my iPod Video 30GB with the correct firmware
06:40:06reverendnathanAnd then ./ipodpatcher said it did everything OK
06:40:16reverendnathanSo I Menu+Select to restart, and...
06:40:27reverendnathan...nothing? Just goes back to the iPod firmware
06:41:07scorcheare you holding the menu button or have the hold switch on while booting?
06:41:21reverendnathanNope, it reboots
06:41:34reverendnathanApple logo shows up and comes back to the original firmware
06:41:48scorchewhere did you get ipodpatcher from?
06:42:28spkydid you install the bootloader?
06:43:15reverendnathanscorche: I got it from
06:43:19 Join nave7693 [0] (
06:43:42nave7693is the sansa's cpu theortically good enough to decode divx and friends?
06:44:09Mouser_Xnave7693: Probably not, but I'm not the one to ask.
06:44:28Mouser_X(I don't have a Sansa, nor do I know that much about them.)
06:44:35spkyi'm trying to patch with, it only has 1 failed hunk in firmware/target/arm/ipod/button-clickwheel.c, can someone please tell me where i'm suppose to paste the second hunk?
06:44:52scorchereverendnathan: did you properly eject/umount the device before unplugging it?
06:45:17reverendnathanscorche: No, I don't know how to umount an iPod without umounting everything
06:45:20reverendnathanWhich is annoyin'
06:45:32reverendnathanSo I just pull the plug out
06:45:35reverendnathanNo dice?
06:45:49scorchespky: there are clues in the patch file and rejects file
06:45:57scorchereverendnathan: yeah...that isnt a good thing
06:47:02scorchereverendnathan: do the whole process over again, but dont pull the plug out till you can correctly umount it (umount /dev/<devicenamehere>)
06:47:06spkyi'm lookin at the patch and the source, i'm not sure where it's suppose to be adding the hunk, the data above and below isnt in the source
06:48:00scorchespky: well, this is just part of syncing patches to current have to figure out where things go and fix them
06:48:32reverendnathanscorche: Should I redo dragon dropping the .rockbox folder to /ipod/? Or just redo ./ipodpatcher
06:48:34spkyi'm sorry i pasted the wrong link
06:48:50scorchedragon dropping?
06:48:59scorcheand you likely have to redo everything
06:49:17reverendnathanDrag 'n' drop
06:49:21reverendnathanDragon drop
06:49:28reverendnathank redoin'
06:50:02spkypieso patch not custom splash
06:50:38scorchewell, whichever patch it is, my answer is the same :)
06:51:18reverendnathanumount: /mnt/ipod: device is busy −− Gah, what do I do here?
06:51:37spkymake sure everything accessing the device is closed
06:52:54reverendnathanwhoops, had a konqueror window open
06:53:23reverendnathaniPod is still saying "Do not Disconnect", terminal says it disconnected...
06:53:43reverendnathancomputer doesn't recognize it
06:53:49reverendnathanShould I pull the plug?
06:54:09scorchedid you successfully umount it?
06:54:43reverendnathanI'm pretty sure
06:54:59scorchethen go for it
06:55:00reverendnathanNot showing up in /dev/
06:55:14psycho_maniacfor some reason my ipod wont unmount when ever i transfer a video file to it.
06:55:15reverendnathanA ha!
06:55:25reverendnathanDidn't even have to reboot
06:56:13reverendnathanNow... I have to redo my music due to the way amaroK organized them?
06:56:23Mouser_XMost likely.
06:56:32scorcheyou can do whatever you want..
06:57:22 Quit DataGhost ("NTOSKRNL.EXE caused a buffer overflow in System Idle Proce")
06:59:26 Join stevenm [0] (
06:59:36stevenmHey, MIDI now sort of has pitch bend depth.. I think
06:59:54stevenmStill need to listen to all the test files and all..
07:00:26reverendnathanCool, I've wanted rockbox for a long time, but I've been a lil' b' about getting it
07:00:39reverendnathanI loved my iAudio UI, and I always found the iPod's really lacking
07:01:29stevenmBut there's this inexplicable clicking every 0.25 sec. I wonder if it's just my sim being dumb. target tiem
07:02:29 Part nave7693
07:10:54 Quit JdGordon ("Konversation terminated!")
07:10:55***Saving seen data "./dancer.seen"
07:12:55 Join JdGordon [0] (
07:15:22 Join atsea-34 [0] (i=atsea-@gateway/tor/x-1edb7fe9060a6a1f)
07:17:18 Join DataGhost [0] (
07:19:24JdGordonlinuxstb: (in case you care)... i didnt see barry mention the load speed diference with the ppbl trick.. its less than 1 second (timed it with a stop watch from power to the rockbox logo), so really i dont think it should be available in sansapatcher incase people stuff it up
07:21:16 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
07:25:05 Quit stevenm ("Connection reset by beer")
07:27:02 Quit jhulst (Remote closed the connection)
07:27:12JdGordonLlorean: scorche: you guys around?
07:27:30JdGordondid you end up testing those patches i mentioned eariler?
07:27:47scorcheJdGordon: ah...i thought jhMikeS was on it, so i didnt go through with it
07:27:51LloreanI don't remember any patches, so probably not. Been a bit hectic.
07:28:03 Quit karashata ("Leaving.")
07:28:17JdGordonit flips the volume setting so going clockwise increases the volume like in the wps, but im just trying it out now and not sure if its more intuitive than svn...
07:29:05psycho_maniacJdGordon: do you mean it would do this in the rockbox menu?
07:29:20JdGordonno.. the volume setting
07:29:57psycho_maniacunder sound settings>volume?
07:30:19LloreanI assume it flips the list, not the direction the highlighter bar moves, right?
07:30:32JdGordonso mute is at the top
07:30:42psycho_maniacok that makes sense because they already flipped this before i believe.
07:30:57psycho_maniacthe scrolling not the menu
07:31:15LloreanI wouldn't mind it.
07:31:30psycho_maniacwhat is the patch number?
07:32:02JdGordonits 956 but that depends on another patch
07:32:15JdGordonwrong url">
07:33:09JdGordonLlorean: you tihnk that sounds good?
07:33:13JdGordonthen ill commit it
07:33:31LloreanI like the idea of clockwise rotation always increasing the volume.
07:44:26jhMikeSy yo tambien
07:44:49JdGordonque' ?
07:45:22jhMikeSNo entiendes lo que yo dije?
07:45:36JdGordonque' ?
07:45:49jhMikeSbah, conyo
07:45:55psycho_maniacJdGordon: is it in svn already?
07:46:36psycho_maniachow long ago?
07:46:54JdGordon2 commits?
07:47:58 Join kubiix [0] (
07:48:35psycho_maniacthought i wasnt quick enough. i tried patching my source but it said it was alredy in there.
07:50:31jhMikeSJdGordon: will the list go up now when I move cw or is the order reversed too?
07:51:05toffe82jhMikeS: debria continuar en espanol ;)
07:51:43JdGordonjhMikeS: the order is reversed.. so you move down the list
07:52:34jhMikeStoffe82: ay, porque? no esta apropiado en una canal ingles!
07:52:37 Quit Mouser_X (Nick collision from services.)
07:52:53toffe82jhMikeS: para molestar los otros :)
07:52:54 Join Mouser_X [0] (
07:52:57 Join flak [0] (
07:53:26jhMikeStoffe82: :) yo disfruto haciendo eso!
07:53:36toffe82jhMikeS: any luck with the info on the lcd driver ?
07:54:06jhMikeStoffe82: It does partially match. Some regs are referenced that don't seem to be in that datasheet.
07:54:09toffe82jhMikeS: where did you learn spanish ?
07:54:53jhMikeStoffe82: Mi novia es puertoriquena !
07:55:17toffe82jhMikeS: mine is chilena :)
07:55:28jhMikeSaha! :)
07:55:36toffe82jhMikeS: I think the datasheet is not exactluy the one of the lcd, it is a close reference, I couldn't find the one on the dtasheet of the lcd
07:55:51flakHey guys thinking about putting RB on my H10, would you advise backing up all data on it?
07:56:02jhMikeSno, not exactly, but it gives a clue. I wonder if something is closer though.
07:56:08toffe82I went on the nec site and there is no datasheet available
07:56:34scorcheflak: you shouls always have a back up, but rockbox wont wipe your device
07:57:22jhMikeStoffe82: I'd at least like to know a few more regs. Did you check all of them?
07:57:28flakscorche: Thanks mate, once rockbox is uploaded its just standard drag and drop to upload files?
07:57:43scorcheor syncing program, or however you like
07:58:08flakThanks heaps mate ill go for a backup now :)
07:59:29jhMikeStoffe82: I'm willing to go looking around when I'm done with the heap of stuff I'm doing now.
07:59:30toffe82jhMikeS: I don't have time and perhaps not the knowledge :(
08:12:11amiconnJdGordon: Speaking about non working things - I think I discovered another ui bug, but I have to verify
08:12:45JdGordonapparently the list buttons in the player are backwards or something also?
08:13:33JdGordonamiconn: can you confirm ?
08:14:00amiconnIf you have icons enabled (just using default), and start something that uses a viewer from the browser, after exiting the viewer, the icon of the current file disappears.
08:14:16amiconnMoving one line down or up makes it reappear
08:14:41amiconnObserved on H300 with mpegplayer
08:14:47*JdGordon checks
08:14:50 Join Rob2222 [0] (
08:15:23amiconnIt's the first time I noticed this, so it might very well be related to my driver work
08:15:40JdGordonits fine here... is that witht he default fonts?
08:16:24amiconnUsing Nibus-12
08:16:31 Join daurnimator [0] (i=daurn@unaffiliated/daurnimator)
08:16:34 Quit jhMikeS (Read error: 104 (Connection reset by peer))
08:16:37JdGordonbah, sorry, i meant icons
08:16:41 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
08:17:11JdGordoni cant reproduce it with svn
08:17:24 Join RaRe [0] (
08:17:35amiconnAh, it is due to the line selector
08:17:43 Join miepchen^schlaf [0] (
08:18:00amiconnLemme check sth
08:18:23JdGordonworks fine with cursor and full line
08:18:34amiconnI use solid color
08:18:37 Join cendres [0] (n=ashes@2001:5c0:8fff:ffff:0:0:0:4d)
08:18:52amiconnThe icon already disappears immediately when starting the file
08:18:59 Quit ashes (Read error: 104 (Connection reset by peer))
08:19:22amiconnYou'll notice if the hdd is sleeping and has to spin up when clicking the file
08:19:22JdGordondeja vu.. im sure i remember having this prblem AGES ago... but i cant reproduce it
08:19:26JdGordonand im off for 30min or so
08:20:35amiconnIt *only* affects solid colour
08:20:45amiconnNot pointer, not inverse, and not gradient
08:20:48 Join GodEater_ [0] (n=bryan@rockbox/staff/GodEater)
08:21:30amiconnNow it's gone... so something inconsistent...
08:21:55amiconnLooks like some variable doesn't get initialised properly
08:25:05 Join ender` [0] (
08:28:58 Join lee-qid [0] (
08:29:09 Join sheppard [0] (n=sheppard@
08:29:24sheppardI recently updated rockbox, updated the bootloader, etc.
08:29:56sheppardbut it seems sometimes when I power my ipod on, I get the stock ipod interface, and other times I get rockbox
08:30:14sheppardand the past few times turning it on it seems to be hooked on booting the default ipod interface
08:30:35amiconnAh, found the conditions...
08:31:26amiconnJdGordon: It only happens when (1) using the solid color or gradient selector _and_ (2) the current file line needs to scroll
08:32:06 Quit Rob222241 (Read error: 110 (Connection timed out))
08:32:18sheppardI just put it down
08:32:23sheppardand rockbox just loaded
08:32:54amiconnWhen clicking the file, the selector extends to the left margin of the screen, probably due to altering the drawing margins, overwriting the icon that way
08:32:56jhMikeSstill noone's converted the last PP502x to packed samples?
08:33:14amiconnjhMikeS: We're waiting for kkurbjun. Iirc he has a mini G1
08:33:27psycho_maniacwell the only way that the OF should load is when you have the hold button on. so i would think maybe your hold button is loose?
08:33:38jhMikeSok, last time I asked about who owned one noone said anything
08:34:04 Quit flak ()
08:34:09psycho_maniacjoin the club
08:34:21 Quit RaRe` (Read error: 110 (Connection timed out))
08:34:49amiconnI had one for testing earlier this year, from a colleague. But unfortunately that one is broken now (hardware wise, unrelated to rockbox)
08:35:21amiconn..and not by me
08:35:44GodEater_methinks amiconn doth protest too much....
08:36:02jhMikeS /me thinks the clocking setup
08:36:47amiconnjhMikeS: He broke the clickwheel wiring while changing the battery
08:37:04*GodEater_ thinks that the humour there was lost...
08:37:40psycho_maniachaha wow
08:38:15amiconnJdGordon: Re the forum post, I know that a number of settings is "backwards" on the player, but so are some on other targets
08:38:28amiconnSo at one point I stopped complaining...
08:39:09reverendnathanDarn, I shoulda recharged before taking a crack at rockbox
08:39:26reverendnathanI keep wanting to do stuff, but through-charging doesn't provide sufficient power
08:39:44jhMikeSrockbox is all about cracks or crack or whatever you like :p
08:41:02Mouser_XInfoNES plugin patch = very full of warnings...
08:41:26psycho_maniacthis is on the daily page......, These are automated daily builds of the current code. They contain all the latest features. They may also contain bugs and/or undocumented changes... identify your model,.... shouldnt it be on the current build page also?
08:41:32Mouser_X(I'm probably wrong, but I wouldn't be surprised if every line gives off some sort of warning.)
08:42:16GodEater_psycho_maniac: it shouldn't ALSO be on the current build page, it shouldn't be on the daily page AT ALL :)
08:43:02psycho_maniacmaybe on the front page?
08:43:23Mouser_XI think you missed his meaning.
08:43:24*jhMikeS is including the definition of NULL from stddef.h and many plugins using struct struct opt_items and complaining since NULL is ((void*)0) not 0.
08:43:43jhMikeS*are complaining
08:45:07jhMikeSwhat the proper value for no ID anway?
08:47:12GodEater_psycho_maniac: what about the front page ?
08:47:47psycho_maniacmaybe that link should be on the front page?
08:48:03*jhMikeS can't believe he hasn't earned a few "lazy points" so he can ask a stupid question here instead of finding out by searching. ;)
08:48:41GodEater_psycho_maniac: what link ?
08:49:06amiconnjhMikeS: The voice id is *not* a pointer!!
08:49:21jhMikeSamiconn: some plugins are using NULL for it though
08:49:23*amiconn wonders why so many devs get that voice id wrong
08:49:38psycho_maniacoops. the device chart page.
08:49:45amiconnThat's a bug and needs to be fixed. No id is -1, ___not___ NULL
08:49:48jhMikeSwhen I included a proper definition of NULL as ((void*)0) they complain
08:49:56GodEater_psycho_maniac: the device chart in the wiki ?
08:50:06 Quit reverendnathan (Remote closed the connection)
08:50:16amiconnapps/settings.h, lines 42ff
08:50:18jhMikeSone is reversi
08:50:33amiconnI fixed that in several plugins already, but didn't check all
08:50:34GodEater_psycho_maniac: it's a little out of date already though
08:51:28jhMikeSanything defining NULL as merely 0 should be fixed as well
08:52:05 Quit psycho_maniac (" HydraIRC -> <- *I* use it, so it must be good!")
08:52:16amiconnAccesses to the MBAR area (e.g. GPIO) do profit from buffered writes - but that means all delay loops need to be checked and recalibrated
08:52:19 Join psycho_maniac [0] (
08:52:55psycho_maniacGodEater_: what player do you have?
08:53:18amiconnAnd the greyscale coldfire target's lcd writes too - sigh
08:53:54JdGordonamiconn: bah ok, sounds like a jobb for nico then :p
08:54:00*JdGordon qiuck to pass the buck :D
08:54:08GodEater_psycho_maniac: I have four
08:55:04jhMikeSamiconn: what buffers writes anyway re: ports? I'd think that would delay some hardware aspect being set when you expect it.
08:55:45amiconnThe bus controller has a store buffer. Accesses are still serialized
08:55:57amiconnJdGordon: I think I already found the bug
08:57:07amiconnAH, no
08:57:29 Join RaZorbacK_ [0] (
08:58:46 Part RaZorbacK_
09:00:12jhMikeSamiconn: If a GPIO write is buffered, when is it actually changed? that's what I'm talking about.
09:00:33amiconnIt's changed a few cycles after the instruction finishes
09:01:20amiconnBut a read following immediately won't yield the old value
09:04:26amiconnJdGordon: Looks like I really found it now
09:05:57amiconnYup, verified. A 2-line fix in lcd-16bit.c
09:08:22 Join petur [0] (n=petur@rockbox/developer/petur)
09:08:46 Part toffe82
09:10:59***Saving seen data "./dancer.seen"
09:13:20GodEater_anyone have a problem with me uploading an e200rpatcher binary to the wiki ?
09:13:21 Nick fxb__ is now known as fxb (
09:13:30JdGordonfor which?
09:13:45GodEater_someone complained in the forums that we don't have one
09:13:55JdGordoni would think that linux users should be able to compile it themselves..
09:14:00JdGordonbut if you want
09:14:11GodEater_they should - but this new generation of ubuntu users.....
09:14:17JdGordonhaha yeah
09:14:27JdGordonand OY! im a ubuntu user ! evil glare*
09:14:39 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:14:45JdGordonhey Zagor
09:14:55Mouser_XHe did say "new generation." Perhaps you're old enough.
09:15:05JdGordoncan you stick some files on for the e200r?
09:15:41JdGordon and the 3 attached to
09:16:28GodEater_JdGordon: that would now be FOUR files :)
09:16:35JdGordonthat was quick :)
09:16:43GodEater_I already had it built
09:17:13Zagorin a new bootloader/sandisk-sansa/e200r-patcher dir then?
09:17:24JdGordonsounds good
09:18:45JdGordonanyone want to volanteer on updating the e200 manual instrausiont?
09:18:49JdGordoninstructions even
09:19:33GodEater_I would - but not being a e200r owner I'd be nervous of missing something out...
09:19:56JdGordonyeah, seems i stuffed up the instructions in the forums :(
09:20:07JdGordone200rpatcher, install the vanialla of, sansapatcher
09:20:37GodEater_yeah I saw that
09:22:22 Join ddalton [0] (
09:22:47JdGordonwhats the url of the master mirror?
09:22:47ddaltonJdGordon: why don't you like my patch?
09:22:57ddaltonI even got a comment from sdoyon :-)
09:23:09JdGordonddalton: i told you why when you first brought it up
09:23:56ddaltonwell you don't use voice and 3 blind users like it.
09:24:21ddaltonAnd I don't think LinusN and pondlife disagree with it completely.
09:24:44JdGordonhows usb going?
09:24:56 Part ddalton
09:24:59GodEater_I was just going to ask that :)
09:25:19 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:25:37Zagorslow, currently. I've hit a snag where I fail to activate interrupt on incoming bulk transfers.
09:25:44Mouser_Xpixelma: You were right. Rockbox+GBS+MOD+NES emu is done. Thanks for you help.
09:26:01Mouser_X*for your help.
09:26:22pixelmanice to hear, you're welcome :)
09:26:49Zagoror, I activate it but I don't get any
09:26:58Zagorvery frustrating
09:27:07 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
09:28:17 Quit hcs (Client Quit)
09:28:55 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
09:29:33JdGordonGodEater_: you didnt link libusb statiaclly?
09:29:51amiconnJdGordon: On the Player, all numeric settings are backwards. That includes all sound settings now, e.g. "+" decreases volume, "-" increases :(
09:30:17JdGordonthat shouldnt have flipped for player :(
09:30:31JdGordonok, ill fix that in a sec
09:30:37GodEater_JdGordon: no I didn't
09:31:17JdGordonok, wiki fixed with new linjks
09:31:51amiconnpetur: Didn't you commit the navigatable credits plugin?
09:32:15amiconnNow 'credits' can't be stopped anymore on Player...
09:32:27JdGordonkill the port.. :D
09:33:52peturamiconn: it can't?
09:35:18amiconnCONTEXT_KEYBOARD isn't defined for player, so it's mapped to the standard context
09:35:36amiconnAnd the standard context will *never* fire ACTION_KBD_ABORT
09:35:46JdGordonwait.. why are you using CONTEXT_KEYBOARD anyway?
09:35:55amiconnAsk petur, not me...
09:36:54peturwe need some general navigation context, I reused the keyboard one out of laziness
09:37:29amiconnIt used to be breakable with any button...
09:37:29JdGordonCONTEXT_STANDARD should be used,
09:37:48amiconn...and I don't see why that should change
09:37:51JdGordonthe credits plugin doesnt need anything more advanced han std_next std_prev and std_ok
09:38:34amiconnHmm, that was already gone before petur added navigation...
09:38:37JdGordonamiconn: all int settings are backwards?
09:38:45amiconnJdGordon: Looks like it
09:38:53 Join rvvs89 [0] (n=rvvs89@unaffiliated/rvvs89)
09:38:57*amiconn mentions the simulators
09:39:12amiconnThey're called *UI*simulators for a reason...
09:39:20JdGordonok, it looks like the charcell hack/fix wasnt put in when i reaarnged the code a while ago
09:39:29JdGordonyeah yeah...
09:39:50*petur has a new pc at work and installed ubuntu in a virtualbox - now for the compilers :)
09:40:19amiconnImho the sims are more useful for their original purpose (simulating the UI) than anything else
09:41:03amiconnLow level stuff needs to be checked on target anyway, and higher level non-ui stuff is usually the same for the whole class of targets (hwcodec or swcodec)
09:41:07JdGordonwait... they are usefull for something else?
09:41:22psycho_maniacanybody ever see this page? how to install rockbox on the gigabeat f/x
09:41:26amiconnBut no developer has *all* targets available, so UI stuff can be checked in sims
09:42:30amiconnJdGordon: Some devs seem to use them for debugging things like codecs, dsp stuff etc.
09:42:37psycho_maniacsorry ment
09:42:39 Join CaptainSquid [0] (
09:43:26*amiconn wonders whether we could simulate the wheel for wheel targets using the touchpad (on laptops) ;)
09:43:51hcsmouse gestures-ish?
09:43:52JdGordonI thought we didnt want the mouse patch built for the sim?
09:44:19 Join Crash91 [0] (n=evil91@
09:44:32 Part Crash91
09:44:33 Quit Mouser_X (Read error: 110 (Connection timed out))
09:48:12JdGordonamiconn: with svn, is the battery capaciy setting backwards also?
09:49:11linuxstbJdGordon: Barry said that boot time was now 5-6 seconds and thought that was about 50% of what it was previously. I don't think it does any harm in sansapatcher - ipodpatcher has similar advanced options which are not documented in the manual.
09:49:35amiconnJdGordon: yes
09:49:35JdGordoni tried it with a stopwatch and yeah, about a 1 sec speed up
09:49:52JdGordonit takes nearyl 2s to start loading anything which is annoying
09:50:15JdGordonok, so at least i'v fixed some of it
09:50:38 Join spiorf [0] (
09:51:06LloreanIt doesn't bypass the Sandisk logo?
09:51:14peturbah, it looks like SVN uses a port that isn't opened by our firewall at work :(
09:51:48JdGordonLlorean: it does, it jus displays nothing for the first secod or so
09:52:10LloreanIn that case, I think how much time you save might depend on which OF you're coming from.
09:52:25LloreanI'd been led to believe "newer" OF has this annoying, long "Lil' Monsta'" animation that the older OF didn't have.
09:52:27JdGordonno, the of shouldnt be run at all now
09:52:43JdGordonexcept when you want it to boot the of
09:52:48JdGordonwhich stil takes age
09:52:56LloreanThat's what I meant though, the newer original bootloader might be longer than the older original bootloader
09:53:30JdGordonamiconn: all fixed, commit, or you want to test first?
09:53:45*jhMikeS thinks SanDisk threw that in there just to annoy _us_. :)
09:54:22JdGordonthe new OF has the stupid monsta thing from the c200's? wow /me not upgrading!
09:54:50amiconnpetur: svn can also use https iirc, but I don't know whether that needs to be enabled on the server..
09:55:02LloreanJdGordon: So I've heard, I haven't tried it.
09:55:07jhMikeSlet's kindly ask them to throw in a database update bypass too
09:55:28JdGordonor better, if they throw docs our way :p
09:55:48jhMikeSwhile they're at it ... schematics
09:56:09hcsmaybe some cash
09:57:44peturamiconn: looks like it. https doesn't respond and http gives me a 'Method Not Allowed'
09:58:11GodEater_JdGordon: with a bit of help from linuxstb I now have a statically linked e200rpatcher
09:58:20peturI'll have a chat with out network admin, just need to figure out a reason to get that port open
09:58:20JdGordonpetur: do you want me to fix credits?
09:58:39JdGordonGodEater_: stick it somewhere so Zagor can replace it
09:58:42peturJdGordon: if it's urgent, yes ;)
09:58:51JdGordonits not, but im bored :)
09:59:33 Quit linuxstb ("Leaving")
09:59:40 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
09:59:44 Join davina [0] (
09:59:55jhMikeSpetur: it's costing the company lots of cash not to? ;)
10:00:36GodEater_JdGordon: I've stuck it on the end of the same wiki page
10:00:48peturwe only have customers with cvs, none with svn it seems :(
10:01:11jhMikeSconvince one of them they should switch
10:01:31ZagorGodEater, JdGordon: should I replace the linux version with this static?
10:02:13GodEater_thanks :)
10:02:26 Join barrywardell [0] (
10:02:58*GodEater_ hides the attachment
10:03:11JdGordonpetur: is the scrolling your code?
10:03:12GodEater_you going to edit the wiki JdGordon ?
10:03:17JdGordonyou can :_)
10:03:29peturJdGordon: about half of it
10:03:44peturit was a patch in the tracker that I modified
10:04:25peturerrr wait, the scrolling or the navigation?
10:04:46JdGordonnavigaition, i tinhk
10:04:46GodEater_JdGordon: I'm getting 404s for all the links - how long does it take to get to all the download mirrors ?
10:04:56JdGordonsome time?
10:05:59 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
10:10:23 Quit qweru ("moo")
10:11:44 Quit Chronon ("ZZZZZzzzzz")
10:12:23pondlifeAnyone else testing Nico_P's work?
10:13:43GodEater_c'est magnifique
10:13:50pondlifeYep, seems good in the main
10:14:13GodEater_Sunday's commit seems to have got rid of the "randomly stop producing sound" issue I had on Friday
10:14:27GodEater_or more accurately - a commit in between probably did
10:14:43pondlifeI'm getting an occasional hard hang today though
10:14:53GodEater_not had that yet
10:14:57pondlifeSeems to be related to codec change
10:15:04JdGordonCONTEXT_STD may not have been the best one to use :p
10:15:10pondlifeI'm testing with alternating MP3/WMA tracks
10:15:13GodEater_pondlife: I only have two codecs on my player
10:15:21GodEater_and I didn't get a problem with switching
10:15:28GodEater_let me try again
10:15:42pondlifeIt's not a reproducible crash (at least, not yet)
10:16:18GodEater_nope - just switched fine
10:16:20pondlifeThe only other problem is that selecting an unbuffered track stops playback for longer than SVN
10:16:26GodEater_MP3 -> SPX here
10:16:36GodEater_that's true
10:16:41GodEater_I said that to Nico too
10:16:58pondlifeIt's very noticeable with crossfade on
10:17:17pondlifeUnder SVN I can make pretty much any transition smoothly
10:17:19GodEater_ah - I don't use that either
10:17:29 Join Benoitb [0] (
10:17:32pondlifeWith MoB, it almost always gaps
10:18:19*GodEater_ turns it on
10:18:20pondlifeGetting close though
10:18:30BenoitbI can't apply the latest version of the mpeg_resume patch fs#7487
10:18:38 Join ddalton [0] (
10:18:39pondlifeOops, just locked again
10:18:53Benoitbeven if I try to apply it before any other patch on a fresh svn
10:18:55ddaltonIs it ok if I close p7705?
10:19:01ddaltondoes anyone still want it?
10:19:03pixelmaBenoitb: probably because it was committed?
10:19:24JdGordonwell this sucks... autoscroll will only stop after 1 full screen is finished being displayed
10:19:32Benoitbit just came into my mind after I wrote it on the chan
10:19:42ddaltonok im closing it...
10:20:51 Quit Siku ()
10:20:59GodEater_pondlife: has Nico's work changed much in playback.c yet ?
10:22:18ddaltonhmmm how do I close it?...
10:22:29GodEater_pondlife: shame it hasn't quashed that bug stripwax showed us :)
10:22:48pondlifeTrue, but it should now be possible to fix bugs more easily
10:23:02GodEater_so I'll be expecting your fix to it any minute then
10:23:13pondlifeMore easily than impossible
10:23:57JdGordoncredits on the player is a bit scray :D
10:24:15 Join Peter15 [0] (
10:24:37psycho_maniacJdGordon: why do you say that?
10:24:38pondlifeIf it's not too cheeky, I'd like to (a) wait until MoB is stable and committed, then (b) review the playback.c code and see if we can put some more comments in... then it'll be fixy-time
10:25:10pondlifeAh, yes and move the voice thread out of playback.c perhaps.
10:25:18JdGordonpsycho_maniac: 2 scrolling lines going pretty quickly, for 300 lines or something
10:25:30JdGordonpetur: all fixed
10:26:24amiconnPutting almost correct lcd waitstates for X5 makes it work when boosted, and speeds up almost everything, but slows down standard lcd updates when boosted a little bit
10:26:42amiconnUsing less waitstates works, but I don't like exceeding specs that much
10:26:58 Join [1]psycho_maniac [0] (
10:27:25amiconnThe HD66773R manual states a minimum cycle time of 200ns. It works down to 100ns
10:28:22amiconnMy first test used 5 waitstates as on H300 (which has HD66789R with a minimum cycle time of 150ns), resulting in already out-of-specs 130ns...
10:28:58 Join reverendnathan [0] (
10:31:23reverendnathanOK, so I just got rockbox, trying new themes for my iPod Video. Using themes with Album Art, but the art won't show!
10:31:36reverendnathanUsing fusion build
10:31:50JdGordonwe dont support custom builds in here
10:32:27peturJdGordon: thanks
10:33:34reverendnathanOh OK
10:34:25amiconnAh, on H300 even 3 waitstates don't work... then it's due to dma vs. cpu transfer
10:35:01amiconnI thinkl I'll settle with 4 waitstates for X5. Normal lcd update when boosted stays almost the same speed then
10:37:00 Quit reverendnathan (Remote closed the connection)
10:37:09 Quit spiorf (Remote closed the connection)
10:41:25 Quit psycho_maniac (Nick collision from services.)
10:41:25 Nick [1]psycho_maniac is now known as psycho_maniac (
10:47:58 Quit homielowe (Read error: 110 (Connection timed out))
10:48:12*amiconn goes hunting delay loops
10:50:31GodEater_pondlife: cross fading is working ok here too
10:51:42pondlifeI doubt that's related to my crashing, but it does make the slower rebuffer more noticable.
10:51:57pondlifeOr maybe there's a lack of yielding during rebuffers now?
10:52:24pondlifeDoes your file browsing go slower during buffering, or was that more obvious with SVN ?
10:52:58pondlifeI wonder if that's more noticeable with a scrollwheel...
10:53:13GodEater_pondlife: that's hard to say
10:53:17JdGordonwhere do i get his latest source?
10:53:28GodEater_the file browser feedback has always sucked during buffering
10:53:52pondlifeGo for the latest snapshot
10:53:57pondlifeOr use git?
10:54:25GodEater_I'm using the code from the last commit Nico made
10:54:43pondlife"git checkout master; git branch -D mob; git checkout -b mob origin/mob" works for me. You may only need the last bit..
10:55:01pondlifeSo browse-during-buffer still sucks?
10:55:08GodEater_as far as I can tell
10:56:41pondlifeHhard crash again... :/
10:56:52GodEater_still working fine here :)
10:56:59pondlifeCaused by skipping backwards during buffering I think
10:57:01amiconnjhMikeS: I found a few delay loops (in fact less than I expected, just in 6 places). And it looks like they need fixing, e.g. pcf i2c is a bit unreliable on my X5 right now
10:57:13amiconn(time occasionally disappearing in wps status bar)
10:57:26*pondlife hopes his reset switch survives the day
10:57:48GodEater_pondlife: I wouldn't say the performance of the file browser is any better / worse than svn during buffering
10:57:50 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
11:00:16pondlifeOK, I can crash it every time now, I think
11:00:21JdGordonno problem with voice which is nice :)
11:01:15pondlife1) Set up a decent-sized playlist (mine's about 9000 tracks, although I doubt you need that many).
11:01:35pondlife2) Go via the WPS context menu into playlist viewer
11:02:01pondlife3) Repeatedly go up/SELECT
11:02:19pondlifei.e. alternate so you're skipping backwards
11:02:19amiconnJdGordon: Umm, now + increases the value and - decreases it, but the list is still backwards
11:02:23 Join Thundercloud [0] (
11:02:38JdGordonin which?
11:02:42amiconnI always expect + to move down. Now it moves up, where the higher values are
11:02:43jhMikeSamiconn: I guess that's to be expected if the i2c is oversped :)
11:02:49amiconnjhMikeS: yup
11:02:50 Join reverendnathan [0] (
11:03:08 Join miepchen^schlaf [0] (
11:03:18amiconnNeed to make a tiny measing loop and then fix the delays
11:03:26JdGordon+ shuold move down the list?
11:03:31JdGordonnow im really confused
11:03:34reverendnathanQuestion, whenever I connect my iPod to the computer, it charges, but doesn't mount... how can I get it to?
11:03:47reverendnathanAs in, the computer ain't recognizing it as a drive
11:03:50amiconn+ always moves down the list
11:03:56JdGordonpondlife: track change to an unbuffered track is at least as fast as svn on my sansa
11:03:57psycho_maniacreverendnathan: do you have a default build ?
11:04:04amiconn(or should, and does, e.g. in menus and the browser)
11:04:05jhMikeSamiconn: how about the serial number chip too?
11:04:17reverendnathanGah, darn it I don't
11:04:24pondlifeJdGordon: Definitely slower here, can take about 5 seconds
11:04:28pondlifeOn H340
11:04:30JdGordonpondlife: and data abort doing the up+select thing
11:04:31amiconnIn sound settings it moves up. That gives correct behaviour for the values, as those are also reversed
11:04:35JdGordonah well that would xplain it
11:04:47amiconnBut it's backwards logic...
11:05:18JdGordonwait a sec.. opening the player sim
11:05:18amiconnImho it should move down, and the higher values should be further down the list
11:05:32amiconnBattery capacity is now correct
11:05:40 Part ddalton
11:05:59JdGordonso volume is now wrong?
11:06:13JdGordon+ makes it louder.. isnt that correct?
11:06:16 Join obo [0] (n=obo@rockbox/developer/obo)
11:06:26amiconnHmm, the list in numeric settings is backwards on all targets. Why is that??
11:06:29 Quit lee-qid (Read error: 110 (Connection timed out))
11:06:42pondlifeJdGordon: + = louder = move down list
11:06:47 Join webguest23 [0] (i=4b498002@gateway/web/cgi-irc/
11:06:47pondlifeShould be, I mean
11:07:01JdGordonwell when its only 2 lines up/down doesnt really exist
11:07:26GodEater_pondlife: creating an enormous playlist now
11:07:43amiconnJdGordon: it does...
11:07:46pondlifeI can crash the sim using this recipe too, but maybe a different reason
11:08:00pondlifeI get a segfault in tagtree_unbuffer_event
11:08:06amiconnjhMikeS: Serial number seems to work correct, but I have that file on my list too
11:08:59amiconnJdGordon: Why are numeric settings backwards on all targets?
11:09:02 Quit hcs ("Leaving.")
11:09:05JdGordonbackwards how?
11:09:22JdGordonlower at the top?
11:09:26psycho_maniacthis seems really confusing :S
11:09:34JdGordoni mean, higher at the top
11:09:35amiconnNo, lowering when going down the list...
11:09:56JdGordonumm... thats what was decided?
11:10:03JdGordonbut yeah, that really doesnt make sense anymore
11:10:33amiconnIt used to be the other way 'round, and I can't remember voting about that
11:11:00***Saving seen data "./dancer.seen"
11:11:26 Quit webguest23 (Client Quit)
11:11:36 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
11:11:40JdGordonnote to nico: poweroff while buffering data aborts
11:12:27JdGordonamiconn: no, it makes sense... higher values at the top of the list
11:12:27 Quit Jon-Kha (Read error: 131 (Connection reset by peer))
11:12:57amiconnNot really...
11:13:13JdGordonit just doesnt feel right with wheel targets, and the player
11:13:22jhMikeSwhy volume should be a list is beyond /me
11:13:26amiconnOr rather, it depends on how you look at that list
11:13:39JdGordonpressing up on the iriver makes sense to be going up in values
11:13:44amiconnIf you see it as a list of values (as I do) , they should increase downwards it was, and wouldn't cause inconsistencies between targets
11:14:03JdGordonand to everyone else.. thats backwards :)
11:14:06*linuxstb agrees with amiconn
11:14:28JdGordondinner.. we'll pick this up in 30
11:14:58amiconnIf you see numeric settings as some kind of slider, it makes sense to have higher values at the top, but then we get inconsistencies between directional moving targets and wheel targets+the player
11:15:01 Quit Peter15 ()
11:15:13 Join CaptainSquid83 [0] (
11:15:47jhMikeSa vertical slider would make sense there though and with lots of room to display min/max + actual value
11:16:17amiconnThere already was a slider experiment. It looked way ugly
11:16:25*amiconn likes the list with its fine control
11:16:27LloreanInstead of showing a list for volume, could we just show the "current" value, much like it is in the status bar?
11:16:45LloreanThen up/down makes sense, since it's not a list.
11:16:45jhMikeSah, but that's the art part and it _can_ look good
11:17:00psycho_maniacmaybe this is the time to use the (forgot the name) but the setting name and actual setting were on the same line.
11:17:17pondlifeGodEater_: brb, time for a cuppa
11:17:42 Quit CaptainSquid (Remote closed the connection)
11:17:56*jhMikeS hopes his player's disk isn't biting the dust...don't need that now
11:20:09 Quit idnar (Nick collision from services.)
11:20:11 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
11:22:04GodEater_pondlife: how many times do you do this till it crashes ?
11:22:33pondlifeAbout 3
11:22:44GodEater_and do you wait for it to start playback
11:22:51pondlifeFast as you can
11:22:52GodEater_nor me
11:22:58GodEater_I've gone back about 30
11:23:00GodEater_no crash so far
11:23:06pondlifeMight be database related
11:23:18GodEater_I don't use it
11:23:27pondlifeSpecifically, my playlist is Database > Tracks
11:23:46pondlifeSim segfault is database related
11:24:01pondlifeI'll try a file instead..
11:24:42 Join Jon-Kha [0] (
11:25:32pondlifeWooh, a crash with a splash! "No codec for ./Pulp/Different Class/03 Common People.mp3"
11:25:40*GodEater_ tries a database load
11:25:41pondlifePin time
11:27:35pondlifeHmm, a file browse crashes too here
11:27:47pondlifeCouldn't repro that splash though
11:29:07 Join bluebrother [0] (i=PnYDE8b9@rockbox/staff/bluebrother)
11:31:03GodEater_pondlife: can't reproduce in the DB either
11:31:42 Quit psycho_maniac (" HydraIRC -> <- *I* use it, so it must be good!")
11:31:55amiconnjhMikeS: The pcf i2c is probably why you saw better performance with the spc codec
11:32:03pondlifeGodEater_: Must be timing critical.
11:32:07amiconnpcf i2c is running all the time for reading buttons
11:32:25jhMikeSamiconn: I was using H120 for that so I doubt it :)
11:32:40amiconnHmm, not all the time, due to the detection interrupt
11:32:50amiconnH100 has a bitbanging adc driver too...
11:34:34*JdGordon back
11:34:53JdGordonthe slider looked terrible
11:35:07JdGordonI tink higher values at the top makes sense for all but the wheel targets
11:35:10GodEater_pondlife: you're clearly faster than me :)
11:35:12amiconnThat reminds me - I wanted to make that one adapt to cpu clock like the pcf driver does
11:35:27amiconnCould save a percent or so when not boosted
11:36:08jhMikeSThere's other fw and software where a slider doesn't look terrible so that makes no sense to me that it's the slider at fault.
11:36:19*amiconn needs to measure clock cycles for GPIO access with vs. without buffered writes
11:36:36amiconnjhMikeS: A slider has some fundamental problems
11:36:43JdGordonthere was an idea for a triangle slider (like the volume display in the statusbar), but using up/down for it feels wierd
11:37:03JdGordonamiconn: i disagree... a list is just as fine-grained as a slider
11:37:08amiconnIf it's horizontal, its operation is inconsistent with the other settings, confusing blind users, or blind operation
11:37:39amiconnIf it's vertical, giving the extra information for precision looks horrible and takes a lot of screen estate
11:38:27amiconnJdGordon: A list may be as fine grained, but without also printing the numeric values, it's not as precise to operate
11:38:53JdGordonoh, anything we do would have to display at least the current value
11:38:57amiconnThat was btw one motivation of the rockbox port to the archos recorder v1 (!)
11:38:58JdGordonand shhould show min and max
11:39:06LloreanI still don't see why it needs to display a full list.
11:39:08JdGordonwhich was?
11:39:13bluebrotherwhy does a horizontal slider need left / right to change values? It's just a graphical representation of a value
11:39:17LloreanOr a slider.
11:39:24jhMikeSman, the archos is just going "cha-chunk, cha-chunk" trying to unzip a build to it. :(
11:39:38amiconnThe archos OF uses sliders, which are sort-of fine grained, but never reproducible, since they didn't show the numeric value
11:39:53pondlifeLlorean: I like that we have as few UI types as possible. Would prefer a list to a bespoke UI..
11:39:55JdGordonbluebrother: having a horizontal slider but pressing up/down feels very wierd
11:40:08Lloreanpondlife: When the user clicks on "Volume" it could just display "Volume: -27db" and scrolling or up/down could change the value.
11:40:13bluebrotherJdGordon: I disagree completely. It's just a graphical representation.
11:40:19amiconnSo you e.g. had to remember a bass setting like: "Between 3rd and 4th notch, a little right from the middle"
11:40:20Lloreanpondlife: Essentially mirroring volume operation in the WPS, except it's a volume-specific screen.
11:40:23JdGordonits also counter-intuitve
11:40:28amiconn(hope that's understandable)
11:40:34JdGordonif you see a horizontal slider, you instinctivly press left right
11:41:16JdGordonis there a fill triangle lcd function?
11:41:18pondlifeLlorean: I'd still prefer a single structure for all settings. Makes code simpler, makes blind naviagation simpler. Although your proposal would be identical blind I guess.
11:41:26amiconnJdGordon: In the plugin lib there is
11:41:37JdGordonsmall enough to put in the core?
11:41:44amiconnWhy would we want that?
11:41:46jhMikeSeasy to get a HD for a player?
11:41:51Lloreanpondlife: It would be identical blind to the current implementation, and I thought numeric lists were already declared differently in the code (and this would pretty much be useful for any numerical list), though I could be wrong.
11:42:01amiconnjhMikeS: Just standard 2.5" notebook hdd
11:42:14 Quit reverendnathan (Remote closed the connection)
11:42:37 Join spiorf [0] (
11:42:41jhMikeSthis one's 2.5" standard scraping noises
11:42:47*bluebrother remembers having done some work on FS #5990 a year back or so ...
11:42:51pondlifeLlorean: Yep, probably. I quite like the scroll bar though.
11:42:58amiconnjhMikeS: Checked the batteries and battery contacts?
11:43:12jhMikeSit is plugged. will that still affect it?
11:43:20Lloreanpondlife: Not quite sure which use of "scroll bar" you're using here. :)
11:43:20pondlifeJdGordon: What was the original problem? Scrollwheel was anticlockwise to increase?
11:43:33pondlifeLlorean: haha... on screen!
11:43:39Lloreanpondlife: Do you mean "like the sliders in the EQ" or "on the left side of long lists"?
11:43:42amiconnThe archos charger is *just* a charger, it doesn't provide enough juice to run the player directly
11:43:45*bluebrother pasted the wrong link ... :o
11:43:48pondlifeThe left side of lists
11:43:59JdGordonpondlife: yes
11:44:19amiconnThe archos never draws more than ~350mA from the charger, but the hdd needs almost 1A when spinning up
11:44:23 Quit Thundercloud (Remote closed the connection)
11:44:38JdGordonbluebrother: yeah, but that wont look good on small displays, i.e the nano
11:44:43pondlifeJdGordon: Why not revert your changes for non-scrollwheel targets?
11:44:53JdGordonwhich change?
11:45:06pondlifeThe numeric list direction...
11:45:11amiconnbluebrother: Now imagine that with a larger font...
11:45:11jhMikeSI don't think this one's archos issue...and dammit, I thought it was plugged-in and it wasn't. lol
11:45:21JdGordonpondlife: nothing changed in that commit for non wheel targets
11:45:22pixelmaJdGordon: the Nano's display has a higher resolution than the H100's...
11:45:29JdGordonit does?
11:45:35bluebrotherJdGordon: iirc I extended that patch to auto-swap to a horizontal view once the space gets too less
11:45:41amiconnjhMikeS: Also, if you didn't use the player for a while, the batteries might be discharged
11:45:44Lloreanpondlife: I think some of us don't like the idea of the list being different ordered on different targets
11:45:52pondlifeAh, ok
11:45:53pixelmabluebrother: I also don't like that you can't see the possible "steps"
11:45:55jhMikeSoh, bloody hell...plugged it and it just wrote fine. omfg.
11:45:56amiconnThose NiMHs are nasty, unless you get som enew-generation one
11:46:03LloreanJdGordon: 176x132 for Nano, 160x128 for H120 if I recall
11:46:14pondlifeI would rather it was the best UI for the input device
11:46:28JdGordonLlorean: I disagree with that.. lists shuold be highr at the top for button targets, and lower at the top for wheel and player
11:46:29pondlifeOn a joystick, up = vol up makes sense.
11:46:42Lloreanpondlife: Well, if the list is actually displayed, scrolling clockwise to go "down" the list makes sense, and if lower values are down, no problem.
11:46:48LloreanOr rather, if higher values are down, no problem
11:47:07amiconnJdGordon: Sort of... depends on how you think of the volume setting
11:47:28amiconnBut with a directional cross, it does at least make some sense to have higher values at the top
11:47:31bluebrotherpixelma: that's indeed a drawback.
11:47:36amiconnOn wheel targets and player, it does not
11:47:51JdGordonthats what im saying
11:48:05jhMikeSI guess it doesn't kindly do hw shutdown when low eh?
11:48:15pondlifeI think we all agree... no?
11:48:25JdGordonyes we dont agree
11:48:27amiconnjhMikeS: No, because that would either be too unreliable, or waste a lot of juice
11:48:29JdGordonor no we do kinda
11:48:41pondlifeMaybe not
11:48:52pondlife"That's not an argument.."
11:48:55jhMikeSno, we agree that yes we don't
11:49:04pondlifeI disagree
11:49:17jhMikeSI agree with that
11:49:19amiconnNiMH cells have a very flat discharge curve, and differences between brands exist, plus differences in adc calibration between individual targets
11:49:28*pondlife heads back to work and testing MoB
11:49:44jhMikeSamiconn: are they just AA NiMh or some custom thing?
11:49:47 Quit Rob2222 (Read error: 104 (Connection reset by peer))
11:49:48amiconnANd since deep discharge doesn't destroy NiMHs, we decided that only LiIon+LiPo targets get low-bat shutdown
11:49:53amiconnJust AA
11:49:57 Join Rob2222 [0] (
11:50:06*JdGordon works on getting rid of the need for the fliplist bandaid
11:50:08amiconnI'd try to get some new generation ones if you want to replace them
11:50:31jhMikeSsure, easy to get that I think
11:50:39amiconnJdGordon: Am I right in thinking that it's just your last change in option_select.c that should be reverted?
11:50:58pondlifeWould anyone apart from me like a WPS virtual LED for CPU boost?
11:51:18amiconnjhMikeS: Avoid those top-capacity ones. They have higher capacity than the new generation, but still the high self discharge rate problem
11:51:27amiconn(ofetn even worse than old ones)
11:51:42Lloreanpondlife: Doesn't serve much purpose in the WPS, does it?
11:51:48JdGordonamiconn: no, this has been a problem for ages... its just that commit brought it to light again
11:52:02pondlifeNah, I'd just like to keep an eye on buffering and boost..
11:52:20*jhMikeS sees 14, 15, 14 in the Bass setting on player ??
11:52:21Isolinearpondlife: I like the idea.
11:52:39amiconnjhMikeS: Range problem... ask JdGordon ;)
11:52:43jhMikeSwtf, the menu's all out of order
11:53:00jhMikeS10, 9, 10, 11,
11:53:38JdGordonwhisky tasngo foxtrot
11:56:41jhMikeSfoxtrot uniform charlie kilo delta uniform papa!
11:57:49pondlifedelta alpha papa?
11:58:35jhMikeStango hotel echo mike echo november uniform sierra
11:59:09JdGordonso, have we agreed that up == increase on button targets, and clockwise == increase on wheel, and right == increase on player?
11:59:21JdGordonfor all int settings
12:00:00*jhMikeS thought we agreed to disagree but finds that agreeable.
12:00:40amiconnJdGordon: I think I can live with that, if the list increases downwards on player and wheel targets
12:01:06*JdGordon 's current desktop if anyone craes :)">
12:01:59JdGordonjust to prove i do use the player sim :)
12:02:40jhMikeSJdGordon: cleanup your desk!
12:02:40 Join jonashn [0] (i=573e6a35@gateway/web/cgi-irc/
12:02:51amiconnJdGordon: What? Just 10 open app windows? ;)
12:02:58bluebrotheruse virtual desktops!
12:03:02*amiconn has 21 windows atm
12:03:05*JdGordon hates virtual windows
12:03:23amiconn...and 29 tabs in ff
12:03:34JdGordonthats a bit... excessive
12:03:44amiconnNah, it used to be more...
12:03:48 Join Thundercloud [0] (
12:03:54*jhMikeS has 6 tb tabs...doesn't want more
12:04:43*JdGordon curses bloody negative Db values
12:05:08jhMikeSdo only windows sims come with the neato player bitmaps? why don't I see those?
12:05:23LloreanjhMikeS: rockboxui −−background, if I recall
12:05:25amiconnUse −−background
12:05:36JdGordonjhMikeS: ,y desktop looks like windows?
12:06:33jhMikeSJdGordon: I thought it was :p...perhaps Vista or something which I refuse to get
12:06:49JdGordonhehe silly bugger... its kde
12:07:42bluebrotherlooks like kubuntu
12:08:27JdGordongot it in one
12:10:31*jhMikeS never thinks about desktop comp stuff and so stays ignorant
12:10:45 Quit jonashn ("CGI:IRC (EOF)")
12:13:18 Join agm3nt [0] (
12:13:34jhMikeSThat slider looks pretty good
12:16:55 Join kubiix [0] (
12:17:10 Join RockBox [0] (
12:18:11JdGordonok, anyone want to test the patch?
12:18:30bluebrotherwhat patch?
12:18:31jhMikeSslider patch?
12:18:50JdGordonto make the int settings consistant on each target">
12:20:23JdGordonamiconn: so you dont get cross after i commit it? ^ :D
12:21:41 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:21:58*bluebrother curses this stupid frame thingy
12:22:18JdGordonyeah, my bloody isp doesnt allow incoming on port 80
12:22:35JdGordonhey Nico_P
12:22:43JdGordoni finally got to try your mob code
12:22:53JdGordona[art from the 2 data aborts, it works nicely
12:23:12Nico_PI just saw a mail from pondlife about a data abort
12:23:34JdGordonshutting down while buffering got one
12:23:47JdGordoni assume pondlife tolkd you abuot the skiping in the playlist viewer
12:24:03pondlifeYes, that's the same one
12:24:07amiconnA whole typo collection ;)
12:24:10Nico_Pah yes
12:24:26Nico_Pdidn't read closely enough ;)
12:24:34jhMikeSshouldn't abs() be in there somewhere?
12:24:35Nico_Ppondlife: thanks, it should be easy to track down
12:24:48pondlifeHope so
12:25:13pondlifeDo you still support the scrobbler callbacks, by the way?
12:25:32Nico_Ppondlife: I'm not sure but there's a chance
12:25:39pondlifeAh, one more thing - the buffering is a lot more relaxed now compared to SVN
12:25:51JdGordon[20:23:58] <amiconn> euh <- was that to the typos or the patch?
12:25:54pondlife(it's very obvious in the sim)
12:25:56Nico_Ppondlife: which means ?
12:25:57amiconnthe typos
12:26:02JdGordonok :)
12:26:47pondlifeNico_P: It means that playing an unbuffered track is rarely seamless with a 2 second crossfade on
12:26:57pondlifeSVN buffers faster
12:27:10pondlifeI suspect you've just got to play with sleeps a bit
12:27:36amiconnWhat's crossfade? :P
12:27:54pondlifeIt's something on those new-fangled SWCODEC targets :p
12:27:55jhMikeSif any issues exist with kernel object speed those will be resolved in not too long
12:27:59JdGordonI didnt check with crossfade on, but had no noticable gap (no bigger than svn anyway) skipping to an unbuffered track on my sansa
12:28:05Nico_Ppondlife: what case is that in ? skipping to an unbuffered track with corssfade on ?
12:28:28pondlifeYes, but crossfade isn't needed ... it just makes it obvious as I'm used to it being seamless.
12:28:48pondlifeYou may need a slow-ish target to see the difference.
12:29:43pondlifeI was using the database (same as the back-select crash) if that's relevant
12:30:44pondlifeWooh, sim segfaults immediately I resume playback too
12:31:08Nico_Plots of bugs :p
12:31:18pondlifePossibly the same one
12:31:32 Join jba [0] (
12:31:36jhMikeSother playback bugs are going to swarm like locusts because of the timing changes
12:32:04Nico_Pamiconn: thanks for the bufix btw
12:32:17pondlifeIf you're interested..
12:32:23Nico_Ppondlife: I sure am !
12:32:42Nico_PjhMikeS: what timing changes ?
12:33:11JdGordonthis should give a small green delta as a nice side effect :)
12:34:43jhMikeSchanging stuff around in playback is like whacking a hornets nest with a stick :)
12:34:43 Join stewball`ghost [0] (n=WTFOMGBB@
12:34:59JdGordonif its a big enough stick, it will work though
12:35:56pondlifeThat's the only hope, really :/
12:35:57jhMikeSNico_P: just things with threading that relied on yields in certain places instead of actual sync
12:36:10pondlifeJdGordon: "min = sound_max(setting_id);" isn't obvious... !
12:36:19 Part agm3nt
12:36:23pondlifeI'm sure it works, but...
12:36:29Nico_PjhMikeS: I'm trying to do better... not sure it actually is better though
12:36:32jhMikeSbig enough stick = fusion nuke
12:37:08Nico_PjhMikeS: what do you think of audio_yield_codecs ? I have a similar one for the buffering code
12:37:09jhMikeSNico_P: It looked better to me quite a bit but there's all the other code too
12:37:39jhMikeSI've always wondered why that's nescessary
12:39:13Nico_Pit interrupts the buffering loop if there are new messages in the queue
12:39:28Nico_Pand if the pcm buffer is low on data too
12:39:50maxkelleyheh.. my friends whacked a hornet's nest with a stick :)
12:40:05Nico_Pin my code it allows not having to buffer all of a track before being able to treat the nex queued message
12:41:05jhMikeSNico_P: queue_wait will yield completely and process a message if needed
12:42:20Nico_PjhMikeS: yes but I can't call queue_wait in the buffering loop, can I ?
12:42:40jhMikeSqueue_wait_w_tmo you can
12:43:14jhMikeSsince it's both a sleep and message grabber
12:43:31RockBoxcan any1 help me out with rockbox here?
12:44:27RockBoxunder database, all tracks, how come all my songs are untagged?
12:45:15GodEater_because all your songs are untagged ?
12:45:30Nico_PjhMikeS: but then how do I get the event in the main buffering thread loop ? it will have been consumed already, won't it ?
12:46:04RockBoxhow do i tag it?
12:46:43GodEater_RockBox: with a good tagging program
12:47:12pondlifeRockBox: Good name, btw..
12:47:29bluebrotherit's kinda strange talking to Rockbox ...
12:47:35RockBoxi change my nick
12:47:53bluebrotherlike talking to the Compiler (who was in once ...)
12:48:01pixelmaRockBox: names some
12:48:03Nico_Ppondlife: I couldn't repro the bug you mailed me about with a non-dataase playlist
12:48:07RockBoxwhat program?
12:48:30amiconncf timing is weird....
12:48:59jhMikeSNico_P: return it to the main loop by passing the event struct pointer?
12:49:12RockBoxbut i transfer like 2000 songs to my ipod, its possible thats its all untagged?
12:50:16RockBoxwhich tagging program u guys recommend?
12:50:20jhMikeSNico_P: you can stay in the function if EV_TIMEOUT is received but return an actual event if one comes up.
12:50:31 Quit rjg (Remote closed the connection)
12:50:38Nico_PjhMikeS: makes sense
12:51:35pondlifeNico_P: I get another crash with file-based browsing, but not reproducible yet
12:51:46*amiconn hopes his H1^80's USB socket will survive all this testing :/
12:52:07 Join ddalton [0] (
12:52:12Nico_Ppondlife: even with a database playlist, I haven't had a crash by doing UP and SELECT repeatedly (gigabeat)
12:52:14 Part ddalton
12:52:19Nico_Pbut audio didn't start
12:52:53*Nico_P feels a little overwhelmed...
12:53:01Nico_PI should note things down
12:53:06*pondlife knows that feeling
12:53:26pondlifeWe need to have a little reserved bit of Flyspray for MoB...
12:53:41Nico_Pmaybe it's time to open a task
12:54:11barrywardellhow do people feel about displaying the rockbox logo in the bootloader (mainly sansa for now)
12:54:26GodEater_badly usually
12:54:34GodEater_it was removed from the Gigabeat bootloader for example
12:54:47GodEater_the feeling is the bootloader should do nothing more than is necessary to boot
12:54:57GodEater_at least, that was how it was explained to me
12:54:59barrywardellI think that was different, it loaded the logo from a file, didn't it?
12:55:09barrywardellI'm talking about a compiled in logo
12:55:12GodEater_it did yes
12:55:31amiconnI now think that a compiled in logo makes sense, but then it should apply to all targets
12:55:56amiconnHowever, there's a problem for colour targets where the bootloader resides in flash: the logo is quite big
12:56:05barrywardellI have made the changes necessary for e200 and it works fine, no noticable slowdown
12:56:14amiconnOn H300 the bootloader probably won't fit with the logo built in
12:56:40barrywardellamiconn: ah, how much space is available?
12:56:45amiconnMy idea was to introduce compressed bitmaps for colour targets, and decompress
12:56:58amiconnIirc it's 64KB, and the bootloader already uses 40-ish
12:57:08amiconnThe logo is ~29KB for H300
12:57:23Nico_Pamiconn: maybe a full screen logo isn't necessary
12:57:35amiconnNico_P: A jumping logo would look very odd
12:57:46Nico_Phmm that's true
12:57:49barrywardellthat's not full screen, it's just the same one shown already on normal boot
12:57:49amiconnEither the same as what rockbox shows, or none
12:58:09amiconnbarrywardell: Yes, but that's already 29KB for a screen width of 220 pixels
12:58:32barrywardellyeah, gigabeat is ~94KB
12:58:37amiconnFullscreen would be 76KB
12:59:04barrywardellhow quick is the compression you were thinking about?
12:59:04amiconnI tried ucl compression on the logo (just using the .o), and got it down to ~8.5KB
12:59:17 Join rjg [0] (
12:59:30amiconnThe cut-down ucl decompressor needs ~1KB of code
13:00:37amiconnDecompression is quite fast. I didn't measure on cf or arm yet, but on 11MHz SH1, decompressing ~200KB takes less than a second
13:01:14amiconnAnd since ucl decompression uses bitshifts, it should be faster on cf and arm...
13:01:23barrywardellthere's also a problem on PP that the bootloader currently needs to fit inside IRAM
13:01:56amiconnI would expect the logo decompression to take around 50ms
13:02:06amiconn(on H300)
13:02:17amiconnVery coarse estimation of course
13:02:19barrywardellthe ucl code is already in svn?
13:02:27barrywardell50ms is very acceptable
13:02:55amiconnThe ucl compressor is also in svn, built for all archos targets
13:03:15 Quit RockBox ()
13:03:19amiconnThe stripped-down decompressor only supports compression method 2e, and only a single block
13:03:19barrywardell <-that's my patch without compression
13:03:49 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
13:03:51amiconnSo the compressed data must be made using uclpack −−2e −−best, and must not exceed 256KB
13:04:03amiconn(the −−best is optional)
13:04:19barrywardellthe stripped down one is in firmware/decompressor?
13:04:45jhMikeSbarrywardell: why do we need to fit it into IRAM? If we don't load RB, any DRAM can be discarded. Of course some things might need setting and resetting to load retailos.
13:05:10barrywardellthe first thing the bootloader does is copy itself to IRAM
13:05:19amiconnYes, and in flash/bootloader/bootloader.c
13:05:39amiconnIt's the very same function
13:05:52jhMikeSsure, but nothing says that _has_ to be the case. I can't see any reason to stick to that.
13:06:31barrywardellamiconn: thanks, I'll look into that
13:06:56barrywardelljhMikeS: but then loading the code to dram will overwrite the currently executing bootloader
13:07:27amiconnbarrywardell: Maybe it's a bit too early to do that now. It would be easier after introducing bitmap headers
13:07:39jhMikeShasn't it loaded rockbox already and then it jumps to whatever fw is picked?
13:07:52jhMikeSthe rockbox bl more accurately
13:08:30barrywardelljhMikeS: no, it copies itself to IRAM, then loads either rockbox or of the DRAM_START
13:08:38barrywardellamiconn: are they planned?
13:09:30jhMikeSbarrywardell: ah, so we're first and then we run OF bl or rockbox fw?
13:10:37barrywardellcurrently it runs OF bl first, then Rockbox bl, which then either loads OF or Rockbox to DRAM_START and jumps there
13:11:02***Saving seen data "./dancer.seen"
13:11:09amiconnbarrywardell: In fact compressed bitmaps would make sense on colour targets in main rockbox too, for large, infrequently used logos
13:11:12jhMikeSright, so then we only put parts of the rb bl that can be overwritten when no longer needed. A bitmap image should only be needed once.
13:11:16amiconnRockbox logo, and usb logo atm
13:11:54barrywardelljhMikeS: yes, true. the bootloader could me modified to work like that
13:11:59amiconnbarrywardell: We could run from dram, load OF or rockbox into a buffer, and then only have a tiny copy-and-run routine in iram
13:13:22amiconnThat's what RoLo does, on SH1 and coldfire at least.
13:13:56amiconnLoad the new firmware into dram, and then call a small iram function that copies the new fw into place (overwriting the old one), and runs it
13:13:58jhMikeSIf we want we could try to lock it in the cache. Gigabeat can do that and perhaps PP. Needs poking sessions.
13:14:24amiconnColdfire can also lock cache lines
13:14:32amiconn(icache only of course)
13:14:52barrywardellyes, all good options. but the current bl works and if aint' broke...
13:15:04amiconnBut why fiddle with that if we can use iram?
13:15:06jhMikeSI think gigabeat may require that for a RoLo
13:15:30amiconnGigabeat also has iram. Even if it's tiny, it should suffice for things like rolo code
13:15:34jhMikeS...if it ain't broke...make it less broke :)
13:15:37barrywardellif the bootloader grows bigger, then it will be necessary
13:17:22amiconnOkay, buffered writes shorten the execution time of instructions writing to gpio by 6 cpu cycles
13:17:27jhMikeSamiconn: I'm aware of that. I'm under the impression there was some trouble with it.
13:17:54amiconnHmm, but pcf i2c still seems to glitch...
13:18:04jhMikeSamiconn: as opposed to what?
13:18:33jhMikeShow many cycles was it before?
13:18:41amiconnDepends on the instruction
13:19:38amiconnThe atomic bit manipulation (and.l / or.l) needs 21 cycles with unbuffered writes, and 15 cycles with buffered writes
13:19:54 Quit FOAD ("I'll be back")
13:21:16jhMikeScould buffered writes cause changes in the GPIO level to have bursts of fast activity?
13:21:58barrywardellwhy is an .o file ~1/3 the size of the corresponding .bmp?
13:22:29jhMikeSit's the native format?
13:22:45barrywardellah, yeah
13:23:24amiconnShould be 2/3 for 16 bit targets
13:23:42amiconn(assuming 24 bit bmp)
13:24:35jhMikeSwell starting with a small-sized 32-bit .bmp with a big BITMAPINFOHEADER to a native 16-bit color format with no header could easily go to 1/3
13:24:39barrywardellyeah, two thirds, I read wrong
13:24:50barrywardell19677 vs 28566
13:25:11amiconnThe bmp header is fixed size, and rather small
13:25:12barrywardellI was looking at the wrong .o file
13:27:49 Part Llorean
13:33:38pondlifeNico_P: Shall I just create a MoB testing wiki?
13:33:46Nico_Ppondlife: please do :)
13:34:00*pondlife does
13:34:23Nico_PI'm implementing jhMikeS's suggestion and then I'll try to investigate your bug reports
13:36:07Zagoranyone want to lend a few moments to look at my usb code? I'm doing something wrong with the bulk transfer setup, but I can't for the life of me figure out what.
13:36:23Zagorsome extra eyes might help
13:37:04Zagorand even just the teddybear effect might
13:40:28pondlifeNico_P: Do you need to wait differently if it's the same file, or another? I'd think that if lots of buffering is needed, it's not th end-of-file that should determine breaks...
13:40:36 Join teddybear [0] (i=c27f0812@gateway/web/cgi-irc/
13:40:51teddybearZagor: now I'm here. Shoot.
13:40:58barrywardellZagor: yeah
13:41:01Nico_Ppondlife: are you talking about the commit I just pushed ?
13:41:03pondlifeNico_P: Or maybe I've misunderstood "/* Don't stop buffering if the request is to buffer the same file */"
13:41:37ZagorI'll throw up a patch in the tracker
13:41:41jhMikeSZagor: teddybear effect? you mean attention?
13:41:47Zagorteddybear: :-)
13:42:02pondlifeWarm fuzzy feeling, from sharing, I'd guess.
13:42:11ZagorjhMikeS: nah, I mean explaining the code to anyone (anything) sometimes help you understand the issue
13:42:12pondlifeMmm, midnightbeerfeast
13:42:14Nico_Ppondlife: prior to this commit, the buffering loop would stop everytime an event was recevied. if the event was Q_BUFFER_HANDLE for the same handle, it's a waste of time to break the loop, only to call buffer_handle again for the same handle
13:42:26pondlifeAh, ok
13:42:41Nico_Pthat's why this check is here
13:42:41pondlifeI assumed it would carry on buffering while there was buffering to be done
13:42:46*jhMikeS petur just came up with a nice marketing scheme for an aging population while saving the teddybear industry from extinction
13:43:01amiconnHmm. Too bad our i2c expert won't be around this week
13:43:04Zagorthere's a legend an IT support center somewhere had a big teddybear outside their office with a sign "explain your problem to me first", which solved half their issues...
13:43:26pondlifeScary stuff
13:43:40jhMikeShmmm...usually a smoke does it for me
13:43:57Nico_Ppondlife: do you think the comment is unclear ?
13:43:57pondlifeMaybe the bear had a packet of cigarettes?
13:44:11pondlifeNico_P: Maybe you could explain it a little more....
13:44:41 Quit safetydan ("Leaving")
13:44:44Nico_Pthere is a mention of this in the comment on top of the function though (inc ase you hadn't seen it)
13:44:45jhMikeSAs soon as I talk about threading code in the kernel IRC usally becomes a teddybear :p
13:44:58*amiconn guesses there's a delay missing somewhere, already out of specs in svn but only showing up with buffered writes
13:45:17pondlifeNico_P: If you don't mind, at some point I'd like to go through playback.c (and buffering.c, although I don't think that'll need so much work) and make sure there are plenty of comments. I'd like you to be involved before you suffer terminal playback.c fatigue.
13:45:21amiconnCould be in start, stop, or acknowledge, have to read up on i2c :/
13:45:31pondlifeNico_P: Which is a very real problem round here :)
13:45:31Nico_Ppondlife: no problem
13:45:40jhMikeSamiconn: I thought we usually pushed it to the limit SVN
13:45:42teddybearZagor: so I'm all ear
13:46:20jhMikeSteddybear: and the other? or did a dog chew that one off?
13:46:46pondlifeteddybear: GodEater_ keeps beating me at arm wrestling, do you know any way I could cheat
13:47:59teddybearpondlife: drink more beer is the answer!
13:48:07pondlifeGood answer
13:48:28teddybearpondlife: you mean, good teddybear?
13:48:34pondlifeOf course
13:49:04pondlifeBut mainly good beer, bear.
13:49:10jhMikeSteddybear: I didn't teddybears talked back. Am I going insane?
13:50:00 Quit iamben (Read error: 104 (Connection reset by peer))
13:50:02Zagorsee #7962
13:50:06*pondlife went insane way back when, but is better now thanks to his friend, teddybear.
13:50:25 Join teddybeer [0] (i=54bd45af@gateway/web/cgi-irc/
13:51:33jhMikeSteddybear: can you write me a prescription for a good phychoactive pharmaceutical?
13:51:37*pondlife imagines that the office at Contactor is full of soft toys, all working hard
13:51:47barrywardellZagor: only for c200?
13:51:58Zagorsuspects: firmware/target/arm/usb-drv-pp502x.c, prime_transfer() and/or init_queu_heads()
13:52:02 Quit teddybeer (Client Quit)
13:52:17Zagorbarrywardell: only tested on c200 at least. may work for e200 too
13:53:12jhMikeSZagor: what should it do up to this point?
13:53:38*jhMikeS supposed he'll just look at FS
13:54:04Zagorit registers with the host and sends all its' descriptors. so you get the device in lsusb. but it doesn't receive the mass-storage commands (no interrupt triggers) so I'm not getting any further right now
13:54:50Zagorbasically it's about the same stage as the svn usbstack, but I'd like it working a little better before committing
13:55:53amiconnThat patch seems to consist almost completely of removals ;)
13:55:54*jhMikeS sees the patch is almost 100% "-" lines
13:55:56Zagorfirmware/usb_core.c does all the top-level stuff (answer config requests), firmware/target/arm/usb-drv-pp502x.c does the chip-level stuff. firmware/usb_storage.c does (will do) the storage stuff
13:56:09Zagoramiconn: yeah it's quite a lot of - ...
13:57:10Nico_PjhMikeS: would you mind taking a quick look at;h=253f0e81c936f42d1472b0cfa2efe661532242df ?
13:58:54Zagorthe patch file is probably quite messy to read due to all the removals. I recommend setting up and applying it to a svn repo
13:59:52*Zagor ponders adding automatic 'diffstat' to flyspray
14:00:05jhMikeSNico_P: yeah, looks good but isn't 10 ticks a long time to pause?
14:00:25 Quit qwm (Remote closed the connection)
14:00:36 Join qwm [0] (
14:00:49pondlifejhMikeS: He's just reduced it to 1, no?
14:01:22Nico_PI have
14:01:24*jhMikeS sees 10 in there
14:01:34Nico_Pit's in another commit
14:03:04 Join Mouser_X [0] (
14:03:29*Nico_P now has over 100 commits on the mob branch
14:04:17 Join Llorean [0] (
14:05:19 Join rogelio [0] (n=rogelio@
14:05:59barrywardellZagor: so usb_drv_int() doesn't get called when it should?
14:07:06 Part rogelio ("Kopete 0.12.4 :")
14:07:36Nico_Ppondlife: have you seen my wiki page edit about how to pull from a rebased branch ?
14:08:05amiconnJdGordon: Now the sound settings are correct on Player, but battery capacity is backwards...
14:08:07pondlifeI'm just editing a testing page...
14:09:14*JdGordon cant win a trick
14:09:24amiconn(- goes up and + goes down as it should, but the values are in the wrong order)
14:10:09JdGordonit seems i may have a genetic desease that means i cant test properly :p
14:10:49JdGordonamiconn: pressing right increases the value...
14:11:04JdGordonweather it goes "up" or "down" is irrelevant when its only 2 rows
14:11:15*JdGordon didnt stuff up
14:11:17amiconnNot here...
14:11:32*jhMikeS wonders if some tick tasks polling a USB register could be flubbing things up
14:11:37amiconn+ decreases the battery capacity here
14:12:01amiconn..since the list is flipped, which higher values at the top
14:12:16JdGordonunless 2450 is lower than 2400 its working corrrectly
14:12:40amiconnNo, 1950 is lower than 2000
14:12:57amiconnAnd other int settings are also reversed, e.g. scroll speed
14:13:16JdGordonone at a time... this makes no sense...
14:13:33JdGordonim pressing right and the cursor is changing to the higher values in the battery capacity setting
14:13:36 Join kugel [0] (i=kugel@unaffiliated/kugel)
14:13:43amiconnAgain - not here
14:13:48amiconnAnd I am at svn head...
14:14:11amiconnOr not..lemme check again
14:14:13JdGordonthats fucked!
14:14:26amiconnI just svn upped, compiled and put it on
14:14:52JdGordondefinalty running r15120?
14:15:39JdGordonZagor: was your usb code tested on a e200? or just c?
14:15:41*amiconn fully rebuilds
14:16:11 Quit jba (Read error: 110 (Connection timed out))
14:16:48barrywardellJdGordon: <Zagor> barrywardell: only tested on c200 at least. may work for e200 too
14:17:59Nico_Ppondlife: nice, thanks :)
14:18:26pondlifeI'm just adding a commit time/date too
14:19:18JdGordonhis patch doesnt apply cleanly
14:19:58barrywardellJdGordon: same problem for me here
14:20:03Nico_Ppondlife: in case you want to refer to a specific git commit, the date-time or SHA1 hash won't do (because of the rebasing), use the title instead
14:21:25pondlifeHow's that?
14:21:25jhMikeSNico_P: It's possible to download a full diff of MoB from SVN?
14:22:20pondlifeNico_P: How do I access your buffering debug screen?
14:22:25Nico_PI can provide test diffs
14:22:33Nico_Ppondlife: system > debug
14:22:46pondlifeIs it only in a debug build?
14:22:59pondlifeI see the usual View audio thread..
14:23:04pondlifeBut nothing new
14:23:15Nico_Pshouldn't be, I have it on my gigabeat
14:23:48 Join jba [0] (
14:23:57pondlifeMaybe the debug menu has limited capacity? It's not using the Menu API...
14:24:02jhMikeSNico_P: that would be a slick setup to have a "download repository as patch to SVN" link :)
14:25:03Nico_PjhMikeS: actually,;h=mob;hp=master;hb=mob;hpb=master
14:25:05GodEater_pondlife: are you going to make a forum posting referencing that page too ?
14:25:14pondlifeNot yet
14:25:21pondlifeI think it's more one for devs, not users.
14:25:33GodEater_fair enough
14:26:07 Join FOAD [0] (
14:26:12JdGordonamiconn: ?
14:26:23pondlifeNot that I normally want to subdivide, but I don't think Nico will want lots of iPodn00bs reporting intermittent problems at this point...and complaining that rockbox sux0r
14:26:38pondlifeHmm, I just found a nice bug too
14:26:55pondlifeWill update my build and retest
14:27:09barrywardellJdGordon: you can ignore the failed bits - they're all just removed files
14:27:42JdGordonexcept the e200 config needs fixing
14:28:08JdGordon... or not
14:28:12Nico_PGodEater_: have you seen that I found how to pull from a rebased branch ?
14:28:34barrywardellyeah, it does
14:28:45barrywardelljust the same as the c200 one
14:28:48amiconnJdGordon: Now the order is correct, and I'm definitely at r15120 - *But*:
14:29:06amiconnNow the start position in the list is wrong, i.e. it's not the current value (!)
14:29:45amiconnLooks like the start position is counted from the bottom, while it needs to be counted from the top
14:30:51 Join tictoc [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
14:34:18JdGordonand fixed
14:35:24 Quit jba ("Leaving")
14:37:27JdGordonwell stuff happened
14:38:20JdGordonis there a way to clear the dmesg log?
14:38:36*teddybear was distracted by another talk and lost the course of discussion
14:39:32*Nico_P wants to know who teddybear is
14:40:13teddybearNico_P: I'm the one you should explain your problem to in order to solve it
14:40:55JdGordonhow thoughtful
14:41:21tictoc2002-05-03: SOUND! Linus' experimental MAS code has played our first 4 seconds of music.
14:42:12amiconnHmm. Our bit-banged i2c seems to bee indeed somewhat qirky regarding timing
14:44:00 Quit teddybear ("CGI:IRC")
14:44:04Nico_Ppondlife: you're depressing me
14:44:28Nico_Ptoo many bugs ;)
14:44:34pondlifeI suspect you've already fixed one of them
14:44:42Nico_Pwhich ?
14:44:52pondlifeThe hard-lock with long tracks
14:45:02pondlifeCan't repro in the sim
14:45:12Nico_PI was just about to say I didn't have it on my gigabeat
14:45:13pondlifeAbout to update my H300
14:46:19pondlifeMight be dircache related?
14:46:47pondlifeI have that enabled
14:46:54Nico_Pme too
14:46:57pondlifeAh, ok
14:47:19pondlifeNo, still locks up on H340
14:47:55pondlifeI'm using "Slow Riot For New Zero Kanada" for my test tracks there if that helps..
14:48:24pondlifeBut any two MP3s that total >= buffer size should do
14:48:49Nico_PI'm trying to repro with exactly that setup
14:49:14Nico_Pwell not the same test tracks but two bit MP3s
14:49:50Nico_Ppondlife: how long in the second track do you skip backwards ? and do you have playlist repeat enabled ?
14:49:51Zagorbarrywardell: sorry for running away. yes usb_drv_int() does get called, for ep0 and reset and everything. it just doesn't get called for ep1 (the bulk endpoint).
14:50:48Zagorhence I suspect an error in the transfer descriptor setup. but i've gone over them dozens of times, and I'm not seeing the error.
14:51:03pondlifeNico_P: About 8 seconds into each track.
14:51:11pondlifeRepeat: All
14:51:44pondlifeI think the playlist reshuffle issue's been fixed...
14:51:51pondlifeBut I keep getting hard locks...
14:51:54 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
14:53:03pondlifeAlso the new build has a slient moment shortly after playback starts...
14:53:18pondlifei.e. it plays the first 3 seconds, then pauses while buffering
14:53:43pondlifeMaybe I should put my config/WPS up somewhere?
14:54:08Nico_PI don't think it's needed
14:54:21pondlifeWell, just in case you can't repro
14:54:31pondlifeCrossfade might be required
14:54:38Nico_Pah I'll try that
14:54:44pondlifeIt's locking up on every skip now
14:54:54pondlifeEven when there should be no buffering
14:55:25pondlifei.e. Start playback of an album, wait for HD to spin down, then skip forward one track :(
14:56:10Nico_Pouch I just experienced the audio cur after three secs, but only with whit enoise
14:56:31Nico_Pit only happends with crossfade enabled
14:57:24preglowwe use dma for audio on x5 too, yeah?
14:59:19amiconnLooking a lot better now :)
14:59:30pondlifeHmm, not heard whitenoise yet..
14:59:32*amiconn wonders how that timing mess had worked
14:59:33Nico_Ppondlife: have you rebooted your player ?
14:59:35barrywardellZagor: just looking over lsusb now...
14:59:48 Quit Thundercloud (Remote closed the connection)
15:00:01pondlifeNico_P: Well, I updated it in USB Bootloader Mode..
15:00:13pondlifeAnd I've hit reset many times ;p
15:00:18pondlifeAnd again!
15:00:32amiconnjhMikeS: I some cases we did flipped SCL twice in succession, without any delay besides the instructions itself
15:00:32pondlifeIt locks up just playing now
15:00:39amiconn*did flip
15:00:54pondlifeNico_P: Last night's code played pretty well for hours this morning.
15:01:40Nico_Ppondlife: you seem to attract bugs
15:01:56pondlifeThat's a skill of mine, I feaer
15:01:59pondlifefear, even
15:02:12pondlifeI can repro a crash in the sim now, just gdb-ing
15:02:13Nico_Pcrossfade is working quite well here, apart from the whitenoise issue
15:02:16barrywardellZagor: those descriptors look fine to me
15:03:04Nico_Ppondlife: the whitenoise seems to occur only when crossfading to a track that wasn't buffered
15:03:11Nico_Pooh, crash
15:03:35pondlifeSim message : failed to allocate space for metadata Postbuffer:2/5
15:03:47pondlifeIt refuses to play long track 2
15:03:50Zagorbarrywardell: I mean the internal transfer descriptors used by the chip to handle transfers. the ones set up with prime_transfer() in usb-drv-pp502x.c
15:04:11Nico_P"failed to allocate space for metadata" is normal, that's what tells us the buffer is full
15:04:25pondlifeYes, but it should still allow me to skip to the track, no?
15:04:46Nico_Pindeed it should
15:04:50pondlifeSkip back gives the same tagtree_unbuffer_event segfault
15:04:58pondlifeSo, maybe these are all the same problem.
15:05:19pondlifeBasically tagtree_unbuffer_event expects the metadata to be around, but id3 = NULL.
15:05:28Nico_Pdo you use database on sim ?
15:05:45pondlifeI have it enabled, but I was doing a file browse this time.
15:05:47Nico_Ppondlife: and could you pastebin the backtrace ?
15:05:55pondlifeYup, looks the same again though
15:06:08Nico_PI have but the lines seem strange
15:06:59barrywardellZagor: but they're working fine for transferring the device descriptors
15:07:29pondlifeSeems like you're unbuffering the metadata before the callback... audio_clear_track_entries() maybe?
15:07:30Nico_Ppondlife: yeah, that's just a missing check
15:07:43pondlifeHmm, shouldn't need a check
15:07:53pondlifethe point of the unbuffer is to wave bye-bye
15:07:58Zagoryes. but those are transfered on endpoint0, which is of type control. mass-storage goes on endpoint1, of type bulk
15:08:02pondlifei.e. it shouldn't pass id3=NULL
15:08:15pondlifeThis is needed for the database run-time stats
15:08:25Zagoreverything works fine with ep0, but not ep1
15:08:39Nico_Ppondlife: on line 2335, tracks[last_idx].id3_hid can be == 0
15:08:47Nico_Pin that case bufgetid3 will return NULL
15:08:53pondlifeWhich file?
15:10:17pondlifeThis happens when the track has already been unbuffered?
15:10:31Nico_Ppondlife: it probably never has been buffered
15:10:39Nico_Pnot even the metadata
15:11:05pondlifeWell, that's fair enough then... I suspect that SVN rebuffers at that point
15:11:06***Saving seen data "./dancer.seen"
15:11:18Nico_Pthat's lousy code translation on my part. in the original code the check wasn't needed because metadata could almost always be assumed to be there
15:11:39pondlifeHmm, need to check all callbacks maybe?
15:11:57pondlifeIs there any point in doing a callback with id3=NULL ?
15:12:10Nico_PI doubt it
15:12:46Nico_PI think the fact it crashes kinda proves it :)
15:12:54pondlifeIf the metadata's not around, shouldn't it still have the pathname set up?
15:13:04pondlifei.e. an id3 structure should still exist.
15:13:29pondlifeI think that's how SVN works (from memory)
15:13:42Nico_Pthe metadata not being there means the buffer was so full there was no room for a struct mp3entry
15:14:01Nico_Pin SVN the struct mp3entries were static
15:14:13pondlifeThink of the initial display of next track info - often a WPS will display the next track filename if the actual title wasn't available...
15:14:26Nico_Pthat's handles by audio_current_track
15:15:10pondlifeI think we need to have an artificial struct for this purpose too. :(
15:15:22pondlifeNot sure how it can be achieved though.
15:16:09Nico_PI've pushed a preliminary fix
15:18:49 Join desowin [0] (
15:20:10pondlifeIt does seem odd that we're calling an unbuffer callback for a track that has either never been buffered, or already unbuffered...
15:21:20barrywardellZagor: it does receive on ep1 according to the logf. have you checked what it receives?
15:21:47ZagorI get nothing. what does the logf line say for you?
15:21:57pondlifeNico_P: Will test after lunch
15:22:00Nico_Ppondlife: I agree it's strange. I'm looking
15:22:13barrywardelluprime 1(2)rx 10117040
15:22:45Zagorbarrywardell: that's the priming, i.e. assigning of receive buffer. that doesn't not mean it received.
15:22:47barrywardelljust after usb_recv 1 512 14V8V8
15:23:34Zagorwhat you want is "usb comp 2"
15:23:40barrywardellah, ok
15:23:42Zagoror what I want, anyway :)
15:24:25barrywardellso it sets up to start receiving, then doesn't actually receive anything
15:24:29pondlifeNico_P: Sorry, but that still crashes on H340.. :( Now I really must eat.
15:24:32Zagorbarrywardell: are you running this on e200?
15:24:44Zagorbarrywardell: exactly. no receive interrupt for ep1
15:24:55Nico_Ppondlife: I've pished another similar commit for after your lunch ;)
15:25:30barrywardellZagor: I'm slowly starting to understand how the code works
15:25:34 Join iamben [0] (
15:25:35amiconnI wonder whether the quirky i2c timing for pcf might even explain the button glitches on X5, M5 and H300
15:25:59 Join Siku [0] (
15:26:55peturamiconn: and the (very sporadic) eeprom corruption on h300?
15:27:12barrywardellZagor: do you know if the interrupt never happens, or does REG_ENDPTCOMPLETED just not get set?
15:27:15amiconnpetur: Is the eeprom hooked to the same i2c bus as the pcf?
15:27:28peturamiconn: I think so, yes
15:27:32amiconn(i.e. is it using the same bitbanging driver?)
15:27:36Zagorbarrywardell: good point, haven't verified that
15:28:01*Zagor adds yet more logf
15:28:17amiconnThe interesting thing is that these quirks weren't introduced with the asm; the C version has them too
15:30:59peturamiconn: I did however close FS #5412 because it was no longer reported
15:31:59amiconnI'll referify the timing (drawing diagrams), then check the pcf specs again. Perhaps the timing has to be done even more different
15:34:02barrywardellZagor: gtg now. I'll let you know if I spot anything
15:34:51ZagorI found we get a crapload of "start of frame" and "suspend" interrupts I didn't know about. could be to something
15:36:11barrywardellyou also don't check for bus reset when priming, dunno if that's important
15:36:56Zagorwe should do that, but it's not the problem here.
15:36:58 Nick Tanuva|Zzz is now known as Tanuva (n=tanuva@
15:39:24*amiconn _really_ wonders how this could have worked
15:41:59 Join Rincewind [0] (
15:42:16RincewindNico_P: are you here?
15:42:38amiconnThe pcf driver in svn has several ill-timed spots. When sending, the GETACK SCL high pulse is too short. And when receiving, there are 3 problems in a transfer. (1) The SCL low pulse between START and INB is short. (2) ACK sends the SCL->high transition directly after setting the data line. (3) the SCL low pulse between ACK and STOP is short
15:43:18RincewindI tried to use your mob branch and I got a few conflicts. Your tree isn't up to date with current svn, right?
15:43:30Nico_PRincewind: no, not quite
15:43:53Nico_PRincewind: it's based on r buffer_callback
15:44:00Nico_Poops, I meant r15095
15:44:12Rincewindhm ok
15:44:34Nico_PRincewind: the raw text diff is here:;h=mob;hp=master;hb=mob;hpb=master
15:44:36Rincewindthen I have to make a branch based on this revision, and pull your remote branch in there
15:44:51Nico_PRincewind: are you using git ?
15:44:55RincewindNico_P: yes
15:45:07Nico_Phave you cloned my repo ?
15:45:33RincewindI cloned your repo, and used git svn rebase from then on
15:45:54barrywardellZagor: you could also maybe check ENDPTSTAT to see if the priming succeeded
15:45:57Nico_Pwell even though you've rebased, a simple git checkout -b my-mob origin/mob should do
15:46:38Zagorbarrywardell: good idea
15:48:27*barrywardell is really going this time
15:48:33Nico_PjhMikeS: here ?
15:48:41 Quit barrywardell ("using sirc version 2.211+KSIRC/1.3.12")
15:51:01RincewindNico_P: I used 'git branch my-mob origin/mob' && git checkout my-mob && git pull. Then I got the conflicts. I tried your method now and it seems to be working
15:51:43Nico_Pno need to pull after checking out (which is what you did with your first git branch call)
15:51:53Nico_Pthough it's strange you got conflicts
15:52:17Nico_PRincewind: was there some time before you did git pull ?
15:52:36Rincewindbut a pull shouldn't do any harm, either
15:52:52RincewindNo, only the time needed to type it in
15:53:20Rincewindnow a pull says 'already up to date'. It seems to be ok now
15:53:53Nico_PI have no idea why you got conflicts
15:54:47Nico_PRincewind: to be on the safe side in case I rebase, you should probably pull with git pull origin +mob:my-mob and then do git remote update
15:55:12markunZagor: do you know how I can setup a category for the wiki?
15:55:46RincewindI set −−track on the checkout, it should pull from your repo automatically.
15:56:40Rincewindif you rebase it gets messy, you're right.
15:56:46Zagormarkun: it's only search patterns. for example CategoryFrontpage is %SEARCH{"^CategoryFrontpage:.*\[.*Documentation.*]" regex="on" type="regex" nototal="on" nosearch="on" format=" * [[$topic]] - $pattern(.*?CategoryFrontpage:([^\[]*).*)" }%
15:57:00Zagorheh, "only". but you get the idea?
15:57:18markunWell, I did "view raw" and didn't see any of this
15:57:19Nico_PRincewind: I need to rebase from time to time... if you have tracking set up you might need to ass a plus sign somewhere
15:57:35Nico_Ps/ass/add... my typos are getting worse
15:57:53markunZagor: or is it not defined in the wiki itself?
15:58:10Nico_Pmarkun: it's in the page you want to add the category items in
15:58:13Zagormarkun: CategoryPlugin is just an explanation. look at PluginIndex instead.
15:58:53RincewindNico_P: If this happens, I am sure that git knows a way to solve it. It is really amazing
15:58:57markunZagor: ah, ok
15:59:59Nico_PRincewind: it won't know what to do if I rebase, it'll only know if you tell it the branch may be rebase, and that's with the plus sign
16:00:31Nico_PRincewind:, the first note
16:00:37Rincewindwhat I meant was, there is always a way to solve these issues with git
16:00:51Nico_Pthat's true
16:02:18 Join n1s [0] (
16:02:20 Join webguest55 [0] (i=5620ba17@gateway/web/cgi-irc/
16:02:30Zagorgotta go
16:02:31 Quit Zagor ("Client exiting")
16:02:31amiconnpetur: The eeprom does indeed share the pcf i2c bus on h300. Go figure...
16:02:40webguest55does anybody know if there is something for ipod nano 2nd gen?
16:03:01webguest55i only found 1st gen
16:03:18linuxstbwebguest55: No there isn't. It's completely new hardware and no-one is working on porting Rockbox to it.
16:03:20 Quit webguest55 (Client Quit)
16:03:28peturamiconn: I remember guessing it had something to do with the eeprom corruption
16:03:31amiconnpetur: Do you have a way to test eeprom access reliably?
16:03:57peturthere's the eeprom dump
16:08:10 Quit CaptainSquid83 (Remote closed the connection)
16:08:20amiconnLooks like the pcf i2c timing is rather forgiving. The minimum clock half-period is 1.25 us (400kHz "high speed") as per specs. The svn driver sometimes generates pulses as short as 0.4 us
16:08:24 Part Benoitb ("Kopete 0.12.5 :")
16:09:09amiconnOnly going further down (by enabling buffered writes) made it start glitching with a rate high enough to become visible
16:10:14amiconnHmm, the start->inb transition was even shorter
16:10:16 Join scorche|w [0] (n=8dc5049d@rockbox/administrator/scorche)
16:13:19*amiconn thinks that fix is worth a separate commit
16:14:28*amiconn also wonders whether we could even remove the special filtering from those button drivers
16:15:28 Join karashata [0] (
16:20:45 Join daurn [0] (i=daurn@unaffiliated/daurnimator)
16:21:12 Quit atsea-34 (Remote closed the connection)
16:23:12markunwas adding files to the wiki disabled?
16:23:29markunI can't seem to find it anymore
16:23:49Nico_Pmarkun: "attach", down the page
16:24:06markunNico_P: how could I have missed that :)
16:27:02 Join Gursikh [0] (i=khalsa@tremulous/officialdevannoyer/khalsa)
16:27:13 Quit desowin ("use linux")
16:27:15markunpicture from the New York meeting:;filename=nyc-20071011.jpg
16:27:29GursikhHI, Just popped in to say "ZOMG Rockbox I love you guys! Keep it up!"
16:27:33 Part Gursikh ("woot rockbox")
16:28:26linuxstbmarkun: Who is who?
16:28:35markunLTR: Nick, Davide, Melba, Rob, Marcoen
16:28:40 Join atsea-34 [0] (i=atsea-@gateway/tor/x-ee5f4b6122e93a3f)
16:29:17linuxstbWho are Nick, Melba and Rob?
16:29:22*linuxstb is bad with names...
16:30:04markunEvilNick, LambdaCalculus379's gf, LambdaCalculus379
16:30:24linuxstbOK ;)
16:31:02linuxstbAre you planning on writing a report from the SoC conference, or did I miss it?
16:31:03markunOrigins LTR: UK, Italo american, philipino american, cuban american, german dutchman
16:31:25markunlinuxstb: I think it's in the forums. I'll copy paste it in the wiki
16:31:37markunalso, I want to make Tower of Rockbox wiki
16:31:47markunwe made 2 while I was in the US
16:33:01peturmarkun: nice choice of beer ;)
16:33:12markunstella and chimay
16:33:26scorche|wmarkun: in the forums?
16:34:06markunlinuxstb, scorche|w: oops, the SoC doesn't have a report yet
16:34:12scorche|wlinuxstb: i was planning on writing a wiki page about it and what we could do better for next year in the next few days
16:34:29markunalthough petur posted something to the mentor mailing list (right?)
16:34:43peturthere is a report in the wiki
16:34:59scorche|win our wiki?
16:35:03 Quit daurnimator (Connection timed out)
16:35:11 Join PaulJam [0] (
16:36:01scorche|woh...i was referring to a report on the GSoC mentor summit
16:36:12scorche|wand i think the others were as well
16:36:16linuxstbI was.
16:36:26*petur walks to the corner
16:36:31*markun joins
16:36:35*petur bumps GodEater away
16:36:41*markun is still too tired. Slept till 14:30 today
16:36:45linuxstbI'm mainly interested in the things we should do better/different next time.
16:36:52scorche|wmonday mornings....
16:37:02markunscorche|w: +jetlag
16:37:08*linuxstb feels lonely in the middle of the room and joins everyone else in the corner
16:37:10scorche|wwell, since markun and i are here, i suppose we could have an irc chat about it
16:37:33 Join fed [0] (i=63e82710@gateway/web/cgi-irc/
16:37:35*Tanuva sits in the opposite corner
16:37:42linuxstbscorche|w: So what things should we do better/different next time? ;)
16:37:47markunsome of the tips were related to strictness towards our students
16:38:08markunif they are not making progress after 2 weeks, just tell them it was nice working with them and that they failed
16:38:12scorche|wwell, if we are going to talk about it, i think we should move in chronological order
16:38:27markunscorche|w: ok, I'll leave you to it :)
16:38:27linuxstbmarkun: But then what happens - we lose a project?
16:38:32markunlinuxstb: yes
16:38:59scorche|wlinuxstb: they also mentioned that if a student isnt up to snuff in the first few weeks, they will try and replace it for another
16:39:01fedI am trying to get the simulator compile, but I keep on getting an error: /usr/bin/ld: warning multiple definitions of symbol
16:39:02markunbut they found out from experience that most of the people which don't do anything in the beginning don't do anything the rest of the time as well
16:39:23markunscorche|w: we should also have some kind of selection for our students
16:39:30scorche|wthe general theme is that they should follow the typical phrase "commit early, commit often"
16:39:35markungive them a simple task so we know they can do something
16:39:39fedI get this for many symbols. Any idea what I did wrong?
16:39:42linuxstbfed: 1) Did you apply any patches? 2) Try deleting your build directory, and building again from scratch.
16:40:13scorche|wone group had an online php test for their prospective students...many people quite liked that
16:40:18linuxstbmarkun: Basically what ffmpeg did? Speaking of which, do you have any idea how ffmpeg's SoC went?
16:40:28 Join lazka [0] (
16:40:37fedI did apply patches. I did try to rebuild and no luck. I try again without the patches.
16:40:58scorche|wi think at the very least, have a dev environment set up...other tests we cna figure out later if we wish for more
16:40:58markunlinuxstb: yes, what they did. We were thinking about having them modify a hello world plugin, compile it for a target and send us the .rock file
16:41:38markunI think all of their projects were a success
16:41:42scorche|was well, one thing that was highly beneficial was, one group emphasized having an intake interview with all students...either on IRC or skype
16:41:45 Quit cendres (Read error: 104 (Connection reset by peer))
16:41:45 Join ashes [0] (n=ashes@2001:5c0:8fff:ffff:0:0:0:4d)
16:42:02markunalso, we think it would be good to have more than 1 mentor per student next year
16:42:23scorche|wa get-to-know-the-community thing and it also serves to see if the student tryly knows what they are talking about
16:42:53fedSame errors with the rebuild. I'll now to start from scratch withtout the patches.
16:42:56scorche|wthey are looking at having more than one mentor able to be selected per student int he web app next year, but at least 2 mentors per student was nice as well
16:43:00markunthere were some projects which had a interview with the student using VoIP, but I'm not so sure I like that
16:43:19scorche|wmarkun: i said, i think IRC would be fine
16:44:03 Join toffe82 [0] (
16:44:07markunlinuxstb: they also said that if you are very very excited about a project you shouldn't mentor it
16:44:20scorche|wconflict of interest
16:44:24markunso that would have excluded me from text to speech
16:44:32linuxstbI don't understand that logic...
16:44:44markunI also didn't completely get it
16:44:50linuxstbIsn't open source all about scratching your own itches?
16:45:14scorche|wbecause then you cant objectively mentor the want them to succeed so much, you make too many breaks and give them too many chances when they may not be doing anything
16:45:19markunlinuxstb: yes, but GSoC isn't exactly open source
16:45:26Nico_Plinuxstb: maybe you'd be tempted do do the work yourself instead of letting the student do it ?
16:45:33markunI mean, people get payed and have deadlines
16:45:39linuxstbYes, but we don't...
16:45:41scorche|whowever, i think it would be fine to ignore that if we paired 2 mentors up woth a student
16:45:52markunscorche|w: true
16:46:04scorche|wNico_P: what i said above was their logic
16:46:25markunI think most projects had some problems with people with poor comunication skills. I still don't know what to do with those.
16:46:46scorche|wit is more of a case-by-case basis
16:47:02 Quit JdGordon (Remote closed the connection)
16:47:39scorche|woverall...the big thing was, the whole purpose of GSoC is to get more long-term developers for your open-source project...i think that is a very handy thing to know, especially as we were wondering about their intent
16:47:44markunthere was also a lot of talk about svn vs git at the summit (and even the night before in the hotel)
16:47:51scorche|wwell, at least their stated intent ;)
16:48:08linuxstbmarkun: I think that's simply handled by having some strict rules stated at the start - e.g. to be available on IRC for at least X hours per day Monday-Friday.
16:48:21markunlinuxstb: yes, good idea
16:48:45scorche|wlinuxstb: yes...a thing about "our expectations for the summer of you" at the front was another thing that would be good
16:48:46markunand maybe a fixed time every week
16:49:22markunlinuxstb: overall I think we will be much better prepared next year, don't you think?
16:49:22Nico_Pa weekly status report can be done by mail IMO
16:49:40markunNico_P: of in the wiki
16:49:54linuxstbWould you want to mentor again?
16:50:01GodEater_I think it would be ideal (where possible) to ensure at least one of the mentors for each project is in a similar timezone to the student too
16:50:10GodEater_otherwise keeping up to date on IRC is very hard
16:50:23scorche|wmarkun: i definitely think so, and am planning on watching over it more closely next year
16:50:58scorche|wGodEater_: that may help, but there is always the 2 mentors per student thing for that...there will always be some overlap
16:50:59markunGodEater_: we could also have a #rockbox-soc wih separate logs
16:51:12markunbut maybe not :)
16:51:13scorche|wmarkun: i dont knwo about that
16:51:17GodEater_yeah - perhaps not
16:51:24scorche|wwe want them to be entrenched int he community
16:51:35GodEater_only two from last year are
16:51:44GodEater_and both of them were already here :)
16:51:57 Join Thundercloud [0] (
16:52:05scorche|wafter all, as they said, the main purpose is to get us new devs, so we should act that way :)
16:52:09GodEater_tellingly perhaps - they're the projects which did best
16:52:47GodEater_I think the test idea is a good one
16:52:59GodEater_although I think a hello_world plugin is a little trivial...
16:53:08markunlinuxstb: also, I think it's better to have non rockbox devs as students
16:53:15scorche|wwhat would you suggest?
16:53:32GodEater_scorche|w: off the top of my head I have no idea - it bears thinking about
16:53:52markunGodEater_: it's just to make sure they have a dev environment setup and can start working on their code
16:54:00scorche|wi think that would be a decent test
16:54:06scorche|wwe dont want to go overboard
16:54:15GodEater_ah well
16:54:22 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
16:54:31markunalso, if they are not selected by us they shouldn't have wasted too much time
16:55:03scorche|wthat test demonstrates that they can compile, can write code, etc
16:55:17GodEater_although the howto of the hello_world.rock ought to be pulled from the wiki in that case ;)
16:55:25GodEater_too easy to cut'n'paste otherwise
16:55:40scorche|wGodEater_: a modification of it to do something would work
16:55:50scorche|wnot necessarily just writing hello world
16:55:58GodEater_10 Print "markun is cool" ;
16:56:00GodEater_20 goto 10
16:56:53markunperhaps we could just give them the hello world plugin as an example and ask them to use their imagination to make a nice plugin
16:56:59Tanuvayou would have to take the helloworld example offline for the time the studend should build that proof of dev-environment :D
16:57:15GodEater_isn't that what I just said ?
16:57:15scorche|wi dont see why we would..
16:57:37markunme neither
16:58:01TanuvaGodEater_: sure, I just continued that... argh, whats the word in english.. "the path of thinking"...
16:58:02linuxstbBeing able to find things in the wiki is a test in itself.
16:58:04markunif I would make a plugin I would also look at the code of existing ones
16:58:07pixelmawouldn't help much... with the google cache (if they are clever enough)
16:58:43GodEater_linuxstb: true enough :)
16:58:51linuxstbTanuva: "train of thought" ?
16:59:11scorche|wthis isnt about making hello world...this is just about seeing if they can modify code and compile....they can make the text green, or make it scroll around...anything, really
17:00:16Tanuvalinuxstb: yes, sounds good :)
17:01:00fedI get the same errors even now that I have done a new svn download into a new directory. This time, it didn't stop at the errors, it is still continuing to compile.
17:01:18linuxstbIIRC, ffmpeg had a few "qualification projects" for their SoC. For example, for someone developing a new codec, we could suggest they implement a codec that plays back raw PCM files (with a hard-coded samplerate, channels etc).
17:01:36GodEater_now that's a proper test
17:01:52*GodEater_ would need to do some serious googling to pass it
17:01:54linuxstbFor someone starting a new port, maybe get the simulator compiling...
17:01:58markunlinuxstb: we could start a wiki for next year and port our ideas there
17:02:17Tanuvalinuxstb: which is _really_ easy
17:02:30linuxstbTanuva: I mean the sim compiling for their new target.
17:02:39Tanuvaah. okay. :D
17:03:48linuxstbmarkun: I think there already is an SoC 2008 page.
17:03:58 Join Crash91 [0] (n=evil91@
17:04:41scorche|wi dont think that the test should be that hard, but you never know...i think the simple test (which would have helped for one student this year..) and an IRC interview to see if they know what they are talking about/will mesh well with people would be decent before accepting them...i could see a more strenuous test after they are accepted and during the introductory period where we can still ask for a replacement
17:05:49fedit looks like the errors I noted previously weren't responsible for the compiling to stop. The last error was /usr/bin/ld: Undefined symbols: _lcd_enable
17:06:09fedAny I dea how I can fix this?
17:07:05fedI suppose the duplicate multiple definitions isn't a problem. Is it?
17:07:32n1sfed: are you building a clean svn source now?
17:07:39Rincewindfed: what platform are you on?
17:07:48 Join mf0102 [0] (n=michi@
17:07:52fedmac osx
17:08:14RincewindI don't know anything about that :(
17:08:22fedThe clean svn gets the same multiple definitions, but not the missing one.
17:08:39markuncan anyone reproduce this? (I can't)
17:08:53bluebrotherfed: have you ran make clean first?
17:09:02fedI did.
17:11:07***Saving seen data "./dancer.seen"
17:11:11n1smarkun: that sounds really weird
17:11:40bluebrothermaybe there are old files left from a patch you applied earlier?
17:11:54fedWhat. The multiple defs or the missing one?
17:12:06bluebrothermultiple definitions.
17:12:19fed(what is wierd, that is)
17:12:19bluebrotheryou could remove the checkout tree and do a fresh svn co
17:12:50fedWhat is the checkout tree?
17:12:53|Rain|n1s: by the time I got a chance to try your midi player patch, it no longer applied cleanly
17:13:02bluebrotherhow did you obtain the sources? From svn?
17:13:14fedYes. SVN
17:13:35n1s|Rain|: it seems like stevenm comitted it to svn so you could test with an up to date svn source
17:14:05fedI tried to use fink to install the sdl. It didn't work, so I installed it manually.
17:14:11|Rain|n1s: wait −− are you sure? only one hunk failed
17:14:12bluebrotherwell, remove the whole tree and fetch a new one from svn. Maybe there are old files left −− svn revert can't remove added files
17:14:42n1s|Rain|: he did it today
17:14:53|Rain|aha, okay
17:15:01|Rain|did you get an e200 benchmark yet?
17:15:04fedwhat do you mean by the whole tree?
17:15:13n1snope, no pp benchmarks at all
17:15:25n1sthat's why I held off on committing myself
17:15:42|Rain|looking for the right revision so I can get a good before/after
17:15:45bluebrotherthe whole sources you checked out from svn
17:15:55 Quit atsea-34 (Remote closed the connection)
17:17:31linuxstbmarkun: Any update on your SoC student? Is that project officially dead now?
17:21:15fedWhen I did the clean svn, I still got the multiple def error, but it compiled fine, and the simulator worked. I then applied my patch of all the extras I was using, and the same error that stopped the compiling occured.
17:21:44fedIs there a problem with the multiple def errors if it still compiles?
17:23:08markunlinuxstb: I guess it's dead. I didn't even ask about it anymore.
17:23:50linuxstbmarkun: Oh well... And we still haven't solved the GPLv3 issue?
17:24:34 Quit Rincewind ("Verlassend")
17:24:58markunkarl suggested using a LGPL api file to link the plugins against, but at least amiconn was against it
17:25:57linuxstbBut wouldn't that open the door to closed-source plugins?
17:26:17amiconnYeah, that's what I'm afraid of
17:26:33 Quit Crash91 ("Bye Bye!")
17:28:14fedI removed lcd_enable from my patch, and now I still get the multiple def errors, but it compiles fine.
17:28:34fedI guess there is a problem with lcd_enable in the sim.
17:28:48fedShould I make a bug reprt about this?
17:29:00markunlinuxstb, amiconn: but what's preventing anyone from adding a LGPL api file now and adding some closed source plugings now anyway?
17:29:17n1sfed: since the problem is caused by a patch you should report it to the patch authort
17:30:09markunI personally don't care much about people using rockbox together with closed source plugins/codecs, but I'm a BSD person anyway.
17:31:15linuxstbmarkun: I'm the opposite... A GPLv3 person...
17:31:20fedOne last thing, whould I be trying to fix this multiple def thing, or not worry about it since it seems to work in the end?
17:32:06 Nick fxb is now known as fxb__ (
17:33:51bluebrotherfed: if there is a problem with svn code the build table would show it.
17:34:13markunbluebrother: not in the sim builds
17:34:15linuxstbbluebrother: Maybe not if it's an OS X specific Sim issue...
17:34:31bluebrotherhmm, ok. Missed the os x part ;-)
17:35:56*linuxstb goes home
17:35:59 Quit linuxstb ("Client Exiting")
17:36:53fedwhere is the build table?
17:41:16fedI cannot seem to get onto that link.
17:41:31peturworks fine here
17:41:36 Quit kugel (Read error: 104 (Connection reset by peer))
17:42:33 Quit Thundercloud (Remote closed the connection)
17:44:08 Join lee-qid [0] (
17:45:45fedWell, thanks. I'll work on this for a bit more.
17:47:08 Quit TMM ("Ex-Chat")
17:49:05 Quit fed ("CGI:IRC (EOF)")
17:49:38preglowcould any people knowledgable about coldfire dma and audio chime in here:
17:49:46 Join Davide-NYC [0] (
17:49:52 Quit Davide-NYC (Client Quit)
17:53:35 Join atsea-34 [0] (i=atsea-@gateway/tor/x-eed9c07106e9dcd4)
17:57:01 Join TMM [0] (
17:57:42amiconnUrgh, the fm radio driver has a similar problem...
18:04:57 Quit stewball`ghost (Read error: 104 (Connection reset by peer))
18:09:08 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:10:57 Quit petur ("work->home")
18:17:54 Join haemmy [0] (n=stefan@
18:19:43 Quit obo ("bye")
18:21:05 Join Thundercloud [0] (
18:22:02 Join bertrik [0] (
18:22:11 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
18:26:27amiconnWee, fixed. Now radio control works reliably even at 124MHz on H300 :)
18:26:39amiconnThat is, with buffered writes of course
18:27:12n1sHmm, it seems all the arm builds that are slightly bigger than the rest comes from which uses gcc 4.0.4 and ld 2.17
18:27:45n1sand the bigger sh builds seems to come from but that uses the same gcc and ld as the others
18:28:09n1sah, you already knew that?
18:28:24rasherHow clever is the buildserver? could it just be set to not send those types of builds to those specific hosts?
18:28:37amiconnThe arm "problem" is obvious. deepthought needs to be brought to the official recommended versions
18:28:49amiconnHowever, the SH "problem" is strange...
18:29:09 Join pondlif1 [0] (
18:29:10n1samiconn: why do we recommend 4.0.3 and not .4 ?
18:29:19 Part pondlif1 ("disconnected has pondlife")
18:29:36amiconnBecause that's what was available when switching to 4.0.x
18:29:45n1srasher: I'm pretty sure it's possible
18:30:06amiconnI guess 4.0.4 isn't well tested for arm rockbox. 4.1.x is known to cause problems
18:30:55n1samiconn: it seems like all the servers will need to rebuild arm-efl-gcc to properly support mrobe and gigabeat s builds so it might be an opportunity to update if that's wanted
18:31:21bertrikI am running a build compiled under cygwin with arm-elf-gcc 4.1.1 (from for sansa e200 and I haven't seen problems yet
18:31:23 Quit jhulst ("Konversation terminated!")
18:31:33 Join linuxstb [0] (
18:32:10bertrikonly did about 10 builds until now though
18:32:38amiconnI don't remember if there were stability problems, or if it was just speed
18:32:46amiconnI didn't perform the tests myself
18:33:09amiconnContrary to what the gcc team promises, newer gcc versions do not always produce faster code
18:33:28n1samiconn very true
18:33:42n1sso who owns the deepthought server?
18:33:43bertrikok, haven't done any benchmarks
18:34:01n1sbertrik: that would actually be quite interesting to see
18:34:16 Join WalterEgo [0] (
18:34:30bertrikplease tell me how and i'll run one
18:34:35 Join styleism [0] (
18:34:40|Rain|I had minor stability problems with rockbox compiled with gcc 4.1.x, and I couldn't compile a working bootloader at all due to gcc bugs messing up the FAT code
18:34:41pixelman1s: it's ender`s and I stated the same findings here in the channel maybe 2...3 weeks ago
18:35:00ender`what's the problem?
18:35:38n1spixelma: ah, ok, maybe we should just ask Bagder to not hand out arm builds to it then...
18:36:07WalterEgoHmmkay.. While I know this is a 'patch' question and thus not supported, I still believe there might be someone on who knows the answer, so.. : Does the album_art patch scans the folder for a .bmp file everytime a new song is played, or everytime the WPS requires it to?
18:36:47ender`n1s: what's the problem with my server?
18:37:04pixelmaender`: your build server is running a different arm-elf-gcc (not recommended version) producing slightly bigger binaries
18:37:18ender`hmm, let me check
18:37:23n1sender`: you also use a different ld version
18:37:51ender`which versions should be used?
18:38:12n1s4.0.3 and 2.16.1 are recommended
18:38:37ender`ok, i'll install that. are any patches needed?
18:38:59n1sthe thumb-interwork patch for gcc
18:40:10 Quit pondlife (Read error: 110 (Connection timed out))
18:40:46amiconnH1x0 radio works too, as does X5 radio :)
18:41:09 Join petur [0] (n=petur@rockbox/developer/petur)
18:41:10n1sis the sever Slasheri's ?
18:41:23n1sSlasheri: ping?
18:42:07 Quit PaulJam (".")
18:46:51ender`ok, binutils installed, gcc is compiling
18:47:23 Join ilgufo [0] (
18:49:28*amiconn still wonders how this i2c could have worked...
18:49:41 Join pondlif1 [0] (
18:49:44 Part pondlif1 ("disconnected has pondlife")
18:50:00n1samiconn: what was the problem with it?
18:50:15amiconnWrong timing in several places
18:50:31amiconn(PCF driver for H300, X5 and M5, and FM radio driver for H1x0 and H300)
18:51:06amiconnThe buggy pcf driver might have been the reason for the occasional button glitches, as well as the occasional eeprom corruption on H300
18:51:50n1sbutton glitch == strange lockups when pressing buttons with backlight off and high load on cpu?
18:52:13ender`ok, arm-elf-gcc 4.0.3 installed, too
18:52:21n1sender`: great
18:52:29amiconnbutton glitch == wrong button event triggered
18:52:46n1snever happened to me
18:53:12amiconnIt depends on the individual units
18:53:15 Join homielowe [0] (
18:53:21amiconnSome are more susceptible than others
18:54:17 Quit homielowe (Client Quit)
18:57:10 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
19:03:51 Quit karashata ("Leaving.")
19:04:00 Join PaulJam [0] (
19:05:50 Nick tictoc is now known as imred (i=tabac@gateway/gpg-tor/key-0xB9002659)
19:05:59 Part imred ("Leaving")
19:06:02PaulJamamiconn: i think the button glitches are ( at least on H300) a hardware issue. see also here:
19:06:24amiconnYes, partially, but most probably not only
19:07:49amiconnpixelma tried my fix on her M5 (which used glitch a lot) before I committed it, and it improved things
19:09:31|Rain|n1s: 72 misses (r15111) -> 69 misses (r15112)
19:10:36n1s|Rain|: thanks, that is a much smaller improvement than for coldfire but at least it didn't get worse
19:10:53pixelmaPaulJam: I think especially with the X5/M5 joysticks people reported this effect as being stronger than in the original firmware (can't compare personally as I didn't use the OF long enough and there is no dual boot and not much motivation to put the OF back just for comparing this)
19:11:07 Join desowin [0] (
19:11:08***Saving seen data "./dancer.seen"
19:11:33pixelmabut this fix really helped :)
19:11:52|Rain|n1s: it's possible the pitchbending changes decreased performance as well, my test midi has a lot of them
19:12:33n1s|Rain|: ah, I only got a very slight decrease in speed on mine but maybe they don't use it too much
19:13:19 Join Bagder_ [0] (
19:17:36 Quit Nico_P (Remote closed the connection)
19:18:13 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
19:18:18 Join lazka_ [0] (
19:20:08 Join Limer [0] (i=5ae5c0e4@gateway/web/cgi-irc/
19:20:09 Quit lazka (Read error: 113 (No route to host))
19:22:44 Join _Bagder_ [0] (
19:23:08 Nick fxb__ is now known as fxb (
19:23:58 Nick Limer is now known as Limerr (i=5ae5c0e4@gateway/web/cgi-irc/
19:24:36LimerrAnyone up to date on which player is the best for rockbox today if i want good support and lots of storage?
19:26:14n1sdepends on what you want to do with it
19:26:46 Join obo [0] (n=obo@rockbox/developer/obo)
19:26:59LimerrI want to listen to music, and need much space(60GB+)
19:28:54LimerrI have used rockbox for a couple of years on my ipod photo now, but its batterytime is getting really low now :(
19:29:28n1sthere are replacement batteries available, you know.
19:29:33linuxstbThe only 80GB device Rockbox supports is the old ipod video (NOT the new classic). If 60GB is enough, then you'll probably be happy with any of them.
19:30:06 Quit Bagder (Read error: 110 (Connection timed out))
19:30:27n1siaudio x5, gigabeat f/x are available in 60 GB
19:31:19Limerri would also like the player to be quite new, cause it is really hard to find those old players :)
19:31:51Limerrthe iaudio one seems quite ice
19:33:25 Quit PaulJam (".")
19:34:17 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
19:35:20Limerrunfortunetly i cant find any iaudio x5, they do not appear to be produced anymore :(
19:35:21 Quit DataGhost (Read error: 110 (Connection timed out))
19:37:16GodEaterapart from the sansa e200 (and possibly c200) range, no rockbox targets are in production any more AFAIK
19:39:27Limerrthat sucks
19:39:36GodEaterit's the way of the world sadly
19:40:10 Quit Bagder_ (Read error: 110 (Connection timed out))
19:40:32LimerrIs there any other opensource project with support for newer players?
19:41:26 Join DataGhost [0] (
19:41:55LimerrOr are all manufactors encrypting their firmwares?
19:42:17GodEaterthe only other firmware(s) we know of (OSS wise) are iPL, and archopen
19:42:28GodEaterneither of which are supporting "current" models
19:43:55LimerrThat is quite sad tbh :(
19:44:07GodEaterblame the manufacturers
19:44:13GodEaterthey don't want people running homebrew
19:44:21GodEaterso they make it harder and harder to do so
19:44:49n1splayermodels are only available new for about a year and porting to a new platform often takes at least that long
19:44:55LimerrWhats up next, do we have to chip our mp3 players like we did with our xboxes?
19:45:54GodEatereven that doesn't help these days Limerr - look at the XBox360
19:45:55n1smaybe an open source friendly dap manufacturer turns up one day
19:46:14scorche|wwell, those can be "chipped" because they are very similar to, say, a regular computer...on an embedded device such as a DAP, it is a bit different..
19:46:17 Nick parafin|away is now known as parafin (
19:46:21n1sneuros has taken the first step but they didn't get all the way
19:46:51rasherGodEater: and openneo
19:47:06rasher(a Rockbox fork)
19:47:17GodEaterrasher: isn't that long since dead ?
19:47:43rasherGodEater: Not really.. last update in May
19:47:58rasherEh, February.. damn US dates
19:48:11GodEaterthat's pretty dead if you ask me :)
19:48:26rasherI think their users would disagree
19:48:50GodEaterlots of people disagree with me sadly
19:49:57LimerrIt looks like i wont be able to find any swedish retailers having the iaudio, any idea wich model to try next? :)
19:50:32GodEaterI'd go with the ipod 5.5G if I were you
19:50:39GodEaterI think it'll be the easiest to find
19:51:14LearI've seen iPod videos available recently, at least on the web.
19:51:29Limerrhow do i tell the differens between the 5.5 and the 6?
19:52:02GodEaterthe 6 is made from metal all over I believe
19:52:08LearVideo vs Classic should be safe.
19:52:14GodEaterthe 5.5G is plastic on front, and metal on the rear
19:52:15rasherLimerr: also, the 6G is called "Ipod Classic"
19:52:24rasherThe 5.5 is "Ipod Video"
19:52:38Limerrso this should be safe
19:53:06GodEaterLimerr: yes
19:53:09GodEaterthat's a 5.5G
19:53:44Limerrgreat :)
19:54:13Limerrhopefully it has better batterytime than the photo :D
19:54:56*GodEater couldn't comment, not owning a photo
19:55:12GodEaterbut I find the battery life on my 5.5G more than adequate
19:55:50Limerrthe batterytime is around 40 minutes while in iPL on my photo
19:56:14Limerrbut in rockbox it is better not sure how much better though
19:56:40GodEaterI get 10 hours on a typical day with mine...
19:56:41Limerrid bet on around 3 hours in rockbox
19:59:55Limerrrockbox estimates 4 hours on my fully loaded photo
20:00:01 Join Arathis [0] (
20:00:42scorche|wthe estimate is not accurate
20:02:26Limerrsince i am using r12 i guess the estimate isnt any good in any way :P
20:02:38 Quit desowin ("use linux")
20:03:17*GodEater thinks r12 didn't run on ipods
20:03:42*bluebrother wonders if r12 even run on archos
20:03:47GodEaterdoubt it :)
20:03:48 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:03:56Avewhat a massive rollback it would be if nano build uses 75 MHz, whole 2 lines against head svn
20:03:58GodEaterbluebrother: helloworld maybe ?
20:04:01Averegarding the fs#7510
20:04:05bluebrotherI wonder why we make up such nice version strings :(
20:05:11GodEaterAve: That isn't a fix though - it's a work around
20:05:35Aveyeah I'm aware
20:05:36Avebut it works
20:05:54Avecurrently some percentage of nano users are stuck with a very borked software
20:05:57bluebrotheroh, r12 is even only a cvs keyword fix :)
20:06:24GodEaterAve: but if that workaround is left in the code, no-one will have any incentive to do the fix properly
20:06:36Avelast night I built svn head with the said 75 MHz alteration and I'm spiff so far
20:06:42Aveyeah ok thats true
20:07:07AveI would certainly do something if I knew what, this is too deep or some such
20:07:28AveI tried the clock skip stuff but that didnt help, atleast not for me
20:07:31GodEatermail your nano to a dev
20:07:52Avethere was atleast one parson on the tracker offering his already
20:08:04Avesuppose I could if he bails out
20:08:21GodEaterwho was it ?
20:08:26GodEaterand did anyone accept the offer ?
20:08:43 Join fm2 [0] (n=chatzill@
20:08:50fm2Hello. Which daps are affected by the FM driver fix? All irivers Hxxx?
20:09:36bluebrotherthe h100 also uses fmradio_i2c.c, so yes.
20:10:08*bluebrother wonders what "iriver radio" means :o
20:10:20Ave[comment] Comment by Florin Popescu (florinp3) - Thursday, 27 September 2007, 11:24 GMT+2
20:10:28AveI live in the EU and I'm willing to lend my iPod if transport fees can be arranged. I tried contacting amiconn and received no answer, I'll try the IRC tonight.
20:11:17Avethats about it
20:11:35GodEateramiconn very rarely reads the forums if that's how he tried to contact him
20:11:52GodEaterI don't recall seeing florin come in here either
20:12:17scorche|wGodEater: tracker ;)
20:12:31GodEaterscorche|w: does he read that? :)
20:12:37 Part fm2
20:13:08*scorche|w shrugs
20:13:44 Join Domonoky [0] (
20:13:44 Quit Limerr ("CGI:IRC (EOF)")
20:14:23Aveyeah my lastlog shows nothing
20:14:34Aveand I've been here quite some time
20:14:47 Join [Tm]Limer [0] (
20:14:59 Nick [Tm]Limer is now known as Limerr (
20:15:15Limerrfinally a real irc client
20:16:20GodEateryou mean you don't like telnet for IRC ? :)
20:16:47Limerri used the webclient before :P
20:17:16Limerri use mirc now, not sure if id call it a real client but anyway ;)
20:17:32scorche|wi wouldnt...
20:17:57Limerri like xchat, but since it isnt free for windows i use mirc
20:18:13GodEaterah - you've missed the freebie windows port...
20:18:14bluebrotherthere are free binaries for windows around
20:18:17scorche|wsearch for "silverex"
20:18:23scorche|wbut this is getting offtopic...
20:18:35 Join Rincewind [0] (
20:18:45Limerrfree binaries, woho!
20:19:21 Join BigBambi_ [0] (n=Alex@rockbox/staff/BigBambi)
20:21:23Limerroh shit, remember i told you i had full battery on my ipod photo?
20:21:32Limerrwell, it is out now :P
20:21:45Limerrand i didnt even use it
20:21:50Limerrjust put it on :S
20:22:51 Quit qwm (Remote closed the connection)
20:23:02 Join qwm [0] (
20:24:37 Join kubiixaka [0] (
20:25:05 Quit kubiix (Read error: 104 (Connection reset by peer))
20:25:44 Quit petur ("switching")
20:25:53 Join petur [0] (n=petur@rockbox/developer/petur)
20:26:47 Quit Nico_P (Remote closed the connection)
20:29:07LimerrDoes rockbox support compressed video playback?
20:32:22Limerrwow, real mpeg2. I would still have to scale it down to the the ipods screensize, but i suppose it would be allot faster converting to mpeg2 then the iPL mvpd.
20:34:34LimerrDo i have to limit frames or can the ipods cpu handle the frameskipping by itself?
20:35:01GodEatervideo playback on the ipod video sucks unde rockbox sadly
20:35:09GodEaterthe ipod cpu is very poor at keeping up
20:35:21GodEaterwe recommend the OF for video watching
20:35:25amiconnGodEater: Photo isn't Video
20:35:40GodEateramiconn: he's thinking of purchasing a video to replace his dying photo
20:35:53Limerrwell my photo is quite dead, so i was talking bout the video ;)
20:36:18Limerri think the battery is completely dead now :(
20:36:20amiconnAnd the Photo already has half-decent video playback, and will become even better in a couple of days, hopefully
20:36:22GodEateramiconn: did you see the mention earlier of someone sending you one of these buggy nanos ?
20:37:03amiconnThe G5 will also become better, but not as much as the Color, and starting from a lower level
20:37:42 Quit BigBambi (Read error: 113 (No route to host))
20:39:09 Quit Rob2222 (Read error: 104 (Connection reset by peer))
20:40:15LimerrThe video only have 1 year warranty, that is quite crappy.
20:44:08*GodEater 's video is over a year old now and still going strong
20:44:58Limerrhehe, but you never know
20:45:27GodEaterI've also abused it chronically
20:45:32GodEaterit's rock solid
20:46:10 Join merbanan [0] (n=banan@
20:46:11 Join Limer [0] (
20:46:25 Quit Limerr ()
20:46:31 Nick Limer is now known as Limerr (
20:46:42LimerrFinally i got a real client ;)
20:48:53 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
20:48:54GodEateryou said that already
20:49:04GodEaterhey Nico_P
20:49:16Nico_Phi again
20:49:35GodEaterMoB build has behaved ok today
20:49:38GodEaterno random silences
20:49:45GodEaterbut rebuffering has got even slower =/
20:50:11GodEaterswitching from one album to another took nearly 8 seconds
20:50:45Nico_P:/ what commit is it based on ? "Rebuffer after playlist changes" ? and I forgot which target it is
20:52:08GodEaterer - the last one you made on sunday
20:52:55Nico_Pprobably that one then
20:53:14GodEateron the other hand
20:53:22GodEaterI can't get the ipod to crash like pondlife can :)
20:53:33GodEaterand I tried REALLY hard :)
20:53:36Nico_Phehe :)
20:54:10Nico_Pwhat was it slow for ? loading an unbuffered track to play immediately ?
20:54:32GodEaterI was playing one album (all buffered already), and then switched to another
20:54:32 Quit jhulst (No route to host)
20:54:39Nico_Pit's one of the things I'm thinking about ATM
20:55:02GodEaterit needed a codec switch too
20:55:05Nico_PI need to be off for a short while, bbs
20:55:58 Nick BigBambi_ is now known as BigBambi (n=Alex@rockbox/staff/BigBambi)
20:59:51 Join qweru [0] (
21:03:50*amiconn is tempted to revert the mpegplayer menu + resume commit ... :|
21:04:57amiconnWould be better if someone with enough mpegplayer clues would fix it though
21:05:44amiconnThere is one more bug in addition to the already known ones: When loading a video, and immediately quitting again via the (annoying) menu, backlight stays on forever
21:07:40bertrikand boost stays on it seems
21:07:58 Quit kubiixaka ("Miranda IM! Smaller, Faster, Easier.")
21:09:23bertrikload elephants dream, exit mpegplayer, then check CPU frequency in debug menu: F=80MHz, boost_counter=1
21:09:57bertrikat least on my sansa e200, tested after a fresh reboot, current build from
21:10:44bertrikI'll submit a bug report
21:11:10***Saving seen data "./dancer.seen"
21:16:17Nico_PGodEater: I'm back with a question
21:16:38pixelmabertrik: probably valid for all players with mpegplayer (found the bug on M5, has been confirmed on X5, c200 and yours...)
21:16:48pixelmameaning: swcodec
21:17:09Nico_PGodEater: how do you find the first buffering compared to earlier mob and compared to svn ?
21:20:48peturif nobody is fixing that mpegplayer issue, I found it...
21:22:21amiconn1 out of 3, 2 to go... ;)
21:22:38*amiconn is busy checking lcd timings
21:22:48peturboth the backlight and the boost issue are the same
21:22:49amiconnThe mpegplayer menu has at least 3 bugs
21:23:01amiconnAh, so 2 out of 3, 1 to go
21:23:20peturthe funny thing is, the comment above it says not to exit like that beyond that point ;)
21:23:25amiconnmpegplayer loads the video, then displays the menu, and when starting, buffers a second time
21:25:08bertrikalways distrust code/comments that tells you to do something without explaining why
21:25:36peturit clearly explains why, the person implementing the menu just ignored it
21:25:50bertrikah ok
21:28:40 Join Rob2222 [0] (
21:30:04 Quit Limerr (Read error: 104 (Connection reset by peer))
21:32:04bertrikheh, that's quite a quick response, petur!
21:32:43peturmy rockbox time was just starting and I thought it was a nice one to warm up ;)
21:33:10peturbut now for that file system corruption bug :/
21:33:59peturI also didn't fix the issue amiconn mentioned although I saw it happen
21:34:19*amiconn thinks the menu shouldn't be there :/
21:35:21amiconnImho there should only be a resume request if the file was played before. If not, it should start playing right away
21:35:28maxkelleycontinuity would be nice.
21:35:36pixelmaand the introduced "Loading..." splash brings back the problem with using mpegplayer and voice
21:36:27pixelma(at least I saw it on the c200 again)
21:36:30amiconnI guess that's not due to the splash, but due to the voice disabling now being either in the wrong place, or even being reverted by roolku's commit
21:37:05pixelmathat was discussed on friday (IIRC)
21:37:47*petur has no use for mpegplayer anyway
21:38:34amiconnWell, it's only an add-on for a dap, but if it's there, it should work properly
21:40:14 Join Buschel [0] (
21:47:14 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:47:48n1s~10 fps is actually quite fine for watching cartoons
21:48:03 Quit w0rd54 (Read error: 113 (No route to host))
21:49:57amiconnI think we'll soon have 24fps without skipping on all targets but the Video
21:50:12 Join Jelemonde [0] (i=4a0dd5e8@gateway/web/cgi-irc/
21:50:34 Join w0rd54 [0] (
21:50:56n1samiconn: h300 too?
21:51:03amiconnHopefully, yes
21:51:17amiconnI get 75fps in lcd_yuv_blit() with buffered writes
21:51:18n1sthat would be a _major_ improvement
21:51:21Jelemondehuh... I have a question: Is there a way to open .bmp image with Rockbox (or i need to always convert to .jpg)
21:51:41amiconnIn combination with the asm optimised idct, this should be enough...
21:51:59amiconnBuffered writes still have some quirks to be ironed out
21:52:02n1sJelemonde: you can open it with rockpaint but i don't think it supports images with larger res than the display
21:52:22amiconn(like pixel columns vanishing/duplicating on H1x0 and M5 when boosted)
21:52:22n1samiconn: impressive
21:52:30Jelemondei tried but it dont work so i cant open online with rockpaint?
21:52:50 Join Lambuntu [0] (
21:52:51bertrikbuffered writes?
21:53:14bertrikI saw some TODOs in the sansa LCD driver about possibly speeding things up by swapping buffers instead of copying
21:53:55 Quit Jelemonde (Client Quit)
21:54:23 Quit n1s ()
21:54:51amiconnThat's something else... I'm talking about coldfire targets here
21:55:01Buschelamiconn: did you port the asm-routines from sansa/gigabeat for this?
21:55:04amiconn(iriver H1x0, H300, iAudio X5 and M5)
21:55:15amiconnBuschel: Not yet, but I will
21:55:30Buschelamiconn: this will supersede my work then ;o)
21:55:43bertrikOK, i won't mind experimenting a bit with that
21:55:57amiconnAlthough, for Video and Color they won't be usable as-is, as the update order can't be switched to dual-line zig-zag
21:56:23amiconnSo the chroma needs to be buffered for the second line
21:56:33Buschelamiconn: nevertheless, just wrote a comment to #FS 7951 about further potential in the asm-routines. maybe you could also take this into account
21:56:38amiconnPerhaps it can be switched, but we don't know how
21:56:45 Quit ToHellWithGA ("You know you'll miss me a lot.")
21:57:14amiconnOn Color with type 1 lcd, Nano, and the H10s we do know how, and I will port the routines
21:57:20amiconnI already used them for c200
21:58:09Buschelamiconn: the video's LCD-driver has kind of lag as it will need to wait for 14ms before accepting the next screen update
21:58:48amiconnI know
21:59:09amiconnI have a Video here for testing (still)
21:59:26 Part styleism
21:59:42Buscheli searched the web for compatible lcd-controllers 2 evenings... didn't find anything :/
22:00:02amiconnYou won't find the docs for the Broadcom chip
22:00:31amiconnBroadcom is very closed regarding docs for their chips
22:00:39Buschelas PP...
22:01:14amiconnWhat would be also nice to find (and perhaps a bit easier) would be the type 0 controller of the iPod Color
22:01:29scorche|wwell, the broadcom headquarters is a few min away...
22:01:31amiconnUnless they used the PP's internal controller, that is
22:02:11 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
22:03:01Rincewindscorche|w: If you go to the broadcom headquarters, give them a kick from me because of the bcm43 wlan chips ^^
22:03:08amiconnscorche: You need arguments... ;)
22:03:30krazykitamiconn, yes, 9mm of argument ;)
22:03:47linuxstbBagder: I thought you were in China?
22:03:59amiconnRincewind: Even their ethernet controller chips are secret... afaik the Linux driver for their gigabit chips is based on pure RE
22:04:15 Nick Tanuva is now known as Tanuva|Zzz (n=tanuva@
22:04:43 Quit Lear ("ChatZilla [Firefox]")
22:04:44Rincewindamiconn: I know, I spent a few days getting my wlan working in my laptop :(
22:04:51markunlinuxstb: is he there for work?
22:05:05linuxstbI assume so - he went with LinusN
22:05:06scorche|win fact, we (fluor) traded them a few buildings, so i was right across the street from the huge broadcom sign a week or so ago
22:05:08amiconn(the Tigon3 driver - saw that one quite often lately)
22:05:10markunah, found the blog
22:05:13 Join ]RowaN[ [0] (i=a2b0y@
22:05:55preglowamiconn: you know why we can't do more than 16 bits with coldfire and dma for audio?
22:06:17amiconnWhat do you mean?
22:06:28linuxstbmarkun: That doesn't explain much to me...
22:06:29scorche|wbah...linus's blog is all in fun
22:06:34amiconnThe reading side of things, or writing to the i2s buffers?
22:06:50preglowamiconn: i don't really know exactly where, that's why i'm wondering
22:07:16amiconnWell, these are 2 different things
22:07:31amiconnWith DMA, coldfire only allows 16 bit samples. That's unavoidable
22:07:46]RowaN[Guys I saw this commit of sansapatcher that mentioned overwriting the bootloader in the PPBL section of the firmware. Does this now mean that running the latest PP5022.mi4 rockbox bootloader will now rid me of the "Sansa" splash at boot time? Or have I missed something?
22:08:06amiconnMeh, and for reading audio, we use AA
22:14:21 Quit amiconn (Nick collision from services.)
22:14:27 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
22:17:02 Quit nicktastic (Read error: 104 (Connection reset by peer))
22:17:08 Join steiner44 [0] (
22:17:18 Part steiner44
22:18:09 Quit _Bagder_ (Read error: 110 (Connection timed out))
22:19:46 Join Alonea [0] (
22:21:33Aloneahey guys. my dad wants a new player for Christmas and I was thinking of getting him a Sansa. Is the e280 compatible since the page says e200 series for rockbox?
22:21:41GodEaterAlonea: yes
22:21:58DogBoyyea I use one
22:22:14AloneaGodEater: sweet. yeah, he seems to like the looks of it so far.
22:22:40bertrikonly difference in e200 series is memory size AFAIK
22:23:01 Quit Xerion (Read error: 104 (Connection reset by peer))
22:23:07AloneaHow does everyone else like the sansa e280? He doesn't need a monster 20-60 gig player, but needs enough space to put a bunch of music on.
22:23:10 Quit ilgufo ("So Long, and Thanks For All the Fish -")
22:23:24dionoeaHello. Anyone have comments about before i commit it ?
22:23:28 Join Xerion [0] (
22:24:11GodEaterAlonea: me personally, I don't own one. But I've seen one, and I don't like the buttons.
22:24:48Aloneahmm, for people who have them, does anyone have any trouble with the buttons? My dad has huge fingers.
22:24:48krazykitAlonea, i didn't like the buttons at first, but i got used to it
22:25:00DogBoyme either
22:25:12DogBoyI got big fingers but it's manageable
22:25:18DogBoyand I got used to it
22:25:19bertrikAlonea: I've had mine for 4 days, it's my first mp3 player and so far i like it
22:25:22Aloneaheh. I hated the gigabeat cross until I figured it was more touch and not press.
22:25:33bertrikindeed buttons are a bit hard to press because of the scrollwheel
22:26:13DogBoybut you get used to it
22:26:15bertrikthe other thing i don't like is the USB/charge plug. I prefer a simpler more standard USB plug
22:26:18DogBoyI did anyway
22:26:21Aloneahmm. well, he shouldn't have too much trouble. He generally just listens to music and thats about it.
22:26:28amiconnpreglow: DMA Auto-Align
22:26:39Aloneabertrik: oh? no ac power charger?
22:27:23bertrikAlonea: no ac charger indeed, mine came with just the USB charger plug
22:27:33RincewindAlonea: you can get a 3rd party ac-usb charger
22:27:46AloneaRincewind: about how much?
22:28:21RincewindI got one from amazon for 12€ (in germany)
22:28:26preglowamiconn: right, yes
22:28:26markundionoea: didn't look at the code, but sounds ok to me
22:29:03 Quit mf0102 ("dormire maintenant, desolee")
22:29:07pixelmadionoea: if it works as advertised I'd like it, because I found with the current implementation that there is a difference if the track to be moved will be placed before or after the hovered track (new placement) depending on if you move up or down
22:29:17dionoeaIt works as advertised :)
22:29:54dionoea(at least it does when i test with my dummy alphabet playlist)
22:30:52*preglow reads the last line of mrh's dma doc with puzzlement...
22:31:03AloneaRincewind: ok, thanks. I might look one up for him.
22:31:14*dionoea commits it then
22:32:03preglowit's an anagram for reverse engineering!
22:32:06*preglow solved it! \o/
22:32:18jhulstI posted my first bug fix, #FS7831, I'd be interested in any comments or suggestions on it
22:32:49krazykit]RowaN[, it appears to not be the case. the sansa splash still comes up, and i just compiled the bootloader from svn.
22:33:09]RowaN[hmm thanks. mysterious times
22:33:28 Join stevenm [0] (n=stevenm@
22:33:39 Join Xerion_ [0] (
22:34:08stevenmhello, people
22:34:12 Quit Xerion (Read error: 104 (Connection reset by peer))
22:34:13amiconndionoea: Did you test in a player sim (just curious how it feels compared to the old method - I rarely use the playlist viewer, and even more rarely for editing the list)
22:34:53*amiconn is pondering to commit buffered writes for coldfire
22:34:55dionoeaWell too late for testing :) it's commited
22:35:05markunamiconn: what's keeping you from doing so?
22:35:12amiconnSeems to work nicely now, with all sorts of I/O seeing considerable speedups
22:35:32preglowamiconn: ata?
22:35:40amiconnmarkun: I don't wnat to break something or cause instability
22:35:48markunjhMikeS: could the gigabeat benefit from write buffering?
22:35:48Aloneahow odd, you can't get an extended warranty if you get the combo with a free case...
22:36:06amiconnpreglow: Works, verified with my 300MB test file method at 124MHz on H180, H340 and X5
22:36:44preglowamiconn: any noted performance gains?
22:37:06amiconnI didn't compare ata performance, but there should be some
22:37:27amiconnI probably should...
22:37:29*amiconn is curious
22:37:38preglowwell, i think so... :)
22:37:59preglowanyway, if it's not too complictad and has no other side effects, i'd just say commit
22:38:01amiconnLCD profits considerably. lcd_yuv_blit() on H300 goes from 54fps to 75fps
22:38:04preglowbut sure would be nice to have some performnac efigures
22:38:34pixelmawhat I wanted to say ;)
22:38:42amiconnH300 lcd_update() doesn't profit, as that uses DMA. On X5, lcd_update() profits too
22:38:48preglowoh, that reminds me, i should check out bushel's newest lcd_yuv_blit patch
22:39:02amiconnIsn't that just C?
22:39:13pixelmapreglow: Buschel is around at the moment...
22:39:55Buscheljust added some comments on the yuv-sutff
22:40:23amiconnI'll do an ata perf comparison on coldfire, then commit buffered writes
22:40:39preglowhe is indeed!
22:40:45amiconnAfter that, I'll take a look at asm idct for mpegplayer on coldfire
22:41:15amiconnNext will be lcd_yuv_blit() in asm for H10
22:41:27preglowBuschel: did you ever hear anything more from moos?
22:41:43amiconnThen color/video. After that, I'll play with the color lcd bridge.
22:41:48Buschelpreglow: he tried to contact me yesterday. but i missed him
22:42:19amiconnThat second colour should be colour ;)
22:42:59preglowamiconn: be sure to make the asm easily portable to nano :P
22:43:11amiconnpreglow: color and nano share a driver
22:43:49amiconnBut if we can't determine what the type 0 color lcd controller is, it might be useful to split them, and make the Nano driver more similar to H10
22:44:16amiconn(Nano is always type1, i.e. HD66789R, which is known)
22:44:22preglowreally, now
22:44:38 Quit ompaul (Read error: 104 (Connection reset by peer))
22:45:55 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
22:46:50Nico_Pdionoea: nice commit. I wanted to test the patch but you were quicker :)
22:47:36Aloneabertrik: about how long does it take to charge? 2 hours?
22:48:08bertrikAlonea: don't know, haven't allowed it to completely run down or charge fully yet
22:48:10Nico_PGodEater: how do you find the first buffering compared to later ones, earlier mob code, and svn ?
22:48:25GodEaterfirst buffering is good
22:48:30GodEaterbetter than svn I'd say
22:48:39Nico_Pok so it's in the rebuffering somewhere
22:48:41Aloneabertrik: ^__^. My gigabeat only takes a couple hours to charge most of the time. probably about the same.
22:50:12dionoeaWoohoo! mega green delta
22:50:35dionoea(except for the player ... but it's not like anyone cares)
22:51:06Nico_Pnice... better usability with less code :)
22:51:35Buschelpreglow: do you have any news from moos?
22:51:36pixelmamarkun: didn't you swap the gigabeat buttons so that usually "Power" is used to quit plugins (not "A"?
22:51:42amiconndionoea: You forgot to bump the plugin min api version...
22:51:45bertrikAlonea: battery is 750 mAh and max charge current is 500 mA, so yeah something like a few hours sounds reasonable
22:51:50dionoeaamiconn: did that 1 second ago :à)
22:52:20Aloneabertrik: alrighty. methinks daddy is getting one for christmas then.
22:52:51*amiconn wonders what a selected lune is and guesses that the removed code was written by JdGordon ;)
22:53:09markunpixelma: I forgot at least a few
22:53:13 Quit Rincewind (Remote closed the connection)
22:54:40pixelmamarkun: yes, e.g. demystify (uses plugin button actions...) - found that while filling out button tables in the manual(s) and seeing that a lot are empty for the gigabeat too
22:55:18pixelmaI try to solve that too where I see it's missing
22:57:41pixelmamarkun: so, should I already state that it is "Power" in advance - or use "A" as it currently is?
22:57:48 Quit merbanan (Remote closed the connection)
22:57:51WalterEgoNico_P : I know you've moved on to other, more important things, but may I ask you for a quick answer to a very easy question (I guess) about the album_art patch..?
22:58:07preglowBuschel: nope
22:58:14markunpixelma: power and fixing code would ideal :)
22:58:26Nico_PWalterEgo: ask away, but I'm not sure I'll be able to answer
22:58:35WalterEgo..namely, if a WPS doesn't use album_art tags at all, does it still access disk, preload image, and such..?
22:59:06Nico_PWalterEgo: I'm really not sure anymore, but there's a chance
22:59:39Nico_Pbest thing to do is look at the patch and right now I really don't have time
22:59:54WalterEgohmmmkay.. Thanks a lot anyway.
23:00:20WalterEgoI'm just looking at saving battery energy wherever I can. :p
23:00:23Nico_PWalterEgo: you're welcome ;) hopefully the AA patch doesn't have long left to live anyway
23:00:40Nico_PWalterEgo: to save power don't use AA ;)
23:00:43amiconnpreglow: Around 14% speedup for ata writes. Reads are unaffected
23:01:33pixelmamarkun: I have no target to test on :P
23:01:53amiconn(measured on H340)
23:01:57WalterEgoNico - That's understood. I guess I'll have to compile my own build to get most wps goodies *without* AA.. Off to reinstall that virtual machine thing..
23:02:06WalterEgoNico - Merci. :)
23:02:43markunpixelma: simulator? :)
23:02:52Nico_PWalterEgo: de rien :)
23:03:17bertrikwould be nice if there was a current sense resistor somewhere to help predict current power usage and remaining battery life
23:03:56pixelmamarkun: and I can't do too much at once. Also I don't want to do much with plugin button actions (and still wait for a discussion/solution/decision about them)...
23:04:24 Join EnterUserName [0] (n=dave@pdpc/supporter/student/GeekZoid)
23:05:13markunpixelma: I'll try to fix it later then
23:05:15amiconndionoea: A better solution would be to nor start at 0 with the rectangles. This makes no sense anyway
23:05:35amiconnI have that locally, but didn't have enough time to check it on my various targets :/
23:05:44dionoeaarf :/
23:05:50dionoeawell i'll let you commit the correct fix then
23:06:03amiconnI mean not start at 0 with the loop
23:06:08Buschelcan anyone with a sansa test my suggestions about yuv-blit speed-up for the arm-based code (#FS 7951)?
23:06:09amiconnSaves the 2 if()'s
23:06:34amiconnBuschel: The sansa arm lcd_yuv_blit() is jhMikeS's work
23:06:38dionoeai'm sure that the time spent in the if()s can be neglected compared to lcd updates :)
23:06:41amiconnjhMikeS: ping...
23:06:41Buschelahem, asm-based
23:06:48amiconndionoea: true
23:07:26jhMikeSamiconn: gnip
23:08:10amiconnicmp redirect to Buschel @ fs #7951
23:08:58pixelmajhMikeS: didn't your X5 also had the famous joystick problem?
23:09:00jhMikeSmarkun: perhaps if it has that sort of write buffering.
23:09:24 Quit haemmy ()
23:09:33jhMikeSpixelma: yes. that's the first problem I had.
23:09:59pixelmaamiconn's fix today really made things better
23:10:03amiconnTry the latest svn... it should significantly reduce the glitching, if not fix it completely
23:10:26*jhMikeS thought pixelma was talking about the fact that they break.
23:10:57pixelmaI didn't know it was broken...
23:11:12***Saving seen data "./dancer.seen"
23:11:39jhMikeSit's not now.
23:13:02jhMikeSamiconn: my glitching was never very bad actually but I'd get an occasional hop up when pushing down
23:13:26 Quit Alonea ("ChatZilla [Firefox]")
23:13:53 Join [IDC]Dragon [0] (i=0c182f0a@gateway/web/cgi-irc/
23:14:00*jhMikeS tests an SVN build
23:14:19amiconnjhMikeS: If you check the logs, you'll see that I wondered (and still wonder) how these bit-banged i2c drivers could have worked... with several wrong timings...
23:14:26bertrikBuschel: can i help testing? i have a sansa e200
23:14:40amiconnBut now I'm about to commit buffered writes :D
23:14:52pixelmajhMikeS: on my M5 "right" was often interpreted as "select", making bubbles very hard...
23:14:54jhMikeSamiconn: should I wait?
23:15:16[IDC]Dragonhello rockbox world!
23:15:19jhMikeSpixelma: that would be very annoying. mine was a bit liveable at least.
23:15:48*[IDC]Dragon went over to the dark side
23:15:53 Join kkurbjun [0] (n=kkurbjun@alamode.Mines.EDU)
23:16:02[IDC]Dragon...and got an iPhone
23:16:09kkurbjunamiconn: are you around?
23:16:24amiconnkkurbjun: Iirc you have a Mini G1?
23:16:42 Part Domonoky
23:17:00kkurbjunamiconn: yes I do
23:17:14kkurbjunwhat were you looking for on it?
23:17:17scorche|w[IDC]Dragon: first you neglect to tell us you were in silicon valley, now you are getting an iphone?....what next?!?
23:17:25amiconnjhMikeS: can explain better, I think :)
23:17:59[IDC]Dragonscorche: calm down, next I buy a Sansa
23:18:00jhMikeSkkurbjun: one more player with a PP502x has to be converted to packed samples ... and that's what you have.
23:18:31kkurbjunOk, so you just want me to change that set of defines? What's the benefit of packed samples?
23:18:47jhMikeS1/2 the interrupts. immunity from channel swapping.
23:18:47kkurbjunand what am I looking for when I change the define?
23:19:16Nico_Pdionoea: seen FS #7967 ?
23:19:22jhMikeSin pcm-pp.c and i2s-pp.c add your player to those lists. hopefully FIFO_FORMAT_LE16_2 will work for it too.
23:20:02Buschelbertrik: in fs #7951 i mentioned a possible speed-up of yuv-blit for sansa. maybe you can change and test it?
23:20:13jhMikeSI suppose it should play fine and record fine if it has that. One good test of each should be enough.
23:20:30kkurbjunI mean, what behavior on the player am I checking for? Is it a situation where itwill or it won't work, or is there some stability testing that needs to be done?
23:20:31amiconnOnly play... Mini has no recording
23:20:51bertrikBuschel: ok, you mean the change in lcd-as-c200.S? Is that asm also used for the e200?
23:20:52jhMikeSkkurbjun: yes, it just plain will or won't work.
23:21:26kkurbjungreat, I'll do that when I get a chance, I'm not by my player at the moment - possibly later today.
23:21:26bertrikI don't know how to benchmark it, but I can tell you if it works or not
23:21:55amiconnBuschel: c200 *very* probably won't see any further speedup in lcd_yuv_blit()
23:22:01bluebrother[IDC]Dragon: what are you doing in silicon valley? Or was that only for getting an iphone? ;-)
23:22:13jhMikeSkkurbjun: ok, thanks. Then all that can be collapsed down to #ifdef PP502x
23:22:21 Quit petur ("reboot: ubuntu stopped using my mouse :( *kicks linux*")
23:22:23amiconnThe calculation is already significantly faster than the wait for the bridge to become ready
23:23:00amiconnNote that this applies to the c200 *only*, not any other arm target
23:23:30amiconns/arm/colour pp/
23:23:34Buschelbertrik: as far as i understood, yes. maybe i should just post a untested patch there?
23:25:09dionoeaNico_P: nope. Should i check if it still happens?
23:25:19[IDC]Dragonbluebrother: I'm working, our headquarters are in Sunnyvale
23:25:29bertrikBuschel: I think i can fix it myself without a patch. Amiconn: do you think it's useful to test it?
23:25:30dionoeaNico_P: I think that the list display refresh i added might have fixed it
23:25:46dionoea(Else the displayed list was always one keypress late)
23:25:50[IDC]Dragonthat company bought us in the beginning of 2007
23:25:56Nico_Pdionoea: well maybe tell that to the reporter and ask him to update
23:26:05 Quit stevenm ("Connection reset by beer")
23:26:09bluebrotherah, I knew I read something ...
23:26:20[IDC]Dragonyou did?
23:26:20bluebrotherso the iphone is not the reason ;)
23:26:44[IDC]DragonI was never up to one, actually
23:27:09[IDC]Dragonbut I learned it has WLAN and is good for a couch browser
23:27:31pixelmaBuschel: just pointing out that amiconn is talking about the c200 Sansa and maybe you mean the e200?
23:27:32[IDC]Dragon(I almost got an iPod Touch instead)
23:27:36dionoeaNico_P: done.
23:27:51bluebrotherso ... when do you start porting Rockbox to it? ;-)
23:28:14[IDC]Dragonwhen that guy comes back with the datasheets ;-)
23:28:28Nico_Phasn't an exploit been found to run homebrew on the iPhone and iPod Touch ?
23:29:20[IDC]Dragonit would be nice to go without iTunes
23:29:31[IDC]Dragonthis software sucks
23:29:41Nico_Pis there any chance of it being remotely applicable to the nano 2G and later ?
23:29:49Nico_P(it == the exploit)
23:30:06[IDC]DragonI have everything neatly sorted in folders (even with covers), but it ignores it all
23:30:12 Join ddalton [0] (
23:30:13krazykitnot from what i can tell, nico_p
23:30:26 Part ddalton
23:30:39bertrikBuschel: I think the new .S did not even get assembled
23:30:41[IDC]Dragonand relies on my medicrce ID3 tags only
23:31:17[IDC]Dragonresulting in a real mess
23:31:38Nico_PGodEater: do you remember of a time when buffering wasn't as slow ? I'm thinking of bisecting
23:31:51[IDC]Dragonmaybe there are better 3rd party iTunes import tools
23:34:19bertrikah I see, I should modify lcd-as-memframe.S, right? It has the exact same code sequences
23:35:25Buschelbertrik: just added the patch...
23:35:55Nico_Pamiconn: I just looked, in an SVN sim, sizeof(struct track_info) == 1268. 32 times that is 40576. Divide that by 24 (which is sizeof(struct track_info) in the MoB code) and you get 1690.
23:35:57bertrikBuschel: i have a e200, not a c200 ....
23:36:37*bluebrother gets quite tired all of a sudden and wonders why
23:37:11Nico_Pamiconn: so I'm thinking maybe I could set MAX_TRACKS to 256 or maybe even 512 but that already seems way overkill to me
23:38:12Buschelbetrik: hmm, i think the patch could be ported... but not today. gotta got to bed now, will be a long day tomorrow :/
23:39:06Buschelbertrik: the code sequences seem to be the same at a first glance
23:39:33bertrikindeed, the dithered version seems different though
23:39:35amiconnNico_P: I think it should depend on the amount of target memory then. And of course the buffering code needs to handle the case that the IDs are used up before the whole buffer is filled
23:40:28Nico_Pamiconn: it does, I checked today. It stops buffering nicely and rebuffers when needed
23:40:30bertrikI'll try just the undithered version then. Mpegplayer should show it if works, right?
23:41:01Buschelbertrik: yes. but it's easier to test with the test_fps plugin
23:41:07Nico_Pamiconn: and yes, for now MAX_TRACKS is 32 on lowmem targets and 128 on others. We'll just need to find good comprimises
23:41:18 Join Bagder_ [0] (
23:41:30bertrikBuschel: sorry, not familiar with that plugin, is it in the standard set of plugins?
23:41:54amiconnI'd use 4 per MB RAM, with a lower limit of 16 tracks
23:42:19amiconnThat'd be 64 for X5 and M5, 128 for most other targets, and 256 for the 64MB ipod video
23:42:53amiconnArchos will get 16 when switching playback engines later
23:42:53Buschelbertrik: you will have to enable it in apps/plugins/sources via adding the line "test_fps.c"
23:43:03amiconn(due to the set lower limit)
23:43:07bertriknever mind, i get assembler errors
23:43:31Buschelbertrik: :/
23:43:40Lloreanamiconn: SPCs are always 64kb, if I understand correctly, so it'll be known that if someone shuffles a collection of them you can fit four times as many as the limit you just suggested.
23:44:03amiconnBuschel: Note that the c200 and e200 code don't use the same asm routines yet; c200 only has the non-dithering variant atm
23:44:19Lloreanamiconn: I was thinking maybe we could use them as a lower limit for software codec targets.
23:44:30 Quit ompaul (Client Quit)
23:44:59LloreanKB rather
23:45:17jhMikeSamiconn: things seem much better with the joystick. no glitch yet
23:45:23amiconnThe advantage on c200 will be that dithering will come for free, fps wise
23:45:31amiconnjhMikeS: yay! ;)
23:45:42amiconnDid you see my update to LcdFrameRate?
23:45:59 Join _Bagder_ [0] (
23:45:59amiconn114.5 fps for YUV on X5 at 124MHz :)
23:46:08jhMikeSsay what?
23:46:56jhMikeSso the load now is basically just decoding
23:48:10bertriktoo many bits in constant i think
23:48:43jhMikeSThe framerate there on ED is much better. I didn't measure. Too obvious.
23:49:10amiconnNow we need that asm idct...
23:50:11 Join Davide-NYC [0] (
23:50:17jhMikeSI think it could play that one full speed.
23:50:27Buschelbertrik: don't know :/ maybe someone with bettere knowledge of asm can pick the idea at least...
23:50:53amiconnI am hoping for full speed on H300 as well...
23:51:39 Nick parafin is now known as parafin|away (
23:52:10Buschelbertrik: sorry to waste your time. and thanks for your testing :)
23:52:25bertriknp :P
23:52:39 Quit Buschel ()
23:53:39jhMikeSbertrik: what's that build?
23:54:14Davide-NYCDoes the (very) recent coldfire commit affect recording in any way?
23:54:15bertrikbuild? you mean the pastebin text?
23:54:16 Quit BHSPitMonkey (Connection timed out)
23:54:36jhMikeSbertrik: yes
23:54:45 Join webguest98 [0] (i=4a469ecd@gateway/web/cgi-irc/
23:54:53 Quit lee-qid ("aufwiederbyebientotsayonara")
23:55:05jhMikeSDavide-NYC: it's a system-wide change. It should speed things up if anything.
23:55:07bertrikan attempt to patch lcd-as-memframe.S with the changes from FS #7951
23:55:09webguest98i read somewhere that sansapatcher works on e200r now. Is that true?
23:55:24*Davide-NYC dons his testing helmet
23:55:38markunDavide-NYC: hi!
23:55:54markunjhMikeS: right now we only use buffered writes for the framebuffer
23:56:24Davide-NYCmarkun: very nice meeting you face2face the other day
23:56:26jhMikeSwhy is the Y-16 removed? that's incorrect
23:56:27 Quit webguest98 (Client Quit)
23:56:34amiconnmarkun: Nope, for anything
23:56:43markunamiconn: for the Gigabeat I mean
23:57:28amiconnDavide-NYC: Did you observe problems? I wouldn't expect any...
23:57:32markunamiconn: can I just turn it on for everything without any penalty?
23:57:44bertrikjhMikeS: I don't know, I haven't studied the patch yet, just tried to see if it would apply
23:57:44markunDavide-NYC: yes, I enjoyed it too
23:57:57*jhMikeS doesn't know why the YUV formulas are being altered in that patch. they are correct now.
23:57:58 Quit Bagder (Read error: 110 (Connection timed out))
23:58:02amiconnmarkun: Depends. I don't know the gigabeat architecture. Some delay loops might need recalibration
23:58:12Davide-NYCamiconn: I'm still at compiling
23:58:18Davide-NYC(damn games!)
23:58:35*markun waves good night
23:58:49 Join Bagder [0] (n=daniel@rockbox/developer/bagder)

Previous day | Next day