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 2008-10-07

00:01:40 Quit herrwaldo ("Konversation terminated!")
00:02:08 Join toffe82_ [0] (
00:03:07 Quit bertrik ("Leaving")
00:03:56gevaertsccache already seems to get a full path here
00:05:01 Quit n1s ()
00:07:10 Quit bmbl ("Woah!")
00:08:43 Quit toffe82 (Read error: 110 (Connection timed out))
00:09:54 Quit tvelocity ("Αποχώρησε")
00:09:54peturdoes casting a 32bit to 16bit guarantee that the upper 16 bits are cleared?
00:10:20Bagdernot if signed
00:10:25 Join MethoS-- [0] (
00:10:29XavierGramiconn: is the second hotfix needed on Windows XP SP3?
00:10:55gevaertsI'd say that if you cast a 32bit to 16bit there are no upper 16 bits
00:10:57 Quit jgarvey ("Leaving")
00:11:01amiconnIt is needed if you have DEP enabled
00:11:52peturBagder: a uint32_t is cast to int16_t
00:12:07Bagderthen I believe you can count on it
00:12:21*gevaerts still doesn't understand the question
00:12:22 Quit Xerion (Read error: 54 (Connection reset by peer))
00:12:37Bagderwell, unsigned to signed will then only be 15 bits
00:12:44 Join Xerion [0] (
00:12:50Bagderor something
00:13:04*petur wonders why it wasn't just & 0xFFFF ... jhMikeS must have had a reason...
00:13:19 Quit bluebrother ("leaving")
00:13:27*Bagder goes to bed
00:14:42peturhmmm maybe endianess...
00:15:12gevaertsThat shouldn't change anything
00:16:51pixelmaamiconn: do you know if there is any possibility to build the manual in interix?
00:17:05amiconnBagder: Better?
00:18:39Bagderyeah, the only thing I don't like 100% is the duplication of code between findtool and findpath but we can fix that at a later time
00:19:35 Quit DerDome ("Leaving.")
00:20:41amiconngevaerts: Could you try that patch on your box?
00:20:48gevaertsamiconn: doing that now
00:20:49amiconnIt works for me on Cygwin and Interix
00:21:11 Join krazykit [0] (
00:21:49amiconnpixelma: InterixDevelopment says somewhere near the top that this is not possible
00:22:11gevaertsIt works. I'm now testing again if and how much it speeds things up
00:22:35amiconnIt would require building tetex on Interix. I've found hints that this can be done, but I probably can't be bothered to try that
00:22:49gevaertsI doubt if interix would be much faster than cygwin for latex anyway
00:23:18amiconnVoice building already has a couple of problems I need to work around
00:23:24gevaertsThat's not something that starts lots of processes
00:23:36amiconnNo, latex is quite fast on cygwin
00:24:11amiconnAnd there is nothing that stops you from having both Cygwin and Interix installed (except maybe available diskspace)
00:24:40 Quit XavierGr ()
00:24:43 Quit MethoS- (Read error: 110 (Connection timed out))
00:25:12amiconnSims will require building a mingw crosscompiler for Interix
00:26:11 Quit mcuelenaere ("Zzzzz")
00:26:44 Join sarixe [0] (
00:28:32amiconnThis is something that should definitely be possible; iiuc gimp-win32 is built that way
00:28:49gevaertsThere may be a small speedup from your patch here, on the order of 0-2%
00:29:58amiconnNo arm compiler problems?
00:29:59 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
00:30:30pixelmahow much space does the interix install need? Can't find it on the InterixDevelopment page
00:31:03gevaertsamiconn: no
00:31:16 Quit culture ("Ex-Chat")
00:31:35 Join culture [0] (
00:32:14amiconnMy D:\SFU takes ~1GB, excluding /home/Jens
00:33:54 Quit shotofadds ("Leaving")
00:34:18 Quit Nico_P (Remote closed the connection)
00:38:37 Quit faemir (Remote closed the connection)
00:38:38 Join fdinel [0] (
00:40:12 Join MethoS_ [0] (
00:43:19 Quit ompaul (Remote closed the connection)
00:44:40 Join MethoS- [0] (
00:48:10 Quit MethoS- (Remote closed the connection)
00:49:30 Quit ender (" How long a minute is depends on which side of the bathroom door you are on. -- Theory of relativity")
00:55:02***Saving seen data "./dancer.seen"
00:55:14 Quit perrikwp (" ajax IRC Client")
00:55:21 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
00:55:33*amiconn now has a version of his patch without code duplication
00:56:58*gevaerts gives up looking for speed differences. The variation between different builds is much larger than any possible speed difference with and without the patch
00:57:04 Quit perrikwp (Client Quit)
00:59:08 Quit MethoS_ (No route to host)
00:59:41 Quit MethoS-- (Read error: 110 (Connection timed out))
01:02:52XavierGramiconn: the default home directory on windows is on Documents & Settings. How did you change that on Interix to avoid having spaces in the home path?
01:03:11XavierGralso what is the password for su?
01:03:17amiconnAs described (edit /etc/profile.lcl)
01:03:34amiconnIt's the password of the admin you choose
01:03:40XavierGrhmm I did that but still it says my home is there :\
01:03:53XavierGrah let me try
01:03:57amiconnDid you create the dirs?
01:04:26XavierGrno I didn't, that's probably it
01:04:39 Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst)
01:06:37 Quit jhulst (Read error: 110 (Connection timed out))
01:06:39XavierGryup that was it, thanks
01:11:31 Join sarixe_ [0] (
01:14:03 Quit dark (Remote closed the connection)
01:16:09 Join Waldo000000 [0] (
01:17:38 Quit petur ("Zzzz")
01:19:04 Nick zodttd is now known as zodttd|away (
01:19:27 Nick zodttd|away is now known as zodttd (
01:20:56XavierGrstarted building the crosscompilers, lets see how much time it needs...
01:20:56 Quit Thundercloud (Remote closed the connection)
01:28:12 Join n17ikh|Lappy [0] (
01:28:28 Quit Waldo000000 (Read error: 60 (Operation timed out))
01:29:39 Quit sarixe (Read error: 110 (Connection timed out))
01:34:28XavierGramiconn: do these instructions inlude ccache?
01:39:58 Join kachna [0] (
01:43:21 Quit sarixe_ ("Ex-Chat")
01:44:53 Join sarixe [0] (
01:46:26 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
01:49:54 Quit PaulJam (".")
01:53:28 Nick fxb is now known as fxb__ (
02:03:56 Part toffe82_
02:08:23 Quit culture (Read error: 60 (Operation timed out))
02:09:13 Part pixelma
02:09:35 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
02:12:58 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
02:18:42 Join Emanuel1279 [0] (n=emanuel1@
02:19:04 Part Emanuel1279
02:21:53 Quit einhirn (Read error: 104 (Connection reset by peer))
02:23:29XavierGramiconn: when typing make -version shouldn't it show the version of make?
02:23:40 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
02:23:43XavierGrmine just says illegal option v
02:24:31XavierGrI think I have the wrong version of make, but I deleted the first and installed the new one
02:25:56XavierGramiconn: when you type which make, what is the path to it?
02:26:19 Join ajonat [0] (n=ajonat@
02:29:29JdGordonXavierGr: hey, I dont understand your comment on the scrolling patch re 255 and impossible to overflow?
02:30:09 Quit scorche (Read error: 54 (Connection reset by peer))
02:30:49 Quit SUSaiyan` (Read error: 60 (Operation timed out))
02:32:28 Join tessarakt2 [0] (
02:32:30 Quit tessarakt (Read error: 110 (Connection timed out))
02:32:48 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
02:32:48 Join SUSaiyan [0] (
02:33:22XavierGrJdGordon: well I tested it with a maximum character filename and it didn't overflow
02:33:55XavierGrJdGordon: the aray that holds the string is quite larger so overflowing is not possible
02:34:05XavierGrat least how I understand it
02:34:10JdGordonthat code is used throught the list which allows up to MAX_PATH strings which iirc is 290
02:36:45XavierGrsee SCROLL_LINE_SIZE define in scroll_engine.h
02:36:57XavierGrit is more than MAX_PATH
02:37:41JdGordonah ok
02:37:51 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
02:38:07XavierGrI tried once more to do it without padding the string but it is a mess with current code
02:38:32XavierGroffset, width, startx etc all take acount the string length
02:40:28XavierGrJdGordon: so what is your opinion, make it 1/4 of the viewport or setting?
02:42:02XavierGramiconn: I got worse time in interix vs Cygwin
02:42:24XavierGrH300 build in Interix took 2.18 while in Cygwin 1.58
02:42:37XavierGrsorry I meant 2.13 in Interix
02:46:53 Join massiveH [0] (
02:47:40JdGordonXavierGr: yeah, i think 1/2 width is a bit much
02:50:13*Llorean thinks it needs to be configurable if it's going to be changed at all.
02:50:14XavierGramiconn: 2nd test Recorder - Interix: 1m14.656s while Recorder - Cygwin: 1m3.656s
02:50:56XavierGrAs I am the guy that wants as much settings as possible in Rockbox, I will have to agree with Llorean
02:51:01LloreanI don't want to have to wait for a bunch of extra blank characters to be drawn before it shows me the beginning of the track name again if I've just gotten to the end, especially if I just turned the backlight on to see something at the beginning of the string.
02:51:29LloreanAnd it doesn't make sense to hardcode it in a way that workarounds are impossible when it's currently hardcoded in a way where you *can* workaround without C knowledge.
02:52:00XavierGrLlorean: yeah but the workaround is a pain :\
02:52:23Llorean"A pain" is better than "Impossible" which is what you'd be offering people who are happy with the current behaviour by changing the hardcoding.
02:52:43XavierGrI will try to make it a setting, maybe something like, defualt (minimum) 1/4 viewport, 2/4 viewport, 1 viewport?
02:53:30LloreanWhy not just a number of spaces using the current font?
02:54:07XavierGrThat could be done too and be more easy, it is just that the user won't be sure about how much it will end up due to different sizes of fonts
02:54:28LloreanAnd with your method, different lines will have different spacings depending on viewport size.
02:54:39XavierGryes indeed
02:54:40LloreanWhich will be particularly noticeable in multiple-viewport WPSes and such.
02:55:03***Saving seen data "./dancer.seen"
02:55:08LloreanIt makes more sense, to me, to just have them pick either "a number of spaces" or "a number of pixels" rather a than a percentage of a variable they don't know the value of in advance, and may be different for different lines.
02:55:22XavierGrok I will make it that way, it's way more easy anyway
02:56:01XavierGrprobably limit it to something sane too, we don't want a huge value for it
02:56:15LloreanLimit it to screen width / 2 maybe?
02:57:51XavierGryeah seems logical, though different proportions, screen width = pixels while the setting will be in characters
02:58:50LloreanMay as well just pick a number then.
02:59:58XavierGrwill try it tomorrow, unfortunately today interix didn't manage to replace Cygwin. It seems that Cygwin will still keep its throne for a long time in Windows Rockbox develpment (at least for me)
03:06:28 Quit perrikwp (" ajax IRC Client")
03:08:59 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
03:11:55JdGordonthe setting will need to be "insert a minimum of X pixels" but that will get messy if the first few are width/something
03:12:46LloreanI think just a flat number of pixels, maybe in 10s so like 10, 20, 30, 40, 50, 60, 70, 80, 90, 100
03:13:13LloreanSo the first one is "a character and a half" or so for the very smallest fonts, while the last one is ~1/3 of the screen on the very largest screens.
03:39:24 Join reacocard [0] (n=reacocar@WL-65.CINE.HMC.Edu)
03:43:25 Quit krazykit ("later")
03:49:04 Join SETH_HIKARU [0] (
04:01:39 Join reacocard_ [0] (n=reacocar@WL-65.CINE.HMC.Edu)
04:01:59 Quit reacocard (Read error: 104 (Connection reset by peer))
04:02:03 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
04:03:41 Join aarcane [0] (
04:04:05aarcaneis this the right place to ask about ipodpatcher binary?
04:04:27Lloreanaarcane: We recommend you use RBUtil these days for Rockbox installs.
04:04:31LloreanWhat was your question?
04:05:11aarcaneLlorean, I'm trying to get .ipod files from official apple firmware images, but I don't want to go through the hassle of installing them to my ipod to get the .ipod image. can ipodpatcher do that somehow ?
04:05:30LloreanThat's not what it's designed to do, no.
04:05:59aarcaneLlorean, can you suggest a tool that IS capable of doing that?
04:06:03 Join JdGordon|phone [0] (
04:06:25LloreanYou'd more or less need to adapt your own.
04:06:38LloreanThere's really never been a use for such a tool in association with Rockbox.
04:07:27 Nick zodttd is now known as zodttd|away (
04:08:19aarcaneLlorean, I had figured it would be a natural extension to ipodpatcher, but I an see how it's not directly related. so you're not aware of any package or program that can do that currently ?
04:08:49Llorean".ipod" isn't exactly a real file format or anything
04:09:35LloreanIn fact I think it's just a header added to a normal iPod firmware image. There's really no reason other people would be making tools to work with a format that's really mostly there just for the Rockbox tools and bootloader to recognize.
04:09:37 Join Ac-town [0] (
04:09:38 Quit SETH_HIKARU ("Leaving")
04:09:52LloreanBut don't quote me on that.
04:10:13 Quit JdGordon|phone (Remote closed the connection)
04:11:32aarcaneLlorean, the only reason I need it is so I can switch between ipodwizard'd firmwares on the fly
04:11:45LloreanWell that's something you should be asking the iPodWizard guys, then.
04:11:53Lloreanmaybe they'll write you a tool that packages up their modified firmwares.
04:13:16aarcaneLlorean, I suppose I could ask..
04:14:00LloreanWell, there's nothing really else for you to do. This channel is for Rockbox-related stuff.
04:15:50aarcaneLlorean, I had just hoped that rockbox's firmware modification tool could create it's own .ipod files from official firmwares. the ipodwizard stuff doesn't really do anything except change s ome basic resources in the file :-S
04:16:11LloreanWe can create our own .ipod files, from the firmware on the device as a means of backing it up.
04:16:14 Quit XavierGr ()
04:16:38LloreanRockbox's tools are, as I said, designed for installing Rockbox safely. Not facilitating some other project's random program.
04:20:09 Nick reacocard_ is now known as reacocard (n=reacocar@WL-65.CINE.HMC.Edu)
04:21:47synergistah, the spirit of giving is alive and well in rockbox tonight ;)
04:22:11*synergist is crabby too, his copy of photoshop loves to wait until he's right at the crucial moment just before a save... then pounce with a crash
04:24:46Lloreansynergist: This channel is on-topic.
04:25:00LloreanIt's not an issue of being crabby, it's an issue of simple guidelines we ask people to follow.
04:32:49 Join _emp [0] (
04:36:16 Quit HBK (Read error: 104 (Connection reset by peer))
04:36:17 Join HBK- [0] (
04:37:20soapOn the "Undefined Instruction at xxxxxxxx" "3G" user...I'm down to believing misinformation from the client or hardware failure.
04:38:12Lloreani was about to ask "you mean rbutil or the guy"?
04:39:11LloreanI'm wondering if I should double check on that rockbox.ipod thing. This could be a case of "he doesn't know what root is" now too.
04:41:46 Join blkhawk- [0] (
04:43:47soapYes, I didn't mean to imply intentional misinformation, but if the original firmware works in all likelyhood something we were told is not so.
04:43:55_emprbutil is on version 1.0.6, right?
04:44:41Lloreansoap: Yes, that unfortunately is very difficult to diagnose. I'm hoping maybe a restore with iTunes (including associated reformatting) will clear things up.
04:45:08Llorean_emp: I'm not sure I'm understanding. Why can't it be built from the tagged SVN revision that was used to create the version we distribute to people?
04:45:09_empthanks, Llorean, and that is the version in the 3.0 file?
04:45:42LloreanThe 3.0 file only represents the current build of _Rockbox_ and the utility release version will be different.
04:45:50_empLlorean, I just need a dist file at the end of the day with the source. It's how the port tree system works.
04:46:11LloreanSo it cannot work from a project that doesn't provide separate tarballs, period?
04:46:39 Quit HBK- ()
04:46:54_empif a tarball exists with a frozen version of rockbox utility source, then that would be a better option since the source is smaller
04:47:03LloreanThat doesn't really answer my question.
04:47:40LloreanRight now, versions of RBUtil are tagged in SVN. There aren't separate tarballs of it, and there's no tarball in existence representing a specific version of it.
04:48:04LloreanThe one in 3.0 is likely to have a few bugs that have long since been fixed (as well as added features such as "The ability to install 3.0").
04:48:24_empthe ports system for the bsds are setup to pull down a file, typically a tarball, with source, extract it into a working directory, do a fake build in that directory structure, generate a PLIST as to where the files will be installed after a make fake, and then install the files per the hierarchy specified by the OS.
04:48:55Llorean_emp: So the "pull down a file" stage can't be replaced with, for example, a version control checkout?
04:48:56aarcanewhere's the table of reported battery life in rockbox on different ipods ?
04:49:05Lloreanaarcane: IpodRuntime, I think
04:49:26_empLlorean, nope, since the process also does a checksum on the file to make sure it's a good file as part of the process of pulling it down.
04:49:56_empthe port system is not designed to work with subversion trees or any other trees
04:50:11Llorean_emp: Then wouldn't it make the most sense for you to archive a 1.0.7 rbutil archive, and have it use that?
04:50:36_empLlorean, it would if I had a place to host the 1.0.7 file.
04:51:06Llorean_emp: It seems reasonably, since we host the necessities for installing it on other OSes, that you could just talk with Bagder about it when he's around...
04:51:44LloreanThat way an attempt can be made to make sure oBSD users (it was that, right) can actually be using the same version as the rest of our users.
04:51:52_empLlorean, thanks. My patches are working to install the 3.0 file some gcc'isms and some mounting changes, but that's pretty much it.
04:52:02_empLlorean, yep
04:52:06LloreanDon't install the 3.0 file.
04:52:16_empno prob; I'll wait
04:52:34LloreanIt's not any specific version of RButil, and it probably can't even install Rockbox 3.0, since that feature was enabled in the util after the release, not at the time of.
04:52:59 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
04:53:03_empI originally built this off of the rockbox.7z file about a month ago.
04:53:06LloreanIf there are changes to rbutil itself, you should go ahead and post them to the tracker.
04:53:10saratogadid anyone take a look at that updated album art resize patch?
04:53:19_empLlorean, my patches are all ready up on the tracker
04:53:21Lloreansaratoga: Haven't heard anyone mention it.
04:53:40Llorean_emp: Ah, that's good then. Maybe call Bluebrother or domonoky's attention to them (when they're around, if they haven't commented yet)
04:53:59saratogai'm skimming it now, but technically i'm supposed to be studying for an exam so i can't really dig into it tonight
04:54:01_empLlorean, one patch to make it work with gcc-3.3.5 and the rest to get it to mount to under the oBSD stuff
04:54:26Llorean_emp: And neither will break other builds of it on the OSes we already support?
04:54:49_empLlorean, let me double check them; it's been a while since I did them.
04:55:05***Saving seen data "./dancer.seen"
04:55:24Llorean_emp: A lot of time we expect patches on the tracker to stay there until they don't require any extra work on the committer's part (or minimal extra work) before inclusion, so it's helpful to try to make sure they're "ready to go" as it were. :)
04:55:54Unhelpful_emp: did you end up having to change the OSX mount stuff much at all?
04:55:56_empLlorean, hey, no problem. Two of them have a potential issue, but I can fix those up next week.
04:56:05Llorean_emp: If you think they are "done", you can also send a message to the -dev mailing list, since that asynchronous method of communication can better help you get in contact with the people who'd most likely look at those patches.
04:56:24_empUnhelpful, oBSD doesn't mount like osx
04:56:44_empLlorean, I'll tweak the two problematic ones. Those had to do with the geometry.
04:56:59Unhelpfuli seem to recall talking to you about developing this... was it the mount enumeration that was similar?
04:57:55 Quit blkhawk (Read error: 110 (Connection timed out))
04:58:01 Nick blkhawk- is now known as blkhawk (
04:58:14_empusb devices typically mount (r)sd(x)(a-z), where r is for raw if needed, x is a number, a-z is the partition; typically c for fat32 systems
05:01:20_empI'll spend some time this week and next week getting this patches, so they don't break anything else.
05:01:40 Join Tetracomm [0] (n=nicholas@
05:05:10 Join miepchen^schlaf_ [0] (
05:05:38saratogawhy hasn't the ipod accessory stuff been added to SVN yet?
05:06:00saratogaseems fairly ridiculous to still be passing out unofficial builds at this point, particularly when we are inviting people to test and use it
05:07:10Unhelpfulsaratoga: part of the "conspiracy" to "cripple" official builds?
05:07:27Lloreansaratoga: Soap wanted a week more just to make sure it wasn't doing anything obviously strange
05:07:48saratogaah ok
05:07:53LloreanThat was a week and a weekend ago, though
05:07:58 Nick zodttd|away is now known as zodttd (
05:08:02LloreanSo we're about at the time where someone should\ commit it.
05:08:13saratogastill seems pretty safe, its basically a driver as I understand it
05:08:20saratogashouldn't have effects beyond PP
05:08:41saratogai usually just commit those things and wait to see if people report bugs
05:09:09LloreanOne thing to check
05:09:22LloreanAccessory power will reduce battery life if it defaults to on.
05:09:36LloreanBut will cause some devices to fail for a non-obvious reason if it's off.
05:09:42LloreanSo we need to decide which way to err.
05:10:37Unhelpfulno way at all to tell when a device is plugged in and turn it on then? is that something the OF can do that's not documented?
05:10:48LloreanUnhelpful: I believe that's the case, yeah.
05:11:18saratogadid we ever agree that the hit was real from turning it off?
05:11:27ameyershrug, make a menu option?
05:11:35ameyerif there's a real hit?
05:11:36saratogaBuschel claimed some ridiculously small difference from turning it on, but he tested via battery bench
05:11:45saratogasorry from turning it on
05:11:47Lloreanameyer: A very real hit.
05:11:58Lloreansaratoga: I thought it was like an hour or more less runtime.
05:12:05saratogai remember some skeptisism that he could measure such small differences
05:12:19saratogaI believe it was much smaller, on the order of minutes not hours
05:12:31soapI get disgustingly repeatable runtimes, FWIW.
05:12:31LloreanI could've sworn accessory power was one of the bigger costs.
05:12:52*ameyer just thinks that if there's a significant difference, it should be easy to disable
05:13:01 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:13:09ameyeralso, would this be all PP or just iPods?
05:13:13soapbut I still round to 15 minutes as I'd rather be have a higher noise floor than false signal.
05:13:27soapameyer, it is easy to disable.
05:13:28Lloreanameyer: Just iPods.
05:13:46saratogaan hours difference would mean it uses ~ 2ma, which seems like an enormous amount of power for a supply thats not actually powering anything
05:14:00Lloreanameyer: The accessory protocol is the Apple firmware talking to Apple licensed accessories.
05:14:16ameyerI figured
05:14:24saratogaand since apple apparently does not turn it off, its hard to imagine it uses much power
05:14:29saratogaotherwise they'd be sure to switch it on and off
05:14:34Lloreansaratoga: Are we sure they don't?
05:14:40soapdoes apple not turn it off or do they auto-detect?
05:14:50saratogaI believe Buschel said they didn't turn it off, but i could be mistaken
05:15:11LloreanWould it make sense to maybe turn it on once every second (or two), and check if an accessory is attached?
05:15:27LloreanThen there'd be a slight detection lag, but it wouldn't be on constantly
05:16:13 Join toffe82 [0] (
05:16:46saratogaspeaking of buschel, he needs to stop by and multithread his MPC decoder, I want to see what kind of battery life he can get at 8 MHz or whatever he wants to run at
05:17:32 Quit saratoga ("CGI:IRC (EOF)")
05:18:35Lloreansoap: Well he can run it at 24 with the Rockbox UI overhead.
05:19:24*Unhelpful wonders if it would be possible to hack up an apple accessory adaptor for other players... useful only under rockbox, of course.
05:20:36soap8 is disgusting. ;) Speaking of which - is there objection to boost on UI, or just the current implementation. My Video has poopy scrolling since MP3 went multicore. MP3 playback w/o boosting appears to leave just too little CPU left over for consistent smooth scrolling through long lists.
05:21:08Lloreansoap: Personally, I really like the idea of "boost on UI"
05:21:16LloreanBut I'm not sure what holds it back / shoots it down.
05:21:33LloreanIt seems to me like it'd make sense on *all* portalplayer targets
05:21:44LloreanOr really *any* target where you can tell during the UI when the CPU boosts/unboosts.
05:22:35soapIt doesn't seem to be much of an issue on the Nano, but the 5G and its large screen and low clock really shows it.
05:22:57 Nick zodttd is now known as zodttd|away (
05:23:01LloreanI haven't tried the Nano with the new CPU patch
05:23:21LloreanBut back before multicore MP3 I could see a difference in scrolling when boosted/unboosted.
05:23:35LloreanAnd UI boost would let us drop to 24/80 with almost no user-visible effect.
05:23:45LloreanIt should average out to better battery life for a lot of people.
05:25:45*ameyer goes to find his sansa to see if soap's issue exists for him too
05:25:59 Join druidu [0] (n=d5a4e705@gateway/web/cgi-irc/
05:26:17druidua quick arm-assember question... anybody? :)
05:27:31 Quit druidu (Client Quit)
05:28:44soapis normal/boost speeds something which could be configured from within a Rockbox installation (needing reboot I would assume) or is that not possible?
05:29:47Lloreansoap: I'm pretty sure we're very much against that, though.
05:32:18 Quit massiveH ("Leaving")
05:32:38soapit makes sense so long as there is disparity in codec speeds, does it not? Let primarily mpc users run @ 8Mhz and mp3 users run @ 24.
05:33:21Unhelpfulcould make it load-based, or is full-on scaling too heavy and/or difficult to integrate into current codebase?
05:34:08 Join Makuseru [0] (
05:34:53MakuseruHi, in a WPS, if i wanted to move a part of it (the peak meters) from the center of the screen, to the bottom of the screen, how would i do that?
05:36:14Lloreansoap: IIUC 24 is some sort of ideal speed (something's native speed), we'd actually be getting rid of one of the states by using it. I'd be curious exactly how much slower we could even get before basic WPS rendering starts causing it to demand a boost for PCM playback.
05:42:02 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
05:42:41saratogai have some ideas about how I think boosting should work
05:43:15saratogarather then boosting on user input, I would prefer it if the button drivers could set a flag indicating that the user interface is active
05:43:42saratogawhile this is set, the normal speed would be adjusted upwards somewhat, perhaps to 45MHz
05:44:00saratogawhen it is not set, the idle speed would be set to 24MHz
05:44:27saratogai prefer this approach to boosting on input because
05:44:41saratoga1) its wasteful to put the CPU at 80mhz just to handle input
05:45:12saratoga2) you're more likely to notice problems in plugins if you ping pong between 30 and 80
05:45:16 Nick zodttd|away is now known as zodttd (
05:46:02saratogathis would also open up the way for additional optimizations on players like the sansa that do not have a passive LCD, because we would know when WPS updating was needed and when it was not
05:46:44saratogaSoap: regarding the Video, I think this problem could be mitigated for MP3 with better balancing of the COP and CPU
05:46:56saratogaI probably put too little on the COP
05:47:04saratogai just haven't had time to work on finding a more optimal balance
06:04:01 Quit robin0800 (Remote closed the connection)
06:04:16 Join robin0800 [0] (
06:05:38 Quit Seed ("cu, Andre")
06:06:17 Join Seed [0] (
06:20:11TetracommHow is hunting and gathering for this player going?:
06:20:44LloreanTetracomm: You can see all the information there and in the thread in the forum.
06:20:53scorcheTetracomm: you might want to check out the corresponding forum thread for it in the new ports section
06:21:43 Join AndyIL [0] (i=AndyI@
06:24:06 Quit saratoga ("CGI:IRC (EOF)")
06:30:39 Quit _emp ("Leaving")
06:34:57 Quit AndyI (Read error: 110 (Connection timed out))
06:36:01 Join nicholas__ [0] (n=nicholas@
06:36:29 Quit Tetracomm (Nick collision from services.)
06:36:43 Nick nicholas__ is now known as Tetracomm (n=nicholas@
06:37:37TetracommAre you still working on it? Or you're not doing anything?
06:37:41TetracommI own it and am wondering how long it has been taking to gather information on it.
06:38:36 Join unstable [0] (i=unstable@
06:39:04unstableI don't want to look through a lot of themes, is there a popularity contest for themes? What is the most popular theme?
06:39:06JdGordonsoap: Llorean: wasnt there a patch which boosted if too many buttons got queued up? its wasteful to boost every time a button is pressed
06:39:13unstableI want to configure the rockbox with the most popular settings.
06:39:19scorcheTetracomm: *i* am not working on it and never have...rockbox as a group doesnt do ports...individuals who come forward with the device itself, time, and capability to amke the port itself do
06:39:44JdGordonsoap: also, yeah I tihnk it could be done so the user could choose the unboosted cpu speed (from a set list though, not arbitrarily)
06:40:09TetracommI wonder what i can do to help?
06:40:13advcomp2019unstable, it is really person taste on the themes
06:40:18scorcheunstable: well, themes are subjective, so what you like, another person probably does not...just search thrpough them and pick what you like
06:40:22unstableadvcomp2019: There is no consensus/voting?
06:40:30unstableyea, I don't want to look through 1000 themes.
06:40:37scorchethere arent 1000
06:40:59Tetracommand no one here actually got to see the device?
06:41:09advcomp2019yea, there is no voting or consensus
06:41:27unstableWhat theme do you guys use?
06:42:06advcomp2019i use "plain text vp"
06:42:08scorcheunstable: there are 138 people in here...please dont ask questions like that...just pick for yourself..
06:44:32unstablehmm, crap
06:44:45unstableI went to "rockbox default" and I have weird squares on the screen
06:46:36unstableHow can I fix this?
06:46:44advcomp2019where are you getting the themes from then? there are broken ones
06:47:30scorcheunstable: is the song you are playing containing foreign characters like chinese, cryllic, etc?
06:47:48*scorche cringes at the grammar of the above sentence
06:47:53 Quit Nibbler (Read error: 60 (Operation timed out))
06:50:44unstableIs there a mega pack of all the themes..for ipod video? that I can download all at once?
06:52:07unstableI have to download them all individually?
06:53:08 Join vitja [0] (n=vitja@
06:53:29unstablescorche: Which one do you use?
06:53:40vitjawhat's usual ata model name?
06:53:51Lloreanunstable: Is it so hard to look at the screenshot and just say "hey, that one looks good"?
06:54:07vitjaI'm now writting code that fills ata_identify for ata-nand-telechips and want to choose name, now it's "TNFL"
06:55:06***Saving seen data "./dancer.seen"
06:56:01unstableI'm looking, I just can't get the full feel of it.
06:56:14unstablethemes aren't just how they look right? it's how the buttons react etc?
06:56:22LloreanThemes are *just* look
06:56:31scorchebuttons dont change
06:56:33LloreanThey do not change Rockbox's functionality in any way.
06:57:11unstableIf I'm playing a song, how can I go back up one parent directory?
06:57:24unstableor even see the directory the file is I'm playing
06:57:28unstablewith one button push
06:57:31Lloreanunstable: read the manual, please.
06:58:52unstableI looked at it, you know it's 175 pages?
06:59:20LloreanA lot of people spent a lot of time writing it, because we'd have to spend hundreds and thousands of times as much time if we answered everyone who didn't want to look at it.
06:59:44LloreanSo, if you have any respect for the time people put in to make Rockbox, spend a very tiny fraction of that time yourself and do some reading first.
06:59:45unstableWhat if that is my last question for the night, then I go to sleep?
06:59:49 Quit Zarggg ()
07:00:01Lloreanunstable: It doesn't matter if it's your only question, ever, the channel has guidelines.
07:00:04scorcheunstable: then you can take the manual to bed with you?
07:03:40unstablehmm, there are 1000 themes. I wish there was a voting system, so I just had to go through the top 10 list.
07:04:14scorchei told you...there are not near 100 themes
07:04:31Lloreanunstable: There are not 1000 themes, and even if there were your opinions are not dictated by those of the rest of the world.
07:04:32scorchewhat one person likes, other people probably dislike somewhere
07:04:48LloreanIn the time you've spent complaining about how many themes there are, you could've looked at the screenshots for every single theme on that page.
07:05:26aarcaneunstable, what there should be is a database which shows previews of themes and is cross-referenced with a compatibility matrix, so you know if it'll work on your ipod or music player
07:05:50JdGordonsomething like the new themes site maybe ( scorche wink wink :D )
07:05:55Lloreanaarcane: What does that have to do with his complaint?
07:06:12Lloreanaarcane: He's complaining that there's no record of their popularity, not compatibility.
07:06:32aarcaneLlorean, if he could see a preview of what the theme would look like on his player, he wouldn't need a voting system to help him decide what he likes.
07:06:50Lloreanaarcane: You mean like the previously mentioned, currently available screenshots?
07:07:22aarcaneLlorean, if they have a screenshot for each applicable player that it's compatible with, then certainly.
07:07:23scorcheaarcane: that is why it is called WPS *gallery*
07:07:44Lloreanaarcane: Themes are compatible with a specific resolution. They will look exactly the same on all players of that resolution.
07:07:49LloreanWhy should there be multiple duplicate screenshots?
07:07:57aarcanebesides, I just popped in with about 5 lines of back text. I didn't here about galleries.
07:08:25scorcheso you proposed an improvement without knowing the existing system?
07:08:33aarcaneLlorean, well there don't need to be a different screenshot for different models of the same resolution and color depth obviously.
07:08:34Lloreanaarcane: It would be really, REALLY helpful if you wouldn't try to "pitch in" when you neither know what the topic of conversation is, nor know enough about some of its subjects to be able to make valid input.
07:09:08LloreanIt really doesn't help anyone when that happens.
07:10:03 Part aarcane ("I kinda liked rockbox, but with such a moody community, there's no chance it'll make a permanent home for itsself on my devic)
07:12:10amiconnvitja: You do not need to mimic the real ata driver as fast as 'identify' is concerned
07:12:36amiconnThe Sansa e200/c200 (v1) SD driver doesn't, and the Ondio MMC driver doesn't either
07:12:49vitjaamiconn: what do you mean?
07:13:30vitjawhat's wrong with it?
07:14:11amiconnIt's not necessary.
07:14:12vitjawe need it for usb-storage
07:14:41amiconnThe sansa (v1) does usb storage without it
07:14:44vitjayeah it uses it to get flash size
07:17:36vitjaamiconn: what ata driver does sansa use?
07:18:54vitjait seems that doesn't have ata_get_identify() at all
07:19:55vitjabut it have card_get_info_target
07:20:14amiconnIt implements struct tCardInfo, defined in hotswap.h
07:21:00vitjawhat is the best way to go with ata-nand-telechips ?
07:21:11amiconnBtw, this is something that's still not optimal (there is bit of unnecessary copying going on, it still contains an unnecessary field, and the struct name violates the coding guidelines)
07:21:17 Nick jhulst_ is now known as jhulst (n=jhulst@unaffiliated/jhulst)
07:22:26vitjaamiconn: I think while we don't have propper NAND layer ata_identify hack is ok
07:23:03amiconnThere are targets which will need the nand driver for internal storage, and an sd driver too, right?
07:23:18vitjad2 for example
07:23:23amiconnThe necessary multiplexing/translation layer will become interesting
07:24:33*amiconn thinks we should perhaps define consistent data structure carrying public information from the "ata" driver (whatever that is internally)
07:25:22vitjaamiconn: fully agree with you
07:25:37amiconnAny driver (including the real ata.c) would only return data using this structure, independent from how it works internally
07:25:42vitja...and remove ata_get_identify()
07:26:15vitjathat should be nice, how will it work in multi-volume case?
07:26:21 Quit soap (Read error: 110 (Connection timed out))
07:27:06amiconnThere would be several such structures, one per drive (not volume, as several volumes can reside on the same driver)
07:27:44vitjaso it should be like get_storage_info(int storage)
07:27:54amiconnIf the driver is the same, it's trivial, and if the drivers are different, the multiplexer would call the appropriate driver's function depending on the drive number
07:28:21vitjahmm what's drive number?
07:28:38amiconnNo, drive
07:28:53amiconnYou can have several partitions (volumes) on the same drive
07:29:38vitjaor you can have one partition on several drives;)
07:29:49vitjasorry not rockbox case
07:30:47vitjaok may be we should use ata_get_identify inside this function, that fills info for begining?
07:31:05amiconnA volume in rockbox is a partition (fat file system). The typical use of multivolume is for targets with more than one storage "block" (internal storage and card slot on flash targets), but there are uses for multivolume on single-drive ata targets as well
07:31:58 Part toffe82
07:32:00vitjaas I understand now we are mostly speaking about phyisical drives, right?
07:34:02 Join massiveH [0] (
07:37:14 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:37:35 Join F28 [0] (
07:38:19amiconnvitja: Yes, 'drive number' refers to physical drive. It does not necessarily match the volume number
07:54:39 Join Tsukasa-Ujiie [0] (
07:59:53 Quit Tsukasa-Ujiie (Client Quit)
08:00:06 Join Tsukasa-Ujiie [0] (
08:03:31 Nick colesterol_dog is now known as m0f0x (
08:06:03 Nick m0f0x is now known as colesterol_dog (
08:19:59 Join imthedan [0] (
08:21:22imthedanwill rockbox play DRM WMA's?
08:21:46imthedanthat sucks =x
08:22:12massiveHno it does not, don't use drm and you will be fine :)
08:22:30Tsukasa-Ujiieyou can dycrypt the drm if you ahve the licenses in windows media player, then convert them to mp3
08:22:39scorcheimthedan: perhaps you should buy music rather than buy a licence to use it at their discretion
08:22:53Tsukasa-Ujiielol nice
08:22:58imthedanor stop jumping to conclusions before you make yourself look like a fool.
08:22:58scorcheTsukasa-Ujiie: getting rid of the DRM is offtopic here
08:23:03imthedanI did have the CD.
08:23:06Tsukasa-Ujiiesorry, i ment legally
08:23:17imthedanIt was a friends CD that I no longer have.
08:23:30imthedanI was just asking a question.. why do you guys in here get all pissy about random shit.
08:23:53scorchei was responding to your "that sucks" comment
08:24:27scorcheyou would be hard-pressed to find an open source project that supports DRM period...the two concepts simply arent compatible
08:25:29imthedanAgain, you responded and assumed that I was downloading music.
08:25:39imthedanregardless of what you were responding to.
08:27:02GodEaterI don't see any mention of the word "download"
08:27:27imthedanimthedan: perhaps you should buy music rather than buy a licence to use it at their discretion
08:27:39 Quit BigBambi (Read error: 113 (No route to host))
08:27:40imthedanOne must not be a genius to infer what he meant.
08:27:47Tsukasa-Ujiienot really scorche's place to say that lol
08:28:22GodEateragain, no mention of downloading
08:28:47massiveHwhy would you rip music with DRM on it, well, "how" is more a question than why?
08:29:16scorcheimthedan: one doesnt need to download to get DRM encumbered this goign to get constructive any time soon, or are we just going to continue off-topic rambling?
08:29:27imthedanall I did was put the CD in my drive and use WMP to rip it.
08:29:43Tsukasa-Ujiiehe didnt rui music with drm on it, wmp adds drm to wma u rip from disks by default
08:29:43massiveHimthedan, by default, WMP does not put DRM on files it rips
08:30:02Tsukasa-Ujiieit does, the older versions do by memory
08:30:32scorchesooooo...what exactly is the issue now?
08:30:33massiveHI've used WMP9 and WMP10, neither rip with DRM
08:30:34imthedanWhy would I lie?
08:30:40Tsukasa-Ujiieno isue, discussion
08:31:08scorchethen lets stop is doing nothing but clog up the logs with nonsense and off-topicness
08:31:39 Quit massiveH ("off")
08:31:44Tsukasa-Ujiiekk wtva...
08:31:45imthedanSHouldn't have started in the first place. I just asked a question and you guys ran with it like I was trying to put rockbox down for some unknown reason.
08:32:07imthedanBy the way, massiveH, WMP9 does put DRM on their files when ripping.
08:32:13imthedanI can guarantee you.
08:32:16Tsukasa-Ujiieto late he left lol
08:32:18scorchehe left
08:32:23 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:32:27imthedanah good point.
08:32:45scorcheTsukasa-Ujiie: please read the channel guidelines...we ask that you use real words here and not words like "kk" "wtva", etc
08:32:48 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:32:56Tsukasa-Ujiiei know =D
08:33:01 Quit jhulst (Read error: 113 (No route to host))
08:33:11scorchethen why not follow them?
08:33:38Tsukasa-Ujiiemeh, forgot, one of the few channels i go onto that hav that rule
08:33:55scorcheplease try to remember...really..
08:34:17imthedanlaugh out loud.
08:35:35 Quit miepchen^schlaf_ ()
08:36:00 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:39:09 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
08:44:05 Quit Tsukasa-Ujiie ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
08:44:31 Join ender` [0] (
08:45:46 Quit tessarakt2 ("Client exiting")
08:46:11imthedaneh sorry for being a headache today. Just took what was said wrong and over reacted. have a good night.
08:46:17 Quit imthedan ()
08:49:55 Join Rob2223 [0] (
08:51:32 Join fragilematter [0] (n=fragilem@
08:55:09***Saving seen data "./dancer.seen"
08:57:28 Quit Rob2222 (Read error: 110 (Connection timed out))
09:00:34 Quit kachna (Read error: 113 (No route to host))
09:10:45 Quit perrikwp (" ajax IRC Client")
09:11:01 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
09:18:20 Quit Kopfgeldjaeger ("Serverwechsel")
09:19:03 Quit Tetracomm ("Leaving")
09:19:10 Quit BHSPitMonkey ("Ex-Chat")
09:24:08 Quit pixelma2 ("-")
09:25:12 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
09:31:43 Quit reacocard (Read error: 110 (Connection timed out))
09:31:46 Nick zodttd is now known as zodttd|away (
09:51:32*linuxstb points anyone interested in bmp resizing to review and comment on
09:52:39B4gderoh cool
09:55:42linuxstbYes, it adds a whole new "read_bmp_resize" function, rather than adapting the current read_bmp function.
09:56:18LinusNwith quite few comments...
09:56:26JdGordonand some #if 0's
09:56:41linuxstbBut I guess it's just a quick first hack
09:57:43 Join culture [0] (
09:59:09pixelmasince I read it was once more, the current manual *seems* to be quite long but if you actually look at the content - the pages are quite empty and what makes the manuals of targets with big screen resolutions so long are the big screenshots. I wonder if it could be done in a way that does not scare someone just lloking at the number of pages (although I wouldn't like screenshots to be small from what I imagine at the moment)
09:59:19*B4gder posted a comment to the patch
10:00:13Lloreanpixelma: We probably could just do a "quickstart guide" that *just* has the first two chapters. Most people will probably read that and learn the controls at least.
10:00:22LloreanMost often people asking things that are in the manual are asking about controls
10:00:38LloreanOr Mpegplayer/Doom/Gameboy, I suppose.
10:02:31pixelmado you offering the manual in two parts, or one "quick start" and a "full" one?
10:02:41LloreanYeah, basically
10:03:39 Part fragilematter
10:07:10scorchei would be in favor of that...most devices nowadays have both a quickstart and a full-manual
10:07:16pixelmamaybe there is also some potential in the page layout. If I remember correctly, I still get a decent size if I print two pages on one A4 and the margins seem big to me too. There is also another reason I think about this, the width defined for the button tables isn't that big and I was wondering if there is even one where an additional column would fit (in the pdf)
10:08:02pixelmafor targets with remote/M3 as discussed at DevCon
10:15:02pixelmaI think I like the quick start guide idea too (same as scorche said), now the question is if buildibg it could be done easily
10:15:36*pixelma wonders what's up with typing today
10:21:17 Join Waldo000000 [0] (
10:24:25 Quit jeffdameth (Read error: 110 (Connection timed out))
10:25:35 Join jeffdameth [0] (
10:27:44 Join kachna [0] (
10:29:30JdGordon... I was saying this a few days ago... we should have a generic quick start on the website... we dont actually need a target specific one... all of them install with rbutil and "putting music on your player" is the same
10:31:04Waldo000000JdGordon: agreed. also, i think rbutil should provide an obvious way to choose either the latest stable release or current build, and have as default, the latest stable release
10:32:47 Join massiveH [0] (
10:34:50JdGordonanyone know roughly how many ticks the button repeat pause is?
10:36:12 Quit linuxstb (Read error: 110 (Connection timed out))
10:37:11 Quit jhulst (Read error: 113 (No route to host))
10:39:39*JdGordon waves the magic wand to get Nico_P in here :p
10:39:55 Join vitja_ [0] (n=vitja@
10:40:34 Quit vitja_ (Client Quit)
10:40:41 Quit vitja (Remote closed the connection)
10:41:02 Join vitja [0] (n=vitja@
10:41:51 Quit GodEater (" ajax IRC Client")
10:42:23JdGordondoes anyone know if the wps parser has any mechanism to allow tags to act as conditionals and also have parameters?
10:48:28amiconnJdGordon: Quickstart is not all the same. Several targets need additional steps during installation. Also, buttons are very different
10:48:42amiconnAnd the button repeat interval is dynamic, it speeds up over time
10:49:58JdGordonquickstart could still be done in a generic way.. and you missed what I was asking about, but found the answer in the code..
10:50:17LloreanJdGordon: This talk of quickstart is to address the fact that people feel the manual is imposing.
10:50:37LloreanIt wouldn't be much use if it didn't contain button mappings, since that's one of the major things people ask about after being scared away from the manual.
10:51:02LloreanA "Quick Install Guide" would be something separate entirely.
10:51:26JdGordonyou will still scare people off seeing a page of links...
10:51:54JdGordonhmm... umm... yeah ok, buttons are too dissimilar to anything about that
10:53:16 Join reacocard [0] (n=reacocar@
10:55:12***Saving seen data "./dancer.seen"
10:56:02JdGordonany objections to FS #9460? (volume button tag)
11:00:58vitjaamiconn: may be we need generic disk layer, or we can still live w/o it?
11:02:12 Part pixelma
11:02:42 Quit massiveH ("Leaving")
11:02:47 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
11:03:35vitjawith all related oop
11:05:12 Join gregzx [0] (
11:05:36 Quit culture (Connection timed out)
11:09:39 Quit JdGordon ("Konversation terminated!")
11:11:39 Join JdGordon [0] (
11:12:30 Part Waldo000000
11:19:47 Join PaulJam [0] (
11:35:49 Join fragilematter [0] (n=fragilem@
11:37:39 Join einhirn [0] (
11:39:43 Nick zodttd|away is now known as zodttd (
11:43:54 Join fyre^OS [0] (
11:53:39 Join Jabone [0] (
11:54:18 Join petur [50] (n=petur@rockbox/developer/petur)
11:54:18 Join kushal_12_27_200 [0] (n=kushal@
11:54:43 Nick zodttd is now known as zodttd|away (
11:57:42 Quit fyrestorm (Read error: 110 (Connection timed out))
11:57:58kushal_12_27_200Hi, I have a sansa c250 with the latest build of rockbox. I think I have run into a problem with the database. Songs are listed tens of times in the database where there is only one song. Is it because I enabled auto update of database? Any ideas?
11:58:11 Join Waldo000000 [0] (
11:58:55 Join Xerion_ [0] (
12:00:56 Quit fragilematter ("Leaving.")
12:01:09PaulJamkushal_12_27_200: could you put the duplicates in a playlist and open the playlist in a texteditor to see if the paths differ?
12:01:36kushal_12_27_200sure thing
12:02:46 Quit JdGordon (Read error: 104 (Connection reset by peer))
12:03:50 Quit Xerion (Read error: 104 (Connection reset by peer))
12:03:50 Nick Xerion_ is now known as Xerion (
12:04:29 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
12:05:26 Join JdGordon [0] (
12:05:58kushal_12_27_200same files,
12:09:42PaulJamhmm, then i have no idea. i use auto update without such problems. you should be able to fix this by re initializing the database (export your database changes if you use the gather runtime feature).
12:09:51kushal_12_27_200Initializing database now ...
12:11:30kushal_12_27_200Worked! Sorry for the trouble.
12:11:46kushal_12_27_200(had to restart the player)
12:12:03 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:12:45JdGordonhey Nico_P
12:13:17JdGordondoes the wps parser have any mechanism to let conditionals have parameters?
12:13:37JdGordonlike.... %?v|1|2|<fd...|..|..>
12:14:08Nico_Pmaybe evaluate_conditional could be modified to allow it
12:17:17 Quit PaulJam (
12:17:17 Quit reacocard (
12:17:17 Quit jeffdameth (
12:17:17 Quit Ac-town (
12:17:17 Quit Twst (
12:17:17 Quit mak3 (
12:17:17 Quit Nimdae (
12:17:17 Quit homielowe (
12:17:17 Quit super (
12:17:17 Quit DismalArcadia (
12:17:17 Quit jfc (
12:17:17 Quit HellDragon (
12:17:17 Quit Bagder (
12:17:17 Quit tarbo (
12:17:21NJoinPaulJam [0] (
12:17:59 Join linuxstb [0] (
12:18:01NJoinreacocard [0] (n=reacocar@
12:18:01NJoinjeffdameth [0] (
12:18:01NJoinAc-town [0] (
12:18:01NJoinTwst [0] (
12:18:01NJoinmak3 [0] (
12:18:01NJoinDismalArcadia [0] (
12:18:01NJoinsuper [0] (
12:18:01NJoinBagder [0] (
12:18:01NJoinjfc [0] (
12:18:01 Join homielowe [0] (n=homielow@unaffiliated/homielowe)
12:18:01NJoinNimdae [0] (
12:18:01NJoinHellDragon [0] (n=jd@Wikipedia/HellDragon)
12:25:10 Join fragilematter [0] (n=fragilem@
12:27:20 Join homielowe_ [0] (
12:30:29 Quit homielowe (Read error: 104 (Connection reset by peer))
12:32:51 Join moos [0] (
12:40:04 Quit gregzx ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
12:41:59 Part Waldo000000
12:43:32 Join nuonguy [0] (
12:45:54NJointarbo [0] (n=me@unaffiliated/tarbo)
12:46:15 Quit fragilematter ("Leaving.")
12:48:10 Quit Nimdae (Remote closed the connection)
12:48:13 Join Nimdae [0] (
12:50:47 Join mak3_ [0] (
12:51:20 Join Twisty [0] (
12:51:22 Quit reacocard (Connection timed out)
12:51:24 Quit mak3 (Remote closed the connection)
12:52:23 Join super_ [0] (
12:53:18 Join massiveH [0] (
12:55:14***Saving seen data "./dancer.seen"
12:55:17 Join reacocard [0] (n=reacocar@
12:56:36 Join jeffdameth1 [0] (
12:57:53 Quit jeffdameth (Read error: 110 (Connection timed out))
12:59:45 Join druidu [0] (n=d5a4e705@gateway/web/cgi-irc/
13:01:16 Quit Twst (Read error: 110 (Connection timed out))
13:03:20 Quit super (Read error: 110 (Connection timed out))
13:05:57druiduhello there! anybody here familiar with arm platform?
13:08:04markunyes, a few of us are
13:09:12markundruidu: any specific question?
13:12:27 Quit Ac-town (
13:12:27 Quit DismalArcadia (
13:12:27 Quit HellDragon (
13:12:27 Quit Bagder (
13:12:27 Quit jfc (
13:12:29GodEaterdionoea: you about ?
13:12:46 Join cactusy1 [0] (i=xnte@gateway/gpg-tor/key-0xB9002659)
13:13:00 Join Tsukasa-Ujiie [0] (
13:13:09NJoinAc-town [0] (
13:13:09NJoinDismalArcadia [0] (
13:13:09NJoinBagder [0] (
13:13:09NJoinjfc [0] (
13:13:09NJoinHellDragon [0] (n=jd@Wikipedia/HellDragon)
13:13:30Tsukasa-Ujiiewow alot of people just joined
13:14:04massiveHlooks liek a netsplit Tsukasa-Ujiie
13:14:19Tsukasa-Ujiiei jsut joined aswell... that mgiht ahve been it
13:15:12scorchethat *is* it...
13:16:12 Quit massiveH ("Leaving")
13:21:42druidumarkun: so, I'm trying to get started on a TCC8200 platform
13:21:45druiduiRiver X20 that is
13:22:30druiduand I managed to use usb boot mode to upload a firmware and I know a sequence of instructions to reset the usb driver which disconnects the usb, so I can no something is happening and when
13:23:05druiduthe thing is, if I write the code in assembler in a .S file −− GNU arm-elf-gcc, it works
13:23:19druiduif I link it with a .c file and write the coresponding code in C, it doesn't
13:23:35druiduso I'm not sure how to interface assembler with C code
13:23:41druiduI just use BL function
13:25:13druiduvoid function( ) { *(volatile unsigned long *) 0x80004018 = 0xFFFFFFFF; *(volatile unsigned long *) 0x80004018 = 0; }
13:26:00druiduI even tried to compile with -S and put all the code in the function there, no BL instruction, and it worked
13:26:12druiduso there must be something with the function prologue or something
13:28:30B4gdertried objdumping the C compiled version?
13:28:46B4gderI mean to compare with the working asm version
13:28:59druiduI loaded the generated file in IDA
13:29:18druiducode looks ok, should work
13:29:36druidulemme try objdump
13:30:31 Join LambdaCalculus37 [0] (
13:30:52druiduI have:
13:30:57druidueb000140 bl 200005a8 <kernel_init>
13:31:07druidu200005a8 <kernel_init>:
13:31:13druidue1a0c00d mov ip, sp
13:31:15linuxstbdruidu: Please use pastebin for pasting example code...
13:32:58linuxstbdruidu: Have you set up the stack pointer? If you're branching to a C function, the C function will probably expect a stack to be there.
13:33:40druidulook at
13:33:49druiduaddress 20000084
13:34:03druiduI copied the 0xdeadbeef code from rockbox
13:34:55druiduI'm 100% sure the code after the BL gets executed
13:35:06druiduif I put the usb reset code after the BL, it still runs
13:35:52druidubasicly, kernel_init should write value 0xFFFFFFFF to address 0x80004018 then write 0 to same address
13:38:15 Join J-23 [0] (
13:38:29druiduI just don't see a reason why that shouldn't work
13:39:12 Join dany_21a_ [0] (
13:39:28linuxstbCan you paste the .S and .C code? That may be easier to follow.
13:39:32 Quit Tsukasa-Ujiie ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
13:39:42linuxstbAnd also whatever Makefile you are using to compile/link/convert to binary.
13:42:16druiduah, wait
13:43:26 Quit Ac-town (
13:43:26 Quit DismalArcadia (
13:43:26 Quit HellDragon (
13:43:26 Quit Bagder (
13:43:26 Quit jfc (
13:44:04druiduincludes the too
13:44:29linuxstbdruidu: I'm not sure if I understand your problem. Are you saying that if you put the C code in your assembler file, USB is being disabled, but when it is in the C function, it doesn't work?
13:44:47JdGordonany ideas which file i need to fiddle with to get rid of "make[1]: *** No rule to make target `/home/jonno/rockbox-sdl/kaiser/firmware/crt0.o', needed by `/home/jonno/rockbox-sdl/kaiser/librockbox.a'. Stop." (trying to build firmware for a sdl non sim target)
13:45:16linuxstbJdGordon: I guess you need to edit firmware/SOURCES to exclude crt0.S
13:45:38NJoinAc-town [0] (
13:45:38NJoinDismalArcadia [0] (
13:45:38NJoinBagder [0] (
13:45:38NJoinjfc [0] (
13:45:38NJoinHellDragon [0] (n=jd@Wikipedia/HellDragon)
13:46:04JdGordonlinuxstb: hmm, is there any way to see which line in SOURCES its getting it from?
13:46:26druidulook in crt0.S, I added the code in asm that works and does the reset
13:47:19druiduso, If I uncomment that code, I get a reset, otherwise no reset
13:47:59linuxstbJdGordon: "crt0.S" (without a path) only appears once in SOURCES...
13:48:55JdGordonah yes.. got it
13:53:18 Quit LambdaCalculus37 ("Ka-chunka")
13:53:53 Join Bagder_ [0] (
13:54:03druiduI also tried adding the assembed code from kernel.c directly into crt0.S and compiling without kernel.c and it also works
13:54:18druiduthere is something happening with the function call / prologue
13:58:46linuxstbdruidu: Referring to this patch - - what's in memory locations 0x20004ec, 4f8 and 4fc ?
13:58:57JdGordon/home/jonno/SDL-1.2.6/include/SDL_rwops.h:73: error: expected specifier-qualifier-list before ‘FILE’ :(
13:59:21JdGordonthe sdl includes cant find the system stdio.h?
13:59:37linuxstbOr they're finding the Rockbox stdio.h first...
14:00:01linuxstbI'm assuming you haven't defined SIMULATOR for this build?
14:00:12JdGordoni havnt
14:00:21 Quit Bagder (No route to host)
14:00:21JdGordonI've butchered up the Makefile a fair bit
14:01:34JdGordonhmm.. seems the wrong compiler is being used anyway
14:03:33 Join soap [50] (n=soap@rockbox/staff/soap)
14:04:47 Quit Ac-town (
14:04:47 Quit DismalArcadia (
14:04:47 Quit HellDragon (
14:04:47 Quit jfc (
14:04:59NJoinAc-town [0] (
14:04:59NJoinDismalArcadia [0] (
14:04:59NJoinjfc [0] (
14:04:59NJoinHellDragon [0] (n=jd@Wikipedia/HellDragon)
14:05:58druiduthere are pointers to stack_start and stack_end
14:08:25JdGordonarg, yeah I've broken something and sdl isnt finding the right .h :'(
14:09:09JdGordoneven adding -I/usr/include/ to GCCOPTS doesnt help?
14:11:33 Join ownxor [0] (n=453eba4c@gateway/web/cgi-irc/
14:12:47linuxstbJdGordon: If I was you, I would define SIMULATOR, and then once you get something working, work through Rockbox changing it appropriately.
14:14:05ownxorjoin #rockbox-community
14:14:07JdGordonI'm trying to build a "bootloader" build and it seems to get thourgh firmware/ without a problem untill the sdl includes start
14:14:32*scorche hands ownxor a "/"
14:15:37JdGordonseems the Makefile in uisimulator/sdl rearranges the includes so hopefulyl i just need to copy that somehow
14:15:37 Join webguest49 [0] (n=db8540c8@gateway/web/cgi-irc/
14:15:47 Quit webguest49 (Client Quit)
14:15:48linuxstbJdGordon: Why a bootloader build?
14:17:14JdGordonbecause the hard part is getting firmware to build... I guess once I have a build working with simple lcd then switching to a full build shouldnt be hard
14:17:20linuxstbdruidu: IIUC, stack_begin and stack_end are the same value - the word after your function. So the stack would be overflowing into kernel_init()
14:17:22JdGordonsame as any new target :p
14:18:26 Quit dany_21a_ (Remote closed the connection)
14:18:36druiduno theyaren't, it's 200025d4 vs 200005d4 :)
14:19:03linuxstbdruidu: OK, I'm going blind...
14:19:22druiduit's hard to see... anyway, SP is initialized to 200025d4 so there's plenty of space
14:19:32druidunow I tried this: jumped in the middle of the function
14:19:48 Quit Ac-town (
14:19:48 Quit DismalArcadia (
14:19:48 Quit HellDragon (
14:19:48 Quit jfc (
14:20:37 Join dany_21a_ [0] (
14:20:49*B4gder identifies the "most complex" functions in rockbox
14:21:02B4gderwith pmccabe
14:21:11B4gderand the winner is...
14:21:16B4gderapps/plugins/brickmania.c(1267): game_loop
14:21:30B4gderand the to firmware function is firmware/mpeg.c(1171): mpeg_thread
14:21:35JdGordonand one that isnt in plugins/ or codecs?
14:21:36B4gderthe top even
14:21:40 Part dany_21a_
14:21:51druidulook at 200000a0... so it looks like the code in the function is bad, it doesn't write 0xFFFFFFFF at address 0x80004018 and then 0 at same address
14:21:56B4gderthe top swcoded firmware funtion: firmware/common/dircache.c(553): dircache_do_rebuild
14:22:09linuxstbJdGordon: I guess I don't see the problem the same as you. I think it would be relatively easy to get the apps, uisimulator and firmware code compiling for your target (as a sim of a faked target). So I would do that first, and then once I have something running, gradually work on changing things so it's no longer a sim, but a target in its own right.
14:23:42 Join MarcGuay [0] (
14:24:05 Quit ownxor ("CGI:IRC")
14:24:41NJoinAc-town [0] (
14:24:41NJoinDismalArcadia [0] (
14:24:41NJoinjfc [0] (
14:24:41NJoinHellDragon [0] (n=jd@Wikipedia/HellDragon)
14:25:16linuxstbdruidu: I'm not sure what you're seeing - the compiled C code looks fine.
14:25:55druiduok, i just realized what must happen
14:25:57MarcGuayAny thoughts on how to modify the manual to make RBUtil the recommended install method? Having links to the GraphicalInstall or RockboxUtility wiki pages doesn't seem so hot, but possibly inevitable. Unless we can have direct download links from latex, which we probably can...
14:26:02druidumy code must erase the other code or smth
14:26:11druiduthe bss init code and stack init code
14:27:11druidu all before line 35
14:27:22druiduthat's where the problem should be
14:29:24 Quit vitja (Remote closed the connection)
14:29:42 Quit trisiak ("Changing server")
14:29:54 Quit Ac-town (
14:29:54 Quit DismalArcadia (
14:29:54 Quit HellDragon (
14:29:54 Quit jfc (
14:30:03NJoinAc-town [0] (
14:30:08 Join HellDragon [0] (
14:30:11NJoinDismalArcadia [0] (
14:30:15NJoinjfc [0] (
14:30:31 Join pfwsmbkwt [0] (n=pfwsmbkw@
14:31:01 Join trisiak [0] (
14:34:48 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
14:35:45druidui'm new to arm assembler, in lines 30-33 which instruction updates the counter ?
14:35:59 Part J-23
14:36:53 Quit Xerion (" ")
14:37:15 Quit petur ("*plop*")
14:38:06 Quit AndyIL ()
14:43:08MarcGuayManual stuff: "[The original BL] is also responsible for the disk mode on your \dap{}." This doesn't apply to the Sansas, would "any emegency, recovery mode, or disk modes" be broad enough?
14:43:08 Quit colesterol_dog (Read error: 60 (Operation timed out))
14:43:48*MarcGuay searches for a clean simplification.
14:44:27 Join colesterol_dog [0] (
14:47:16 Quit Ac-town (Read error: 113 (No route to host))
14:47:47LambdaCalculus37MarcGuay: I believe that "any emergency, recovery mode, or disk modes" should cover the Sansas.
14:48:33MarcGuayLambdaCalculus37: And leave room for future developments. Merci!
14:49:47LambdaCalculus37MarcGuay: You're welcome. :)
14:50:16 Join AndyI [0] (i=AndyI@
14:51:02 Join olegfink [0] (
14:52:13olegfinkhi, I've got a few questions about VX767 and rockbox on it
14:53:07olegfinkfirst, seems that here in russia we have a localised oem version of vx767 called qumo vision, but apparently it doesn't have a touchscreen.
14:53:23olegfinkbut its firmware is named VX767.hxf
14:53:38 Join Xerion [0] (
14:53:59olegfinknext, I haven't quite got what's the status of rockbox on vx767
14:54:49olegfinkand, after all, the most silly question: is rockbox a port of linux kernel or is it a completely independent piece of software?
14:55:00 Quit kushal_12_27_200 ("Leaving")
14:55:18***Saving seen data "./dancer.seen"
14:55:20 Join J-23 [0] (
14:55:42LambdaCalculus37Rockbox is not a port of Linux; it is its own kernel. There is a little POSIX influence in the kernel, but that's about it.
14:56:25olegfinkaha, so basically you port everything you want to run on rockbox?
14:56:49LambdaCalculus37Read this:
14:57:49olegfinkgot it, not that shiny but still something
14:58:35MarcGuayLambdaCalculus37: How familiar are you with latex? I'm seeing "\index{Speech}" in the installtion section but no indication that it's actually creating an index for Speech. ?
14:59:25olegfinkdo you have, say, an fb2 viewer?
14:59:31MarcGuay\subsection{blah} seems to be the way to add things to the indexes...
14:59:42olegfinkfor such a luxury screen, vx767 got just a lame txt viewer
15:00:53 Quit cactusy1 (Nick collision from services.)
15:00:56 Join Maha` [0] (i=xnte@gateway/gpg-tor/key-0xB9002659)
15:01:17LambdaCalculus37MarcGuay: I guess my skills in LaTeX would be "intermediate", but I'm not sure what the "\index{Speech}" function is for. You could ask pixelma; she might know what those do.
15:01:34LambdaCalculus37Where are those located?
15:03:45MarcGuayLambdaCalculus37: Check the getting_started/installation text. Another weird one is \index{Bootloader} next to the Rockbox bootloader section. Seems to no longer be indexed as such anywhere I can see...
15:06:51LambdaCalculus37MarcGuay: I'm looking through the file now.
15:07:59LambdaCalculus37MarcGuay: There is one thing I've noticed... getting_started/installation.tex still has a blurb about Rockbox 2.5. Want to change it to 3.0?
15:09:03LambdaCalculus37And also change the \opt{swcodec} under the "Release" section to also mention v3.0.
15:09:12 Join dany_21a_ [0] (
15:10:52linuxstbolegfink: There is only a lame text viewer in Rockbox as well (at the moment,...)
15:12:28GodEaterwhat's an fb2 viewer anyway ?
15:12:43linuxstbGoogle tells me "fiction book 2.0" - I'm assuming some kind of ebook format.
15:13:04olegfinkyep, an overengineered (as all open source tends to be these days) format for ebooks
15:13:25olegfinkthere are quite a lot books in it, and many implementations for all major desktop and mobile platforms
15:14:09 Quit robin0800 (Read error: 104 (Connection reset by peer))
15:15:37olegfinklinuxstb: at least there's some hope I can write my own for rockbox
15:16:47linuxstbSo a ".fb2" file is an XML document?
15:17:19*B4gder disagrees about open source tending to be overengineered
15:17:32*linuxstb thinks XML is a perfect format for structured documents
15:18:24olegfinkoh well
15:18:35olegfinkdo you have an xml parser for rockbox?
15:19:57linuxstbolegfink: The biggest problem will probably be that the Rockbox LCD drivers currently only support loading a single font face. So that aspect probably needs fixing first...
15:19:58B4gderif there's a lean and mean one around I suppose it shouldn't be too hard to port
15:20:11olegfinkso first write one, make sure it runs on all your devices with reasonable parse time, then we'll speak about how good xml is.
15:21:51linuxstbAnother problem is available RAM - a Rockbox plugin only has access to between 32KB (on low memory targets) to 512KB (on most targets) without stopping audio playback. And accessing the hard disk uses power, so you would need to minimise disk access.
15:22:16linuxstbThat includes both code and data for the plugin.
15:22:37 Join robin0800 [0] (
15:23:34 Nick Maha` is now known as Maha (i=xnte@gateway/gpg-tor/key-0xB9002659)
15:28:22amiconnlinuxstb: The number of fonts has nothing to do with the lcd drivers
15:28:52markunolegfink: I think mculenaere is the one to talk to about the vx767 port. He
15:28:53amiconnIt's the font loader that would need to be extended
15:29:15markunolegfink: He's usually here every day.
15:29:22linuxstbamiconn: I was obviously simplifying it...
15:30:57markunolegfink: are you planning to help out with the port?
15:35:04*amiconn wonders what happened to markun's font caching rework ideas...
15:36:23 Join miepchen^schlaf [0] (
15:36:38 Join janne_oksanen [0] (
15:36:45janne_oksanenhey guys
15:36:54janne_oksanenI have a brand new 80GB 5.5th gen here
15:37:06janne_oksanenand I'm having trouble installing the bootloader
15:37:23janne_oksanenit says "ipod can't be opened" or something like that
15:37:34janne_oksanenany idea as to what I'm doing wrong?
15:37:51JdGordonrun as root or admin
15:37:59janne_oksanenoh right
15:38:12linuxstbAlso, it it definitely a 5.5th gen, and not a "Classic" ?
15:38:19 Part LinusN
15:38:26janne_oksanenyes, it's definitely 5.5th gen
15:38:45janne_oksanenshiny front panel with sharp edges
15:39:29LambdaCalculus37Is it a thin iPod, or a little fatter?
15:39:47janne_oksanenpretty fat
15:39:57janne_oksanen13mm maybe
15:40:00JdGordonisnt it a different error if its not a recognised ipod?
15:40:33LambdaCalculus37Are you using Rockbox Utility, or the manual install method?
15:40:54janne_oksanenrockbox utility in ubuntu
15:41:09janne_oksanenand running it with sudo seems to have done the trick
15:41:56druidulinuxstb: is it possible that objcopy -O binary messes sections up so that addresses don't match anymore?
15:42:46janne_oksanenyea baby yea! rockbox in the house :D
15:42:50janne_oksanenthanks guys
15:46:00 Quit Seed ("cu, Andre")
15:49:10janne_oksanenmy USB hard drive is crapping out on me now
15:49:42janne_oksanenI think the ipod is pulling too much juice and is messing up the bus or something
15:51:48B4gderI doubt that
15:52:00B4gderyou're still using apple's usb mode
15:52:18 Quit moos ("Rockbox rules the DAP world")
15:53:05janne_oksanenshould I do something about it?
15:53:47B4gderyou can't do anything about it, but my point was more that Rockbox didn't change anything of that
15:54:02janne_oksanenah I see
15:54:06janne_oksanenI wan't saying that
15:54:15B4gderso the "crapping out" was more likely due to something else
15:54:44janne_oksanenwhat I meant is that my crappy motherboard can't handle me charging the ipod and using the USB hard drive at the same time
15:55:23B4gderbut is this the first time you charge your ipod over usb then?
15:55:35janne_oksanenI just got this thing
15:55:37B4gderah ok then, then I see
16:03:02linuxstbdruidu: Disassembling the binary should show if you that's the problem though. Does it work if you remove the loops that zero the bss and writes 0xdeadbeef to the stack?
16:14:54 Join obo [0] (n=obo@rockbox/developer/obo)
16:23:28 Nick fxb__ is now known as fxb (
16:24:09 Quit Maha (Remote closed the connection)
16:26:09 Quit robin0800 (Remote closed the connection)
16:28:52 Join funman [0] (n=fun@
16:29:34olegfinkmarkun: for now, I'm considering if this is worth bying
16:30:01olegfinkIf I figure out it is, I will certainly have no option other than use rockbox, so yes.
16:31:58 Part unstable
16:34:05markunolegfink: the purchase of my first player (iriver h120) was also influenced by a rockbox port being in progress :)
16:34:22janne_oksanenin case anyone cares turing the power off/on on the external HDD fixed my issue
16:35:14 Quit Zagor (Read error: 110 (Connection timed out))
16:35:43funmanthe PL180 datasheet mentions that it uses MCLK clock, but I can't find any setting for this clock in the AS3525 datasheet
16:36:03 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
16:36:35funmanThe AS3525 also mentions I2SI_MCLK and I2SO_MCLK (I2S in / out)
16:36:52 Join meven [0] (n=meven@
16:37:41funmanIn the PL180 datasheet I see this relation: fp >= 3/8 fm , where m = MCLK and p = PCLK; but no mention of what f is (frequency?)
16:37:53 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
16:38:32 Join {phoenix} [0] (
16:38:57olegfinkmarkun: can you point me to any piece of code in SVN being a port of something POSIXish?
16:39:12olegfinkI want to estimate the effort of porting.
16:40:45markunolegfink: ok, let me take a look
16:41:04 Quit J-23 ()
16:42:32markunolegfink: look at some of the header files in here:
16:44:32olegfinkI may be mistaken, but for me that looks like ANSI C headers
16:44:33 Part janne_oksanen
16:45:35markunyes, rockbox is mostly written in ANSI C. To be quite honest, I dont't know exactly what the POSIX standard contains.
16:46:33*markun googles
16:46:52olegfinkthreads, sockets, and other stuff you'd expect from some unix variant
16:46:56B4gderansi and posix overlaps quite a bit
16:47:11Zagorolegfink: we have none of those parts of posix
16:47:35olegfinkin a nutshell, everything that makes it impossible to port GNUish software to sane OSes
16:47:35 Quit amiconn (Nick collision from services.)
16:47:41 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
16:48:05Zagorour posix compatibility is limited to a number of common functions
16:48:26olegfinkaha, so I can't take, say, mplayer and compile for rockbox?
16:48:30 Join robin0800 [0] (
16:48:35Zagorolegfink: hehe, no
16:48:57B4gdermplayer of course being a good example of something that uses a lot of stuff beyond posix...
16:49:23B4gderrockbox is a small OS for tiny embedded things
16:49:26olegfinkwell, the only thing that justifies rockbox is absence of network connectivity on most players :-)
16:49:45B4gderrockbox is crafted to be a music and media player OS
16:50:07B4gdersee how the ipodlinux fails completely on ipod
16:50:36olegfinkwell, it makes sense after all
16:50:46markunB4gder: but the newer ipods are also based on a POSIX like system (OSX), right?
16:50:53olegfinknot that I'm a big fan of making more OSes
16:51:07B4gderit's already made
16:51:17olegfinkI'd start with something like plan9, but, well, that, irrelevant.
16:51:20B4gdermarkun: yes, so they say
16:51:26olegfinkright, so, you don't have some reader?
16:51:33B4gderolegfink: read up on the history and you might learn why
16:51:34markunwe have some reader
16:52:19olegfinkand it does better than 'the' reader on onda?
16:52:23 Join toffe82 [0] (
16:53:02markunI've never seen 'the' reader, but I assume it's not any better.
16:53:03olegfinkfor a 4.3" screen, the very same piece of code that is used on all 1.5" cheap mp3 things looked a bit... disgusting.
16:53:22markunIt's just a simple text viewer with support for some codepages and bookmarks
16:53:32 Part B4gder
16:53:52olegfinkwell, the thing about rockbox is that I have the source.
16:54:57olegfinksomeone here said about 32K limit for application memory. Is that a hard limit imposed by rockbox or just an experimental fact?
16:55:00markunolegfink: maybe with the help of the rockbox reverse engineering you could post mips linux to the onda, but then again, maybe it's easier to buy a linux based DAP.
16:55:22***Saving seen data "./dancer.seen"
16:55:25markunpost -> port
16:55:53linuxstbolegfink: 32KB is for the lowmem targets. Most targets have 512KB. That's an arbitrary limit (can be changed easily per-target), but Rockbox is designed to maximum the space available for buffering audio.
16:56:31 Quit robin0800 (Remote closed the connection)
16:56:39 Join Phalangees [0] (i=8a31a842@gateway/web/ajax/
16:57:04 Join robin0800 [0] (
16:57:38 Quit Zagor ("Client exiting")
16:57:49olegfinkaha, seems vx767 has 64M ram
16:58:04markunquite a lot
16:58:09olegfinkI think something will still be available for apps.
16:58:32markunsince it's a flash player you don't need much RAM for audio buffering anyway
16:58:38olegfinkwell, sure I can buy a linux pmp, or better just a nokia n810
16:58:48olegfinkbut that onda thing is _very_ cheap
16:58:54PhalangeesHello. I just bought a giagabeat s60 and I was wondering where the most up-to-date rockbox installation instructions are.
16:59:34funmanPhalangees: the front page quotes "Toshiba: Gigabeat X and F series (not the S series)"
16:59:56markunfunman: but he knows that we are working on a port :)
17:00:24funmanah sorry :)
17:00:49LambdaCalculus37Phalangees: Here's the Gigabeat S install page:
17:01:28 Quit meven (Remote closed the connection)
17:01:45PhalangeesThere is a part about "loading code from windows" on here too:
17:01:57PhalangeesI was just wondering which one is more accurate.
17:02:16 Quit druidu ("CGI:IRC (EOF)")
17:04:08 Quit pixelma2 (Read error: 110 (Connection timed out))
17:04:16 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
17:06:06MarcGuayPhalangees: I'd follow the installation instructions page. Keep in mind you can't dual boot and rockbox can't charge (without appying a patch), so it's a bit of a suicide case.
17:06:39PhalangeesI'll just build it with the patch. Thanks for your help!
17:07:04markunMarcGuay: why is the patch not in svn? Is it a bit of a hack?
17:07:30MarcGuaypixelma2: Could you help me understand the use of \index{whatever} in the manual? There are a few good example is in gettingstarted/installation.tex.
17:07:39MarcGuaymarkun: Not sure. Maybe a WIP?
17:07:56 Join Seed [0] (
17:12:42MarcGuaymarkun: It's got some issues, yes. It pulls a charge when connected via USB but the energy is spent because the disk is constantly spinning. AC charging is fine.
17:13:34 Quit funman ("leaving")
17:13:51MarcGuayPerhaps it's better now that FS #8943 is in.
17:16:51 Join bughunter2 [0] (n=Jelle@
17:17:53 Quit perrikwp (" ajax IRC Client")
17:18:21 Join dabujo [0] (
17:19:24 Join cactusy1 [0] (i=xnte@gateway/gpg-tor/key-0xB9002659)
17:32:26 Join DerDome [0] (n=DerDome@
17:32:42 Quit DerDome (Client Quit)
17:35:51 Quit Phalangees (" ajax IRC Client")
17:36:30 Join J-23 [0] (
17:41:13 Join n1s [0] (n=nils@rockbox/developer/n1s)
17:42:28 Join druidu [0] (n=d5a4e705@gateway/web/cgi-irc/
17:42:56druiduok, I've been thinking... let's say I have an ARM processor, on the TC8200 platform
17:43:14druiduwhat are the steps to get it started so that I can call a C function?
17:46:46markundruidu: what we do in our crt0.S files is call a C function 'main' without any arguments and then call the other C functions from there.
17:48:21olegfinkmarkun: is there any chance of mculenaere in a next couple hours? that is, is it too early or too late for them?
17:50:25markunolegfink: it's 17:50 here. Maybe he's at university.
17:51:15 Quit sarixe ("Ex-Chat")
17:51:28 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
17:51:35markunolegfink: there he is :)
17:51:52mcuelenaereolegfink: you had some questions regarding the VX767?
17:51:56mcuelenaere(haven't read the logs yet)
17:53:53 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:55:12 Quit druidu ("CGI:IRC (EOF)")
17:57:14olegfinkmcuelenaere: yeah, those questions can be packed in one: "tell me about vx767 and its affair with rockbox"
17:57:31mcuelenaereit's just one step behind the vx747 and rockbox
17:57:37mcuelenaereit doesn't have a lcd driver
17:57:47mcuelenaerebut the vx747 rockbox port isn't very advanced though
17:57:48markunyet :)
17:57:52mcuelenaereyet :)
17:57:59mcuelenaereso in total, there are some good fundamentals
17:58:11olegfinkso the kernel boots?
17:58:12mcuelenaerecurrently the blocker is the Flash Translation Layer
17:58:14markunmcuelenaere: is everything in svn?
17:58:22mcuelenaerewell about everything
17:58:29mcuelenaeresome testing stuff is in my local tree
17:58:55olegfinker, what's FTL?
17:59:06mcuelenaereFlash Translation Layer ;)
17:59:23olegfinksure, but what does that mean feature-wise?
17:59:25 Quit linuxstb (Read error: 110 (Connection timed out))
17:59:32 Join petur [50] (n=petur@rockbox/developer/petur)
17:59:56mcuelenaereit's the layer which translates physical NAND blocks/pages to sectors (on which the FAT partition is located etc)
18:00:14mcuelenaereNAND chips have some limitations because pages wear out after some time
18:00:21mcuelenaereand the FTL needs to keep track of that
18:00:26olegfinkaha, so you basically can't access NAND on vx767?
18:00:31mcuelenaereon the vx747
18:00:44mcuelenaereI haven't tested the vx767 because I don't have one
18:00:49mcuelenaerebut it should work
18:00:54mcuelenaerethere's just no working LCD driver
18:00:59olegfinkdoes someone have it?
18:01:07mcuelenaereand you
18:01:09olegfinkI'm considering getting one
18:01:13olegfinkso not me for now
18:01:15mcuelenaereah ok
18:01:26mcuelenaerebut we can access NAND
18:01:29mcuelenaerejust not what's on it
18:01:41markunolegfink: do you think it would be more fun to port MIPS Linux?
18:01:54mcuelenaereand the other drivers shouldn't be that hard to port as there's source code
18:02:20olegfinkmarkun: I don't have any idea of vx676's hardware
18:02:25mcuelenaereUSB already works (compared to other new ports where it doesn't)
18:02:34mcuelenaereolegfink: have you read the wiki pages?
18:02:46olegfinkon vx747 and vx767? yes.
18:02:56mcuelenaerethe hardware description is there
18:03:12olegfinkoh, so you can already do some debugging over usb?
18:03:23wpyhmcuelenaere: yeah?
18:03:26mcuelenaereerr yes could be
18:03:33mcuelenaerewpyh: you have a vx767, right?
18:03:46 Join Schmogel [0] (
18:03:51olegfinkcool, could you tell in short how much does it rock?
18:04:00mcuelenaerethe OF doesn't
18:04:06wpyholegfink: it doesn't rock−−- yet
18:04:08 Quit Bagder_ (Read error: 110 (Connection timed out))
18:04:09mcuelenaerebut that's rather off-topic
18:04:38mcuelenaereolegfink: join #rockbox-community if you want to exchange user experiences
18:04:50olegfinkaha, will do
18:05:08mcuelenaerebut on topic: do you have any embedded hardware skills?
18:05:13mcuelenaereor programming skills in general?
18:06:08olegfinkthe latter - yes, the former - I was participating (and probably still do) in linux for hp jornadas HPCs project, but didn't code anything for the hardware
18:06:38mcuelenaereok, but you could help out if willing?
18:07:56olegfinkyes, my plans for using vx767 are a bit perversive, so I'll have to do some coding anyway.
18:08:14mcuelenaerecool :)
18:08:14 Join AndyIL [0] (i=AndyI@
18:08:16 Quit AndyI (Read error: 110 (Connection timed out))
18:08:36mcuelenaerekeep in mind though that Rockbox doesn't have good video capabilities nor text reading capabilities
18:08:45mcuelenaereso you'll have to do quite a bit of porting for those if you'd want them
18:09:05mcuelenaere(and don't forget the dual-boot option, so it would be possible to use the OF for watching videos)
18:09:48 Join faemir [0] (
18:09:56markunmcuelenaere: but the vx767 is quite powerful, so it might make sense to improve our video player
18:10:02markunmore codecs, scaling
18:10:10mcuelenaerethere's even source code for that ;)
18:10:52 Quit Rob2223 (Connection reset by peer)
18:10:59markunI had the impression that even the mplayer hackers don't like the source code very much :)
18:11:44 Join DerDome [0] (
18:11:56olegfinkyeah, I really want text stuff
18:12:07mcuelenaeremarkun: you mean the Ingenic code or mplayer in general?
18:12:07olegfinkbut that doesn't employ too many embedded skills, right?
18:12:12mcuelenaerenope :)
18:12:20mcuelenaere_if_ Rockbox would work of course
18:12:37markunmcuelenaere: ah, didn't know there was Ingeic mplayer code
18:12:43olegfinkhehe, otherwise I'm left with a good player with crappy firmware.
18:13:02mcuelenaeremarkun: that's what I meant with source code ;)
18:13:06mcuelenaereIngenic provides quite a lot
18:13:10mcuelenaereeven hw accelerated ode
18:13:19olegfinkbut mplayer pulls its own copy, while VLC prefers shared libraries
18:13:22olegfinkat least on linux
18:14:49 Quit Schmogel (Read error: 104 (Connection reset by peer))
18:15:15MarcGuayNo current build for the Archos 8MB?
18:17:28 Join perrikwp [0] (i=98212ec9@gateway/web/ajax/
18:17:32 Join MethoS- [0] (
18:17:43 Join Tetracomm [0] (n=nicholas@
18:19:15 Join Schmogel [0] (
18:23:15 Quit perrikwp (" ajax IRC Client")
18:23:16LambdaCalculus37MarcGuay: There's one:
18:23:41MarcGuayLambdaCalculus37: Not listed on the current build page, though.
18:24:11LambdaCalculus37MarcGuay: Strange... I see it listed here.
18:24:18MarcGuayBut there is on the 3.0 page. The Archos models confuse me.
18:24:57MarcGuayArchos FM Recorder 8MB?
18:25:33LambdaCalculus37MarcGuay: The Archos 8MB Recorder builds are specially built for a hardware mod that people have done on the JBR to bump the internal RAM from 2MB to 8MB.
18:25:39 Join bertrik [0] (
18:27:59MarcGuayLambdaCalculus37: Okay. But do you see how there is a build for the Archos FM Recorder 8MB on the 3.0 page but not the current build? Is that normal?
18:29:09LambdaCalculus37MarcGuay: Ahh, now I see what you're talking about:
18:29:16*LambdaCalculus37 wonders about that...
18:29:48 Join herrwaldo [0] (
18:31:27 Join nicholas_ [0] (n=nicholas@
18:31:53 Join fragilematter [0] (n=fragilem@
18:32:13 Quit Tetracomm (Nick collision from services.)
18:33:21 Part fragilematter
18:33:55 Nick nicholas_ is now known as Tetracomm (n=nicholas@
18:39:37 Join vitja [0] (n=vitja@
18:42:20 Join thegeek [0] (
18:45:25 Join tvelocity [0] (
18:46:28 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:49:12 Quit kachna (No route to host)
18:49:18 Quit {phoenix} ("Konversation terminated!")
18:52:00 Quit thegeek_ (Read error: 113 (No route to host))
18:55:15 Join Kopfgeldjaeger [0] (
18:55:24***Saving seen data "./dancer.seen"
18:56:36 Quit cactusy1 (Remote closed the connection)
18:57:36 Join Thundercloud [0] (
18:57:37 Quit Thundercloud (Success)
18:57:52 Join Thundercloud [0] (
18:57:55 Join Rob2222 [0] (
18:59:36 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
18:59:56 Join cactusy1 [0] (i=xnte@gateway/gpg-tor/key-0xB9002659)
19:00:59 Join culture [0] (
19:04:20 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:06:26MarcGuayHrm. Anyone else having problems building manuals?
19:07:45MarcGuayenumitem.sty not found...
19:08:15bertrikseems to work fine here (built for sansa e200)
19:08:52MarcGuaybertrik: Thanks. Odd that that one file would be missing.
19:09:42bertrikenumitem.sty is part of texlive-latex-extra on my ubuntu 8.4
19:10:28MarcGuayI've built before without problems... But have been through some compooter issues since then...
19:13:24 Join gregzx [0] (
19:15:08 Quit robin0800 (Remote closed the connection)
19:15:22 Quit MethoS- (Remote closed the connection)
19:17:16 Quit pixelma2 ("-")
19:17:26 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
19:20:55 Join peerlessdeepak [0] (n=peerless@
19:23:11 Join Nibbler [0] (
19:24:00 Quit culture (Read error: 104 (Connection reset by peer))
19:24:27 Quit pixelma (" reboot")
19:24:34 Join culture [0] (
19:24:36 Quit culture (Client Quit)
19:24:41 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:24:58 Join culture [0] (
19:26:25 Join perrikwp [0] (i=9821448c@gateway/web/ajax/
19:29:05 Quit tvelocity (Read error: 60 (Operation timed out))
19:30:12 Quit wpyh (Read error: 110 (Connection timed out))
19:33:17 Join meven [0] (
19:34:43 Join gregzx_ [0] (
19:35:00 Quit gregzx (Nick collision from services.)
19:35:02 Nick gregzx_ is now known as gregzx (
19:35:22 Quit Thundercloud (Read error: 104 (Connection reset by peer))
19:35:27gevaertsI think we need to do multi-volume multi-driver ata soon, so that it's certain to be stable by the next release
19:36:15 Quit culture (Read error: 104 (Connection reset by peer))
19:36:30gevaertsor wait until after that, which is not that nice for ports that need it, like d2
19:37:58 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
19:38:45gevaertsMy current thinking is to prefix all implementations of functions in ata.h, and to make wrappers with the real names. Those wrappers could (and should, if we do it properly) get compiled away at least on single-volume targets
19:39:03 Join Thundercloud [0] (
19:39:30gevaertsAm I overlooking something, or should I just get started?
19:39:31shotofaddsgevaerts: I think preglow had some ideas about this while he was working on the D2 SD code. I don't know if anything was implemented though.
19:39:57gevaertspreglow: ping
19:40:01 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
19:40:01amiconngevaerts: I'd just leave ata_* as ata_*, and change the other driver
19:40:17amiconnmmc_*, sd_* etc.
19:40:40amiconnThe abstraction layer could then use storage_*
19:40:53amiconnFor single-driver targets, they would just be #define'd
19:40:59gevaertsGood idea.
19:41:32amiconnBtw, for the abstraction to work nicely, the ata driver needs some work
19:41:35*gevaerts will change ata_ temporarily too, to make sure he catches all callers
19:41:47gevaertsthe identify stuff?
19:42:01 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
19:42:15amiconnCurrently, ata_read_sectors() and ata_write_sectors() are limited to 256 sectors (65536 sectors when compiling with lba48 support *and* the drive actually uses lba48)
19:42:26 Join tvelocity [0] (n=tony@
19:42:27 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
19:42:31amiconnIt should be extended to allow an arbitrary sector count
19:42:36gevaertsah, yes. I seem to remember something about that
19:42:48amiconn(i.e. split internally into multiple ata transfers)
19:43:05 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:43:07amiconnThe MMC driver is safe; not sure about the sansa SD driver
19:43:14gevaertsIs there actually code that would exceed that?
19:43:26amiconnNot now, but I'd rather play safe here
19:43:44amiconnThe filesystem currently limts accesses to 256 sectors - for *all* targets
19:44:10amiconnThis code could be thrown out if the ata driver is extended to handle arbitrary sector counts
19:45:26gevaertsI'll start by renaming things and adding the storage_* wrappers.
19:46:11*gevaerts also doesn't really like #define NUM_VOLUMES 2 in ata.h
19:50:33preglowwhy didn't we just use doxygen?
19:50:47preglowgevaerts: i'm here
19:50:55amiconngevaerts: Multivolume/multidrive support needs restructuring
19:51:10amiconnWe should be able to support both, independently
19:51:32gevaertspreglow: do you have ideas/opinions about ata? (what was discussed the last 15 minutes or so)
19:51:52gevaertsamiconn: indeed. Still, one step at a time I think
19:52:19preglowah, no, what i have implemented so far is hacky
19:52:30 Quit gregzx (Read error: 104 (Connection reset by peer))
19:53:23preglowwell, we should just make some generic block functions, then define a table of operations for each driver
19:53:39preglowif there is only one driver, #define wrapper for the generic stuff that just uses ata_* or whatever
19:54:02preglowlimits to have fancy we should do it
19:56:04*gevaerts spots the first small hurdle : ata_spinup_time
19:56:10 Join Strife89 [0] (n=michael@
19:59:04amiconngevaerts: Did I say ata needs rework?
19:59:33amiconn*iirc*, ata_spinup_time was only made public for the debug menu
19:59:34gevaertsamiconn: I seem to remember something like that :)
19:59:52preglowwrap ata_spinup_timer in a function
20:00:07LambdaCalculus37Strife89: Beatbox is decrepit code at this point, IIRC.
20:00:14preglowit is indeed
20:00:33gevaertsmpeg.c and pcm_record.c use it
20:00:54 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
20:00:55Strife89This is one of those times when I'm kicking myself for not knowing C.
20:01:08amiconngevaerts: Ah yes, for calculating the watermark
20:01:31amiconnThen it needs to stay, and will just be a constant for non-ata targets
20:02:28gevaertsamiconn: it's not that simple. On multivolume the number can be different for different volumes
20:02:46bill2or3ata_spinup sounds vaguely related to a possible reason my H300 won't boot. (unless I go into USB Bootloader mode first)
20:02:54amiconnHmm, then it might need an access function
20:03:26gevaertsOf course you could make it the worst-case of all volumes
20:04:11 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
20:05:30bill2or3as best as I've been able to determine there must be some difference in the ata init code when you exit the usb-bootloaded, as compared to a regular boot.
20:05:36bill2or3if anyone has ideas, speak up.
20:07:21 Quit XavierGr (Read error: 104 (Connection reset by peer))
20:07:33 Join gregzx [0] (
20:07:50 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
20:08:44 Join meven_ [0] (
20:09:19preglowwhy not make it an access function? access functions are not evil
20:09:35preglowif it varies from driver to driver, then it should do so also in the api for quering it
20:10:27 Join janne_oksanen [0] (
20:10:37gevaertsAgreed, but suddenly mpeg.c and pcm_record.c will need to know which volume they are using. Yet another item for the todo list (but unavoidable I guess)
20:10:40janne_oksanenok, I may have a problem
20:11:16janne_oksanenafter about 4 hours of charging the new ipod and simultaniously uploading music I was able to transfer all my music in
20:11:48janne_oksanenthen I played around with some games and couldn't figure out a way to get out of the solitaire game
20:11:54 Quit XavierGr (Nick collision from services.)
20:12:05 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
20:12:05janne_oksanenso I shut down the ipod and not it's not responding to anything at all
20:12:55amiconngevaerts: Which drive, not which volume
20:13:10amiconn(but maybe those can be combined in a nice way)
20:13:13gevaertsAaarghh. This does indeed need cleanup!
20:13:19gevaertsamiconn: right :)
20:14:03gevaertsjanne_oksanen: try setting hold, then unsetting hold, and then resetting (menu+select for a while). After that, look for the manual ;)
20:14:11amiconnI would like to see the ability to specify how many volumes per drive to mount independently of how many drives to support
20:14:32gevaertsWe do want that, yes
20:15:05amiconnRight now, these values are combined. On Ondio we support 2 volumes, in order to make the MMC accessible
20:15:42amiconnBut if someone makes 2 partitions on the internal flash, the MMC would no longer be mounted on plug-in (due to no free volume descripotrs)
20:17:19janne_oksanengevaerts: there we go, thanks
20:17:24gevaertsIt's always easy to say "don't do that, then", but the same problem exists on the hotswappable cards, and there saying that gets a bit harder
20:17:40janne_oksanenfwiw I did read the battery FAQ
20:17:52gevaertsjanne_oksanen: I meant about exiting plugins :)
20:17:56amiconnThe cards aren't affected that much, because they don't come first
20:18:11gevaertsThat's true, but it's still annoying
20:18:54*gevaerts thinks that if we ever do usb host, we will have to make sure that this code is really clean before even starting
20:19:27amiconnThe typical case *is* just one volume per drive, and that's what we should try to mount in the official builds. But we should allow for special cases, and rockbox shouldn't mount 2 partitions of drive 0 and then have no more free descriptors for drive 1
20:20:31 Join lasser [0] (
20:22:05 Quit meven (Read error: 110 (Connection timed out))
20:23:25 Join kachna [0] (
20:26:05*n1s thinks he figured out what causes the obscure reversed-alphabetical-dir-sorting-after-deletinga-a-folder-which-we-later-try-to-resume-from
20:26:24*gevaerts gives n1s a beer
20:26:44n1sooh, cheers! :)
20:27:00n1sI'm not sure how to fix it though :/
20:29:09RyoSrewrite rewrite *SCNR*
20:31:08 Quit Nibbler ("Ex-Chat")
20:31:41 Join Nibbler [0] (
20:31:45gevaertsWhat does HAVE_FLASH_STORAGE mean? : (a) there is no ata_get_identify(), (b) there is no disk_spindown, (c) don't worry about buffer margins, (d) show a different disk activity icon, (e) all of the above
20:31:59n1sRyoS: the fix will require a bit of rethinking at least, probably not so much rewriting
20:32:34bertrikgevaerts, I think I've run into that one too, it's only defined for one (currently unsupported) target IIRC
20:33:04pixelmagevaerts: I believe that's only used for the ifp port, don't know why though
20:33:11gevaertsbertrik: no. That one is defined for lots of targets. The ifp7xx uses ata_flash.c, which is a different thing
20:33:48gevaertsAnd the answer is (e)...
20:33:51amiconnCan you spell 'mess'?
20:33:52Strife89gevaerts: I want to say (E),
20:34:04gevaertsm...e...s...s ?
20:34:17Strife89Well, heh. :)
20:34:38*bertrik was confused with HAVE_FLASH_DISK, only defined in config-ifp7xx.h
20:35:08*pixelma too
20:35:46 Quit Thundercloud (Remote closed the connection)
20:36:00gevaertsOTOH none of the tcc targets declare any HAVE_ATA_whatever. Slightly annoying
20:36:47pixelmagevaerts: but progress, (d) is only true since DevCon, before only the Ondio had it... ;)
20:36:52gevaerts(except HAVE_ATA_POWER_OFF, which I don't believe)
20:37:04*n1s feels a bit guilty since he introduced HAVE_FLASH_STORAGE...
20:39:26gevaertsActually it makes some sense in a single-drive world. However, I think we need to either get rid of most of them, or replace them by something like HAVE_SLOW_ACCESSTIME
20:40:11gevaertsOr HAVE_DISK_STORAGE
20:40:37*gevaerts guesses that when n1s added them, disk targets were still the majority
20:40:52n1salthough when i introuduced it, all it did was disable the disk spindown and anti-skip buffer options on flash based targets
20:42:20*n1s isn't sure wht the problem is...
20:43:58gevaertsThe main problem is (a), which is my fault (usb_storage.c). A smaller, possibly irrelevant problem is that a target may have both flash and disk, and #ifndef HAVE_FLASH_STORAGE will disable stuff that should stay
20:44:23 Quit ajonat (Read error: 110 (Connection timed out))
20:44:39amiconnThis also raises the question what disk icon to display on mixed-storage targets
20:44:51gevaertsThere is no such target currently, so it's not an issue right now. I'd still like to invert the logic though
20:45:14amiconnBtw, the beginnings of a port to such a device are already in svn...
20:45:17*n1s has no objections
20:45:21gevaertsamiconn: easy! Waste RAM and compile in both icons!
20:45:40amiconnThe Tatung Elio is HDD based and has an SD slot
20:46:11LambdaCalculus37That's true.
20:46:20amiconnBut it's a linuxstb port.. ;)
20:46:55 Join AlexStacey_ [0] (
20:47:05 Part AlexStacey_
20:48:12gevaertsSo how do I call the define meaning you need to take spinups and related delays into account ? HAVE_DISK_STORAGE?
20:49:47amiconnIf the beast's performance scales linearly with CPU clock (and it probably does for stuff that fits into the cache), I'm now at 99% realtime for ape -c5000 at full clock (measured 49.5% realtime at 264MHz)
20:50:00*gevaerts finds that there are 81 different HAVE_* defines in config-*.h
20:50:30amiconnSome of them really shouldn't be there
20:50:59LambdaCalculus37amiconn: That's setting a new record with APE! :)
20:51:08*LambdaCalculus37 hands amiconn an entire crate of beer
20:51:36gevaertsJust a few percents more, and you'll have realtime!
20:51:39amiconn(those which define stuff that only depends on the actual target, and redefining it for that target would make no sense)
20:52:33pixelmaMarcGuay: I can't tell you how the \index work, not much of an LaTeX expert either, maybe try asking bluebrother... ;)
20:52:54*gevaerts expect the monkeysaudio people will introduce -c6000 as soon as they hear that a portable DAP can handle -c5000
20:53:17amiconn-c2000 became a little bit slower for some reason, but -c3000, -c4000 and -c5000 all became faster
20:53:20MarcGuaypixelma: Okie doke. Thanks anyways. I have a feeling they're leftover from another organization attempt but want to be sure...
20:53:32pixelmacould be
20:54:03*Strife89 finds this whole -c1000, -c2000, etc. thing irritating.
20:54:19 Quit jhulst (Read error: 104 (Connection reset by peer))
20:54:44LambdaCalculus37Strife89: It's Monkey's Audio (APE); amiconn is working on speeding up the codec as much as he can.
20:54:51 Quit herrwaldo (Remote closed the connection)
20:55:27***Saving seen data "./dancer.seen"
20:55:41Strife89LambdaCalculus37: Yeah, I knew ya'll are discussing APE.
20:55:50pixelmait's the quality setting for it, there are also alias names for -c1000 etc (don't ask me though)
20:55:53bluebrotherMarcGuay, pixelma: \index{} add something to the index. You need to print that index though
20:56:14gevaertsLambdaCalculus37: the point is that there's no such thing as -c4950 if -c5000 is just too slow
20:56:19bluebrotherand you can substructure with ! −− as in \index{Installation!Fonts}
20:56:23*Strife89 can't be bothered to try APE at the moment.
20:56:32MarcGuaybluebrother: So does it currently get used or is it theoretical?
20:56:38bluebrotherthis will make an index entry "Installation, Fonts"
20:56:54bluebrothernot yet. There was a previous attempt of using it but it got never finished.
20:57:12 Quit Kopfgeldjaeger ("Serverwechsel")
20:57:27 Join herrwaldo [0] (
20:57:29bluebrotherI had an index on my personal copy for testing, but unless all items that should get indexed use \index there isn't too much point in it −− it'll be rather empty
20:57:39MarcGuayI see. The current index is created using section/subsection/item, etc.
20:57:55MarcGuayOr is it meant as an Appendix?
20:57:56bluebrothercurrent index? No, that's a TOC ;-)
20:58:21MarcGuay:) When words collide.
20:58:23bluebrotherindex is meant as words you look after −− like "what pages deal with fonts"?
20:58:50MarcGuaybluebrother: I call that an appendix. Oh my, English is fun.
20:58:56bluebrotherIf you're interested I can look after my previous attempt and put it somewhere.
20:59:14gevaertsmcuelenaere: why HAS_BUTTON_HOLD instead of HAVE_BUTTON_HOLD?
20:59:20 Quit perrikwp (" ajax IRC Client")
20:59:23 Join Kopfgeldjaeger [0] (
20:59:27mcuelenaeregevaerts: hmm where?
20:59:28bluebrotherwell, appendix (in the sense of latex formatting) is everything that comes after the "book". Like the config file options being an appendix
20:59:37mcuelenaeremust be a typo
20:59:48MarcGuayI was more interested in getting rid of it. :) I see how it could be useful for people who refuse to search, though...
20:59:49gevaertsmcuelenaere: creative and onda ports
20:59:52bluebrotherappendices are numbered using letters, while chapters are using latin numbers (well, usually ;-)
21:00:10BigBambitrue that
21:00:14bluebrothergetting rid of the nonexisting index?
21:00:14 Join jhulst [0] (n=jhulst@
21:00:32bluebrotheror of the TOC?
21:00:57BigBambiAn index would be a good idea IMO
21:01:03mcuelenaerehmme /me doesn't know where he got that HAS_BUTTON_HOLD
21:01:15mcuelenaereI must've copied it from somewhere
21:01:19*bluebrother would like to see a full index too
21:01:19MarcGuaybluebrother: Shows how much I know about my native tongue, you're right about appendix/index.
21:01:41mcuelenaeregevaerts: what gives if you do a grep -R HAS_BUTTON_HOLD on config-* ?
21:01:49mcuelenaere(haven't started my Linux vm up yet)
21:01:53MarcGuayAn index could be useful, but imagine the work involved. I think the links to relatedl sections are good enough.
21:02:02gevaertsmcuelenaere: only your targets
21:02:02n1sI think the define HAS_BUTTON_HOLD is the one we use for hold
21:02:20*Strife89 would just like to see a .txt manual that can be read on the device. :)
21:02:37pixelmagevaerts: I believe that define was called this way for a long time, something, as n1s said...
21:03:01MarcGuayStrife89: Take a look at the text bible that someone made. Something similar could be useful.
21:03:29mcuelenaeregevaerts: backlight.c:#ifdef HAS_BUTTON_HOLD
21:03:38Strife89MarcGuay: Was that in the Wiki or the forums?
21:03:45n1sgevaerts: other targets have it in their button-target.h file in the target tree
21:03:49mcuelenaereseems like it's defined in all button-target.h files
21:03:49MarcGuayStrife89: And then integrate it into the context menus for on-the-fly assistance.
21:03:52 Quit gevaerts (Nick collision from services.)
21:04:01 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
21:04:09bluebrotherStrife89: make manual-txt ;-)
21:04:19*gevaerts apologises to mcuelenaere for the false alarm
21:04:27Strife89bluebrother: I have no Latex.
21:04:32*mcuelenaere is going to correct it anyway
21:04:51Strife89bluebrother: And I don't know ho to get it or set it up.
21:05:31MarcGuaybluebrother: I suppose a script could be created that searches for keywords (i.e. Fonts) and adds a \index{Fonts} next to it. A simple find->replace even.
21:05:33bluebrotherStrife89: well, I can't help you much with that. But it's already possible creating a txt version
21:06:02MarcGuayStrife89: ManualHowto wiki page knows.
21:06:28Strife89The Wiki knows all........ ;)
21:06:51 Quit bertrik (Read error: 113 (No route to host))
21:07:00bluebrotherMarcGuay: that would be good as a start, but a good index needs to get maintained manually. But as you spread the keywords around the text this is rather easy −− when adding / changing text just put in the correct index item if it should show up there
21:07:36Strife89MarcGuay: It knows no more, technically. :)
21:08:14MarcGuaybluebrother: I think that's asking a lot of maintainers. They'll need to know all of the keywords and remember to add them.
21:08:30MarcGuayStrife89: I don't understand.
21:09:06MarcGuaybluebrother: IMO it seems like a lot of work for something that can easily be done by searching the PDFs.
21:09:10bluebrotherMarcGuay: I don't think so. You can add anything you like to the index, and if we decide to only add "big" topics this shouldn't be really that of a problem
21:09:19Strife89MarcGuay: I'm searching GeneralFAQ now with no luck.
21:09:37bluebrotherplus, a manually maintained index is usually a let better than a full text search. Given that the index is maintained ;-)
21:09:58 Join bertrik [0] (
21:10:01MarcGuayOne-letter difference.
21:10:03 Quit miepchen^schlaf ()
21:10:30Strife89MarcGuay: Oops.
21:10:43*bluebrother wants a redirect feature in twiki like MediaWiki can do for catching such stuff
21:10:46Strife89MarcGuay: Guess I can't spel today. ^^;
21:11:47MarcGuaybluebrother: The problem there was that there's an out of date page with the one-letter difference spelling.
21:12:24MarcGuayIt was part of the wiki manual - a general How To.
21:13:36gevaertsmcuelenaere: actually, HAVE_BUTTON_HOLD is never checked for in any code
21:13:50 Join miepchen^schlaf [0] (
21:14:11mcuelenaeremeaning, it is an useless commit?
21:14:19bluebrotherStrife89: maybe your distro is already using texlive?
21:14:35gevaertsmcuelenaere: no. Just a not-entirely-correct one ;)
21:14:45Strife89bluebrother: Well, if so, then what? It's Ubuntu 8.04.1
21:15:28bluebrotherthen the package is most likely called differently. Don't know about ubuntu, but all newer distros are changing to texlive as tetex isn't maintained anymore
21:15:56gevaertsmcuelenaere: If I understand things correctly, other targets define HAS_BUTTON_HOLD in button-target.h, and no HOLD related things in config-*
21:16:01*Strife89 fires up Synaptic.
21:16:22mcuelenaerehmm this sounds like it needs a cleanup
21:19:22 Join {phoenix} [0] (
21:19:58 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
21:20:57 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
21:21:07 Part J-23
21:26:59 Nick zodttd|away is now known as zodttd (
21:27:38gevaertsmcuelenaere: didn;
21:27:53gevaertsmcuelenaere: didn't my commit from a few days ago break onda usb?
21:28:21mcuelenaerehmm its been a while since I tested Rockbox on the Onda
21:28:35mcuelenaereanyway, USB on the Onda is hacky and uses it's own code
21:28:39mcuelenaereso not the Rockbox one
21:28:53gevaertsah, ok
21:29:56 Quit tvelocity (Remote closed the connection)
21:36:14 Quit nplus (Read error: 104 (Connection reset by peer))
21:37:26 Join nplus [0] (n=nplus@141.25.Globcom.Net)
21:37:35*gevaerts commits and hopes
21:38:13denesgevaerts: hi, I added FS #9466 , my fisrt attempt at a driver for old meizu m3 lcds
21:38:17denesgevaerts: could you try it?
21:39:46 Quit jhulst (Remote closed the connection)
21:40:01 Quit nplus (Remote closed the connection)
21:40:10gevaertsdenes: sure
21:40:39denesgevaerts: thanks
21:43:09 Join HBK [0] (
21:46:44 Quit amiconn (" boot, re")
21:46:51gevaertsdenes: it doesn't seem to work. I'll try to find out why
21:47:18denesgevaerts: maybe try doubling the delay in lcd_sleep
21:47:29denesgevaerts: but does the backlight switch on?
21:51:49 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
21:52:46 Quit herrwaldo (Remote closed the connection)
21:52:53gevaertsIncreasing the delays doesn't help
21:53:13 Join kachna|lappy [0] (
21:53:16*gevaerts wonders...
21:53:53gevaertsThis uses an external chip, right? Are we sure that that one isn't connected to some power pin we need to turn on?
21:54:22 Quit kachna (Read error: 104 (Connection reset by peer))
21:54:54denesgevaerts: this is a disassembly of the OF
21:54:58shotofaddsmcuelenaere: are you sure r18734 is correct? HAS_BUTTON_HOLD (rather than HAVE_..) is used throughout, including in features.txt. Although for some reason it's usually defined in button-target.h rather than config-xxx.h
21:55:04denesgevaerts: nothing more, nothing less
21:55:16denesgevaerts: it worked for the new lcd
21:55:28mcuelenaereshotofadds: ask gevaerts ;)
21:55:29gevaertsThere is that of course
21:55:53gevaertsmcuelenaere: no fair! I asked the same
21:56:01*mcuelenaere checks
21:56:10shotofaddshmm. actually I don't understand that commit at all. ZVM now defines both!
21:56:35mcuelenaereit does?
21:56:37 Join webguest87 [0] (n=626ebb96@gateway/web/cgi-irc/
21:56:50denesgevaerts: maybe try forcing lcd_type to 0 in lcd_init - in case the detection is broken
21:56:54mcuelenaereno it doesn't
21:57:13gevaertsmcuelenaere: the button-target.h change is fine, but config-* don't need any defines for HOLD
21:57:19 Quit webguest87 (Client Quit)
21:57:29 Quit LambdaCalculus37 (" ajax IRC Client")
21:57:32mcuelenaereok, I'll include that in my next commit then
21:57:53 Join herrwaldo [0] (
21:58:20 Join Ghost [0] (
21:58:29 Nick Ghost is now known as Guest91128 (
21:58:43Guest91128link doesnt work
21:59:07 Quit Guest91128 (Client Quit)
21:59:30gevaertsdenes: if I do that, the backlight doesn't turn on any more
22:00:00*gevaerts adds a delay to check something
22:00:04denesgevaerts: from what change?
22:00:29gevaertsdenes: forcing lcd_type=0
22:00:51denesgevaerts: maybe you just have to wait for the long delay
22:01:04denesgevaerts: or maybe not
22:01:09denesgevaerts: interesting
22:01:25 Quit meven_ (Remote closed the connection)
22:01:29gevaertsVery interesting, as it should get turned on before that
22:01:38denesgevaerts: yes indeed
22:02:03denesgevaerts: is it not possible that the lcd turned black? just guessing
22:02:55gevaertsI don't think so. It looks exactly the same as when powered off
22:03:05denesgevaerts: strange indeed
22:03:06Strife89Eh? I'm trying to make a sim and I'm being told that "SDL isn
22:03:25Strife89't (properly) installed."
22:03:48Strife89Can someone tell me what package I need to resolve this?
22:03:57gevaertsStrife89: libsdl-dev
22:04:23Strife89gevaerts: Does the version number matter? Synaptic lists v1.2
22:04:28gevaertsdenes: if I add a delay just after turning the backlight on but before the lcd_* things, the backlight works
22:04:32gevaertsStrife89: you need 1.2
22:04:45Strife89gevaerts: Marked. Thanks. :)
22:05:04 Join jeroen- [0] (n=jeroen-@wikipedia/jeroenvrp)
22:05:13jeroen-hi folks
22:05:34jeroen-I like to disable my apple firmware from starting up with the hold key on
22:05:44Strife89"Simulator environment deemed little endian." 0.o
22:05:59*gevaerts has text on his screen!
22:06:14gevaertsjeroen-: you'll need to change and recompile the bootloader for that
22:06:31jeroen-gevaerts: oh that sounds dangerous
22:06:51gevaertsjeroen-: on ipod, not very.
22:06:51mcuelenaeregevaerts: \o/
22:07:07jeroen-ok I will remember if I have time
22:07:08gevaertsdenes: I disabled the clock setting...
22:07:20gevaertsSo it's definitely timing
22:07:52Strife89Eh? Another pothole:
22:08:30Strife89Never seen it croak on a bitmap before.....
22:08:31gevaertsStrife89: build an e200 sim, not an e200r one.
22:08:43*Strife89 slaps himself.
22:08:57*gevaerts doesn't have to do that then :)
22:08:59 Nick zodttd is now known as zodttd|away (
22:10:06denesgevaerts: so with the clock settings on, a delay after turning the backlight on, and doubled (or trippled) lcd_sleep - so then the backlight turns on, but there is no visible change on the lcd?
22:10:31gevaertsdenes: looks like it. I went to 4x.
22:11:00denesgevaerts: good. and what happens if you only run lcd_init - so not clearing the screen etc.
22:11:02gevaertsdenes: Now that I can print text, I can print the default clock settings
22:11:20denesgevaerts: there should be garbage then if the init itself works - maybe only the screendrawing part is broken?
22:11:41gevaertsSo just remove lcd_update()?
22:11:59denesgevaerts: yes
22:12:14denesgevaerts: is the lcd_type forced to 0?
22:12:25gevaertsNot anymore
22:12:39*Strife89 looks at the APE Wiki page.
22:12:53denesgevaerts: well it doesn't matter if it's 0 - question is, is it set to 0, or is it misdetected
22:12:59gevaertsdenes: you were right. Garbage
22:13:00*Strife89 thinks Insane (-c5000) is truly insance.
22:13:04denesgevaerts: great!
22:13:08denesgevaerts: so the init works!
22:13:12Strife89insance -> insane
22:13:38Strife893.6% speed on the c200....... 0.o
22:13:40bertrikugh, rockbox has one of those infinite bounce left/right progress indicators
22:13:51denesgevaerts: you mean you saw garbage?
22:13:59Strife89bertrik: It bothers you too, then?
22:14:07denesgevaerts: that's great, I am happy
22:14:17Strife89I definitely prefer a percentage (number).
22:14:37*gevaerts thinks that it takes a special sort of person to be happy about seeing garbage :)
22:14:47denesgevaerts: so if you put everything the way it is in the patch, BUT uncomment lcd_update, you see garbage again?
22:16:05bertrikStrife89, it doesn't really bother me too much, but I prefer one that simply fills up from left to right proportionally to progress.
22:16:06denesgevaerts: because then the next step is to IMHO to change the lcd_sleep(1) in lcd_update to something big, like lcd_sleep(200) - and then you can should be able to see on the display as it is drawn
22:16:44 Part jeroen- ("Leaving.")
22:16:53Strife89bertrik: It drives me a little crazy, the way it is.
22:17:42 Join tvelocity [0] (n=tony@
22:17:50Strife89bertrik: I much prefer having _some way_ of knowing how much has been done, and how much is left to go.
22:19:25bertrikyeah, the left/right bouncy type of indicator says nothing about progress at all, basically just a "I'm still alive" indicator
22:19:55gevaertsdenes: it stopped working somewhere after going to a clean patch. Investigating now...
22:21:40denesgevaerts: ok. (meanwhile I go and have dinner)
22:21:51gevaertsI think I'll just leave the clock for now...
22:22:02RyoSsome x5 user in here ever sent back the player outside warranty and let cowon handle "an issue"? i just wonder how much it would cost me, and they told me to send in the player to tell me
22:22:39RyoSi dont want to pay a new player lol
22:22:41Strife89bertrik: In one case it was doing that but it wasn't "alive".
22:22:42RyoSthats all..
22:23:47gevaertsRyoS: you're probably better off asking that sort of thing on the forums
22:23:50Strife89bertrik: On one occasion, I was copying a file (5.8MB) from internal flash to the memory card while music was playing. 30 minutes later..... it still said it wasn't done.
22:25:53 Quit jeffdameth1 (Read error: 113 (No route to host))
22:26:36Strife89bertrik: Ultimately I killed the power and checked my storage. I wound up reformatting the card due to bad sectors (after archiving my data, of course).
22:26:59 Join jeffdameth [0] (
22:28:34RyoSgevaerts: did that, no answer
22:28:49RyoSi solderd a new battery but it lasts as long as the original, on OF and rockbox
22:29:05RyoSso i guess, something else got to be wrong..
22:30:31ameyerhow long are we talking about?
22:30:49*gevaerts thinks that this is not really on topic here...
22:31:02RyoShowever~ just wanted to throw in the question XD
22:31:21BigBambiPlease don't unless it is on topic
22:34:55 Join _Andrew_ [0] (
22:36:03denesgevaerts: so did you manage to get back the garbage?
22:36:13gevaertsdenes: by not setting the clock
22:37:00denesgevaerts: great. so then did you try changing the lcd_sleep(1) in lcd_update to something big, like lcd_sleep(200) ?
22:37:19denesgevaerts: to see if it draws something on the screen maybe
22:37:37gevaertsIt obviously needs more, as the init doesn't work when I do set the clock. I must have done something wrong earlier
22:37:56denesgevaerts: maybe the delays are too short at 200MHz
22:37:57gevaertsWould knowing the clock register values help?
22:38:32*gevaerts rememers something else...
22:38:35denesgevaerts: if we can get it to work at the default clock than I think we can be happy
22:38:56denesgevaerts: than we can adjust stuff to make it work at 200 MHz (probably by adjusting the delays) - IMHO
22:39:08gevaertsDoes yours work with the default clock?
22:39:56 Join MethoS- [0] (
22:40:52 Quit Strife89 (Read error: 60 (Operation timed out))
22:41:09denesgevaerts: probably. I did mess around with the clock, until I found where I had to insert an lcd_sleep to make it work at 200MHz
22:41:27denesgevaerts: I can try it right now
22:41:36denesgevaerts: sorry I mean I can't try it right now
22:42:11denesgevaerts: so I think we shouldn't worry about the clock right now
22:42:22denesgevaerts: I worry about the lcd_update right now
22:42:36gevaertsOn slow clock lcd_update works too
22:42:44denesgevaerts: great! so you got picture?
22:42:53gevaertsyes (or rather text)
22:42:57denesgevaerts: :)
22:42:59denesgevaerts: great
22:43:24denesgevaerts: did you increase the lcd_sleep(1) in lcd_update, or it was not needed?
22:44:21denesgevaerts: maybe you could try increasing lcd_sleep (in the function itself) 10x - and then try it at 200MHz. just guessing again
22:44:24 Nick cactusy1 is now known as maha (i=xnte@gateway/gpg-tor/key-0xB9002659)
22:44:54denesgevaerts: could you perhaps commit or send me the driver that did work for you at default clock? to see what changes you did?
22:45:10gevaertsAt default clock I need no changes actually
22:45:24denesgevaerts: great, than just increase the lcd_sleep 10 times IMHO
22:45:36 Nick zodttd|away is now known as zodttd (
22:45:39 Join Strife89 [0] (n=michael@
22:46:17 Join bughunter21 [0] (n=Jelle@
22:50:40denesgevaerts: so what I found is that lcd_update can send the data too fast int LCD_WDATA - and if the fifo overflows, the whole machine seemed to crash (maybe because of an interrupt or I don't know)
22:50:59gevaertsI'd guess an interrupt, yes
22:51:03denesgevaerts: so the lcd_sleep in lcd_update is important, and it might be too little
22:51:37denesgevaerts: but it does work as it is on my m3
22:52:07 Quit {phoenix} (Read error: 110 (Connection timed out))
22:52:07gevaertsI'm currently busy getting confused. Most of the changes I make have unexpected results
22:52:09denesgevaerts: so maybe the other delays are too short - they are just educated guesses from the firmware, they could be off by any multiplier unfortunatelly;)
22:53:05denesgevaerts: what kind of changes are you making?
22:53:36gevaertsenabling and disabling the clock setting, doing and not doing lcd_update, and playing with sleep multipliers
22:54:21denesgevaerts: great
22:55:03denesgevaerts: well I gave you all info I know - so now you are on your own I am afraid :)
22:55:11gevaertsNo problem :)
22:55:25denesgevaerts: yes, I wasn't worried of course :)
22:55:29***Saving seen data "./dancer.seen"
22:55:30 Quit mcuelenaere ("Zzzzz")
22:55:33amiconnThose armv6 simd instructions are really useful :)
22:55:57denesgevaerts: so I am looking forward to your commit of a fully working meizu m3 lcd driver
22:56:11gevaertsI'll do my best :)
22:56:21bertrikStrife89, I see that the bouncy progress indicator is only used for directory delete and file copy
22:56:59Strife89bertrik: .....Which I do a fair bit of.
22:57:15bertrikI think for file copy it can quite easily be modified to show true progress, but for directory delete you don't know how much is going to be deleted in advance
22:57:30 Quit nuonguy ("This computer has gone to sleep")
22:57:50Strife89bertrik: I can live with just "file copy progress". :)
22:58:02denesgevaerts: one more thing: AFAICT the OF actually used the timer for the long delays in the old lcd driver (not in the new one). I don't know why.
22:58:23gevaertsdenes: your sleeps seem to be way off (much too long)
22:58:32 Join Waldo000000 [0] (
22:58:58denesgevaerts: ok, cool
22:59:06denesgevaerts: that was the problem?
22:59:18gevaertsIt doesn't work yet...
22:59:49Strife89bertrik: Delete tasks rarely take long anyway.
22:59:49 Quit Strife89 (Remote closed the connection)
22:59:55 Join Strife89 [0] (n=michael@
23:00:24Strife89Lost my signal again.....
23:00:35*petur wonders why his settings code produces options that are the other way around (first option shown last).
23:01:58 Quit domonoky (Read error: 104 (Connection reset by peer))
23:02:17bertrikhmm, never declared an initialised array like that
23:03:00peturjust copy/paste from another entry... /me lazy ;)
23:03:20*petur checks how the other entry works
23:03:59bertriknot saying it is wrong...
23:04:54 Join Dhraakellian [0] (
23:05:19 Quit bughunter2 (Read error: 110 (Connection timed out))
23:06:40peturlooking at the other entries, this seems to be how the settings code works... bloody macro hell
23:07:53Strife89bluebrother: After all that, I have no .txt file. o_o ?????
23:08:23*petur goes looking for a recording source
23:09:48bluebrotherStrife89: you have lynx installed? Also, you need htlatex too ...
23:10:00bluebrotherwhat target you want to build?
23:10:09Strife89bluebrother: I don't have lynx installed.
23:10:17Strife89bluebrother: Target=c200
23:10:41bluebrotherah no, it's links.
23:10:56Strife89bluebrother: No links either. :)
23:11:02Dhraakellianjust did a battery test with 3.0 on my e260
23:11:19Strife89Dhraakellian: How'd it turn out? :)
23:11:30Strife89bluebrother: Is it safe to stop midbuild?
23:11:58Strife89bluebrother: I already entered make manual-txt for the second time and it's working on it.
23:12:33Dhraakellian24' or so less than the r18117+batt_bench_v3 test
23:13:07 Quit tvelocity (Remote closed the connection)
23:13:08bluebrotherhmm. There seems to be an issue with creating feature.tex when not building the pdf
23:13:21bluebrotherat least it seems it doesn't work for the first run. :/
23:13:38Strife89bluebrother: Second run worked as far as the PDF went.
23:13:56Dhraakellianthen again, looking at the Time Left estimates at 100%, this one appears to have been closer
23:14:10Strife89bluebrother: "links is already the newest version."
23:15:11bluebrotherStrife89: you might want to check if building the pdf works.
23:15:29*bluebrother needs to check for features.txt when building the html manual
23:15:45Strife89bluebrother: I did.
23:16:09Strife89bluebrother: With "make manual".
23:16:39Strife89bluebrother: r18730
23:16:58 Quit dabujo (Read error: 54 (Connection reset by peer))
23:17:25 Nick bughunter21 is now known as bughunter2 (n=Jelle@
23:17:51bluebrotherso, running "make manual-html" and / or "make manual-txt" runs?
23:18:27Strife89The latter does, but I don't get a "final output" and I saw a TON of errors (I believe).
23:18:38bluebrotherthe txt version needs the html first, and that in turn requires htlatex
23:18:56 Join miepchen^schlaf [0] (
23:18:59bluebrotherthere will be loads of notices on stdout. The result will get placed in a subfolder "txt"
23:19:14Dhraakellianwould I be right in surmising that Rockbox should now have drastically better battery life than the SanDisk firmware with equivalent mp3?
23:19:15bluebrother(it's not a single file −− it's split up in chapters)
23:19:17Strife89bluebrother: Bah, I haven't built the HTML yet.
23:19:40Strife89bluebrother: Also, I see the folder, with the split files. They're all 0 bytes at the moment.
23:19:53bluebrotheryou don't need to build the html first as make will do that. But it's one step less so you can check if that works
23:20:18*Strife89 runs "make manual-html"
23:23:39*Dhraakellian goes to add the results of his test to the SansaRuntime page
23:24:23denesgevaerts: cool
23:24:37gevaertsdenes: as you see, the colours need some work :)
23:24:56denesgevaerts: but is it a driver problem?
23:25:55gevaertsEither driver or some define somewhere that says how the colours should look
23:26:21peturyes, might be the pixel packing
23:26:41Strife89bluebrother: Now running "make manual-txt".
23:26:51denesgevaerts: i see
23:27:06gevaertsAnyway the driver still needs a lot of work, and we still need to decide if it's worth it going to full 24 bit, so I won't worry about it now
23:27:22gevaertsIt's more than good enough for debug output :)
23:28:09gevaertsdenes: this is still with the default clock
23:31:06denesgevaerts: I am positive with correct timing it should work at 200MHz too in the future
23:31:39gevaertsdenes: of course. I think we want real timings though, not guessed numbers.
23:32:06gevaertsSo we need to find out which lcd controller this is, so we can get the datasheet
23:32:28denesgevaerts: the numbers are not guessed, it's either 10000 1000 or 15000 in the OF - the question is what metric
23:32:41gevaertsExactly :)
23:32:47Dhraakellianmethinks the battery on this one might not be quite as good as the one in the e200 that got RMA'd because of the headphone jack
23:33:58*gevaerts wonders how much more than 100% a codec needs on a full-speed beast (532MHz) to be realtime
23:34:28 Quit ompaul (Read error: 113 (No route to host))
23:34:52denesgevaerts: but I think getting rockbox to work on the m3 is first priority, it's even before prettyfying the driver - for me at least. also, for accurate timing the timer should be used instead of busy loops - I don't know if there is such functionaility in rockbox
23:35:17 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
23:35:40 Quit gregzx ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
23:35:46gevaertsFor init a busy loop could be OK. For the update loop, I think we're missing a ready bit, or an interrupt.
23:36:09 Quit Nibbler (Read error: 113 (No route to host))
23:36:12DhraakellianWhat's the best way to time the OF? hook it up to Audacity, hit record, and boot up the DAP?
23:36:12denesgevaerts: the update shouldn't be a loop in the first place, but should be done through DMA - that's how OF does it
23:36:21gevaertsThat too :)
23:36:24denesgevaerts: but I think this driver will doo for now
23:37:20gevaertsIt's more than good enough to continue working. Once vitja has the usb driver for tcc mostly working, I'll get busy making it work on meizu
23:37:35denesgevaerts: sounds good :)
23:37:50Strife89bluebrother: Bah, it's still not working here. I won't worry about it any more for now; I'm going to go eat. :)
23:37:59Strife89bluebrother: Thanks, though. :)
23:38:09gevaertsAnd once we have usb and lcd, we're done. Who needs more than a remote controllable display anyway ;)
23:38:12denesgevaerts: I think I will try to write the "bootloader". I mean the part that initializes the memory controller, and such stuff
23:38:30 Quit Strife89 (""Gone to eat dinner. :)"")
23:38:37gevaertsYes, we'll need that too
23:39:29 Quit petur ("Zzzzz")
23:39:47denesgevaerts: or maybe I will check the nand reading/writing - I don't know what I will have time for
23:40:14gevaertsDo whatever you find interesting. There's more than enough work for everyone!
23:40:27denesgevaerts: indeed :)
23:42:12 Join krazykit [0] (
23:42:18Dhraakellianwould having the microSDHC card in have made any difference in terms of battery life?
23:43:37 Join fophillips [0] (n=fophilli@unaffiliated/fophillips)
23:43:59fophillipsIs it possible to generate the database on a computer then transfer it over to the media device?
23:45:16 Quit MethoS- (Remote closed the connection)
23:52:37Bagderthat's rather unmaintained since a long time though
23:52:49fophillipsAnything more up to date?
23:53:05Bagderthere have been work at #ifdefs for the actual C code to enable a native tool to get built
23:53:15 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
23:53:18Bagderbut I don't think that's top notch either
23:53:41fophillipsHmm, Rockbox doesn’t like the files produces.
23:54:53 Quit bughunter2 ("bye")
23:58:49n1sfophillips: it produces the old database format which was replaced with tagcache a few years ago...

Previous day | Next day