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 2008-12-15

00:00:07lucentbertrik: okay. I'm pretty sure I can login, but yes there was an error when I registered for the TWiki today
00:00:14kugelmc2739: buffering screen
00:00:16n1sDavide-NYC: likely some makefile is the culprit
00:00:43kugelsystem -> debug (Keep out!) -> View buffering thread
00:01:58mc2739pcm=529200, alloc, real, usefl = 4573136
00:02:01Davide-NYCDo we really have USB charging on the Gigabeat S now?
00:02:09kugelmc2739: too less too
00:02:21 Quit freddy_ (Read error: 110 (Connection timed out))
00:02:22kugelmc2739: we're supposed to have ~5.5MB
00:02:27n1sDavide-NYC: no, but that page is not about what rockbox does but what the player can do
00:02:41Davide-NYCn1s: you must be Nils!
00:02:52 Join freddy_ [0] (
00:02:56n1swe don't support usb host on any target either for example, and yes i am :=
00:03:20kugelmc2739: and once that buffer is filled, my playback crashes and I get the panic
00:04:02mc2739hmm, maybe this is still configured for the clip
00:07:08kugelmc2739: I don't think so. shows expected values
00:09:55mc2739kugel: didn't also show wrong values before the lds file was corrected?
00:10:39kugelmaybe we don't have 8MB, as we expect
00:10:48kugel8MB ram that is
00:10:56 Quit bluebrother ("leaving")
00:11:06 Quit bertrik ("Leaving")
00:11:51mc2739kugel: my rockbox info screen shows 4.91MB buffer
00:12:59n1sDavide-NYC: tools/functions.make:28 seems to do exactly what i guessed, it substitutes ".c" with ".o" in a string and from what i can see not only from the end
00:13:36n1sif you are not up for makefile hacking, file a bugreport and poke Zagor
00:14:04Davide-NYCn1s: thanks
00:16:02kugelmc2739: do you have problems with panics at all?
00:16:28 Quit tessarakt ("Client exiting")
00:16:30mc2739kugel: I get a few
00:16:36kugelupdating my bootloader definitely helped
00:16:54kugelstill, 4.95MB buffer is too low
00:17:41mc2739bootloader update makes sense, because of the changes to lcd and button drivers
00:19:20n1sDavide-NYC: think i found a fix, testing now
00:19:43kugelmc2739: have you tried the changes I told you about? I'm not sure, I get more panics with it
00:19:57kugelbut I also did svn up, either of one causes more panics
00:20:11mc2739I have not tried those changes yet
00:20:52 Join akur [0] (
00:21:06 Part akur
00:21:13 Join slact [0] (
00:21:33kugelmc2739: my driver looks like this now
00:22:28 Quit EspeonEefi ("さよなら")
00:22:38 Join lazka [0] (
00:22:50slactg'day. I stopped by here earlier about the Rockbox TWiking registration errors. The Validation Code page failed badly when entering the emailed code.
00:23:06lucentslact: ditto here
00:23:31slactdo you know if there's a solution yet?
00:23:45lucentI heard we need to poke a Badger about it
00:24:06hobbsslact: it did that for me as well but it didn't seem to actually have any consequence. The registration processed and I was able to edit once someone turned me on.
00:24:20hobbs(sorry for not mentioning that, it was late last night when I did the activation, and I forgot about it)
00:24:36lazkahere is a small patch for the themes site (usability only)..
00:24:49slactah. did you get a TWiki page for your username and all?
00:25:17slacterr... userpage or something...
00:25:24n1slazka: patches belong on the tracker
00:25:29lazkan1s, ok
00:26:23mc2739kugel: where do you lock lcd_mutex?
00:26:30hobbsslact: no, looks like that wasn't created. Perhaps that's what the script is bombing on.
00:26:43kugelmc2739: I used mutexing instead of disabling interrupts
00:26:51slactmaybe. anyway, who activates around here today?
00:27:31kugelmc2739: this is my lcd driver
00:27:58kugelthe diff of it rather
00:28:07kugelbut you see where I involved mutex
00:28:47 Join beta [0] (
00:28:52mc2739yes, thanks - does that seem to help your display?
00:28:57 Nick beta is now known as beta2k (
00:29:04mc2739I am not noticing problems on mine
00:29:12kugelmc2739: the cleaner button driver helped much mroe
00:29:42kugelmc2739: but you have still lcd_update_rect changed to be effectively lcd_update, not? That's what domonoky did
00:29:47kugelI don't have that
00:30:27mc2739yes, my lcd_update_rect uses lcd_update
00:31:07mc2739and I still have a lot of delays in my lcd file
00:31:14kugelI have the strong feeling we're just too slow to do ogg, not only SD driver problems
00:31:23 Quit faemir (Remote closed the connection)
00:31:35kugelmc2739: lcd_delay isn't needed at all on the fuze, probably neither on the e200v2
00:31:50mc2739but, I do not think we are boosting the cpu
00:32:03kugelwe are boosting, but apparently the wrong part
00:32:21kugelcgu at 248MHz doesn't seem to help decoding
00:32:30kugelthe pclk needs to be higher
00:34:04mc2739lol, when it boosts, cpu goes from 31000000 to 24800000
00:34:38beta2kkugel: What are you running at 248mhz?
00:35:20mc2739oops, read it wrong, add 0 to 248
00:35:20 Quit MethoS-- (Remote closed the connection)
00:35:21kugelmc2739: but I still see that my fuze has a hard time filling the pcm buffer, as a results it's about 80% boost ratio
00:35:58lazkan1s, whats the status on themes.rockbox?
00:36:09n1slazka: i have no idea
00:36:17*scorche hides
00:36:42 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:37:10scorcheslact: i alerted those who needed to be...they may not have had enough spare time to look at it yet
00:37:24slactok, cool. thanks
00:38:11***Saving seen data "./dancer.seen"
00:38:43 Quit Schmogel (Read error: 104 (Connection reset by peer))
00:39:06kugelmc2739: it looks much better for flac
00:39:38mc2739yes, flac and wav boost at about 5% for me
00:40:21kugelmc2739: do you have a v1?
00:41:08mc2739kugel: yes
00:41:28kugelwhat I always wondered: are the screens different between v1 and e200v2?
00:41:42kugelthe fuze's screen is a lot better imho, but I don't know about e200v2
00:42:14mc2739I don't see much difference between the two
00:43:57kugelmc2739: in case you like the backlight fading, it's most likely going to work on our v2s too
00:44:47mc2739yes, I have already tested it
00:45:04kugelhehe :)
00:45:40kugelok, my 1h MP3 file is the mortal enemy of my fuze
00:47:09kugelmc2739: have you tried enabling dsp effects?
00:47:50kugelcrossfeed seems to increase panic rate
00:48:20mc2739no, just doing basic functions right now
00:48:51rashermcuelenaere: Very generous commit message
00:49:50 Quit DerDome ("Leaving.")
00:49:51kugelas minimal as the diff
00:50:38kugellinuxstb: ping
00:53:13 Quit ender` (" Be nice to the USA, or they'll bring you democracy.")
00:53:21 Quit herrwaldo ("Konversation terminated!")
00:57:43*kugel wishes someone would commit :(
00:58:49 Join kushal_12_27_200 [0] (
01:00:26*kugel decides to do a little change more
01:01:19 Quit ZincAlloy ("CGI:IRC (EOF)")
01:01:26Davide-NYCthe OLED display on the sansa clip is yellow on the first two lines and blue everywhere else or is it possible to use both colors across the entire display?
01:01:54scorchethe former
01:02:21scorchethough "first two lines" is a bit inaccurate
01:03:01Davide-NYCthat's bizarre. thatnks for the clarification
01:03:19kushal_12_27_200Hello, I have a question about non-existent USB mode in Sansa c250 with V03.02.05A as the OF. And does v03 in sansa c250 mean it is one of the newer sansa for which rockbox is not available?
01:03:36 Quit XavierGr (Nick collision from services.)
01:03:47 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
01:04:36 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
01:07:03lucentDavide-NYC: the OLED on the clip is actually 2 different displays
01:07:18lucentDavide-NYC: each is a single color display
01:07:26n1sDavide-NYC: i have a "fix" for that .c issue now so if you really want to use those dirnames i can post ib but i'm not sure it's nice :)
01:07:52Davide-NYClucent: that's totally bizarre. I wonder why they went for that particular design
01:08:19Davide-NYCn1s: should I post a bug to the tracker so that you can attach your patch there?
01:08:47n1syes, taht's probably bes, so someone who knows this make stuff can look at it
01:09:07lucentDavide-NYC: it's not bizarre at all, OLED technology is brand spankin' new when the Clip was designed, it was frighteningly expensive to implement any other way
01:09:31Davide-NYCn1s: am doing. Just to clarify, this bug applied to dirnames that contain both ".c" and ".o" correct?
01:09:49Davide-NYClucent: ah, all is clear now
01:10:23lucent16 pixel height orange OLED display and uh, however tall the green display is
01:10:31 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
01:10:47scorchewith space in between
01:10:52n1sDavide-NYC: no, i think it would only affect dirnames containing .c or .S
01:11:06lucentyeah, there's a small gap between the displays on the device
01:11:13n1sthose are the two things that get substituted by .o
01:11:46*lucent keeps running into "PANIC SD DATA TIMEOUT"
01:12:37lucentDavide-NYC: if you run the plugin:demos:cube it's easy to see the layout of the display on a Clip
01:12:41 Join killan_ [0] (
01:13:15 Quit killan (Read error: 54 (Connection reset by peer))
01:17:24n1sok patch attached
01:20:05lucentthe mandelbrot plugin "freaks out" after a certain high zoom level
01:20:17lucenttrying to zoom out doesn't restore it
01:20:34lucentis that well known or some kind of overflow bug?
01:20:55kushal_12_27_200thanks, lazka.
01:21:31Davide-NYCre the keymaps well defined for the clip sim?
01:21:44rasherlucent: I think mandelbrot is amiconn's baby, you'll probably want to talk to him about that
01:21:50kugelmc2739: I updated my lcd driver changes here: you might try some of those for the e200v2 as well
01:23:34lucentamiconn: nice mandelbrot plugin :) does it have a known overflow condition at high zoom levels? I zoomed way in and then was not able to zoom back out, the display was not consistent with having zoomed in
01:23:43 Quit kushal_12_27_200 ("Leaving")
01:29:36 Quit lasser (Read error: 110 (Connection timed out))
01:29:41 Quit CaptainKewl (Read error: 54 (Connection reset by peer))
01:30:01 Join CaptainKewl [0] (
01:33:52*kugel put his todays progress on the tracker and feels good now :)
01:34:50lucentkugel: is SD and bank reading still needed? on hold until LCD is fixed?
01:35:04 Join esthar [0] (
01:35:33kugellucent: I think I can have a look at it now. But I'm likely to fail anyway, so I don't promise success by no means
01:36:26lucentkugel: so far so good ;)
01:36:54kugelbut having access to my full 4GB would surely be nice
01:37:03lucentkugel: I have 8GB Fuze and 8GB uSD cards I can try with, and a 2GB Clip
01:37:24kugelI have a 4GB/4GB combo :)
01:37:29kugelno clip though :(
01:37:45lucentwant a Clip? how much to send one to you from Jackson WY USA
01:37:53*kugel hands lucent a µ sign
01:38:11mud-rbis there a common reason that my plugin would compile but not link? i think i have it listed in the right locations (SUBDIRS and CATEGORIES) and I have a .make
01:38:20kugellucent: No idea. soap may have a clue how much it is
01:38:42n1smud-rb: well, the actual error message might tell something
01:39:03mud-rbn1s: there isn't one. it just doesn't have the LD goban.rock line
01:39:35lucentkugel: the Clip is a really nice little player with "just enough" graphical options, good little device for not worrying about breaking an LCD screen but still can tell me what song I'm playing
01:39:35kugelmud-rb: try make V=1 to make make more verbose
01:40:13kugellucent: I'm sure it's a nice player
01:40:39n1smud-rb: does your .make get called, and does it produce a goban.elf file in builddir/apps/pligins?
01:40:54kugelmud-rb: also, make sure you didn't accidently put it into a #ifdef'd part
01:41:01lucentkugel: well, it is inexpensive though, if there is some test I can do which will help you just let me know what it is and I will try
01:41:24lucentor if I can ship it to you and that will help Rockbox support, I would do that
01:41:32kugellucent: I'm not really keen on buying one (1 reason is that I'm a poor guy student)
01:41:42mud-rbn1s: no .elf, no.
01:41:51 Quit moos ("Rockbox rules the DAP world")
01:42:06n1smud-rb: do you get .o files then? inside yor subdir?
01:42:12kugellucent: it'd be great if you would sent it over
01:42:49lucentokay, hm, privmsg me your address where you are in the world, I can see how much shipping charges would be and if I can afford them
01:43:00mud-rbn1s: yeah, the object files are all there kugel: i'm pretty sure it's not in a bad ifdef in SUBDIRS
01:43:10 Quit mcuelenaere ()
01:43:33scorchelucent: how big of a package would you send it in?
01:43:55n1smud-rb: sorry the elf should be in the subdir too, but it isn't linked at all, is it?
01:44:23lucentscorche: it's tiny, I could stuff it inside of an envelope
01:44:47lucentespecially if I don't send the removable spring clip, and it is a standard mini-USB plug to charge
01:44:58lucentso all that needs to be sent is the device itself
01:45:01scorcheexpect $5.40 if you do it right then
01:45:11mud-rbn1s: the only thing in the build dir is the object files themselves, I don't see an .elf anywhere (and not in the parent directory either)
01:45:24n1si suggest checking that you're doing what the other makefiles do or posting a patch so we can see what you do
01:45:55mud-rbokay, i'll check again that i'm doing the same. might have missed something. thanks
01:50:26 Quit n1s ()
01:52:59lucentscorche: thanks mate, what, a small padded envelope?
01:54:35Unhelpfullucent: i've had the same issue w/ mandelbrot. i'd suspect an overflow somewhere, but i'm not sure why it's irreversible...
01:55:25lucentUnhelpful: did you see it on a Clip or a different hardware? I'm curious to know if that is specific to hardware, or possibly a software bug
01:56:01lucentmight be software bug than
01:56:04kugel"The monster"
01:56:07kugelmost likely
01:56:18Unhelpful"bug" or "implementation detail"
01:56:19kugel"The beast"* Sorry
01:57:01kugelUnhelpful: I remember some yield()s where added to mandelbrot recently, not sure if that's related (or am I confusing it with yields added to another plugin?)
01:57:23 Nick Bensawsome is now known as NinJew (n=Bensawso@unaffiliated/bensawsome)
01:57:48rasherkugel: Superdom had yields added
01:57:55lucentkugel: it *could* be a display bug, I don't know for fact that overflow is happening, I am commenting on how it looks as a user
01:58:05 Nick NinJew is now known as [i][B][e][n] (n=Bensawso@unaffiliated/bensawsome)
01:58:19Unhelpfulkugel: no, it's nothing that could relate to yields, i'd think. the thing is, if the zoom is just overflowing when it's stepped high enough, decrementing it should take it back out
01:58:34lucentoh hm yeah
01:58:36 Nick [i][B][e][n] is now known as Bensawsome (n=Bensawso@unaffiliated/bensawsome)
01:58:56rasherBensawsome: Please stick to a nick.
01:59:11lucentit looks like mandelbrot plugin "forgets" where X+Y it is supposed to be and then we're lost somewhere in bogon space
01:59:32lucentcould be minor implementation bug
01:59:39lucentor indicate bad problem
01:59:44Bensawsomelol sry rasher :S
02:00:01Unhelpfullucent: that makes more sense, if "somehow" X/Y is lost it would explain the inability to resolve by zooming out
02:04:00 Join admin1lbo [0] (
02:04:53lucentscorche: USPS quotes me a price USD of $12
02:05:10lucentscorche: seems a bit higher, which courier did you use?
02:05:14scorchethey only give a rough estimate
02:05:39 Quit admin1lbo (Remote closed the connection)
02:05:42scorcheUSPS...i have sent a lot of me ;) will charge you 8.65 max
02:06:30 Quit Thundercloud (Remote closed the connection)
02:08:09 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
02:08:46saratogaI sent a pair of CDs to the republic of Georgia from the US for about 5 dollars this summer, theres no way it will cost more then that for germany
02:08:48lucentanyone else want a chance to play with the Clip? If no one else does, it will go first to Germany (kugel).
02:09:00saratogajust go to the post office and ask how to most cheaply send the thing
02:10:02scorchelucent: you might want to ask that on the dev list
02:10:45saratogaon a side note, Clips in the US are about $15 shipped, so if anyone wants one I can easily mail it
02:11:24kugellucent: there's definitely a few people who would like to have a clip to port rockbox. But actually there's also quite a few devs which already have one
02:16:30 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111319]")
02:17:47lucentyeah, Clip is becoming an inexpensive rockbox target
02:18:21Unhelpfullucent: a thought occurs... presumably the stored X/Y are of one of the screen corners? and since the plugin zooms centered, they'd be adjusted on zoom... and if the zoom depth overflowing breaks the x/y adjustment, something crazy would end up in x/y...
02:19:49rasherDavide-NYC: FS #9642 doesn't belong on the Known bugs list.
02:20:51Davide-NYCrasher: you mean on the release31 page?
02:21:30rasherIt's not a list of all bugs.
02:21:47Davide-NYCRight, how do I revert that?
02:21:59lucentUnhelpful: sounds logical, though I can't know unless I understand the code
02:24:02rasherDavide-NYC: I'll remove it
02:25:20 Quit midkay ("Leaving")
02:25:35Davide-NYCI added the task to the tracker, linked to it from the CFMod wiki page and just thought it would be a good thing to aim for in a 3.1 release.
02:27:40rasherDavide-NYC: All bugs would be fine to fix in 3.1, but this is not important enough to go in there. a) It doesn't affect a lot of people. b) A workaround exists. c) Bootloader changes on hxxx are not likely at any rate.
02:28:08Davide-NYCWhat's the workaround?
02:28:48rasherAs far as I can tell, you're saying that in *some* situations you get the error. Then don't do that.
02:30:21rasherUnless I'm missing something, I don't see this as a terribly grave bug.
02:31:49Davide-NYCNo no. It's not grave. But there isn't a workaround either. If you boot an H1x0 series that has been CFModded while USB connected you will enter disk mode. But on subsequent unmount you will be presented with an ATA -80 error and have to reset the device (twice) to get back to a working player.
02:32:55rasherThe workaround is: Don't boot with usb connected
02:32:56Unhelpfulthey may very well overflow well before zoom would, if they're moved by some multiple of screen width on each zoom
02:33:13Davide-NYCrasher: LOL
02:33:43rasherDavide-NYC: I don't see why that's so funny. Rockbox has a perfectly fine USB mode.
02:34:28Davide-NYCTrue but one of the great things about RB on the H1x0 is the ability to mount the drive from the bootloader in case of a problem.
02:34:40rasherDavide-NYC: you still have that ability.
02:34:47rasherThere's just a slight hiccup afterwards.
02:34:55Davide-NYCabsolutely true.
02:35:06rasherI'm not saying it isn't a bug, just that it's pretty minor
02:35:15Davide-NYCI absolutely agree.
02:35:52rasherAnd hence, it shouldn't go on the list of "the most important, unresolved bugs"
02:35:57 Quit jrsharp ()
02:36:20lucentit's a minor bug that SHOULD be noted on any documentation for that player in combination with use of Rockbox
02:36:34lucentnot a showstopper for the Rockbox release
02:37:01rasherlucent: except it's only exhibited if you've performed a hardware mod. I don't think it should be anywhere but on the pages regarding that hardware mod and in Flyspray
02:37:33lucenthm, okay
02:37:40Davide-NYCRasher is correct. I was overzealous in my placing it on the release31 page.
02:37:45rasher(and even then, perhaps only for specific CF cards?)
02:37:59lucentit doesn't happen on unmodified hardware? or it just doesn't cause an error?
02:38:15***Saving seen data "./dancer.seen"
02:38:36mud-rbi'm still trying to diagnose why my plugin isn't getting linked into a .rock, would someone mind looking at this patch and seeing if anything looks wrong/missing? i get no error message, even with make V=1, it just doesn't seem to be trying to link it. i made the plugin as simple as possible, it's basically a copy of helloworld
02:38:41rasherlucent: I don't believe such a bug exists on a plain h100
02:38:47Davide-NYCno it is only present (AFAIK) on CFModded H1x0 players.
02:39:24Davide-NYCLet me prostletize for fixing this bug in that it has a special audience.
02:40:25Davide-NYCThe optical I/O on the H1x0 series makes it especially attractive to the recording crowd. Harddrive spinup noise is a downside for that same crowd.
02:40:43Davide-NYC(granted not an issue when using optical input)
02:41:09Davide-NYCThe H1x0 series player, when RTC and CFModded become a very very attrictive recording device.
02:41:22*Davide-NYC gets off soap box
02:44:21lucentmud-rb: does your plugin get built?
02:44:55mud-rblucent: yes, there's a .o file built
02:45:29lucentmud-rb: you might need to uh, make clean, and configure
02:45:33lucentjust guessing
02:45:46lucentI'll try it in a few minutes
02:46:02mud-rblucent: yeah...i already tried that. this is actually in a brand new svn checkout
02:55:10 Join ackers [0] (
02:57:38lucentugh, needs to change IMO
02:57:51lucentshould use environment variables
02:58:22lucentI'm trying to remember if there's a bash-ism to conditionally set a variable if it hasn't been already set
02:59:32rasherWhy a bashism? Won't this do: if [ -z "$FOO" ]; then FOO="bar"; fi (I don't think that's a bashism anyway)
02:59:57lucentrasher: doesn't handle the case of setting FOO=""
03:00:07lucentin that case FOO is set, it is set to empty string
03:00:10rasherAnd yeah, I'm not happy about the destination etc. being hard-coded
03:00:21rasherlucent: I'd say that's the user's fault
03:00:31lucentyes and we should raise an error
03:00:42mud-rblucent: look at post number 6 here:
03:01:30lucentat least, in example, Ruby language has foo ||= "value"
03:01:38lucentwhich only applies if foo is not already defined
03:03:59zeman bash, /Use Default Values
03:04:04ze(double spaces for some reason)
03:04:20zeand scroll up 1 line to see the one for that, and the rest available go from there
03:04:32 Part lazka ("cya")
03:05:15lucentoh, yuck! replacement? oh well
03:05:19lucentthanks ze, nice find
03:06:48 Join ron_o [0] (
03:06:58lucentsomething like this
03:07:39zei think you're musing a - for what i'm thinknig you intend
03:07:39lucentprefix="${ROCKDEV_PREFIX:-/usr}" rather
03:07:56ron_oRockbox rocks! I was thinking the other day, 'what if someone actually created their *own* firmware for their mp3 player." :)
03:08:50 Join Guest82979 [0] (
03:09:09 Part Guest82979 ("Leaving")
03:09:23lucentron_o: I'm still waiting for the "what if a major manufacturer actually contributed code to the rockbox project."
03:10:45ron_olucent, no way. There's that 'contributor to those folks who steal music' that might be tagged on a company.
03:10:47ackerswhy don't they
03:11:01ron_othink like them and you'll understand them better.
03:11:07lucentron_o: the blocking point for me is there's no commercially available player with full Rockbox support
03:11:19ron_olucent, and there won't be.
03:11:23ron_olet's face it.
03:11:36ackersthey would make more money
03:11:38lucentSansa AMS-based players are humping along pretty quickly
03:11:48ron_oalso, rockbox itself can't by its very nature give a company an edge over another company.
03:11:53lucentwe might see full rockbox support before the end of the product lifecycle
03:12:35ron_oackers, not really. They make more money for a time, but in the end most people go with what 'looks' best and is easiest to use. So far that's the original firmware.
03:12:42saratogaI think most companies doing DAP design do very little of their own software development, so picking an open source project and developing it to their needs is probably beyond the ability of most of them
03:12:54ron_olucent, I just bought a Sansa by chance.
03:13:04ron_o4GB for $40. What a deal!
03:13:08saratogathey seem to license almost everything and then just add minor changes
03:13:11ron_othe e260.
03:13:26ackersi just bought an 8gb clip!
03:13:42lucentron_o: the 2GB v1 Sansa Clip bootloader works for me, there's some funky issues with embedded SD I/O
03:13:53ackerssansa clip
03:14:10lucentackers: wow, Clip envy here
03:14:40ron_osorry, new here and to mp3 players. Clip?
03:14:41lucentackers: I was annoyed when I bought my 8GB Sansa Fuze, there was an 8GB Sansa Clip for sale at lower cost
03:14:45ron_oSD card?
03:15:02ackersyeah clip's sure are great
03:15:03lucentI really like the design of the Clip, it's damn near indestructable and no proprietary charge interface jack
03:15:26ackersron_o sadly no sd slot
03:15:51ron_ostill not following what "Clip" means.
03:16:03lucentron_o: Clip is literally a model name
03:16:03scorchethis ship is heading straight out of the neutral zone and into #rockbox-community territory!
03:16:11*scorche hides
03:16:12ron_ooh sorry. :/
03:16:12lucentyou have an E2x0 something, we have a Clip
03:16:13ackersit's like an ipod shuffle but better
03:16:28ron_oI see. :)
03:16:50lucentron_o: no LCD screen to crack
03:16:52ron_oI just never thought I'd buy an mp3 player.
03:17:08ron_ojust got one last week.
03:17:37ron_olove technology, but, man −− I got wires everywhere!
03:17:42ron_oI'll get used to it someday.
03:17:51ron_ohave to get more organized and sh*t.
03:18:02lucentrockbox may run on your player
03:18:07ron_oit does.
03:18:12ron_oI am using it now.
03:18:16lucentoh that's neat!
03:18:18ron_olucky me.
03:18:58ron_oI kept seeing 'rockbox' in reviews and only checked it out recently what that meant. You can't learn everythign at once you know.
03:18:59scorchewhat i meant by the above comment, is that social chatter should be directed to #rockbox-community...
03:19:03 Join Chesteta [0] (
03:19:29ron_oscorche, you're right you know.
03:19:36saratogayes he is
03:19:37ron_oit is indeed getty too chatty in here.
03:20:08ron_osorry, my fault. I tend to do that. I'm too used to ##politics, which you can talk about anything.
03:20:44rasherThis channel is supposed to be strictly about Rockbox. Everything else (related or not) can go in #rockbox-community
03:21:09ron_orockbox problems more like.
03:24:42 Quit Chesteta ()
03:30:11 Quit perrikwp (" ajax IRC Client")
03:31:18lucentrasher: okay, I am almost ready with a patch against
03:31:29ron_ocan anyone point me to directions on how to install a theme manually?
03:31:40lucentwhat do you think of environment variable names -> RBDEV_PREFIX RBDEV_DOWNLOAD RBDEV_BUILD
03:31:49ron_oI'm looking in the manual now for it.
03:31:53lucentokay or should use just "PREFIX"
03:32:14rasherron_o: Simply unzip to the root of your device. This *should* be in the manual
03:32:27rasherlucent: I think the RBDEV_ prefix is a fine idea
03:32:39lucentrasher: it's a simple patch, one moment
03:34:25 Join midkay [0] (n=midkay@rockbox/developer/midkay)
03:37:27saratogahmm yes PMing me to ask me to buy a Sansa View and then port Rockbox to it, this is a good idea
03:37:56lucentSansa View? yuck
03:38:14lucentit's a failed product if you ask me
03:38:21lucentSanDisk won't stand behind it
03:38:48lucentthey're pushing forward with the Clip and Fuze, abandoning the View as an interim product that didn't compete well in the market
03:39:04rasherIs View the one that runs Mono?
03:39:22rasherAh no, that's "Sansa Connect"
03:39:43lucentthe Connect runs some variety of Linux kernel on it
03:39:53lucentmore likely to be the Connect not the View
03:40:42lucentrasher: I don't know if I like RBDEV_PREFIX over just plain PREFIX though
03:41:22lucentvery fine detail but it could bring complaints and wouldn't make sense to change after comitting to one
03:41:54 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
03:42:28lucentmost users should be okay with the build dir being /tmp/rbdev-build
03:42:43lucentthe two common options to change would be the download dir and prefix setting
03:42:52rasherlucent: I think it makes sense to use the prefix for all three of none. And "DOWNLOAD" and "BUILD" doesn't make sense
03:43:18lucenthm, I agree
03:43:36lucentit's traditional to export PREFIX though when compiling software
03:43:53lucentoh well if it's wrong someone will have to convince us a better way
03:44:22rasherMaybe even for that reason, we shouldn't do it? Might we end up confusing the compile processes?
03:44:34 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
03:45:32lucentrasher: I like what you say about that
03:48:15rasherLooks good
03:48:31 Part ackers
03:48:56lucentdoes it need to be entered into a flyspray bug to be committed?
03:52:42rasherProbably best. I don't know if changes to peripheral stuff is fine during the freeze
04:01:08rasherlucent: So having it in the tracker is probably the best option, at least it won't get lost if I forget
04:03:36lucentrasher: thanks, I'm registering for flyspray and will post the patch with a new bug report
04:08:27 Join blkhawk- [0] (
04:10:14mud-rbalright, this is kind of random, but how does sudoku not appear in CATEGORIES? (trying to understand the build process for plugins since mine won't link)
04:11:37rashermud-rb: it's in viewers.config
04:13:00mud-rbrasher: i see. so if it's in viewers.config it shouldn't be in CATEGORIES?
04:13:28rashermud-rb: I don't know, to be honest
04:15:49lucentwho can add new player targets to the Flyspray system?
04:16:08lucentlike Sansa Fuze, Sansa Clip, Sansa Ex0 (AMS)
04:16:15rasherlucent: Admins. Limited to Bagder, Zagor and LinusN, I think.
04:20:10lucentrasher: submitted as FS9647
04:25:48Unhelpfulamiconn: ok, what i'm going to do is add a struct that gets passed to image loaders, with a line output callback, and a calc_size function, since the loader is supposed to return the size. if NULL is passed for this, the loader and the scaler (if used) just do what they've always done, so far, and use the built-in native bitmap support. if something gets passed, *that* gets used to calculate sizes, and to output lines.
04:25:56 Quit blkhawk (Read error: 110 (Connection timed out))
04:26:25 Nick blkhawk- is now known as blkhawk (
04:38:16***Saving seen data "./dancer.seen"
04:44:49 Quit midkay ("Leaving")
05:04:44 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:11:11Unhelpfulshould i be putting non-critical scaler dev work on FS during the freeze?
05:11:59 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
05:12:06lucentUnhelpful: it won't hurt if you do, my guess
05:12:34Unhelpfulonly other question is new task or the pre-merge one :/
05:20:18 Quit mc2739 ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
05:21:57saratogafile spray entries are fine during the feature freeze
05:22:23saratogathough if its for the scaler maybe it should go in svn if it fixes anything
05:26:01Unhelpfulsaratoga: it doesn't fix anything except readability, binsize, or speed
05:26:52Unhelpfulfixes i've been pushing straight to svn
05:41:58 Quit amiconn (Nick collision from services.)
05:42:01 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
05:43:55 Quit Horscht ("IRC is just multiplayer notepad")
05:49:54 Quit Tetracomm (Read error: 60 (Operation timed out))
05:52:55 Quit alexbobp (Read error: 104 (Connection reset by peer))
05:55:45 Join __lifeless [0] (n=lifeless@
05:56:10 Quit _lifeless (Remote closed the connection)
05:57:47 Join pixelma_ [0] (n=pixelma@rockbox/staff/pixelma)
05:59:43 Quit pixelma (Read error: 110 (Connection timed out))
06:02:58 Nick __lifeless is now known as _lifeless (n=lifeless@
06:03:06 Join alexbobp [0] (n=alex@
06:05:08 Join Tetracomm [0] (n=nicholas@
06:05:42 Join kremlin [0] (
06:11:29kremlinI was wondering earlier, would it be possible to create a rockbox extension and driver pair which would allow a media player running rockbox to be used as a USB sound card? Seems like it would be an interesting hack if someone was willing to give it a shot.
06:11:48scorchedepending on the device, yes
06:12:43scorchethe idea has been kicked around a bit, but hasnt had much work put into it
06:13:57 Join Xerion_ [0] (
06:21:45 Join Chesteta [0] (
06:22:55Chestetahello; I am wondering about the e200v2 keymap; on the hardware map site it mentions its hte same as on the e200 (v1)... how can I get it working on my e200v2?
06:24:12Chestetaactually i cannot find where it says that however i remember reading it somewhere ?
06:24:23Chesteta(i just looked at the page)
06:28:20 Quit toffe82 (Read error: 110 (Connection timed out))
06:30:23 Quit Xerion (Connection timed out)
06:30:23 Nick Xerion_ is now known as Xerion (
06:33:44 Quit XavierGr ()
06:38:17lucentChesteta: are you referring to Sansa-AMS ?
06:38:21***Saving seen data "./dancer.seen"
06:49:12 Join toffe82 [0] (
06:49:39 Part toffe82
06:50:36Chestetahey lucent
06:51:23lucentChesteta: check a look at flyspray and search for "fuze"
06:51:49lucentChesteta: you should see 4 tasks, one of them has a mention about the e2x0
06:52:18Chestetathanks man
06:54:03Chestetahave you tried it on your fuze?
06:54:21lucenthaven't tried rockbox on -my- fuze yet, no
06:54:39lucentI'm kind of suffering through borrowed wifi access though heh
06:54:45lucentwhen I can I am going to try
06:55:27Chestetaah right on
06:55:44Chestetai hear ya man
06:58:15 Quit Davide-NYC ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
07:03:32 Quit saratoga ("CGI:IRC (EOF)")
07:05:51mud-rbi figured out my build problem, but the answer doesn't make any sense. i sure hate make
07:12:09 Join fdinel [0] (
07:15:39 Quit fdinel (Client Quit)
07:23:57 Join einhirn [0] (
07:36:35 Quit Tetracomm ("Visit:")
07:47:12 Quit Chesteta (Read error: 113 (No route to host))
08:12:54 Quit GodEater (" ajax IRC Client")
08:13:01 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
08:14:22 Join ender` [0] (
08:16:31 Join Rob2223 [0] (
08:23:49 Quit Rob2222 (Read error: 60 (Operation timed out))
08:27:01 Join bertrik [0] (
08:33:38 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
08:36:40 Quit ender` (Read error: 104 (Connection reset by peer))
08:36:56amiconnlucent: The mandelbrot plugin isn't my work originally, but I've done a lot of work on it, and the original author isn't around anymore these days.
08:37:23amiconnWhat you experience when zooming in too far isn't overflow, but an effect of limited fixed-point precision.
08:37:28lucentamiconn: acknowledged.
08:38:24***Saving seen data "./dancer.seen"
08:38:29lucentamiconn: just hoping it's not an indicator of a serious bug
08:38:36amiconnI the beginning this wasn't an issue, since the plugin was slow, and the only targets back then were slow, so not many people zoomed in that far
08:38:56 Join ender` [0] (
08:39:07 Join slact_ [0] (
08:39:36amiconnBut now, with all the optimisations and fast targets, it doesn't take that long to reach those zoom levels
08:39:47 Quit slact (Read error: 104 (Connection reset by peer))
08:39:48 Nick slact_ is now known as slact (
08:41:04amiconnSo far the plugin doesn't limit the zoom level. It probably should.
08:42:07amiconnIf we want to allow further zooming in, we would need to go 64 bit for the coordinates, and then we'd need an optimised 64x64->128bit multiply routine for deep zooming
08:43:15lucentamiconn: I think adding a limit is sensible
08:43:55lucentthe bug I experience is not that it can't zoom in any further, but that it can't zoom back out to get to a sane state
08:43:57 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:46:18lucentBagderr: Hi there, I heard you are the guru to ask about TWiki registration errors?
08:47:24 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:47:50B4gderI could have a look
08:48:55lucentit seems though new users are being created but not added to the list of users
08:49:45lucentstep to reproduce this issue is to register a new user, I don't think there is a special step, it has happened to me and one other user mentioning this, we are registering within the past 3-4 days
08:51:24 Quit BHSPitLappy (Remote closed the connection)
08:52:43 Join erickytam [0] (n=46c555e8@gateway/web/cgi-irc/
08:53:11erickytami have a problem trying to install rockbox on my ipod
08:53:35erickytami downloaded all the necessary files and the ipod patcher, but every time i try to install anything it says the ipod wasn't detected
08:53:42erickytamand i know it's connected
08:53:46erickytamand it's in disc mode
08:53:57erickytambut it won't let me install anything
08:54:05scorchewhat ipod do you have?
08:54:15erickytam80g 5.5g
08:54:38Lloreanerickytam: Out of curiosity, why didn't you try RBUtil?
08:55:05erickytami've tried the 1.0.7 and b versions, are those not it?
08:55:17scorcheare you sure it is not a classic?
08:55:17LloreanThose aren't iPodPatcher.
08:55:37LloreanRButil should report if it's a classic, I think.
08:55:43erickytami've downloaded the ipod patcher seperately
08:55:44 Quit bertrik (Remote closed the connection)
08:56:04Lloreanerickytam: You don't need ipodpatcher if you're trying to use rbutil. What error does rbutil give when you try to detect it?
08:57:01erickytamit goes from Starting bootloader installation/Searching for ipods/No ipods found
08:57:19erickytami set it to detect the F drive and look for the 80gb ipod
08:57:23erickytambut it won't find it
08:57:28LloreanClick the auto-detect button.
08:57:29erickytamand i've tried the 30gb setting
08:57:39LloreanI asked what happens when you try to detect it, not what happens when you manually point to it.
08:57:42pixelmaare you sure it's a 5.5G not an Ipod "Classic"/6G?
08:58:05B4gderlucent: you you're added but not present in the users list now?
08:58:06erickytamyup, i bought it a couple of weeks ago
08:59:02 Quit jhulst (Read error: 148 (No route to host))
08:59:09advcomp2019erickytam, most likely it is a classic
08:59:13 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:59:15LloreanB4gder: The user page for created accounts doesn't seem to appear, but it does react somewhat differently if they try to use the confirmation code agian.
08:59:54B4gderthe TWikiUsers file had no write permisson for the web user for some reason
09:00:45B4gderZagor: can you think of a legitimate reason for that?
09:01:07Zagorrcs perhaps?
09:01:08scorcheB4gder: while you are checking that, and are messed up as well ;)
09:01:23lucentB4gder: pretty sure, I'm Eric Shattow, my TWiki username is EricShattow
09:01:51lucentother weird things today, got rockbox humping along on my Sansa Fuze, but it's stuck in karaoke mode (ha?)
09:02:09B4gderthose *ListBy* pages need some magic
09:02:15ZagorBagder: no, all other topics have write permission so it sounds wrong
09:02:18B4gder"Illegal modulus zero"
09:02:35B4gderZagor: yes, my impression too and I've now done a u+w on it
09:02:37Lloreanlucent: Audio support isn't really finished on it, it may just be improperly initialized hardware.
09:02:56pixelmaBagder/Zagor: could that be the reason why some user names are marked as unknown wiki words on the "recent changes" table even though they should exist since they just esited a page?
09:02:57lucentit's just funny
09:03:04lucentI'm impressed it does even this much
09:03:12B4gderpixelma: I would think so, yes
09:03:34lucentpleased enough with this that I can feel better about donating a Sansa Clip 2GB to Thomas M
09:05:12 Quit erickytam ("CGI:IRC (EOF)")
09:06:11Lloreanlucent: One thing. We ask that you try to use real English where possible in this channel, specifically because English isn't the native language for many people here. Could you try to avoid phrases like "humping along" which aren't something they're likely to be able to understand or find in a dictionary?
09:07:43lucentsee, it's a weird day :)
09:08:13fluxit's not that difficult to figure what it meant, even if I one hadn't seen the expression before :)
09:09:19Lloreanflux: It can be for some users. We have people in here who depend on machine translation even.
09:10:51 Part Llorean
09:11:09 Join Llorean [0] (
09:17:18 Quit parafin (Read error: 110 (Connection timed out))
09:18:32 Join petur [50] (n=petur@rockbox/developer/petur)
09:20:33lucentis a "works for me on XYZ hardware" helpful for flyspray tasks like FS #9645 ?
09:24:05Lloreanlucent: Probably not. Probably should only comment when/if it doesn't work as it's supposed to, as that would indicate hardware variation or something wrong with it.
09:24:11 Part Llorean
09:24:16 Join n1s [0] (n=nils@rockbox/developer/n1s)
09:25:03lucentthen how the hell do you know if it works?
09:28:58Zagorwe never commit stuff without testing
09:29:15Zagorhowever I personally think a comment saying someone tested it with no ill effects is useful
09:30:57 Join Llorean [0] (
09:33:05 Join PaulJam [0] (
09:35:03lucentFS9623 and FS9645 are really nice :)
09:37:32Lloreanlucent: In response to the "how do you know if it works", is there an assumption going on that the person who wrote the patch didn't use it?
09:40:28lucentthey didn't use it on an 8GB Fuze, that's for sure
09:41:49LloreanDo we compile a different build for the 8gb Fuze than the other Fuzes?
09:43:31lucentLlorean: have a nice night
09:49:48 Join Thundercloud [0] (
09:49:48n1sZagor: did you see FS #9644 ? not so sure about my "fix", there should be a nicer way to do this. (or we just tell people to use sane dir names) :)
09:51:59Lloreann1s: "sim.clip" is the folder they're building *in*?
09:52:33 Quit kachna (Read error: 113 (No route to host))
09:52:41n1sand we sustitute .c for .o everywhere in the path and then we can't find the .o files :)
09:52:48B4gder"sorry sir, we detected a silly directory name and refuse to build in there"
09:52:58LloreanWhy do we leave the build directory and substitute outside of it?
09:52:59 Join write_erase [0] (n=Olivier@
09:53:07linuxstbn1s: The bug happens if ".S" or ".c" is in the middle of a directory name?
09:53:14n1slinuxstb: yes
09:53:30n1ssim.clip -> sim.olip
09:53:41B4gderkinda silly bug really
09:54:42 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:54:44 Join robin0800 [0] (
09:54:45write_erasehi ... I compiled rockbox, then I installed rockbox firmware, but I get codec failure... Did I miss something ?
09:54:57B4gderwrite_erase: yes, a correct install
09:55:22Lloreanwrite_erase: Did you do "make zip" then extract the whole zip to your player?
09:55:23B4gderhow did you install it?
09:55:33write_eraseI had an already working rockkbox... I just compiled it from svn and upgrade the firmware
09:55:44write_eraseno make zip :-)
09:55:50LloreanYou can't just copy rockbox.ipod over
09:55:52B4gderthen no candy for you
09:56:13LloreanThe .codec files (and many others really) should or often need to be the exact same version.
09:56:13Zagorn1s: you fix looks good. I'll take a deeper look.
09:56:19write_eraseLlorean, make zip in build folder & then unpack it in .rockbox ?
09:56:45LloreanNot in .rockbox
09:56:51LloreanThe zip will contain a .rockbox folder, so the root of your device.
09:57:11n1sZagor: afaiu it will replace the suffix of any file with .o but maybe that's not a problem
09:57:17write_eraseIt's amazing how well your toolchain works in FreeBSD.... It's oven a nightmare to cross-compile on that OS.
09:57:39B4gderI don't believe that
09:57:44B4gderbut it's OT anyway
09:58:23Zagorn1s: there should be only source files in those lists, so it ought not be a problem
09:58:42write_eraseB4gder, Works from scratch, with (just had to change TMP Dir and destination location so the toolchain is installed in /home/someone/toolchain)
09:59:00B4gderthat's good to hear
09:59:07B4gderwe did put some effort into that...
09:59:16write_eraseSure you did ...
09:59:33write_eraseWhat's the licence for rockbox ?
09:59:41n1sgpl v2 or later
10:00:08n1seasy to find info, it's in every source file :)
10:00:38 Join lasser [0] (
10:02:04 Join Nibbl [0] (
10:08:43 Quit robin0800 (Read error: 104 (Connection reset by peer))
10:11:17write_eraseLlorean, now I have sound ... make zip did it
10:11:49LloreanI thought it might. :)
10:16:14Zagorn1s: "i'm not sure that's a good idea though as it will add all kinds of files to the obj lists" <−− is this something you have observed?
10:16:48n1sZagor: no, just an assumption because it strips any suffix and adds .o
10:17:46Zagorn1s: ok. then I think you should commit this.
10:18:07n1sok, thanks for looking
10:19:12n1shmm, should it wait until after the freeze?
10:19:49Zagorno, this is a bug fix
10:20:31linuxstbCan't you just use the "bar := $(foo:.c=.o)" construction?
10:20:58Zagorlinuxstb: that causes this bug
10:21:14linuxstbI thought the problem was using "subst" ?
10:21:57Zagorno the problem is substituting ".c" blindly, which substitution references (which you link to) does aswell
10:22:33Zagorthe reason I was using subst explicitly instead of subst reference was that the reference had problems with the long pathname subst
10:23:17linuxstbIIUC, the construction I quoted only matches at the end of a word - that's what the documentation says at that link.
10:23:30Zagor$(foo:%.o=%.c) would work though
10:23:46Zagoror perhaps not. I have to test.
10:24:47Zagorindeed it does. then yes, that would work too.
10:29:21 Quit Thundercloud (Remote closed the connection)
10:30:30n1sZagor: will you commit or should i cook up a new patch?
10:30:49ZagorI can commit
10:31:54 Quit Seed ("cu, Andre")
10:38:26***Saving seen data "./dancer.seen"
10:39:44 Join parafin [0] (
10:49:09 Join kharo [0] (
10:51:47 Nick fxb__ is now known as fxb (
10:57:56*JdGordon cant think of any nice way to get a config/wps browser/chooser into the quickscreen :/
10:58:26JdGordonactually... maybe /me can!
10:59:52JdGordonhmm.... nope... :p
11:07:07JdGordonanyone use radio presets? FS #9646 could use a tester...
11:08:27n1sthat kind of patch is a good indication of what stuff devs never use :P
11:08:39JdGordonyeah :/
11:11:12pixelmaI actually saw these problems related to deleting presets recently...
11:12:07JdGordoncare to double check the patch? looks good assuming it doesnt add any more bugs
11:14:58pixelmaok, will try. Might take a bit though
11:17:02JdGordonok great..
11:17:31JdGordonthe patch has one thing that needs changing before commit which I'ev fixed here so let me know if its good and ill put it in
11:22:37JdGordondoes anyone else have an opinion on alot of string buffers in rockbox being char blaa[SOME_MAX_DEFINE + 1]; ?
11:22:43*JdGordon thinks that should stop
11:24:13mud-rbJdGordon: but how would you fix that? I can't think of a way that doesn't just move the 1 to a different place.
11:24:19n1sI guess it's confusion about if the MAX_FOO defines includes the terminating null, which imo it should
11:24:39JdGordonyeah, the max should be including the \0
11:24:50JdGordonits not a big deal, but imo looks cleaner
11:25:47n1sbut you must then make sure that a string you copy to that buffer is not longer thatn MAX_FOO -1
11:26:26 Join robin0800 [0] (
11:27:06JdGordonwhich is what strncpy is used for... unless our one doesnt add the \0 if the src string is the length of the dest...
11:30:02Zagorhowever we do, there will always be a +1 or -1 somewhere
11:30:40Zagorit's just the nature of C string handling
11:31:02B4gderand strncpy() is very often wrong where strlcpy() is right...
11:32:04JdGordonso I should ignore it and we will just use whatver?
11:33:18mud-rbwell, the cleanest way is generally to pick one way to do it and go with that universally..but i doubt you want to go through and make them all the same (and you'd likely introduce a large amount of bugs)
11:33:19n1sdo we have a strlcpy?
11:34:31n1swell, it seems much saner than srncpy so maybe we should add one, and start using it?
11:35:04B4gderI too think that's more often in line with what you actually want
11:35:24B4gderbut switching to that might be a bit of a work...
11:35:57JdGordonor, we could modify strncat to always add the \0 regardless of how many chars were copied
11:36:01JdGordonthat would make it just as safe
11:36:24B4gderstrncat ?
11:36:27B4gderdo we use that?
11:36:35Zagorif we add strlcpy, I say we should remove strncpy.
11:36:37JdGordoncpy i mean
11:36:45B4gderJdGordon: but that would change its behavior/function
11:36:53B4gderwhich may also break things
11:37:01B4gderZagor: I agree
11:37:13B4gderstrncpy()'s evil padding is also widely not understood
11:37:27n1syes, that padding it a bit stupid
11:37:32mud-rbwow, i had no idea strncpy was defined that way (looking at the man page). i can't imagine when i would want that behavior...
11:37:52 Join Darksair [0] (n=user@
11:41:34n1sgreping for strncpy gives just over 200 results in rockbox so, yes quite a bit of work to switch
11:42:08JdGordondoes snprintf have the same issue?
11:42:13ZagorJdGordon: no
11:42:38 Quit parafin ("So long and thanks for all the fish")
11:42:45JdGordonit \0's if the dest isnt big enough?
11:46:10 Join mofux [0] (
11:55:14robin0800is any one able to sync this patch, FS #8363 - Charger configuration for Sansa e200/c200, to make it work with latest svn
11:58:12 Nick fxb is now known as fxb__ (
11:58:18 Quit kharo (Read error: 110 (Connection timed out))
12:00:25 Join kharo [0] (
12:12:04 Quit _lifeless (Remote closed the connection)
12:13:22 Join _lifeless [0] (n=lifeless@
12:19:21 Join Zom_ [0] (
12:19:24 Quit soap (Remote closed the connection)
12:19:35 Join soap [0] (
12:21:28 Quit Zom (Read error: 104 (Connection reset by peer))
12:21:55 Join __lifeless [0] (n=lifeless@
12:29:20 Join moos [0] (i=moos@rockbox/staff/moos)
12:38:21 Join kugel [0] (n=chatzill@unaffiliated/kugel)
12:38:30***Saving seen data "./dancer.seen"
12:40:31 Quit _lifeless (Read error: 110 (Connection timed out))
12:50:32 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:01:00 Join Aurix_Lexico [0] (
13:02:23 Quit Darksair (Read error: 104 (Connection reset by peer))
13:08:30 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
13:09:15pondlifeJdGordon: Don't know if I missed the discussion, but I prefer to see the +1 for all null-terminated strings
13:09:32pondlifeThat could mean you #define MAX_PATH 256+1 and char p[MAX_PATH] though
13:09:51pondlifeYou probably agree with me
13:10:20pondlifeI agree with scrapping stncpy() too
13:10:54JdGordonwell, I think that it should be inplicit that the MAX define includes the \0...
13:10:56Zagorpondlife: the question is not if we see +1, but where. in the variable declaration or its' use.
13:11:31pondlifeOK, I'm not fussed too much, but I like the +1 to be visible, personally
13:12:41pondlifestrlcpy() would be very good
13:13:06B4gderbut maybe we should save the strncpy() => strlcpy() for 3.2
13:13:24pondlifeYes, it's not one for the freeze period
13:13:44B4gderbut it could very well be an early action after the 3.1 release
13:14:06pondlifeIs 3.1 release meant to be 24th Dec?
13:14:18JdGordonI think I just found a dependancy problem.... I commented out a #define in recorder/radio.c and did make bin but it said nothing to be done
13:14:29B4gderpondlife: 25th or 26th I'd say
13:14:45*JdGordon likes the rock-boxing day release
13:14:57ZagorJdGordon: try just "make"
13:15:04pondlifeBit of a belated present
13:15:16JdGordonZagor: recretaed credits.raw...
13:15:28ZagorJdGordon: and nothing more?
13:15:59JdGordonthis is a sim build by the way, but that shouldnt make a difference?
13:16:05ZagorJdGordon: I assume your target has radio support?
13:16:44Zagordo you have apps/recorder/radio.o in the build dir?
13:17:08JdGordontarget build works, sim doesnt
13:17:10*JdGordon checks
13:17:36Zagorwhich target is it?
13:19:47Zagorworks here
13:20:01JdGordongrr... sorry, false alarm... moved the source tree and didnt reconfigure
13:20:07JdGordonwhich is annoying anyway
13:20:29JdGordondidnt it used to use relative paths?
13:22:44ZagorI don't think so. I didn't change anything fundamental in configure when I redid the build system
13:24:58 Join AndyIL [0] (i=AndyI@
13:29:19JdGordoncan someone with a e200 check the fm scanning with svn? does it think every possible station is tuned?
13:35:43 Quit AndyI (Read error: 110 (Connection timed out))
13:46:46 Quit Aurix_Lexico ("Leaving.")
13:48:57 Join Schmogel [0] (
13:50:36 Quit markun (Read error: 54 (Connection reset by peer))
14:10:30 Join kachna [0] (
14:13:02 Join markun [50] (n=markun@rockbox/developer/markun)
14:28:43 Join Strife89 [0] (n=michael@
14:37:02kugelZagor: I don't know if it was supposed to be fixed with the overhaul, but I'd thought so: If you change the target-config, everything is remade properly, but apparently settings_list not
14:37:34Zagoryou have to clean the dir before changing target
14:38:11kugelno, not changing the target, change the config-<target>.h file
14:38:34***Saving seen data "./dancer.seen"
14:39:15kugellets say I remove #define HAVE_RECORDING since I don't like to record, apparently the whole tree is remade, but then I still get compiling errors with settings.h and/or settings_list.c
14:39:43Zagorit rebuilds for me
14:39:56kugelcould be linker errors too, I'm not sure, it's loads of errors
14:40:05ZagorI'll try your specific example
14:40:35 Quit J-23 ("ZNC -")
14:40:56kugelI have experienced with several #defines, it rebuilds all fine, except for settings_list
14:41:49 Join J-23 [0] (
14:46:17 Join gregzx [0] (
14:47:22Zagorkugel: well does it not rebuild or does it not work? there's a big difference.
14:47:42kugelit rebuilds all fine, except for settings_list
14:48:17ZagorI can't repeat that. settings_list builds for me.
14:49:17kugelmaybe have_recording was a bad example
14:49:38Zagorit doesn't matter which define you change. make only looks at the timestamp of the file.
14:49:53kugeleven though I get errors in other files after removing HAVE_RECORDING
14:50:14 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
14:51:27kugelso, then such things wasn't supposed to be fixed with your remake? I still need to make clean after changing config-target.h
14:56:07 Join FireThrower [0] (
14:56:07 Quit kachna (Read error: 110 (Connection timed out))
14:56:41kugelZagor: hm, is it possible that I only got the errors with make bin?
14:56:59Zagorkugel: not likely. it uses the same dependencies.
14:57:17kugelI'm a bit failing to reproduce for some reason.
14:57:30kugelbut I still get errors, just not in settlings_list, but in other files
14:58:02kugele.g. in button-e200.c, after removing the #HAVE_BUTTONLIGHT define
14:58:24kugelthat's a linker error
15:01:02Zagorlink errors are generally not cause by make
15:02:26UnhelpfulZagor: if we're going to have devs using the git mirror regularly, it might be a good idea to commit .gitignore files to svn, or to maintain them in the git mirror
15:02:31kugelZagor: hm, yea, this one isn't properly #ifdef'd
15:02:37FireThrowerI have a Sansa e280 and I plagged it to my computer in MSC mod,I extracted .rockbox to the main Folder of my e280 and then I tried to start the "sansapatcher" program who should install the bootloader on my e280, and the program wrote that it isn't find any sansa's. please help me
15:03:00kugelprobably the others which I had problems with too, I'll have a closer look next time I get such an error
15:03:10Unhelpfulthere's a git-svn tool to generate them from svn ignore props
15:03:33ZagorUnhelpful: officially we don't support git. it's just a fad ;-)
15:03:59 Quit J-23 (Remote closed the connection)
15:04:18 Join J-23 [0] (
15:05:28linuxstbFireThrower: 1) Are you sure it's a "version 1" e280? 2) Are you running sansapatcher as root/Administrator?
15:05:53*kugel wants someone to commit FS #9623 and FS #9645 :(
15:06:22 Quit J-23 (Client Quit)
15:06:48kugellinuxstb: mkamsboot isn't prepared for the new fuze firmware, is it? 1.01.22 I mean
15:07:09 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
15:07:35FireThrowerI just discoverd that my e280 is V2. thank you anyway. can I downgrade my V2 firmware to V1 firmware?
15:07:47linuxstbFireThrower: No, "V2" relates to the hardware, not the software.
15:07:48Unhelpfulunofficially you're doing a lot of other things to support git ;)
15:08:12Unhelpfulputting them in the unofficial git mirror would be pretty unofficial ;)
15:08:13linuxstbkugel: If no-one has added the md5sum, then no.
15:08:17ron_obtw, there are no instunciton in the manual for the e200 for manually installing a theme to rockbox.
15:08:36kugellinuxstb: Ok, I'll prepare a patch
15:08:45linuxstbkugel: No need - I'll add it.
15:08:57Unhelpfulron_o: unzip it on your player
15:09:09ron_oUnhelpful, I found that out now. :)
15:09:24ron_obut someone last night told me it should be in the manual. It's not.
15:09:25kugellinuxstb: ok, but I didn't test if the firmware works
15:09:38Unhelpfulor the util can do it, i think
15:09:56ron_oit can do it but there's problems with automatically doing it ATM.
15:10:28 Join J-23 [0] (
15:10:33kugelyou just drag and drop the unzipped theme on the root of your dap
15:10:34FireThrowerOne last questation: why there is no bootloader for V2 sansa?
15:10:52kugelin most cases, that is, when the zip contains the .rockbox directly at the top level
15:11:22linuxstbFireThrower: It's in development.
15:11:22B4gderFireThrower: the v2s are now called AMS Sansas by us, and they are work in progress. There is no complete or well polished install method yet
15:11:28 Quit slact ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
15:17:01FireThrowerok thank you
15:17:02 Part FireThrower
15:19:23 Quit Strife89 ("I must remove my distractions ... for now. :)")
15:22:39 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111318]")
15:24:45 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:28:10 Join Jaykay [0] (
15:30:13Jaykaytwo questions: how do i attach files (a battery bench) in the wiki?
15:30:31Jaykayand is there any need in unifying(?) the runtime-pages?
15:30:49 Quit ron_o (Read error: 110 (Connection timed out))
15:31:23 Join ron_o [0] (
15:32:54 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
15:34:05linuxstbJaykay: If you're logged in, you should see an "Attach" link at the very bottom of every wiki page.
15:34:35 Quit Nibbl ("Ex-Chat")
15:35:09 Quit lasser (Remote closed the connection)
15:35:23Jaykayand the second question?
15:36:17linuxstbI have no opinion.
15:37:12 Join {phoenix} [0] (
15:37:51Jaykay-.- ok
15:38:52 Join courtc_ [0] (n=court@unaffiliated/courtc)
15:38:57 Quit Zagor (
15:38:57 Quit nplus (
15:38:57 Quit krazykit (
15:38:57 Quit courtc (
15:39:01NJoinnplus [0] (n=nplus@141.25.Globcom.Net)
15:39:17NJoinZagor [0] (n=bjorn@rockbox/developer/Zagor)
15:40:00 Join krazykit [0] (n=kkit@
15:40:26 Quit Horscht (Read error: 110 (Connection timed out))
15:44:26 Join Horscht [0] (
15:51:01 Quit Jaykay (Read error: 104 (Connection reset by peer))
15:54:42 Join lasser [0] (
16:00:25 Nick fxb__ is now known as fxb (
16:03:40 Quit __lifeless (Remote closed the connection)
16:04:00 Join __lifeless [0] (n=lifeless@
16:04:27 Join gregorovius [0] (
16:04:44 Quit CaptainKewl (Read error: 110 (Connection timed out))
16:06:09 Join kachna [0] (
16:13:02 Part pondlife
16:14:40gregoroviusHi. I have a 5g ipod running rockbox. I don't know if it's the database, but when I turn it on the HD gets trashed for a couple minutes, and I can't do anything in the meanwhile... If I try to play some music is gets stuck there for up to 5 minutes. I'm running a couple days' old svn, is it normal?
16:15:54n1sgregorovius: do you have the database auto update turned on?
16:24:44 Quit {phoenix} (Remote closed the connection)
16:25:17 Join parafin [0] (
16:26:58B4gderMAKEFLAGS ?
16:27:12B4gderthat's the one my info page mentions
16:27:57gregoroviusn1s: yes, but it happens both when I've updated my music and when it hasn't changed
16:28:55gevaertsgregorovius: do you have dircache and load database to ram enabled?
16:29:33rasherB4gder: I was close
16:29:52rasherBut basically nothing that should interfere with
16:30:15B4gdernope, I agree with you that it should already just work
16:31:24freddy_Is FS #9617 in svn ?
16:31:43gregoroviusgevaerts: yes. I disabled dircache and it was the same, I'm gonna try disabling auto update
16:32:49gregoroviushm, that's it, with only dircache after 10 seconds or so after starting up the hd spins down
16:33:16gregoroviusin any case, the updating of the db shouldn't trash everything so much...
16:34:05 Join Nibbl [0] (
16:34:52rasherI don't see how else you'd expect auto-update to work, if not scanning the disk
16:35:00 Join toffe82 [0] (
16:37:03Zagorperhaps a little more in the background
16:37:43rasherMaybe the auto-update process should yield more, if that's it?
16:38:01gevaertsgregorovius: and database in ram?
16:38:15Zagorrasher: it sounds like it, from gregorovius' description
16:38:36***Saving seen data "./dancer.seen"
16:39:57 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
16:41:53 Quit __lifeless (Remote closed the connection)
16:42:13 Join __lifeless [0] (n=lifeless@
16:43:01gregoroviusgevaerts: yes, database in ram as well. any tips?
16:43:40gregoroviusI just disabled database in ram and auto-update and in a couple seconds the hd had spinned down
16:48:10 Join evilnick [0] (i=0c140464@gateway/web/ajax/
16:49:10 Quit perrikwp (" ajax IRC Client")
16:51:09 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
16:54:47n1sgregorovius: did you say dircache made no difference?
16:55:54gregoroviusn1s, with auto-update and load-to-ram on, no, at least not anything perceptible
16:57:57n1sdo you have a lot of files on the device?
17:00:19 Join EspeonEefi [0] (i=eefi@SAFFRONCITY.MIT.EDU)
17:00:45 Quit Nibbl (Read error: 113 (No route to host))
17:03:29gregoroviusn1s: 6900, a tad less than 40gb
17:04:13n1shmm, does dircache work when browsing in the filebrowser?
17:04:57gregoroviuswell, browsing is quite snappy, so I guess yes
17:05:54n1swell, then the database update should not take _that_ long with dircache enabled...
17:08:06gregoroviusweird. I just turn on load-to-ram and auto-update, turned it on, and it was done in 15 secs or so
17:08:24gregoroviusbut on the bus home it hung almost 10 minutes
17:08:59n1sit could be some weird timing issue
17:09:24gregoroviusmight be, it has happened 2 or 3 times already
17:09:27 Quit Zagor ("Client exiting")
17:09:53gregoroviusin any case, I have to wait until the hd spins down or playback stops every 15-30 secs
17:20:07 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:23:08 Quit write_erase ("Leaving")
17:24:04 Quit perrikwp (" ajax IRC Client")
17:30:53 Quit ron_o (Read error: 110 (Connection timed out))
17:35:50 Quit __lifeless (Remote closed the connection)
17:36:02 Join _lifeless [0] (n=lifeless@
17:40:31 Join __lifeless [0] (n=lifeless@
17:47:03J-23how to select positions from menu on e200? Scrollwheel?
17:47:05 Join T_Blade_Gangsta_ [0] (
17:47:10T_Blade_Gangsta_Hey ;-)
17:49:28 Quit n1s ()
17:52:03 Quit gregorovius (Read error: 110 (Connection timed out))
17:52:33 Join gregorovius [0] (
17:53:10 Quit T_Blade_Gangsta_ ()
17:53:45J-23nvm, I checked it on simulator
17:55:12LloreanJ-23: There's also the manual.
17:55:38 Quit moos ("Rockbox rules the DAP world")
17:58:40 Quit _lifeless (Read error: 110 (Connection timed out))
18:01:11 Quit petur ("work->home")
18:02:49 Part B4gder
18:07:27 Join kugel [0] (n=chatzill@unaffiliated/kugel)
18:12:20 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
18:21:50 Join _lifeless [0] (n=lifeless@
18:23:29rasherAttention all devs: Please use the following when deprecated translation phrases: tools/ −−deprecate −−id LANG_SET_BOOL_YES −−inplace apps/lang/*lang
18:24:00rasherAnd other langtool switches as well. It makes things easier for translators.
18:24:17LambdaCalculus37rasher: Acknowledged.
18:26:32 Join faemir [0] (
18:27:20 Join bertrik [0] (
18:29:02 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:35:19 Join {phoenix} [0] (
18:35:33 Quit mofux (Read error: 110 (Connection timed out))
18:35:44 Join mofux [0] (
18:38:39***Saving seen data "./dancer.seen"
18:41:21 Quit __lifeless (Read error: 110 (Connection timed out))
18:41:47 Join __lifeless [0] (n=lifeless@
18:42:01 Join miepchen^schlaf [0] (
18:42:11 Quit _lifeless (Read error: 110 (Connection timed out))
18:45:23 Join Seed [0] (
18:58:13kugeldomonoky: hi :)
18:58:45kugeldomonoky: re GPIOA in lcd: the fuze lcd driver has those too, and this is purely based on disassembly
19:00:11domonokyso we should still find out if everthing is needed, and if we can findout what it really does. :-)
19:00:41kugeldomonoky: No! Don't question linuxstb's disassembly! :)
19:01:18kugeldomonoky: and re: buttons on fuze: I got it working nicely, w/o lcd hacks (besides the mutex, but I wouldn't count this as hack)
19:01:55kugelthe lcd weirdness was caused by changing the afsel on the whole port, instead of only the pins we need
19:02:10 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
19:02:56 Part LinusN
19:03:07domonokystrange, but if it works better :-)
19:03:32kugelyes it does, and I don't need to hack update_rect into update
19:04:00domonokygood. i will try it on my e200v2 later this evening..
19:04:53kugeljust copy and paste the whole fuze button driver and insert the mutex into lcd driver like I di
19:07:58 Quit Horscht ("electromagnetic radiation from satellite debris")
19:10:53 Join ZincAlloy [0] (n=d9eec17a@gateway/web/cgi-irc/
19:11:24 Join Thundercloud [0] (
19:15:12 Quit gregorovius ()
19:17:44 Join karashata [0] (n=karashat@
19:21:50 Join thegeek [0] (
19:26:44 Quit thegeek_ (Read error: 60 (Operation timed out))
19:28:25kugeldomonoky: on what are you focusing now? rest of the buttons, or sd, or playback reliability?
19:30:21kugelI'd really like to get my microsd mounted
19:30:54*domonoky focuses on not having to work.. till tomorrow morning.. :-)
19:31:46kugeldomonoky: btw: I'd like to get my button and lcd patch committed
19:32:27 Quit Bagder ("*plopp*")
19:35:08kugelunless there are objections, that is
19:37:22 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
19:37:53domonokythe buttons are probably fine, but i am still not sure if this use of the mutex-locked member is 100% save. perhaps someone with rb-kernel expirience could take a look at it.
19:38:28kugeljhMikeS: ping :)
19:39:14kugeldomonoky: it's defintely better then disabling irqs. disabling irqs can get weird behavior, since not only the buttons are interrupt driven
19:39:28kugelor am I wrong?
19:39:45 Quit faemir (Remote closed the connection)
19:40:07domonokykugel: sure its better then disabling irqs. But we should make sure accessing just the mutex member (not using mutex_lock() ) is 100 % save
19:40:36kugeldomonoky: mutex_lock() does lock, it doesn't return the locked state
19:41:03domonokykugel: thats correct, and it also blocks, which would be bad in a irq.
19:41:25kugelthat's why I just return if it's locked
19:41:39domonokywe would need a mutex_lock-nowait() to make it proper... :-)
19:41:42kugeland not e.g. wait until it's unlocked
19:42:31kugelwhy do you doubt that reading the locked member is unsafe?
19:42:32domonokybut you have to very carefull, when accessing the same var in a thread and in a irq (read-before write and opposite problems)
19:43:57kugelplease note that the buttons can interrupt the lcd, but not the opposite. if the button interrupt is emitted, the lcd has no chance anyway
19:44:48 Join faemir [0] (
19:44:49kugeland checking locked makes sure that the buttons do not disturb lcd transfers
19:45:09domonokyafter a bit of thinking, i think it should be save. but you cant be carefully enough with this things.
19:45:12kugelbut sure, lets ask a kernel/mutex expert
19:45:21 Quit miepchen^schlaf ()
19:46:12kugeldomonoky: it's not that I haven't tested it :) I was listing to my fuze in the tram today just fine (I was also browsing and stuff)
19:46:30kugellistening to my only flac file that is :P
19:47:15 Quit freqmod_qu (Remote closed the connection)
19:47:16domonokykugel: testing isnt enough for this things, you need to be really sure, or else a strange timing or else breaks it. But in this case it should work.
19:48:22kugeldomonoky: have you read this page?
19:50:19domonokykugel: yes, long ago. And i also hacked in other kernels, so i know the normal problems your get there. Luckily with cooperative multitaskting this locking buisness is much easier in rockbox :-)
19:50:59 Nick Zom_ is now known as Zom (
19:54:23 Join miepchen^schlaf [0] (
19:55:29 Quit faemir (Remote closed the connection)
19:55:59kugeldomonoky: I haven't actually tried not checking for mutex.locked at all, I'd be interested in what could/would happen
19:56:03 Quit ZincAlloy ("CGI:IRC (Ping timeout)")
19:57:07*domonoky can tell what will happen. you will interrupt the lcd, and probably cause stray buttons and strange lcd :-) But its worth a try.
19:57:32gevaertsJust checking and then locking won't solve anything
19:57:40kugelI expected similar things :)
19:58:36kugelgevaerts: :?
19:58:53gevaertskugel: you need an atomic check-and-lock
19:58:54domonokygevaerts: the lcd takes and releases the lock. buttons only check if its locked, and doesnt do anything if it is locked at this time.
19:59:46gevaertsdomonoky: sure, but what if it gets locked between checking and locking?
19:59:59*gevaerts hopes he understands the context right
20:00:05kugeldomonoky: it's seems to work just fine
20:00:06domonoky? the button doesnt take the lock, its a irq
20:00:13gevaertsAh, ok
20:00:44bertrikso, what do you do if it's locked, just skip a button reading?
20:00:45*gevaerts shuts up and goes back to his usual corner
20:01:04kugelI'm holding the down button to go down in the list, and it does that just fine, and the selector bar redraws just fine too
20:01:14domonokybertrik: thats what the patch does at moment.
20:01:31bertrikand you absolutely want to use an interrupt for the buttons?
20:01:47domonokybertrik: the tick-tasks run in a interrupt ...
20:02:22domonokyso we dont use a interrupt in the button code, but its called in interrupt context.
20:03:22domonokykugel: you are saying you can interrupt the lcd with the buttons without problems ? so dont need this mutex thing ?
20:03:48bertrikwhat pins are conflicting anyway? what functions do they have for the LCD?
20:04:13kugeldomonoky: seems so. I'm gonna try to remove the mutex alltogether
20:04:28domonokybertrik: lcd-dbop is on this pins
20:04:57gevaertsIf you don't need to read the buttons from interrupt context for other reasons, you can just wakeup a thread from the tick task or something like that
20:05:18kugeldomonoky: as fine as with mutexing...
20:07:01domonokykugel: sounds promising.. if it also works on e200v2 without locking. we dont need to care about which context the tick-task uses :-)
20:10:52 Join hillshum [0] (
20:13:21 Quit miepchen^schlaf ()
20:13:30*kugel goes updating the bootloader before making final claims
20:14:19kugelI claim it works
20:15:44*gevaerts claims that kugel is lucky :)
20:16:40hillshumis the log working fine? (i cant' get past 02:33
20:16:42kugelthe new 1.01.15 OF gave us some more space to insert code
20:17:03kugel150K now, instead of 123K
20:18:10hillshumnow its fine
20:20:30bertrikwow, is it really the case that the fuze lcd controller hardware is unknown, yet we have a driver for it?
20:22:14 Join miepchen^schlaf [0] (
20:23:49 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:24:10Unhelpfula reverse-engineered spec is better than a good name and no spec? ;)
20:26:15bertrikthere is an implementation, not a spec, as far as I know
20:28:28kugeldomonoky: btw: it's cleaner to only reset int_btn for the buttons we're going to read, instead of simply setting it 0
20:29:16 Join _lifeless [0] (n=lifeless@
20:29:30domonokykugel: depends on how we can integrate the other buttons :-)
20:29:51kugelnot really
20:30:05kugelit's cleaner in anyway imo
20:30:50 Join massiveH [0] (
20:30:50domonokywhy ? if you read all buttons in this function, you surly need to reset all buttons.
20:30:57 Quit EspeonEefi ("さよなら")
20:32:05domonokyand if we can read all buttons in this function, we should put everything directly in button_read_device(), like it is done on m200v4
20:32:46 Join akur [0] (
20:32:55kugeldomonoky: sure, reset all buttons you read in that function.
20:33:17kugelif you read all buttons in a functions, you of course reset all buttons :)
20:33:41kugel(but not more)
20:33:42 Part akur
20:33:55kugelanyways, I updated my patches
20:34:12*domonoky decides not to answer.
20:35:52 Quit miepchen^schlaf ()
20:37:13 Join Photoguy [0] (n=Photoguy@
20:37:40PhotoguyHas anybody here used froobi?
20:37:55kugelused? abused
20:38:00Unhelpfuli've got a 6GB sansa from them
20:38:17PhotoguyI ordered a 4gb e200 about 5 days ago..
20:38:26PhotoguyIt's still not shipped!
20:38:39Photoguy6 days
20:38:43***Saving seen data "./dancer.seen"
20:38:45LambdaCalculus37Photoguy: Then take it up with Froobi.
20:38:53LambdaCalculus37We can't help you with slow shipping.
20:38:55PhotoguyCall 'em up?
20:39:16PhotoguyJust wondering if the site was legitimate.
20:39:22gevaertsIt is
20:39:33LambdaCalculus37Photoguy: It is, but that's really off-topic in this channel.
20:39:49 Join miepchen^schlaf [0] (
20:39:50PhotoguyRight..sorry just a quick question.
20:40:31 Quit __lifeless (Read error: 110 (Connection timed out))
20:47:48 Join dany_21a_ [0] (
20:48:08 Join freqmod_qu [0] (i=quassel@2001:700:300:1430:213:d3ff:fee9:5ed0)
20:48:53dany_21a_hi kugel, just a quick note to you fuze-button patch (havent tested so far): you dont need to "GPIOC_PIN(2) |= (1<<2);"... "GPIOC_PIN(2) = (1<<2);" is enough, as the address already maskes away the other bits
20:49:25 Join ibseco [0] (
20:49:27kugeldany_21a_: well, we don't want to write the other bits
20:49:37dany_21a_you wont
20:49:49kugel(at least funman doesn't recommend)
20:50:03kugelif they happen to be not 0, for whatever reason, let them be 1
20:50:21 Quit ibseco (Remote closed the connection)
20:50:30kugeldany_21a_: you write 0 to the other bits on that pin with =
20:50:50kugeloh, I see what you mean
20:50:53dany_21a_In order to write to GPIODATA, the corresponding bits in the mask, resulting from the
20:50:53dany_21a_address bus, PADDR[9:2], must be HIGH. Otherwise the bit values remain unchanged
20:50:53dany_21a_by the write.
20:51:08dany_21a_(from the GPIO datasheet)
20:51:33dany_21a_and with the given macro the correct address is calculated
20:51:34kugelok ok :) but does it matter?
20:52:18mud-rbwell, it's an extra read, right?
20:52:27dany_21a_- not critical in your code, but just wanted to note, so that you have the right picture of the GPIO-data-register in other cases, where it might matter
20:52:32 Join seraph [0] (
20:52:40dany_21a_yes its an extra read... but only in the init-function
20:52:42kugelehh, why do I argue anyway? That was domonoky's part :D
20:53:22kugeldany_21a_: only in the init function? the bits are reset at every button read
20:53:25dany_21a_oh :) - just saw your patch on the flyspray...
20:53:39kugeldany_21a_: it's based on
20:54:08kugeldany_21a_: credits go to domonoky, for getting how to reset the pins correctly
20:54:36kugel(I apparently didn't, I always wanted to reset with = X, and not = 1<<X9
20:54:54 Quit freqmod_qu (Remote closed the connection)
20:55:17 Join freqmod_qu [0] (i=quassel@2001:700:300:1430:213:d3ff:fee9:5ed0)
20:57:16kugelgevaerts: any idea why my microsd doesn't show up in the filebrowser, or any hint on what to enable so that it shows up? It's initialized, and also shows up in the debug menu under partition info
20:57:29dany_21a_i dont understand why reading from the 0x3fc-address dosent work (have tried it once..) - as this should give the whole register (mask=0xff), would be also a speed up
20:59:07 Join mc2739 [0] (
20:59:11gevaertskugel: the two obvious causes would be missing HAVE_MULTIVOLUME and missing HAVE_FAT16SUPPORT of course, but config-fuze.h seems to have them covered.
20:59:36mc2739dany_21a_: 0x3fc does work
20:59:40dany_21a_the PADDR[9:2] from the read/write is used as mask for the GPIO-data registers... so it is (according to the manual) possile to read more than one bit at once
20:59:45kugelgevaerts: I remember I could boot from it back a few month, maybe the interrupt changes broke it?
21:00:09gevaertskugel: it could be a hotplug issue as well
21:00:13*gevaerts greps a bit
21:00:22dany_21a_kugel [9:2] means you have to calculate the addres=mask*4 (the macro does this)
21:00:34kugelgevaerts: hm, it doesn't matter if I insert pre-boot or when rockbox already up
21:00:43kugeldany_21a_: yup
21:01:04dany_21a_and 0xff*4=0x3fc
21:01:23kugeldany_21a_: iirc doing (1<<2|1<<3|...|1<<9) is also the same, without the calculation
21:01:41gevaertskugel: do you have logf? Maybe check if you get at the disk_mount(1) in ata_sd_as3525.c
21:02:11kugeldon't know if I have logf, but it's worth a try
21:02:51kugelgevaerts: but this is only if hotplugged, or am I wrong?
21:03:15gevaertskugel: depends. I don't know how this is handled at boot
21:03:41gevaertsAnyway, there could be other causes
21:04:02gevaertsDo you by any chance have more than one FAT partition on the internal flash?
21:04:10 Quit Schmogel (Read error: 110 (Connection timed out))
21:04:58 Quit freddy_ (Read error: 110 (Connection timed out))
21:05:22 Join fredddy [0] (
21:05:47kugelgevaerts: possible, I don't know what mkfs.vfat exactly did. But only 1 partition shows up when I plug the fuze to the pc
21:06:07gevaertskugel: mkfs.vfat won't make partitions
21:06:54gevaertsHave a look at the output of sfdisk -l /dev/sdX (for both internal and sd)
21:07:12kugelthere was a 4GB partition before, and mkfs.vfat made a 500MB one, I don't know what exactly happened to the other 3.5GB. But fact is, that the microsd is recognized (as the debug menu shows)
21:08:04gevaertsYes, but HAVE_MULTIVOLUME usually means two volumes. I just want to see if there could be an issue there
21:08:15domonokykugel: filesystem, not partition..
21:08:58gevaertskugel: if you just told mkfs.vfat to make a smaller filesystem, the partitions weren't touched, and this is not the issue
21:09:41gevaertsIs the microSD card partitioned, or does it just have a filesystem?
21:09:41 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:09:59gevaertsHow big is it? Is it FAT16 or FAT32?
21:10:16kugel4GB, fat32
21:10:21*gevaerts has no real idea, so he starts by collecting lots of information
21:10:42kugelgevaerts: it shows in my e200v1, so it's "rockbox prepared"
21:11:00kugeland it shows in the OF
21:14:19gevaertskugel: ata_sd_as3525.c needs to have a real card_detect_target() (maybe just a dummy that always returns true will do to get started), and possibly also card_enable_monitoring_target()
21:15:23 Join Photoguy_ [0] (n=Photoguy@
21:15:46*gevaerts clarifies
21:16:05 Quit Photoguy_ (Client Quit)
21:16:18gevaertsYou will need card_enable_monitoring_target() to handle hotswap properly, but it's probably not needed to just access a card that's been there from boot
21:17:16 Join petur [50] (n=petur@rockbox/developer/petur)
21:18:02 Join mc2739__ [0] (
21:18:15kugelgevaerts: hotswap doesn't seem to work.
21:18:47kugelif I put the card in after boot it's apparently not recognized (disk info says "Not found!")
21:18:48 Quit mc2739 (Nick collision from services.)
21:18:50 Nick mc2739__ is now known as mc2739 (
21:18:59 Join vertic23 [0] (
21:19:09gevaertskugel: hotswap needs more. What happens if you just change card_detect_target() to always return true?
21:19:12domonokythe sd-ams-code probably needs more love for that :-)
21:19:27vertic23I'm looking for information whether someone has started to port rockbox for the samsung YP-Tx series
21:21:12kugelgevaerts: \o/ I love you :)
21:21:31kugelok, maybe not
21:21:40gevaertsvertic23: I'm not aware of any, but you might check the New Ports forum or
21:22:20kugelgevaerts: well, it shows up now, and I can browser my music
21:22:45kugelbut once I clicked on a file, rockbox seems to start playing, just to reboot a second later w/o detecting any storage
21:23:54kugellol weird, playing another file didn't cause that
21:23:59gevaertsI'm afraid you're probably in new code territory then.
21:24:15 Quit evilnick (Read error: 60 (Operation timed out))
21:24:54 Quit perrikwp (" ajax IRC Client")
21:25:11kugelgevaerts: it seems to be related to mp3
21:25:23gevaertsMore fun :)
21:25:35 Quit GodEater (Read error: 104 (Connection reset by peer))
21:25:57vertic23which player would be best to buy? I see that the iriver series is not available anymore :/
21:26:05kugelerr, some mp3 cause the reboot, but with properly detected storage
21:26:54 Quit LambdaCalculus37 (Read error: 104 (Connection reset by peer))
21:27:06gevaertskugel: for actual hotplug to work you need to monitor the card somehow (probably GPIO). This monitoring can be started and stopped by card_enable_monitoring_target(). PP does this with an interrupt handler
21:27:14domonokyvertic23: take a look a BuyerGuide in the wiki.
21:27:44kugelgevaerts: there's a sd_present function, can we just call that with drive=1 to monitor the card?
21:27:54vertic23okay thanks
21:28:11kugellogf doesn't seem to work, so I can't tell if sd_present is working properly
21:28:39gevaertskugel: sd_present() won't help you. It looks at initialisation state, and something else needs to set that
21:28:49 Quit Photoguy (No route to host)
21:29:12 Join Aurix_Lexico [0] (
21:29:16 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
21:29:17*gevaerts recommends looking at ata-sd-pp.c
21:30:03kugelgevaerts: GPIOD has something todo with sd stuff, I'll look in the datasheet
21:30:16 Quit jhulst (Remote closed the connection)
21:32:21 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
21:32:46vertic23how can I see whether the ipod I want to buy is the correct version?
21:33:01vertic235/5.5 Gen. <−− which generation is the latest?
21:34:08 Quit Thundercloud (Remote closed the connection)
21:34:34vertic23how long do you expect will it take till 6 is supported?
21:35:02*gevaerts personally expects that to take an infinite amount of time
21:35:05domonoky123,45years.. :-)
21:35:16 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
21:35:22vertic23hehe, well have experience, don't you?
21:35:32vertic23like long did it take from 4 to 5 ...etc.?
21:36:00mud-rbvertic23: it's not that simple. the newer generation is more locked down
21:36:11LloreanAnd has completely different hardware
21:36:13domonokyvertic23: ports are not planned, they are done by people who have the time/device/motivation/skills.. so nothing predictable
21:36:23kugelgevaerts: do you have the datasheet?
21:36:32LloreanThe differences between the 4th and 5th generation are relatively minimal, while 5.5->6 is an entire redesign of everything as far as we know.
21:36:45 Quit linuxstb (Read error: 60 (Operation timed out))
21:36:51gevaertskugel: no. I'm not sure if I want it, you'll just set me to work ;)
21:39:54 Quit mc2739 ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
21:40:02vertic23hmm, does anyone know how to buy a 5th / 5.5 generation ipod?
21:40:30LambdaCalculus37With money?
21:40:57vertic23well to find it
21:40:59 Join akur [0] (
21:41:06 Part akur
21:41:32LambdaCalculus37There are usually a good number of them on eBay.
21:41:56 Quit miepchen^schlaf ()
21:42:30vertic23hmm I see's like the good players which are supported by rockbox are more and more sold out
21:43:27gevaertsUnfortunatly porting rockbox to a player often takes longer than the market time of that player
21:45:06kugelgevaerts: I think I found the GPIO pin
21:45:09 Join bxcracer [0] (n=bxcracer@
21:46:50 Quit soap (Remote closed the connection)
21:50:50 Join akur [0] (
21:51:12kugelgevaerts: so, card_detect_target returns the GPIO pin state?
21:51:17 Join soap [50] (n=soap@rockbox/staff/soap)
21:51:39 Join iTayb [0] (
21:51:39 Part akur
21:51:39gevaertskugel: for PP it does, yes
21:52:16 Quit dany_21a_ (Read error: 110 (Connection timed out))
21:52:20kugelgevaerts: I think A2 indicates an inserted card
21:52:30kugelor, I'm pretty sure
21:53:18 Quit robin0800 (Remote closed the connection)
21:53:27iTaybHi all, i wanted to ask how the development of sansa v2 is going?
21:53:42kugelgevaerts: ok, that works. At least if inserted pre-boot
21:53:52iTaybi have bought an E280 and it apears to be v2
21:54:05LloreaniTayb: The status is in the thread, in the Rockbox changelog, and in the wiki.
21:54:13LloreanDon't worry, we'll post when it's supported and working.
21:54:24iTaybok. thank you!
21:54:45 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:56:58kugelgevaerts: I guess for monitoring I need a isr on that pin?
21:57:26gevaertskugel: If you go the interrupt way (which I think is best), yes
21:57:30 Join miepchen^schla [0] (
21:59:52 Quit LambdaCalculus37 (" ajax IRC Client")
22:02:53 Quit {phoenix} ("Konversation terminated!")
22:04:53 Join aneqrs [0] (
22:09:11 Quit fredddy (Remote closed the connection)
22:09:34kugelgevaerts: hm, maybe I'm dumb...but what should the isr do? I imagine for hotswap, card_detect_target would be called regulary, I'm not entirely sure what my isr should do :/
22:09:50gevaertskugel: have a look at ata-sd-pp.c
22:10:03kugelit enables the isr
22:10:09*kugel has looked at it
22:11:54kugelyep, I'm looking at it
22:12:30kugelit resets the isr, but I'm not sure what about that timeout register
22:13:08gevaertsThat one registers a callback to be called (by the timer tick) after 300ms
22:13:49kugelgevaerts: ah, I think i should queue_post SYS_HOTSWAP_*, right?
22:13:57kugelto sd_queue
22:14:55gevaertskugel: not right away. You only post those if/when you can actually talk to the card, hence the 300ms delay. Check sd1_oneshot_callback().
22:15:21*gevaerts pretends to know all about this
22:15:47 Join herrwaldo [0] (
22:15:48kugelhaha, guess why I asked you about that in the first place ;)
22:16:00gevaertsBecause I'm good at pretending? :)
22:17:24kugelfor sure
22:17:58*hobbs doesn't think The Pretender ever took a kernel programming gig
22:18:10 Quit iTayb ("MULEz SCRIPT: I haven't lost my mind, I know exactly where I left it.")
22:19:48 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
22:20:16 Join faemir [0] (
22:21:35 Join ZincAlloy [0] (n=d9eec17a@gateway/web/cgi-irc/
22:24:39kugelgevaerts: ok, now I'm curious if it works
22:26:11*kugel can hotswap in one direction
22:26:33kugelgevaerts: yep
22:26:44kugelnot in, it gets me a panic
22:27:23gevaertsAt least that means that something is detected
22:27:23 Join evilnick [0] (i=0c140464@gateway/web/ajax/
22:28:12kugelgevaerts: it's the same panic as if I made card_detect_target always return true and boot w/o card inserted
22:29:54 Part seraph ("Ex-Chat")
22:30:13kugelgevaerts: "wait for state failed" about line 600 in ata_sd_as3525c
22:31:23kugelgevaerts: sd_wait_for_state() doesn't really seem to wait
22:33:54 Join Zagor [242] (
22:34:10gevaertskugel: but it does work if the card is there at bootup?
22:34:22kugelgevaerts: yes
22:34:51kugelmy isr is fine, card_detect_target is fine, but it seems wait_for_state() has it's name for a wrong reason
22:38:41 Nick courtc_ is now known as courtc (n=court@unaffiliated/courtc)
22:38:47***Saving seen data "./dancer.seen"
22:47:52 Quit massiveH ("Leaving")
22:48:35 Join ron_o [0] (
22:53:12 Quit _lifeless (Remote closed the connection)
22:54:59kugelgevaerts: any idea?
22:55:43 Join _lifeless [0] (n=lifeless@
22:57:04 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
22:57:50kugelif funman was here :/
22:58:14 Quit Bensawsome (Remote closed the connection)
22:58:29 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
22:58:31gevaertskugel: not really
23:01:00kugelsend_cmd seems to fail
23:01:31gevaertsMaybe it needs more time? Try increasing the callback delay
23:01:38 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
23:01:54 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
23:02:33kugelgevaerts: I made a while(tries−−) before wait_for_state, I get the panic after 20s or so
23:02:41gevaertskugel: my guess is that you need another sd_init_card() on plugin (if you don't already have it)
23:02:59kugeland I also tried that :)
23:05:54gevaertsMaybe remove the sd_enable() calls that are all over the place, and leave it enabled all the time for now?
23:06:34 Join MethoS- [0] (
23:06:51 Quit petur ("Zzzzz")
23:07:01 Quit bertrik ("Leaving")
23:08:42kugelgevaerts: maybe a good idea, btw: I fail in sd_transfer_sectors()
23:09:23kugelbut that one does sd_enable(true); just before my panic
23:14:24kugelit's not initialzed after sd_init_card
23:17:38kugelgevaerts: \o/
23:17:54gevaertsWhat was the problem?
23:17:55kugelat least once, that I get another panic, a very different one though
23:18:06kugelgevaerts: yes and no
23:18:42kugelgevaerts: not sure what funman did, but his check for initialised was just wrong (it checked for < 0, instead of <= 0), so I got the wrong panic
23:19:16kugeland then he didn't even initialize, but issued an instant panic
23:19:44kugelgevaerts: my new panic is "DATA CRC FAIL", after the second insertion
23:21:20 Join gregzx [0] (
23:21:38kugelok, I tried only twice, but it was both times after the 2nd insertion
23:22:05gevaertsDid you wait a while before inserting the second time?
23:22:24kugelyes, several seconds
23:22:39gevaertsok, so it's not an issue with the eject
23:22:45gevaerts(at least not directly)
23:23:04 Quit domonoky (Read error: 54 (Connection reset by peer))
23:23:14 Quit BHSPitLappy (Read error: 110 (Connection timed out))
23:23:55kugelif I have the card inserted at boot I already get it after the first insertion
23:24:43gevaertsMaybe something doesn
23:24:50gevaerts't get cleared properly on eject?
23:25:01kugelI'm wondering if I reset the isr properly
23:25:36kugelgevaerts: disk info shows not found after ejecting, so I think yes
23:26:12kugelgevaerts: also, the tick task forces a reinit after ejection and insertion
23:26:18gevaertskugel: high level it's probably fine, but maybe the controller needs to have some bits cleared?
23:27:04gevaertsCan you post your current work? It's getting hard to follow with the plain svn version
23:27:27kugelthe isr is only cleared upon insertion
23:27:44 Quit bmbl ("Woah!")
23:29:09kugelstill nothing :(
23:33:23 Join Chesteta [0] (
23:33:39gevaertsWhat happens if you reinitialise the controller on eject (i.e. call init_pl180_controller(SD_SLOT_AS3525) again)?
23:34:21*gevaerts has no real idea if this will help
23:36:23Chestetahello; wondering about the ams3525 1-2 gb filesystem issue... on the main "target status" page it shows that only 1-2 gb works; however on the "sansav2" subpage it simply says "yes"... which is it? :)
23:36:59ChestetaI am referring to the "disk/flash" column
23:37:14 Quit Shiny (Read error: 113 (No route to host))
23:40:44kugelgevaerts: haha, you pretender!
23:41:50*kugel got microSD and hotswapping working on his fuze \o/
23:41:54kugelgevaerts: thanks a lot
23:42:13gevaertskugel: now add the sd_enable()s back in and see what happens
23:42:59 Quit karashata ("G'bye everyone!")
23:43:08Bagderspam on the users list
23:45:33 Join ameyer [0] (
23:46:16 Quit XavierGr (Nick collision from services.)
23:46:27kugelgevaerts: looks good
23:46:27 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
23:46:33 Quit HellDragon (Remote closed the connection)
23:47:22gevaertskugel: I'd say if your code is clean, it should go in
23:47:43kugelI think so too :) I'll put the patch for review onto fs
23:48:52kugelhaha, so my gpio irq isr intermezzo was worth it after all, even if it failed where I tried it first :)
23:49:13 Join HellDragon [0] (
23:51:35 Join Nibbl [0] (
23:53:55 Quit bluebrother ("leaving")
23:54:30kugelgevaerts: not exactly userproof yet, if you insert and eject and insert and eject (...) too quickly you will still get a panic
23:54:52gevaertskugel: so what part of the AMS port is userproof yet? :)
23:54:56kugelbut if you're mature it's just fine
23:56:23*Zagor got tools/database to compile

Previous day | Next day