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).

#rockbox log for 2010-10-16

00:00:16 Part meesebyte
00:15:17 Quit bluebroth3r (Ping timeout: 255 seconds)
00:15:21 Join bluebrother [0] (
00:15:21 Quit bluebrother (Changing host)
00:15:21 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
00:16:59pixelmabertrik: you looked for small bugs to fix but maybe you can have a look at the dutch translation update in the tracker instead?
00:24:44bertrikpixelma, I don't know the proper dutch translation for root, the patch in the tracker leaves it untranslated ...
00:33:18 Join grawity [0] (
00:33:38pixelmawell, then there may be no proper translation and most people would use "root" in Dutch context too. There are some things I couldn't really "translate" to German either.
00:35:05grawityHey, does anyone here have a rockbox.ipod for nano2g that is at least one month old, and can email it to me? I want to compare the LCD contrast (seems quite low in today's build, unless it's just my eyes)
00:37:49saratogayou'll probably need to change the entire build, not just the one file
00:38:51grawityWell, the rest are just fonts, codecs, plugins.
00:39:39grawityrockbox.ipod seems to be holding the entire firmware or whatever it is called.
00:40:27pixelmaif you purely want to test the contrast the file would be enough but you won't be able to play music or games
00:41:51grawityTo play music I'd need codecs/ ?
00:42:16saratogato do anything you'll need a working rockbox install
00:42:27saratogai think the main binary will just get you the menus and not much else
00:43:01grawityIf you have an entire build from last month handy, it would be nice... Otherwise I'd settle for rockbox.ipod.
00:43:36pixelmaunfortunately daily builds don't get as far back anymore
00:44:51grawitysaratoga: The entire "install" consists of bootloader, firmware, plugins, codecs, and themes, plus some other files (tagnavi.conf and codepages, they never change), doesn't it?
00:45:08saratoganot the bootloader
00:45:25grawitypixelma: that's why I was hoping someone here is lazy at upgrading.
00:45:27saratogaand tagnavi.conf is your local settings file
00:45:48grawityNo, local is tagnavi_local.config
00:46:15grawityEr, _custom
00:46:42saratogaoh yeah
00:48:25*grawity remembers a backup of three-month-old rockbox somewhere.
00:49:19grawityI guess I could downgrade to that.
00:58:45*grawity wonders if r28190 could be the cause.
01:02:50 Quit grawity (Quit: jmIrc destroyed by the OS)
01:06:36***Saving seen data "./dancer.seen"
01:11:25 Join krabador [0] (
01:11:49 Quit n1s (Quit: Lämnar)
01:17:04 Quit bertrik (Ping timeout: 240 seconds)
01:17:51 Part toffe82
01:18:57 Quit jre2 (Quit: leaving)
01:19:21 Quit ender` (Quit: This space intentionally has nothing but text explaining why this space has nothing but text explaining that this space would otherwise have been left blank, and would otherwise have been left blank.)
01:27:14 Quit user890104 ()
01:32:17 Quit Dreamxtreme (Ping timeout: 240 seconds)
01:34:10 Quit drizztbsd (Remote host closed the connection)
01:42:08 Join Barahir [0] (
01:42:51 Quit Barahir_ (Read error: Operation timed out)
01:43:29 Quit GeekShadow (Read error: Connection reset by peer)
01:43:38 Join Dreamxtreme [0] (~Dre@
02:21:36 Join factor [0] (
02:22:10 Quit krabador (Quit: Sto andando via)
02:39:09 Quit factor (Remote host closed the connection)
02:41:19 Quit DerPapst (Quit: Leaving.)
02:45:24 Join Sudos|C840 [0] (~ZOOP@
02:45:43 Part Sudos|C840
02:47:17 Quit _s1gma (Quit: Leaving)
02:48:40 Join xavieran [0] (
02:49:44 Join factor [0] (
03:06:37***Saving seen data "./dancer.seen"
03:22:07 Quit Judas_PhD (Quit: This is a quitting message)
03:27:12 Join Judas_PhD [0] (
03:27:22 Quit Judas_PhD (Client Quit)
03:29:11 Quit leavittx_ (Ping timeout: 240 seconds)
03:38:52 Join anewuser [0] (kvirc@unaffiliated/anewuser)
03:44:51 Join MethoS- [0] (~clemens@
04:02:19 Quit amiconn (Disconnected by services)
04:02:19 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:02:23 Quit pixelma (Disconnected by services)
04:02:25 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:02:28 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:02:38 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:04:20S_a_i_n_twhen/why are bootloaders released?
04:05:12S_a_i_n_tthere's been a change to the Nano2G code that requires a new bootloader to be built, and as far as I can tell (unless what was implemented is broken which I doubt) RBUtil is still using the old Nano2G bootloader.
04:05:49S_a_i_n_t"change" == enable Nano2G to boot the OF by holding Menu during startup like the other iPods.
04:06:42S_a_i_n_tas of r28237
04:07:29saratogathey can be updated whenever someone feels like tagging new ones in SVN and then pinging the swedes to update the download site
04:07:30 Quit TheSeven (Ping timeout: 240 seconds)
04:08:32S_a_i_n_tAh, right. Well, building the Nano2G bootloader would be great. I'd be happy to pass it through the crypto on one of my Nano2Gs also.
04:08:58S_a_i_n_tI just don't have a dev environment set up presently.
04:11:23 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:12:02 Quit edboyer93 ()
04:20:56 Join Horschti [0] (~Horscht@xbmc/user/horscht)
04:22:39 Join Kohli [0] (
04:24:47 Join Naked [0] (
04:24:57 Join kisak_ [0] (
04:25:03 Join guymann_ [0] (
04:25:20 Quit Hadaka (Ping timeout: 240 seconds)
04:25:20 Quit Kohlrabi (Ping timeout: 240 seconds)
04:25:20 Quit kisak (Ping timeout: 240 seconds)
04:25:21 Quit krazykit (Ping timeout: 240 seconds)
04:25:21 Quit Horscht (Ping timeout: 240 seconds)
04:25:21 Quit guymann (Ping timeout: 240 seconds)
04:25:27 Nick Naked is now known as Hadaka (
04:25:31 Join krazykit [0] (
04:41:00 Join smartboyathome [0] (
04:42:02 Quit Barahir (Read error: Operation timed out)
04:42:04 Join Barahir_ [0] (
04:42:28smartboyathomeSo, I'm thinking about developing a plugin, but I was wondering if there was any standard way to modify the database or (hopefully) the ID3 tags? I'd like to be able to change the rating of a song.
04:42:59saratogasmartboyathome: i'm not sure if thats in the plugin api, but if its not, you could add a function to do it
04:43:52smartboyathomesaratoga: is there any way to modify ID3 tags from within rockbox, or is that frowned upon?
04:44:10saratogawe don't provide code to do that, of course you're welcome to provide your own code in your plugin to do it
04:44:30saratogaor do you mean modify the database?
04:46:12 Quit bluebrother (Disconnected by services)
04:46:13 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
04:46:21smartboyathomeSaratoga: i'm just wondering if my plugin would be frowned upon for doing such a thing. I'd probably start out just modifying the database, but I'd like it if the ratings weren't restricted to rockbox itself.
04:48:07 Quit dys (Ping timeout: 276 seconds)
04:48:24 Join dys [0] (
04:49:34saratogadoing that is fine
04:49:45saratogatheres no real restrictions on plugins as long as they're GPL, etc
04:50:07smartboyathomeok thanks ^^
04:53:58saratogaof course you have to actually make all this work on a couple hundred Mhz (at most) CPU with barely any memory
04:54:07saratogabut thats the fun part
04:54:36smartboyathomeyeah it is xD
04:54:56 Quit MethoS- (Remote host closed the connection)
04:55:01smartboyathomei'll treat it as just another puzzle that needs to be solved
05:06:38***Saving seen data "./dancer.seen"
05:19:42 Quit S_a_i_n_t (Ping timeout: 245 seconds)
05:23:38 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
05:30:10soapgevaerts / Torne. Why not just have the VID / PID pair in either an undocumented obscured menu setting or in an undocumented config file?
05:31:28soapand, gevaerts, "most people not liking file-only settings" is something I don't think is true of this. Many people would kiss your bum for the option to set it w/o a compile. It would be a set-and-forget for many so a one-time file change is small potatoes.
05:34:12 Quit ps-auxw (Ping timeout: 265 seconds)
05:45:17 Join ps-auxw [0] (
06:00:31 Quit alexbobP (Quit: See other host)
06:01:45 Join alexbobP [0] (
06:17:27 Join simonrvn_ [0] (
06:19:13 Quit simonrvn (Disconnected by services)
06:19:13 Nick simonrvn_ is now known as simonrvn (
06:19:55 Quit S00row (Ping timeout: 260 seconds)
06:23:15 Quit GodEater (Remote host closed the connection)
06:26:33 Join krazykit` [0] (
06:26:54 Quit krazykit (Disconnected by services)
06:26:58 Nick krazykit` is now known as krazykit (
06:44:45 Join GodEater [0] (
06:44:46 Quit GodEater (Changing host)
06:44:46 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
06:53:42 Join z35_11 [0] (
07:06:30 Quit timccc (Quit: Leaving.)
07:06:42***Saving seen data "./dancer.seen"
07:11:50 Quit anewuser ()
07:52:11 Join n1s [0] (
07:52:11 Quit n1s (Changing host)
07:52:11 Join n1s [0] (~n1s@rockbox/developer/n1s)
07:59:22 Join Horscht [0] (~Horscht@xbmc/user/horscht)
08:01:04 Join esperegu [0] (~quassel@
08:02:29 Quit Horschti (Ping timeout: 276 seconds)
08:09:46 Join Strife89TX [0] (
08:15:20 Join Barahir [0] (
08:17:15 Quit Strife89TX (Quit: Bed.)
08:18:55 Quit Barahir_ (Ping timeout: 260 seconds)
08:21:17 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:35:11n1sgevaerts: do you think fs#11485 can be closed? it WFM too
08:44:40 Quit bluebroth3r (Ping timeout: 240 seconds)
08:44:44 Join bluebrother [0] (
08:44:44 Quit bluebrother (Changing host)
08:44:44 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
08:48:14 Join Highlander [0] (
08:52:21n1shmm, hold select to charge without mounting isn't working on my c200
08:54:00 Join S_a_i_n_t [0] (S_a_i_n_t@
08:55:38n1swhere is the code that handles that?
09:00:57S_a_i_n_tCan someone look into updating the bootloader for the Nano2G please? r28237 introduced a change (boot OF from startup using Menu, same as other iPods) that requires a new bootloader be built.
09:06:45***Saving seen data "./dancer.seen"
09:08:28 Join Rob2222 [0] (
09:11:58 Quit Rob2223 (Ping timeout: 250 seconds)
09:13:42*S_a_i_n_t decides to be nice to the D2 owners and adds touchscreen defines to the iLike Video theme.
09:14:12S_a_i_n_tHehe, iPod "OF" will certainly look very foreign on a Cowon d@
09:14:22S_a_i_n_t*D2, even
09:14:54S_a_i_n_tgrid mode is just terrible.
09:19:26 Join bieber [0] (
09:19:37S_a_i_n_tSpam of the Forum:
09:19:45S_a_i_n_tOMG! bieber!
09:20:20S_a_i_n_tI believe your skills are needed to re-sync SVN and the themeeditor ;)
09:20:51 Join Kitr88 [0] (
09:23:20bieberI know, I just need to get a little free time
09:23:35 Quit Kitar|st (Ping timeout: 276 seconds)
09:24:06S_a_i_n_twe were talking about abducting you from school ;)
09:25:24 Quit Kitr88 (Ping timeout: 250 seconds)
09:26:27S_a_i_n_tgah...more Forum spam: and
09:28:11 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
09:30:24 Join Kitar|st [0] (
09:32:56 Quit bieber (Remote host closed the connection)
09:34:21pixelman1s: hold select for charging works for me on the c200 (sort of) - it charges and the USB data connection is not established, but XP said it found a new HID device (can't use it though because I'm not in the USB screen which has the buttons mapped)
09:35:31 Join bieber [0] (
09:36:08n1sif i am in wps, press and hold select, the context menu opens, i then connect usb and when i let go of the button i end up in the usb screen and it mounts
09:36:24pixelmaspeaking about keymappings: the A-B marker thing should work but you have to hold Power first and that's a bit "dangerous" because that's also power-off
09:36:46pixelmaI tried it from the menu or file browser
09:36:57n1sjust did, it works there...
09:37:24n1sfrom the wps it constantly fails
09:37:45pixelmaworks for me
09:37:51pixelmain the WPS too
09:39:14pixelmaand I'm still in the WPS context menu but there's no problem going back to the WPS
09:39:31n1sweird, now it works here too...
09:48:26 Quit dys (Ping timeout: 276 seconds)
09:54:03 Quit hondza (Quit: quit)
09:55:08 Join dys [0] (
09:55:54 Quit bieber (Read error: Connection reset by peer)
09:58:11 Join stripwax [0] (
09:59:02 Quit sasquatch (Quit: WeeChat 0.3.2)
09:59:28 Join sasquatch [0] (
10:00:27 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
10:11:03 Quit stripwax (Quit: Miranda IM! Smaller, Faster, Easier.
10:11:36 Join stripwax [0] (
10:11:51 Join esperegu_ [0] (~quassel@
10:13:08 Quit esperegu (Ping timeout: 276 seconds)
10:16:55 Join stoffel [0] (
10:18:17 Join leavittx_ [0] (~lev@
10:18:18 Join wodz [0] (
10:18:25wodzpixelma: ping
10:31:07 Quit factor (Read error: Connection reset by peer)
10:32:28bertrikDo we have anyone in charge of a new release now?
10:35:54S_a_i_n_tpass, I haven't even seen the "we're in freeze for release" announcement...
10:39:57Bagderwe should rotate around and appoint a single person "head release human" for each release we do
10:40:23Bagderor something
10:40:54S_a_i_n_tpfffft! that sounds too sensible and efficient ;)
10:43:17 Join ender` [0] (
10:43:41n1snot if we just appoint some random person, preferably whitout them knowing ;)
10:44:25S_a_i_n_tlook, over there!
10:44:30*S_a_i_n_t appoints n1s
10:45:46n1si should have seen that comming ;)
10:45:59 Quit esperegu_ (Read error: Connection reset by peer)
10:46:49n1sbut seriously, we should freeze soonish, i propose monday and release next monday 25 oct
10:48:01 Join factor [0] (
10:49:14 Join esperegu [0] (~quassel@
10:52:36amiconnpixelma: Sw usb devices enumerate as a hid device in charging mode. This is intended, as it's about the only way to properly enumerate without needing special drivers if we don't want msd
10:53:32 Quit bmbl (Quit: Bye!)
10:54:58wodzit seems to be more effective to scare people that we are about to freeze phase than to freeze actually :-)
10:57:38 Join bmbl [0] (
10:57:39 Quit bmbl (Changing host)
10:57:39 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
11:01:11 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
11:03:31n1si think it's a good time now since whe haven't had any big changes lately and most of the recently introduced bugs have been fixed afaik, if we keep waiting someone will commit a major change that will break stuff
11:04:05bertrikMy android phone seems to register itself as msc, even in "charge-only" mode. Maybe it just reports 0 devices at the SCSI level.
11:06:49***Saving seen data "./dancer.seen"
11:07:07 Join deed [0] (
11:07:13deedhi all
11:10:14n1shmm, test_codec gives me "illinstr at 00000002" on h300, i thought it was something with the profile build when it did that yesterday but it also happens with a regular build
11:10:19S_a_i_n_tdeed: It's not a social channel, usually don;t get any response from "Hi all" etc.
11:10:30S_a_i_n_tIf you have a question about Rockbox, just ask it.
11:10:38n1slikely a null function pointer somewher
11:11:14n1sor maybe not, test_codec hasn't changed in a long time
11:12:42 Join bimbel [0] (
11:12:42 Quit bimbel (Changing host)
11:12:42 Join bimbel [0] (~bmbl@unaffiliated/bmbl)
11:12:47 Quit thegeek (Ping timeout: 276 seconds)
11:12:51 Quit bimbel (Read error: Connection reset by peer)
11:13:40 Part deed
11:14:11 Nick Kohli is now known as Kohlrabi (
11:15:09n1snope, works on arm so that shouldn't be it, maybe something with that plugin_crt0 stuff
11:16:14 Join Judas_PhD [0] (
11:16:51 Quit ender` (Quit: When I was a kid I used to pray every night for a new bicycle. Then I realised that the Lord doesn't work that way so I stole one and asked Him to forgive me. )
11:18:30wodzwhere the hell memcpy is declared?
11:19:45 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
11:21:43n1snp, i often have to click my way through firmware looking for headers since the recent reorganizations
11:27:03amiconnThere's this useful thing called grep ;)
11:28:36 Join ender` [0] (
11:28:55n1samiconn: sure but i rmemeber where the header i want is, it just isn't there anymore...
11:30:17amiconngrep -r −−include=*.h memcpy ~/rockbox/firmware
11:35:34 Join earcar [0] (
11:37:30 Join thegeek [0] (
11:39:46 Join badcam [0] (
11:44:17 Join DerPapst [0] (
11:47:42 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:51:50JdGordonAlexP: you never sent off an email about the rolling stable release? (or did I misremember you saying you were going to?)
11:56:52badcamHi Guys. Do you mind if I ask a question about a theme issue I'm having?
11:58:29CIA-7New commit by nls (r28285): libtremor: tiny optimization 0.5%-1.5% on coldfire and PP
11:59:57pixelmabadcam: if it's a Rockbox theme then just ask
12:00:10CIA-7r28285 build result: All green
12:00:27pixelmaand be patient, not everyone will have an answer
12:05:26badcamGreat. Thanks. I've actually just answered the problem. I realise that there was a bitmap I hadn't copied into the WPS new theme folder. Now, I'd like to take a screenshot, but using system/debug/screendump and then plugging the usb cable in, doesn't work. I can get a screenshot with the emulator, but I'd really like to get it working on my Sansa ClipV2.
12:08:49badcamI'm using Linux Mint BTW :)
12:11:02pixelmasounds like it could be a limitation of the ClipV2 part but I don't know and don't own such a device
12:11:42pixelmaerr.. ClipV2 port
12:12:38S_a_i_n_tafaik screendump should be working...
12:14:10pixelmawell, from the description I don't thiink badcam is doing something wrong though and so I suspect there may be a bug in Rockbox which is why I said that
12:14:18badcamSorry. Port? I'm probably not following the instructions properly. After I have chosen Screendump and it confirms that it's enabled. I presume I just navigate to the screen I want a dump of and connect the USB cable. Correct?
12:15:51pixelmabadcam: in Rockbox for the ClipV2
12:16:31badcamOK. How long should the capture take? Does the Clip give any indication that it has done this?
12:16:45S_a_i_n_tIt should be instant.
12:17:05S_a_i_n_tYou're looking in the correct place?
12:17:15S_a_i_n_tthe files should be on the root of your DAP
12:18:18pixelmathere should be some dump-timestamp.bmp files
12:19:10badcamYes. Root directory. I'm trying a different screen location. Checking now...
12:20:54badcamNope. I'm just seeing the usual folders and files. No .bmp's. Hidden files are switched to on, not that that should make any difference.
12:20:56AlexPJdGordon: No I didn't and yes I did respectively :)
12:21:10AlexPJdGordon: I have to pop out in a bit, I'll send one later this afternoon
12:21:25 Join moos [0] (moos@rockbox/staff/moos)
12:21:52S_a_i_n_tbadcam: Search the tracker to check if this is known behaviour/already reported...if not, file a bug report.
12:22:00badcamI've tried this on my Clip+ as well and no luck there either. Perhaps it's Mint?
12:22:37S_a_i_n_tIf I understand correctly, it just needs to detect power on the data pins, it shouldn;t matter what OS it is.
12:24:08badcamSure. Will do. Pins: I have a mini usb wall charger. I'll try it with that.
12:25:12 Join drizztbsd [0] (~quassel@unaffiliated/drizztbsd)
12:28:05 Quit drizztbsd (Remote host closed the connection)
12:28:57 Join drizztbsd [0] (~quassel@unaffiliated/drizztbsd)
12:32:41 Quit earcar (Quit: bye)
12:36:16 Join hondza [0] (K5wNGm8J@unaffiliated/hondza)
12:39:02 Join teru [0] (
12:43:26 Quit domonoky (Read error: Connection reset by peer)
12:43:38 Join _s1gma [0] (~d.d.derp@
12:45:33 Join s1gma_ [0] (~d.d.derp@
12:47:52 Quit _s1gma (Ping timeout: 240 seconds)
12:48:54badcamNo luck. That's OK though. I'll file a bug report as I can't see any similar reports in there. You probably get it all the time, but I think you guys are the shizzle. Rockbox is AWESOME. I've got some great ideas for some themes, but I've only just been looking at this for the first time tonight. Here's my initial attempt, if anyone is interested. I've used a combo from a couple of themes,mostly Clock. Anyway, goodnight from me and thank
12:49:35badcamSorry about the double link.
12:50:32 Quit s1gma_ (Ping timeout: 240 seconds)
12:55:12-->"seen flyndice" received from funman (~fun@rockbox/developer/funman)
12:57:22 Join funman [0] (~fun@rockbox/developer/funman)
13:05:12CIA-7New commit by teru (r28286): Fix FS #10981: Viewers icons applied incorrectly.
13:06:47CIA-7r28286 build result: All green
13:06:50***Saving seen data "./dancer.seen"
13:07:42wodzAlexP: do You have some time to test new version of png viewer?
13:07:44gevaertsn1s: maybe this is a v2?
13:18:20AlexPwodz: yep, hit me
13:19:27wodzAlexP: I am testing new ver in sim currently, will ping you when ready than
13:19:36 Part funman
13:28:03 Join krabador [0] (
13:28:06 Join _s1gma [0] (~d.d.derp@
13:29:00wodzAlexP: RealLife(TM) comes in action, maybe I'll finish this at night :/
13:29:06 Quit wodz (Quit: Leaving)
13:29:11 Quit krabador (Client Quit)
13:29:24AlexPNo worries
13:29:31 Join krabador [0] (
13:30:30 Quit krabador (Client Quit)
13:30:55 Join krabador [0] (
13:31:12n1sgevaerts: the c200?
13:37:54 Quit pjm0616 (Ping timeout: 276 seconds)
13:39:15 Join pjm0616 [0] (~user@
13:49:31gevaertsn1s: yes
13:49:42n1sthat's a possibility, yes
13:51:06n1si asked on FS
13:59:39gevaertsok. If we don't get an answer in a week or so I guess we can close it
14:02:26gevaertssoap: not an *undocumented* setting...
14:02:53gevaertsWe can always hide it quite effectively by only documenting it in the manual of course :)
14:06:16*n1s found why test_codec crashes
14:06:30n1sit was a null func pointer that's called
14:06:59gevaertsI once did that on my gigabeat F. It just caused a reboot :)
14:08:31n1snow i wonder why the plugin api only has cpucache_(flush|invalidate) if the respective HAVE_CPUCACHE_* is defined but the codec api has them uncondidtionally?
14:09:44 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:10:27 Join stripwax__ [0] (
14:11:24n1si mean either they should be guarded everywhere or nowehre
14:11:32 Quit stripwax (Ping timeout: 240 seconds)
14:14:08*n1s puts in some guards
14:14:33*teru found uisimulator/sdl directory is empty.
14:15:09gevaertsn1s: IIRC kugel suggested to remove all guards a while ago
14:15:27n1sgevaerts: why?
14:17:19n1sok, i'll try that, extra function pointers and some empty calls shouldn't hurt
14:18:04 Quit kugel (Ping timeout: 245 seconds)
14:18:33n1sok, confirmed that fixes it
14:30:10CIA-7New commit by nls (r28287): Remove some inconsistent guarding around cpucache_* functions to fix a test_codec crash on coldfire, breaks plugin api compatibility so sorted and ...
14:32:19CIA-7r28287 build result: All green
14:43:38 Join stripwax [0] (
14:44:44 Quit stripwax__ (Ping timeout: 245 seconds)
14:50:49CIA-7New commit by teru (r28288): fix loding of default remote viewers icons.
14:52:21 Quit krabador (Ping timeout: 240 seconds)
14:52:30CIA-7r28288 build result: All green
15:04:12 Quit T44 (Ping timeout: 276 seconds)
15:06:54***Saving seen data "./dancer.seen"
15:10:03 Quit soap (Quit: soap)
15:10:18 Join soap [0] (
15:10:19 Quit soap (Changing host)
15:10:19 Join soap [0] (~soap@rockbox/staff/soap)
15:19:53 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:19:59soapIs there a known way to update old links to the sandisk forums? We have directed people there for a long time for various reasons, and every link I've tried today is dead.
15:23:14*JdGordon assumes everyone would get very pissy if 11678 got cleaned up and commited tomorow?
15:24:24soapyou mean during the freeze?
15:24:44gevaertsJdGordon: you're the one who announced the freeze that we might or might not be in now! :)
15:25:01JdGordonits a bug fix though :)
15:25:14JdGordonand yes, can someone make a descision already?
15:25:59JdGordonI also want to get 11575 is soon... its RaaA changes only and a bugfix... feedback would be nice
15:26:06gevaertsI think we should assume we're frozen. I don't think there's been any non-bugfix commit the last week or so, so I don't see the point of restarting the freeze
15:30:03JdGordonkugel: ^
15:31:36JdGordonthe linux install paths are very annoying... It really shouldnt need 3 different directories for /usr/local/share/rockbox /usr/local/bin/rockbox/ and /usr/local/lib/rockbox :/
15:59:57 Join kevku [0] (
16:03:08 Quit teru (Quit: Quit)
16:06:08 Join GeekShad0w [0] (
16:08:11 Join GeekSh4dow [0] (
16:09:25 Quit GeekShadow (Ping timeout: 265 seconds)
16:11:50 Join Topy44 [0] (
16:11:57 Quit GeekShad0w (Ping timeout: 255 seconds)
16:12:48soapbadcam, your talk about screendump and working on a new theme made me curious.
16:13:18soapYou are aware that Rockbox offers a simulator which would make theme development much easier than testing on-target.
16:13:50 Quit Topy44 (Client Quit)
16:23:33 Join Topy44 [0] (
16:36:13 Join stripwax__ [0] (
16:38:17 Quit stripwax (Ping timeout: 260 seconds)
16:39:41 Quit stripwax__ (Read error: Connection reset by peer)
16:49:13 Quit moos (Ping timeout: 272 seconds)
17:05:24 Join edboyer93 [0] (
17:05:40 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
17:06:56***Saving seen data "./dancer.seen"
17:08:57bertrikgevaerts, can this task be closed now?
17:11:51soapIn the forums we have around 100 IPs banned, and nearly a thousand hits against said bans.
17:12:10soapIs there really a good defense for a permanent IP ban?
17:12:37n1shits over how long a period?
17:13:51soaplifetime of the forums
17:14:57soapWhat caught my eye was a hit against IP ban
17:15:25n1sthe problem with IP bans is that well, a lot of people have dynamic IP's so it's very likely to hit random other people
17:15:28soapHit this month, but it was banned five months ago.
17:16:04soapReading the bans there I have yet to see a *permanent* IP ban justified.
17:16:08gevaertsbertrik: only the crash part is fixed
17:17:21soapI'm proposing to wholesale remove all permanent IP address bans from the forum, along with a staff memo that in the future permanent IP bans should be justified in the appropriate comment box.
17:17:43soapI'll go work on my roof for the rest of the day, and do this tomorrow if there are no arguments.
17:18:24bertriksoap, sounds reasonable to me
17:18:52 Quit esperegu (Ping timeout: 240 seconds)
17:22:16n1smc2739: is FS #10809 still affecting you(r sansa) ?
17:23:43CIA-7New commit by bertrik (r28289): skin_engine: struct gui_skin can be static
17:25:02AlexPsoap: I only ever ban on account + email anyway
17:25:39CIA-7r28289 build result: All green
17:30:39soapAlexP, I've done temp IP address bans on people who were "hot under the collar" in an attempt to force a cooling off period.
17:31:16AlexPyeah, that can sometimes be necessary, but I don't think I've banned a real person in a long long time
17:31:37AlexPI mean just a non-spam account by real person :)
17:31:44soapTHOUGH, I've never followed up and checked if the IP trigger actually got hit. So that's more of a gut move than a brain move.
17:38:50 Join G4Oblivion [0] (
17:41:46n1swow, this code is weird
17:48:15n1si hate when lots of unctions share global vars
17:48:45bertrikn1s, what part of the code are you looking at?
17:49:36n1stalk code, i wanted to fix FS #10479 by not killing playback if there is no voice file to load
17:49:57n1snot a very important bug but it looks easy :)
17:50:44n1ssome functions share a global file handle
17:51:02n1sand the usage of it depends on a bunch of #ifdefs
17:56:16n1syay, it works!
17:56:26n1snow to see if voice still works
17:56:56soapG4Oblivion, S_a_i_n_t had mentioned last night encrrypting the new bootloader on his Nano2G
17:57:01soapGet with him?
17:57:24soapHe's on New Zealand time.
17:58:28G4OblivionAh, sorry. Not used to IRC. I will try to ask him if I remember.
17:59:49 Quit GeekSh4dow (Quit: The cake is a lie !)
18:01:09soapHe's in the channel, he's been highlighted, he'll respond when he sees that. Just hang around.
18:03:09n1shmm, the music playback skips when loading the lang, i guess that thing never yields, sigh
18:05:13 Quit bmbl (Quit: Verlassend)
18:09:19n1sdo disk functions yield?
18:09:27n1slike read and lseek?
18:11:38bertrikI haven't checked, but I assume that at some point the disk drivers do some kind of wakeup_wait to wait for hardware
18:12:35 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
18:18:11 Join Secondgen [0] (ae5b48af@gateway/web/freenode/ip.
18:18:15SecondgenWhts up peeps
18:18:51SecondgenI just dug up my old nano 2g. Nethin cool thts come about i can put on ?
18:23:29 Quit t0rc (Remote host closed the connection)
18:24:28 Quit G4Oblivion (Quit: ChatZilla 0.9.86 [Firefox 3.6.12pre/20101016051926])
18:24:45n1smeh, i have nfi how this driver code works, somethign is clearly starving the codec thread though
18:26:57 Join GreenKirby [0] (
18:27:39n1shmm, another funky issue, when i unplug usb after usb on my beast, the display remains turned off untill a reboot
18:27:46 Join flan_suse [0] (61675f5a@gateway/web/freenode/ip.
18:27:49GreenKirbyhey guys how do you get the mpeg player to work on the ipod video, when I select it in files and it has the extension .mgp it just goes to a black screen and doesn't play?
18:28:00flan_suseThis disktidy plugin is not deleting any files.
18:28:35GreenKirbyany suggestions
18:28:38flan_suseI added a custom entry that is *.bmark but whenever I run the tool, it says 0 files deleted. And when I check, I see that none of my bookmark files were deleted.
18:29:12gevaertsGreenKirby: does the Elephants Dream video from work?
18:29:40GreenKirbyhmm lemme check
18:30:14flan_suseIf someone else can try the disktidy plugin? It might not be just me actually.
18:33:33gevaertsGreenKirby: at least that one plays correctly on my ipod video with a current build
18:33:41n1shmm, now it doesn't disable voice if i pick a lang without voicefile if i have one loaded...
18:33:45GreenKirbyI went to that page but it doesn't have a download (on the actually page it has and HD which won't work)
18:33:54GreenKirbyI think
18:34:00GreenKirbyand its a big file
18:34:25 Join moos [0] (moos@rockbox/staff/moos)
18:36:59 Quit flan_suse (Quit: Page closed)
18:53:12GreenKirbythanks man
18:53:12 Join MethoS- [0] (~clemens@
18:53:14GreenKirbyit worked
18:53:22GreenKirbythe mpg I was trying to play
18:53:31GreenKirbyisn't quite correct
18:53:46GreenKirby......weird movie
18:54:01 Quit GreenKirby (Quit: ChatZilla 0.9.86 [Firefox 3.6.10/20100914125854])
18:56:53 Join benjamin_ [0] (
18:56:58gevaertshm, if we're going to actively use that "waiting on customer" status in flyspray (which does seem like a good idea), maybe it should be renamed...
19:00:11bertrikit could be a nice help to select old tasks for closure when the last response was (for example) more than 6 months ago
19:00:42gevaertsOnly we don't have customers, and we certainly don't wait on them :)
19:01:29n1sindeed, i actually made a list of things i want to discuss/change about (our) flyspray
19:06:59***Saving seen data "./dancer.seen"
19:10:15 Join aevin [0] (eivindsy@unaffiliated/aevin)
19:19:35 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
19:21:48 Quit t0rc (Client Quit)
19:23:54 Quit Secondgen (Quit: Page closed)
19:24:24 Quit GeekShadow (Quit: The cake is a lie !)
19:24:40 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
19:29:05 Join paulk_ [0] (
19:30:03paulk_I'm maybe going to buy a Sansa Fuze, but I'd like to know if the Fuzev2 USB is now correctly supported by rb, if it'll soon be the case or if there is no work on it ?
19:33:35gevaertsThere are still issues
19:34:12 Join s1gma_ [0] (~d.d.derp@
19:34:39gevaertsI don't think pamaury has given up, but from my experience this really isn't the sort of problem where one can guess how much work is still needed, let alone estimate a timeline
19:35:13paulk_all right
19:37:52 Quit _s1gma (Ping timeout: 276 seconds)
19:43:18 Join DerPapst1 [0] (
19:44:18 Quit jfc (Read error: Connection reset by peer)
19:44:32 Quit GodEater (Read error: Connection reset by peer)
19:44:35 Join GodEater_ [0] (
19:44:36 Quit GodEater_ (Changing host)
19:44:36 Join GodEater_ [0] (~bibble@rockbox/staff/GodEater)
19:44:59 Join jfc [0] (
19:45:10 Quit DerPapst (Ping timeout: 245 seconds)
19:51:45 Join anewuser [0] (kvirc@unaffiliated/anewuser)
19:53:03gevaertsWe must remember to not release 64mb ipod video builds...
19:56:06Dreamxtremei still want ipod classics
19:56:14Dreamxtremerockbox is the ONLY way now
19:56:29Dreamxtremesince the itunes firmware is borked
19:56:30gevaertsBecause they're not needed any more
19:56:51Dreamxtremesoz wrong chan
20:03:04 Join simonrvn_ [0] (simon@
20:06:20 Quit simonrvn (Ping timeout: 240 seconds)
20:06:20 Nick simonrvn_ is now known as simonrvn (simon@
20:13:48 Quit moos (Quit: ChatZilla 0.9.86 [Firefox 3.6.10/20100914125854])
20:14:04 Quit stoffel (Ping timeout: 255 seconds)
20:16:17n1sgevaerts: shouldn't be too hard, configure doesn't let you config for one anymore
20:16:30gevaertsn1s: it never explicitely did
20:17:00n1show did you build them then, hack the makefile by hand?
20:17:36gevaertsoh, wait
20:17:42soapenter the MB of storage you wanted (had).
20:18:22gevaertsYes, it doesn't ask for RAM size any more now
20:18:28n1sbut you can't do that anymore
20:18:50 Quit krazykit (Ping timeout: 240 seconds)
20:19:22kugelbut what if you pass −−ram=64?
20:19:52n1skugel: yes, you can do that
20:20:04 Join krazykit [0] (
20:20:22linuxstbShouldn't we just remove it from
20:20:23n1salthough it always configures for 64MB
20:20:52gevaertslinuxstb: yes, but we decided to do that at release time to avoid interesting issues
20:22:05kugelthe release download page is using right?
20:22:13n1sspecifying ramsizes !=64 doesn't stick in the Makefile
20:22:26gevaertskugel: yes
20:22:33n1sso even if the buildscript does it that way both builds will be identical
20:22:50kugelcan't it just use the from the v3.6 (latest release) tag which would be correct even if a target was promoted or removed?
20:23:04gevaertsn1s: indeed. I looked at the current builds for both, and md5sums match
20:23:17 Join Highlander_ [0] (
20:23:52gevaertskugel: I think the rbutil helper cgis also use it, and those need to know about current builds too
20:24:24gevaertsRemoving a target should be very rare, and promoting a target just shouldn't be done at non-release times
20:24:46linuxstbWouldn't it be useful to keep track of which targets were in which release somehow? Wouldn't that avoid this?
20:24:56kugelI thought about using from 3.6 for releases and from current svn for current builds
20:26:33 Quit Highlander (Ping timeout: 240 seconds)
20:44:08 Part paulk_
20:48:49 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
20:53:36 Join moos [0] (moos@rockbox/staff/moos)
20:55:32bertrikgevaerts, amiconn, jhMikeS, Did you ever experience the problem in ?
20:56:32gevaertsbertrik: I've only ever used my h10 for testing, so no
21:07:02***Saving seen data "./dancer.seen"
21:09:27TheSevennow that 3.7 is approaching, are there any objections to bumping the nano2g to stable?
21:11:24 Join simonrvn_ [0] (
21:11:50 Quit simonrvn (Disconnected by services)
21:11:50 Nick simonrvn_ is now known as simonrvn (
21:12:29kugelTheSeven: I guess it's up to the port maintainer
21:12:44kugelbut, have the various stkovs fixed been fixed?
21:13:09TheSevennot really fixed, mitigated by increasing stack sizes in various places
21:13:37TheSevenand some have disappeared after some fat code rework
21:13:38 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
21:13:51kugelif it can be fixed this way I'd say go for it
21:14:04gevaertsI haven't seen any recent stkov reports for nano2g
21:14:15kugelthe android port needs bigger stacks too, I don't think bigger stacks on a per-target basis is a bad thing
21:14:17TheSeventhere might be some hiding in the corners
21:14:53n1skugel: why does the android port need bigger stacks?
21:15:02kugelI don't know
21:15:03TheSevenfor example, there used to be one in battery_bench, which might have been caught by the fat rework, but i'm currently using a locally increased stack on mine
21:15:30kugelI guess bionic is a bit more bloated than our c library
21:15:53TheSevenbut I haven't hit any stkovs during everyday use since the fat code rework
21:16:11n1sTheSeven: is FS #11097 fixed?
21:16:40TheSevenlet me check...
21:16:56n1snot that it should block a stable status imo
21:17:11gevaertsn1s: actually, I strongly suspect that the mr500 stkovs on screenshot...
21:17:12TheSevenno, that one is still present
21:17:16kugelS_a_i_n_t reported a lot before his downtime
21:17:31n1sgevaerts: ah
21:17:49n1sdisplay size dependant stack buffer?
21:18:04gevaertsNot the best code I think
21:18:19gevaertsI remember having a quick look at it
21:18:32TheSevenshould I commit an #ifdef'ed stack size increase for screendump?
21:19:06 Quit kugel (Remote host closed the connection)
21:19:30gevaertsTheSeven: for a release I think so. For trunk I'm not sure
21:19:49TheSevenwhat i'm more concerned about is the FTL corruption on unclean shutdown issue that apparently went away by itself.
21:20:23TheSevensince i don't have an explanation why it disappeared, i fear that it might just come back at any time
21:20:32soapIs there a safety margin to be "bought" for the release version by blindly increasing the stack size?
21:20:47 Join earcar [0] (
21:21:11gevaertsTheSeven: maybe "DEFAULT_STACK_SIZE + 4*SECTOR_SIZE" in usb.c?
21:21:24TheSevenwhy on earth would it need 4 sectors?
21:21:28gevaertsno idea
21:21:37gevaertsIt's 0x800 now
21:21:56TheSevenwhich is one sector on nano2g
21:22:07gevaertsUnless my brain is malfunctioning, 4*SECTOR_SIZE would leave this unchanged anywhere except on nano2g
21:22:15TheSevenprobably yes
21:22:16n1sthe screendump function shouldn't use more than 1k of stack or so, does it run in the usb thread?
21:22:23gevaertsn1s: it does, yes
21:23:05TheSevengevaerts: do you remember how many on-stack buffers (worst case) the fat code is using?
21:23:15TheSevenIIRC you brought that down to two
21:23:32TheSevenor even one?
21:23:33gevaertsI can't remember if we managed to get it down to one or not
21:24:04gevaertssee r28011
21:24:12TheSevenhm. so where is all this stack space being used?
21:24:36TheSeven352 bytes for the LCD line, 2KB for the sector
21:24:49TheSevenand IIUC we have 3KB of stack
21:24:50gevaertsDEFAULT_STACK_SIZE is 1K, so the USB stack has 3K
21:24:54saratogadoesn't the 2G have tons of ram for a flash player? might as well bump the stack sizes for releases
21:25:03saratogathat way users are unlikely to run into problems
21:25:29TheSevensaratoga: yes, it doesn't really matter if we have 30 or 29.99MB or audio buffer :)
21:25:32n1sTheSeven: screen_dump has a MAX_PATH buffer too so ~600bytes in total for that func
21:25:45n1show big is the usb stack?
21:25:54TheSevenlet me try bumping it to 4K
21:26:31n1sbtw, making that screendump code use a smaller buffer looks like it would be pretty easy
21:27:04TheSevenIIUC (DEFAULT_STACK_SIZE + SECTOR_SIZE + lcd width * lcd bytes per pixel) should be sufficient
21:27:19gevaertsTheSeven: the line buffer is 704 bytes
21:27:24TheSevenwhy that?
21:27:27 Join _s1gma [0] (~d.d.derp@
21:27:32gevaertsIt's 2*LCD_WIDTH in shorts, not chars
21:27:47TheSevenhuh? why the factor of two then?
21:27:53gevaertsoh, wait
21:27:57gevaerts /2 again
21:27:59n1sgevaerts: there's a /2 in the declaration of the array
21:30:25 Quit s1gma_ (Ping timeout: 272 seconds)
21:33:01TheSevenwhat's the mr500 lcd witdh?
21:33:03gevaertsscreendump() uses 624 bytes of stack on the nano2g
21:33:14TheSevenok, so another 1280 bytes for the line buffer
21:33:32n1sin fact, the buffer is entirely useless on the nano2g IIUC, all that code does is process one lcd framebuffer line with htole16 and then write it to disk, but since the nano has LE frame buffer data and LE cpu it could just write the lcd frambuffer directly to disk
21:33:34gevaertsBut on the other hand it has 512-byte sectors
21:34:00gevaertsSo we have 400 bytes margin...
21:34:06*gevaerts investigates further
21:34:23TheSeveni'd like to make the stack size DEFAULT_STACK_SIZE + SECTOR_SIZE if possible
21:35:30gevaertsThat would be ideal, but I wouldn't shrink the stack on any target one week before release
21:35:53TheSevengevaerts: i'd do that after branching, and commit it ifdeffed for nano2g to the branch
21:35:54 Join nerwal [0] (
21:38:59nerwalI just started using RB on my sansa e280, and I can't find the 'shortcuts' application. It should be there, shouldn't it?
21:43:48gevaertsTheSeven: not counting functions like memset() and memcpy(), I count 3124 bytes of stack usage on nano2g from usb_tread via screendump() to create a file, using "sub sp, sp, XXX" from disassembly
21:44:37 Nick guymann_ is now known as guymann (
21:45:10TheSevenso 724 + BMP_LINESIZE + SECTOR_SIZE
21:45:24TheSeveni'm trying DEFAULT_STACK_SIZE + SECTOR_SIZE + BMP_LINESIZE right now
21:46:02*TheSeven wants a faster linux box
21:46:23gevaertsyes, that should be enough on all targets I think
21:51:00TheSevengevaerts: stkov
21:52:44gevaertsWhere did those 300 bytes go?
21:53:04n1syou did note the MAX_PATH buffer in screen_dump, no?
21:53:44gevaertsn1s: I looked at the "sub sp,sp,XX" lines in the disassembly. Those should have everything
21:54:11n1sthey should, yes
21:55:20TheSevengevaerts: stmfd sp!, {...}?
21:56:19 Quit AlexP (Remote host closed the connection)
21:56:51TheSevendon't those also allocate stack space?
21:57:31gevaertsI didn't see those
21:57:37TheSevenit apparently needs about 1.5K + SECTOR_SIZE + BMP_LINESIZE
21:57:58TheSeven0x1c0 instead crashes
21:58:52TheSevendid you just analyze screen_dump() or also the code calling it?
21:58:53 Quit nerwal (Quit: CGI:IRC (EOF))
21:59:30gevaertsIt's called from usb_thread(), which uses 12 bytes
22:03:35*TheSeven checks a post-mortem memory dump
22:03:54gevaertsTheSeven: fat_rename() still seems to be bad, with 4716 bytes...
22:04:17TheSevenwhy should screendump call that one?
22:04:42gevaertsIt won't
22:04:49gevaertsBut it shows that we still have problems
22:12:26 Quit kevku (Ping timeout: 240 seconds)
22:16:43 Join wodz [0] (
22:17:19wodzhow much stack can I allocate from the plugin?
22:18:30n1swodz: plugins run in the UI thread which has 8k stack
22:18:50TheSeven...or 16k on the nano2g
22:19:06wodzgood, so ~4kB of tables is not a problem
22:19:36n1sdepends on how much other things use, etc
22:20:04TheSevenis there any reason why they can't be allocated statically?
22:20:30wodzthey can but than it is ram which is slow on coldfire
22:20:44 Join petur [0] (~petur@rockbox/developer/petur)
22:20:49wodzstack is on iram
22:21:47*TheSeven wonders why ida needs several minutes to import the memory dump
22:22:18wodzida is faaaar from perfect but unfortunately still the best
22:22:51 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
22:24:23 Join robin0800 [0] (~robin0800@
22:39:09 Quit robin0800 (Remote host closed the connection)
22:41:01 Quit benjamin_ (Quit: Ex-Chat)
22:43:26TheSevengevaerts: *facepalm*
22:43:42TheSevenit's the FTL/NAND driver of course, that's using the additional space
22:44:00TheSevenand the FTL seems to be rather hungry
22:51:44TheSevengevaerts: the usb_thread function actually needs 48 bytes because it (unneccessarily) saves a bunch of regs on entry
22:53:41TheSevenscreen_dump needs 656 bytes
22:54:04gevaertsah, right. Push adds things...
22:55:03gevaertsTheSeven: can I assume that stack usage is the "sub sp,sp" stuff plus 4 bytes per push?
22:55:34TheSevenprobably yes, if gcc doesn't allocate things deeper inside the function
22:55:44TheSevenif there are variable-size arrays on the stack it won't be that easy
22:58:06 Quit earcar (Quit: bye)
23:01:50Bagderit would be fun to know how the "stock" ffmpeg codecs compare performance wise with Rockbox codecs on android
23:01:55amiconnbertrik: No, but (a) I rarely ever use my H10 and (b) the report doesn't say whether it happened on a big or small H10
23:03:46TheSevengevaerts: write_long_name is such a case...
23:03:46 Join grawity [0] (
23:04:38bertrikamiconn, I was not aware of different h10 models, but as far as I can see, there is only one button driver for the h10
23:05:36amiconnThe hardware may still behave differently, even if the button part is the same
23:06:06amiconnThe microdrive needs much less spinup current than the 1.8" hdd
23:07:06***Saving seen data "./dancer.seen"
23:10:01grawityIs there a chance commit r28190 (FS #11646) could be undone? With reduced voltage, the contrast is reduced enough to be annoying. (the LCD is ILI9320)
23:13:35 Quit Judas_PhD (Quit: This is a quitting message)
23:14:46 Quit petur (Ping timeout: 265 seconds)
23:16:33 Join thegeek_ [0] (
23:18:07 Join binaryhermit [0] (
23:20:14 Quit Highlander_ (Quit: Quitte)
23:20:55 Quit thegeek (Ping timeout: 250 seconds)
23:21:10 Join G4Oblivion [0] (
23:22:04 Join earcar [0] (
23:26:27CIA-7New commit by bertrik (r28290): Make two local functions static in skin_tokens.c
23:26:38 Quit DerPapst1 (Read error: Connection reset by peer)
23:26:48TheSevengevaerts: i unrolled the post mortem dump in ida:
23:27:00TheSeventhe numbers in parantheses are the amount of stack space allocated by that function
23:27:29TheSevenif the stack would have been 4 bytes bigger, it wouldn't have overflowed
23:28:01TheSeventhread_stkov was the first stack frame that overflowed the stack
23:28:22CIA-7r28290 build result: All green
23:28:22gevaertsSo it's basically the long name that gets in the way
23:28:30 Join DerPapst [0] (
23:28:38TheSevenit only got triggered because switch_thread overwrote the magic number, but nothing more than that
23:28:50 Join casainho [0] (~chatzilla@
23:29:33TheSeventhe long name, and the high number of function call levels in the storage driver
23:30:08TheSeventhe storage driver was eating 408 in that case
23:30:13TheSeven408 bytes*
23:32:14TheSevengrawity: could that be fixed by changing some contrast settings in the lcd controller?
23:33:02 Quit grawity (Quit: Good night.)
23:33:42TheSevenon the LDS176 it seems to be fine
23:34:08*TheSeven just noticed an interesting bug
23:34:42 Join grawity [0] (
23:35:12TheSevenwith incompatible plugins installed, entering credits shows an incompatibility message and a rockbox logo. when you press a key, it returns to the main menu, but with remnants of the rockbox logo still visible
23:36:19grawityTheSeven: I'm just a user, so I don't know about the hardware details
23:36:54TheSevenso apparently the backdrop isn't redrawn in that situation
23:37:30TheSevengrawity: [23:33] <TheSeven> on the LDS176 it seems to be fine
23:37:37TheSevenso i can't test it on mine
23:38:02TheSevenwhat exactly is the problem? black being shown as gray?
23:38:08*grawity just upgraded a month old rockbox, noticed the low-contrast display, and downgraded to a July's backup.
23:39:43grawityWell, black seems to stay black, but overall colors are much lighter, "washed out" or something.
23:39:59 Quit wodz (Quit: Leaving)
23:40:15*grawity can try making photos for comparison tomorrow
23:40:21TheSevenis the brightness too low?
23:40:25 Quit badcam (Ping timeout: 250 seconds)
23:41:04TheSeveni.e. does it look better at maximum brightness?
23:41:47grawity(and tbh, I'm not sure it's that particular revision either; it's just the only nano2g display-related change I could find. Sorry if it's not the one)
23:42:10grawityNo, it looks bad irrelevant of brightness settings
23:44:15G4Obliviongrawity: r28290 without LCD patch
23:45:05 Quit Gabe_G23 (Read error: Connection reset by peer)
23:45:08 Quit scorche (Read error: Connection reset by peer)
23:46:08 Join Gabe_G23 [0] (~gabe@bzflag/player/GabrielG)
23:46:09grawityThanks, will try out tomorrow. (Kinda late here)
23:47:25 Join scorche [0] (~scorche@rockbox/administrator/scorche)
23:48:04gevaertsTheSeven: rename() is still bad... It apparently can use about 8KB of stack in the worst case
23:48:32TheSevengevaerts: which code path is that?
23:48:37gevaertsTheSeven: looking now :)
23:48:44*gevaerts has this all in a database!
23:50:06TheSevenrename itself eats 0x128 bytes
23:54:04 Quit casainho (Remote host closed the connection)
23:54:09CIA-7New commit by moos (r28291): Update the french translation for the upcoming release.
23:55:09*gevaerts thinks that there are some deep call chains in rockbox
23:55:59CIA-7r28291 build result: All green
23:56:36n1slooks like it!
23:56:46TheSevengevaerts: what did you use to figure that out?
23:56:57gevaertsTheSeven: awk, sqlite, and some patience :)
23:57:03gevaertsoh, and objdump
23:57:12gevaertsTrying to clean it up now
23:57:35TheSevenso you're disassembling it and parsing the output?
23:57:52gevaertsI don't take variable arrays and push/pop into account yet
23:58:01TheSevenhm. why can't functions just have an attribute in the elf file that tells how big their stack frame is.
23:58:16gevaertspush/pop should be possible, but variable arrays are basically not doable
23:58:24gevaertsand recursion is of course a problem

Previous day | Next day