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 2011-11-16

00:00:45pamaurythat's a zip file, it contains the source code and the makefile but just send the file and don't try to rebuild it. Then you will need to build mem_reader and perhaps modify it if you are working under windows.
00:01:16pamauryideally you should run something like this: read_util 066f:3770 0xFFFF0000 65536 rom.bin
00:02:23 Join lovasoa [0] (~olojkine@
00:02:49low_lightOk. I'm running out of time, I'll have a look at it tonight or tomorrow.
00:03:30lovasoaI just posted a patch, and as adviced on flyspray, I come here to signal it
00:03:37lovasoait's here:
00:04:03pamaurylow_light: ok thanks
00:04:11lovasoaIt's a very small patch to correct a "bug" in minesweeper
00:04:19rarogoh, this was a feature I missed. :D
00:04:52rarogI used to play minesweeper quite often in my rockboxed ipod.
00:06:13lovasoaI didn't play it with my iPod, because playing with a scrollwheel and a large screen is horrible
00:06:32*gevaerts thinks the game would be better if the starting tile test were removed altogether. Mines are dangerous things, and the game should reflect that! ;)
00:07:22rarogxD this reminds me of peta in the news with the evil rat killing in bf3 and now about the evil mario.
00:07:53lovasoaYou should also remove doom plugin: killing zombies is evil
00:08:27lovasoaAnd snake. Snakes are dangerous
00:08:47gevaertslovasoa: I'm not proposing removing the minesweeper game
00:10:02rarogso removing all the mines would ensure happy people as they would always win. xD
00:10:48lovasoaJust as rockblox1d? You can only win?
00:11:05*gevaerts must not be expressing himself clearly today
00:11:58lovasoaAlso, I'm not a native speaker... I'm french, and not very good in English at school...
00:12:41lovasoaSo, can you include the patch, gevaerts?
00:14:12*gevaerts thinks
00:14:25*lovasoa hopes
00:14:37*rarog scratches his head
00:15:10gevaertslovasoa: I'm wondering if some people might complain about the game getting too easy with this
00:15:36gevaertsI guess those people are already unhappy about not being able to lose on the very first move though
00:15:54pamaurywhy is this a problem that the first tile is adjacent to a mine ?
00:16:11pamauryby avoiding this you are clearly making it easier
00:16:20gevaertsNot really
00:16:30lovasoaBecause then, you have to pick another tile randomly
00:16:35gevaertsYou're making it harder to get into a situation where you *have* to guess
00:17:11pamauryfor big ones, you nearly always have to guess once at least, so guessing right after the first one is not really a problem
00:17:38lovasoaI looked at "gnome mines" source code, and found this comment:
00:17:40lovasoa /* Randomly set the mines, but avoid the current and adjacent locations */
00:17:42lovasoa private void place_mines (uint x, uint y)
00:17:44lovasoa {
00:18:59gevaertspamaury: should the current code stay then? I mean, why can't you lose on the very first move when nothing has been uncovered yet?
00:19:00lovasoaI think that what I implemented is the "implicit normal behaviour" of the game
00:19:09 Quit ender` (Quit: foot, n. a device used for finding Lego bricks in the darkness)
00:19:27Lloreangevaerts: I think there's no point in losing on the first move.
00:19:33pamaurygevaerts: no point
00:19:42Lloreangevaerts: What's happening in an implicit "restart the game until you didn't lose on the first move"
00:19:56LloreanBut any other "hand holding" removes player agency. It makes what you're doing matter less.
00:20:17gevaertsSo why not extend this to the second move where you're *still* just guessing?
00:20:24LloreanIf you could lose on the first move, all that would result in is how long you have to wait (by restarting) until you get to actually play
00:20:30pamaurybut I understand the reason for the patch and I have no objection to it, just stating a potential counter argument
00:21:50gevaertsIdeally I think you'd want a generator where you *never* have to guess (if you understand the game well enough and given enough coffee), but I don't think it's reasonable to expect a simple patch for that
00:21:52 Part low_light
00:22:05lovasoaI played rockbox's minesweeper today, and was annoyed by having to always restart the game after the 2nd move...
00:22:47funmanlovasoa: i'd say avoid the current location but not the adjacent ones
00:22:55 Quit robin0800_ (Ping timeout: 245 seconds)
00:23:05 Quit lebellium (Quit: ChatZilla 0.9.87 [Firefox 9.0/20111109112850])
00:23:16gevaertsfunman: but why are those cases different?
00:23:18pamaurygevaerts: then you need to implement an AI and it's not really interesting
00:23:19lovasoaIt was the behaviour before my patch
00:24:09pamauryexcept if you are able to generate interesting game which never need to be guessed but I'm no sure it's really feasible
00:24:34pamauryanyway, minesweeper is just a mecanimal game
00:24:46gevaertspamaury: the implementation in sgt-puzzles does this
00:25:01funmanlovasoa: ah right, i like this current implementation
00:25:05*gevaerts still wants to port that collection to rockbox one day...
00:25:37pamauryif you do this you are removing all the pleasure of the game: guessing the last square :)
00:25:48lovasoaWhy does everyone wants to annoy me?
00:26:16gevaertsfunman: why is having to guess the first move bad but having to guess the second move good?
00:26:19pamaurylovasoa: my remark was not on your patch, it's fine; it's a general thought about the minesweeper
00:27:03funmangevaerts: 0 information on the first move
00:27:12*pamaury thinks minesweeper is in fact a guessing game since it's too easy to solve everything else
00:27:28gevaertsfunman: if the first move tells you there are adjacent mines, you still have 0 information
00:27:50funmanyes yo21:08 < geal> xtophe:
00:27:51funmanu do
00:28:14 Quit pamaury (Remote host closed the connection)
00:28:25funmanyou know a probability (but ok you knew one at the first move too)
00:28:43gevaertsThe way minesweeper works you can only deduce things based on at least two positions
00:29:08gevaertsI do believe we have a dice rolling plugin if you just want to guess :)
00:29:08 Quit liar (Remote host closed the connection)
00:32:02 Join robin0800_ [0] (~robin0800@
00:35:23gevaertslovasoa: you're alrady in CREDITS, right?
00:35:37lovasoaYes I am :)
00:37:47 Quit bertrik (Ping timeout: 260 seconds)
00:44:58CIA-22New commit by gevaerts (r30994): Don't put mines adjacent to the starting position instead of just not *on* the ...
00:45:56*gevaerts prepares for the flamewar
00:46:30lovasoaThank you ! ;)
00:46:31funmanwe'll just let you walk on a mine...
00:46:53JdGordongevaerts: you wanna commit all the lang updates that are sitting around too? :)
00:46:55CIA-22r30994 build result: 2 errors, 1 warnings (gevaerts committed)
00:47:13lovasoaOh oh...
00:47:39*gevaerts claims that that warning is also JdGordon's
00:48:09JdGordonoh dammit
00:49:19gevaertsWhy is the recorder 8MB build larger than the regular?
00:49:33JdGordonbecause 8mb is the minimu for some features
00:49:42JdGordondatabase? ramcache?
00:50:26gevaertsah, yes. At least dircache
00:51:12lovasoaNow that you commited, I can go to bed...
00:51:18lovasoaGood night !
00:52:10*gevaerts thinks it's time for the archoses to move to a regular bootloader-based system
00:52:21 Part lovasoa
00:59:27***Saving seen data "./dancer.seen"
01:01:50 Quit Scromple (Ping timeout: 240 seconds)
01:01:58gevaertsYes, just dircache
01:03:29 Quit rarog (Quit: ChatZilla 0.9.87 [SeaMonkey 2.4.1/20111011102430])
01:18:28 Join nosa [0] (
01:18:40 Join Viperfang [0] (
01:19:18ViperfangGood evening, FS #12083 no long applies correctly (1 failed chunk) would anyone mind helping my apply it?
01:19:19fs-bluebot Add delay to usb-s3c6400x init to fix usb hang on some ipod classics (patches, new)
01:20:04 Quit nosa-j (Ping timeout: 256 seconds)
01:20:04 Nick nosa is now known as nosa-j (
01:23:45 Part Viperfang
01:26:22 Join ChickeNES-laptop [0] (~ChickeNES@
01:30:45 Quit ReimuHak_ (Quit: Leaving...)
01:43:25 Quit ChickeNES-laptop (Read error: Connection reset by peer)
01:43:46 Join ChickeNES-laptop [0] (~ChickeNES@
01:44:56 Quit ChickeNES-laptop (Read error: Connection reset by peer)
01:46:11 Join ChickeNES-laptop [0] (~ChickeNES@
01:50:56 Join Strife89 [0] (~Strife89@
01:51:59 Quit ChickeNES-laptop (Quit: Computer has gone to sleep.)
02:03:42 Quit Keripo (Ping timeout: 256 seconds)
02:13:03 Join Zarggg [0] (
02:14:01 Join qnm [0] (~qnm@2001:44b8:3110:f300:208:9bff:fec0:179a)
02:18:18 Join Keripo [0] (
02:22:20 Quit Poodlemastah (Remote host closed the connection)
02:27:05 Quit Strife89 (Quit: Leaving)
02:28:41 Join Strife89 [0] (~Strife89@
02:32:56 Quit funman (Quit: leaving)
02:35:02 Quit Strife89 (Ping timeout: 240 seconds)
02:35:50 Quit markun (Ping timeout: 260 seconds)
02:41:41 Quit GeekShadow (Remote host closed the connection)
02:46:22 Quit robin0800_ (Ping timeout: 248 seconds)
02:47:52 Quit Keripo (Quit: Leaving.)
02:59:28***Saving seen data "./dancer.seen"
03:02:23 Join Scromple [0] (
03:07:09 Quit Rob2222 (Ping timeout: 256 seconds)
03:21:48 Join robin0800 [0] (~robin0800@
03:26:36 Join Keripo [0] (
03:26:42 Quit fatline (Quit: This computer has gone to sleep)
04:11:36 Quit robin0800 (Quit: Leaving)
04:18:31 Join nosa [0] (
04:20:04 Quit nosa-j (Ping timeout: 258 seconds)
04:20:04 Nick nosa is now known as nosa-j (
04:21:02 Join fatline [0] (
04:22:10 Join robin0800 [0] (~robin0800@
04:24:32 Join Rob2222 [0] (
04:33:04 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:33:04 Quit amiconn (Disconnected by services)
04:33:12 Quit pixelma (Disconnected by services)
04:33:14 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:33:16 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:33:26 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:35:53 Quit hilbert (Quit: Textual IRC Client:
04:39:56 Join Scr0mple [0] (
04:40:19 Join ReimuHak_ [0] (
04:41:20 Quit Scromple (Ping timeout: 252 seconds)
04:48:19 Join fyrestorm [0] (
04:51:30 Quit [7] (Disconnected by services)
04:51:45 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:59:30***Saving seen data "./dancer.seen"
05:07:51 Quit fatline (Quit: This computer has gone to sleep)
05:08:02 Quit Keripo (Quit: Leaving.)
05:10:14 Join fatline [0] (
05:15:12 Join Rob2223 [0] (
05:18:55 Quit Rob2222 (Ping timeout: 248 seconds)
05:28:22 Join dreamlayers [0] (
05:28:22 Quit dreamlayers (Changing host)
05:28:22 Join dreamlayers [0] (~dreamlaye@rockbox/developer/dreamlayers)
05:31:38JdGordonDoes anyone know why FS #5111 isnt in svn?
05:31:40fs-bluebot Ipod piezo driver (patches, unconfirmed)
05:32:31 Quit MethoS- (Read error: Connection reset by peer)
05:33:21 Quit ps-auxw (Ping timeout: 245 seconds)
05:35:29 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
05:43:03 Quit Horscht (Quit: Verlassend)
06:04:17 Quit fatline (Quit: This computer has gone to sleep)
06:05:05 Quit Zarggg (Quit: Rebooting client...)
06:08:45 Join fatline [0] (
06:21:47 Quit robin0800 (Ping timeout: 240 seconds)
06:34:07 Quit fatline (Quit: This computer has gone to sleep)
06:46:30 Join robin0800 [0] (~robin0800@
06:49:32 Join nosa [0] (
06:50:39 Quit nosa-j (Ping timeout: 260 seconds)
06:50:39 Nick nosa is now known as nosa-j (
06:54:11 Quit robin0800 (Ping timeout: 244 seconds)
06:55:02 Join nosa [0] (
06:58:14 Quit nosa-j (Ping timeout: 260 seconds)
06:58:15 Nick nosa is now known as nosa-j (
06:59:31***Saving seen data "./dancer.seen"
07:29:41dreamlayersToday I did some testing of the PP502x IDE DMA code. It seems like data being read isn't corrupt, but cpucache_invalidate() causes unrelated memory corruption. Even just adding it to ata_transfer_sectors() without enabling DMA causes the same dircache commit crash.
07:30:09dreamlayersSorry, I mean tagcache. It's late. I'll get back to this tomorrow. Goodnight.
07:30:51 Quit dreamlayers (Quit: dreamlayers)
07:48:21 Nick Scr0mple is now known as Scromple (
07:59:11 Join markun [0] (
08:19:32 Join nosa [0] (
08:20:30 Join whereisit [0] (
08:20:49 Quit nosa-j (Ping timeout: 252 seconds)
08:20:50 Nick nosa is now known as nosa-j (
08:22:48whereisitI'm trying to find the Simulator, I have looked and searched under, Docs index + General info + Wiki and I can't see a obvious link to the Sim, could someone please provide me with the URL ?">
08:23:28whereisitthank you kindly
08:23:45JdGordonor"> if oyu want current svn instead of th elast release
08:25:43 Quit whereisit (Quit: CGI:IRC (EOF))
08:37:02 Join bertrik [0] (
08:37:02 Quit bertrik (Changing host)
08:37:03 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:50:08 Quit Scromple (Quit: Leaving)
08:59:34***Saving seen data "./dancer.seen"
09:00:07 Join wodz [0] (
09:05:48 Quit wodz (Quit: Leaving)
09:09:14 Quit bluefoxx_ (Ping timeout: 252 seconds)
09:11:44 Join einhirn [0] (
09:13:47 Join KiwiCam [0] (
09:21:34 Quit bertrik (Ping timeout: 248 seconds)
09:47:41 Join why [0] (
09:48:11whyhow do I get it to display all files so that I can rolo and fleash my H140 ?
09:48:43whythe quick menu doesn't have that handy little menu anymore
09:50:43whyand wtf os root under files, i get "playlist buffer full" flag what does that mean ?
09:53:45whyseriously guys, how do I get to display all files like it used to so that I can use the lastest build and flash my player, anyone ?
09:55:19whyok, back to the old build I go.
09:55:27 Quit why (Quit: CGI:IRC)
10:00:11pixelmahow patient
10:05:32 Quit factor (Read error: Connection reset by peer)
10:12:48 Join pamaury [0] (
10:12:49 Quit pamaury (Changing host)
10:12:49 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:14:01 Join ender` [0] (
10:23:22 Join factor [0] (~factor@
10:37:11JdGordonHow do/should we handle patches which have a really long life time (i.e 5 years), with quite a few people posting patch updates (though probably more than half are just resyncs) where not everyone has their full name on flyspray?
10:39:04LloreanTry contacting any of the ones who wrote copyrwriteable code (IE not resyncs or cosmetic changes)
10:39:36kugelJdGordon: IMO a mere resync doesn't imply copyright
10:39:45JdGordonfor sure
10:40:02LloreanThey should all be associated with at least an email address.
10:40:08 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
10:40:11LloreanIt's probably possible to contact most of the ones who've contributes original code.
10:40:21JdGordonbut 47 task attachments is alot of effort to figure out exactly who shld get credited and who shouldnt
10:40:26Zagordesowin: added!
10:40:53JdGordonthe people who did the major work (i.e the piezo driver) are all commiters so that isnt such an issue
10:41:47JdGordonfreqmod sounds familiar, is he in credits already?
10:41:52gevaertsJdGordon: rewrite it :)
10:42:14JdGordonnot tonight thanks :)
10:43:47JdGordonthere is only one frederik in CREDITS, does that probably mean that is he? :)
10:43:56JdGordonhe's had a few patches accepted so i guess so
10:48:25kugelJdGordon: did you see my question on the ml?
10:48:34JdGordonr12241 is a patch from him
10:48:46JdGordonkugel: yes
10:50:48JdGordonI'm going to change the keyclick setting to speaker/headphone/both/off and commit
10:56:16JdGordonactually, headphone keyclick isnt just an on/off setting so I really have no issues with the two clickers being speerate options
10:57:09kugelJdGordon: some auto setting based on headphone detection would be nice
10:57:29JdGordonI agree with soap on that one
10:57:33kugelI still think the change you added in the last patch should be separate
10:57:49JdGordonand i tihnk that is a waste of time
10:58:03JdGordondo you have an ipod?
10:58:15JdGordonactually, it should work on software keyclick too
10:59:35***Saving seen data "./dancer.seen"
10:59:56kugelso what's the situation in the patch?
11:00:06kugelhave those separate settings or what?
11:01:38kugelrepeat keyclick is even a setting already, why did you implement it?
11:01:49kugeldoesnt that setting apply for piezo?
11:02:08 Join bluefoxx [0] (
11:03:33JdGordonI fixed it so it doesnt keep beeing like crazy when you hold down a button
11:03:40JdGordonso it only beeps on the first repeat
11:04:01kugelyes. that setting exists for keyclicks
11:04:32kugelit should work for both, which is what I meant with integration
11:04:38 Quit Bagder (Ping timeout: 240 seconds)
11:05:10JdGordonwhat we have here is a failure to communicate!
11:05:29kugelpossibly :)
11:05:29JdGordonI'm not sure what you are talkin baout
11:05:45kugelthe "dont repeat clicks if buttons are repeated" one exists in SVN as a setting
11:06:10JdGordonmy change is not that
11:06:25JdGordonwith that setting enabled, every BUTTON_REPEAT event triggers the beep which is horrible
11:06:36JdGordonso i fixed it so it only beeps on the first one
11:06:37kugelthat's why it's a setting you can turn off
11:06:58kugeloff is even the default IIRC
11:07:14JdGordonI dont want it off
11:07:21JdGordonI want it to not be broken
11:07:32kugelyou "fixed" the setting to be ignored?
11:07:57JdGordonreread what i said
11:08:28kugelthe setting is intended to keep clicking as buttons are hold
11:09:08kugelbtw, this is why that very change should be separate
11:09:22kugelit can be discussed independantly of piezo
11:10:09JdGordondo you use the keyclick?
11:10:33kugela while ago
11:10:41kugelnot on raaa :)
11:13:21 Join perrikwp_ [0] (
11:14:30kugelJdGordon: what does the setting do with your change?
11:15:42 Quit perrikwp (Ping timeout: 248 seconds)
11:16:09JdGordondont worry, im pulling it out for the itme being
11:16:13JdGordoncant be bothered arguing
11:17:49kugeli guess I just don't understand what it's doing
11:18:58 Join GeekShadow [0] (
11:20:41kugelJdGordon: btw, great job on the skin-engine-to-buflib one
11:22:18kugelhm, there's another main menu item now. time to restart the setting-system-merge again!
11:24:36JdGordonOH CAN WE>?!
11:24:49JdGordonand thanks, I was really surprised it was so simple
11:24:55JdGordon"simple" :p
11:25:42JdGordonfingers crossed this actually builds for every ipod :p
11:25:45CIA-22New commit by jdgordon (r30995): Finally commit FS #5111 - piezo clicker for ipods! ...
11:26:37pamauryoh, the famous piezo clicker driver which has been lying around for so long
11:26:59JdGordonOpened by Robert Keevil (obo) - Thursday, 13 April 2006, 10:34 GMT+10
11:28:32CIA-22r30995 build result: 2 errors, 1 warnings (jdgordon committed)
11:29:41JdGordonkugel: ok, so the problem is, with the repeat setting enabled, when you hold down a button (i.e seek or poweroff) it clicks constantly which is really irritating
11:29:54JdGordonwith the setting off thugh the scrollwheel doesnt click at all except the first time
11:30:21JdGordonso i changed it so with the setting on it only clicks the first repeat event (so twice in total)
11:30:37JdGordonthough I'm tihnking I'll just change the 'off' behaviour to do what i want which makes a ncer default
11:30:56kugelokay, that seems to make sense
11:31:13kugelyea, changing the behavior of off seems a bit better
11:31:18 Join wodz [0] (
11:31:24kugelor better yet, a "once" option perhaps?
11:31:39JdGordonI also want to fix the lists to stop clicking once you hit the end with wrap off :)
11:32:02wodzJdGordon: I think rockbox headers in files from 'piezo' commit are not compatibile with our current scheme
11:32:11 Quit ps-auxw (Ping timeout: 240 seconds)
11:32:22JdGordonhow so?
11:33:03wodzthey are missing statement about GPL version used
11:33:20 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
11:33:59JdGordonwell they *are* 5 years old :)
11:34:26kugelthey should be changed. right now it makes them gplv2 only
11:34:35wodzthats the difference
11:35:38JdGordon(though technically its not my right to change it)
11:36:50wodzthe obvious solution would be to revert and find original authors/contributors how they see license change -)
11:36:54kugelyes and no
11:37:29kugelwhen we switched to gplv2 or later we assumed all files that are not explicitely v2-only are ok to change, no?
11:37:49kugelwodz: I'm very sure roolku is fine with that
11:37:50CIA-22New commit by jdgordon (r30996): Fix the gpl notice in the previous commits new files, and fix the yelow gevearts naughtily ignored
11:38:13gevaertskugel: yes
11:38:34wodzso it should be ok
11:39:26wodzJdGordon: this time you introduced unrelated change to skin_parser.c
11:39:34JdGordonoh shoosh!
11:40:17CIA-22r30996 build result: 2 errors, 0 warnings (jdgordon committed)
11:41:22wodzand we definitely need to do something with this hanging build errors
11:42:22gevaertsYes, but what?
11:42:46kugelwell, I'd say discontinue recorder, leaving 3.10 the last release it supports
11:42:47wodzProbably that is the question to be rised on ML
11:43:03gevaertskugel: no need for that, really
11:43:10gevaertsIt's only the 8MB build that fails
11:43:35JdGordonand anyone who has a 8mb recoder can do their own build disabling the feautres they dont want
11:43:40wodzthere was a thread on forum which showed that most users use outdated builds on archoses anyway
11:43:44JdGordonbut I'd like to fork out hwcodec anyway
11:43:53gevaertsI'd say we either drop the 8MB recorder build or we disable dircache for it explicitely
11:43:59kugelthe only difference to the 2mb one should be dircache and tc ramcache
11:44:05wodzpixelma and amiconn excluded
11:44:28 Join TomColler [0] (
11:45:02kugelgevaerts: I don't think there's a lot point in starting to disable features
11:45:04JdGordonor moving to a bootloader like gevaerts said last night
11:45:41gevaertsAh, yes. I'd forgotten about that :)
11:47:20wodzI think that branching is more beneficial personally
11:47:54JdGordoncan anyone think of a possible excuse to have keyclick enabled but *not* want the scroll wheel to beep every step? (I can sort of get why you wouldnt want press buttons to keep beeping on repeat thugh)
11:48:36*kugel cant
11:50:04 Join nosa [0] (
11:50:50wodzclicking on wheel step is the most usefull
11:51:22kugelis the binsize statistics discontinued?
11:51:28 Quit nosa-j (Ping timeout: 276 seconds)
11:51:28 Nick nosa is now known as nosa-j (
12:05:47CIA-22New commit by jdgordon (r30997): Change the "keyclick repeat" setting behaviour so when it is off only the *first* repeat will click. ...
12:05:57JdGordonnot the best commit message, but much nicer behaviour
12:06:46 Quit einhirn (Ping timeout: 244 seconds)
12:07:47 Quit Silicium (Ping timeout: 240 seconds)
12:08:07CIA-22r30997 build result: 2 errors, 0 warnings (jdgordon committed)
12:08:50 Join Silicium [0] (
12:11:28wodzJdGordon: You have proposed a change in action.c related to HAVE_SCROLLWHEEL some time ago. Have you tried it? I can't find it now in irc logs.
12:11:56JdGordonI don't remember what it was :)
12:12:35wodzJdGordon: it is about hack in line 255
12:12:58 Quit tmzt (Ping timeout: 260 seconds)
12:13:34JdGordoncrap, I don't remember what it was
12:13:43JdGordoni remember i had an idea though... grr
12:15:19*JdGordon greps his logs
12:15:29 Join tmzt [0] (
12:20:45JdGordonwodz: ah, I may have stumbled on a nice place for a fix
12:23:10JdGordondoes this line look correct?
12:23:22JdGordon((last_button&(BUTTON_SCROLL_BACK | BUTTON_SCROLL_FWD)&button) == 0)
12:23:40JdGordonif last button was one of those two, and button is not one
12:23:57wodzJdGordon: I see two possible solutions (as I consider current 'fix' as a nasty hack) 1) fix lowlevel drivers to emit REL event after some timeout or on direction change 2) craft some more general change in action.c
12:24:27wodzThe second approach is much less work but I try as much as possible to not touch anything above /firmware :-)
12:24:44Misanthroposi recently wanted my clip+ to wake-"me"-up, e.g. alarming me its time go. But while hearing music this does not work. Before I start to get into coding something I'd like to know if a wakeup-sound which would be played is a good idea (or maybe hard to implement)
12:25:57wodzI think that alarm feature is meant to power-up the device on particular time
12:28:28JdGordonwodz: what was the actual issue again?
12:28:39Misanthroposthats seems a bit restrictive - lets say you'd like to be alarmed in 30 minutes - working on somehthing while hearing music... it would just have to play the alarm sound for some seconds
12:29:24Misanthroposhmm... i will have a look into that this evening :D
12:29:36wodzJdGordon: If you remove this hack, after screen change either scrollwheel or buttons stop working (can't remember now)
12:30:02wodzJdGordon: that is because action system waits for REL event
12:30:14JdGordonthe hack fixes it?
12:30:50wodzJdGordon: most of the times - but there are some sporadic reports that something doesn't work as should
12:31:43JdGordonI'm not sure i understand what is happening. buttons only get eaten (waiting for _REL) on a context change which cant happen because of a wheel event
12:31:53wodzand moreover with current code I would need to define HAVE_SCROLLWHEEL and add some fake button code defines to the hd300 scrollstrip driver
12:32:09wodzI don't like defines which are misleading
12:34:25JdGordonit will still be BUTOTN_SCROLL_FWD/BACK no?
12:37:31wodzwell in current hd300 driver it is BUTTON_UP and BUTTON_DOWN
12:38:32wodzit is hard to call it FFW or BACK as this aren't used for this
12:38:59 Join MethoS- [0] (~clemens@
12:39:24JdGordonis BUTTON_UP not a press in a specifc region of the scrollbar thing?
12:39:35JdGordonand BUTTON_SCROLLL_UP being something else
12:41:17wodzno, scrollstrip in hd300 works exactly the same as scrollwheel in ipods and the only difference is the shape
12:42:07wodzif you hold finger on one end of scrollstrip it will generate appropriate button events with some steady freq.
12:49:03 Join casainho [0] (
12:59:37***Saving seen data "./dancer.seen"
13:21:22 Quit nosa-j (*.net *.split)
13:21:22 Quit factor (*.net *.split)
13:21:22 Quit [fred] (*.net *.split)
13:23:04 Join nosa-j [0] (
13:24:01 Join 5EXAAB03Q [0] (
13:24:01 Join factor [0] (~factor@
13:24:01 Join [fred] [0] (
13:25:31 Quit 5EXAAB03Q (Ping timeout: 252 seconds)
13:37:42dfktis there something like "−−dry-run" for "svn up"? i want to be able to review currently implemented patches before they break cause of svn up
13:41:38 Join zaytsev [0] (~zaytsev@unaffiliated/zaytsev)
13:41:51zaytsevhi rockboxers
13:42:13zaytsevanyone know if there were any attempts to circumvent speedo aquabeat 2 to run rockbox?
13:42:28zaytsevgoogleing didn't really bring any results...
13:43:52kugelnever heard of that
13:45:28zaytsevkugel: it's one of the few waterproof players out there. just got one. somehow it said in the description that it would play flac, but I'm not sure, and hey, rockbox is the best either way
13:45:55 Quit ReimuHak_ (Quit: Leaving...)
13:48:03zaytsevkugel: it mounts as an msc device and just expects the files to be dropped there. dead simple player. I wonder if I can demonstrate enough skills & dedication to try to port rockbox. but I guess if nobody have approached it I will have to open the player and for something that has been waterproofed it doesn't sound like a terribly good idea
13:49:49pamauryzaytsev: is there an update/recovery mode like holding a button and plugging usb ? Then can you tell the VIP&PID of the device ?
13:50:45zaytsevpamaury: in the manual it says that the hardware reset can be performed by holding M button for 10 seconds, but I coudlnt' find any mentions of a firmware upgrade facility
13:51:47pamauryzaytsev: on google, "speedo aquabeat 2 firmware upgrade" gives a result, it should apply to your device
13:55:45zaytsevpamaury: oh I was too stupid to google for that. there is no mention on the official website that I can find, but one of the google results is a website in russian which has a zip file with an alledged firmware upgrade
13:56:25 Join openstandards [0] (~openstand@
13:56:59zaytsevpamaury: in the PDF though it has the photos of aquabeat 1 and not aquabeat 2 which is what I have. I'm affraid this upgrade might not be applicable
13:57:16pamauryyes, if you have the device, dont't try to firmware upgrade but just plug it after reset and holding stop (as described) and report the VID&PID. If by chance it's a known one, it can mean that the hardware is known to us
13:57:21JdGordonkugel: something is broken in svn with the clip, trying to load a theme
13:57:56JdGordonusing repro from
13:59:07JdGordoncrash during core_shrink()
14:00:39pamauryzaytsev: usually the procedure to enter recovery mode is of this form: shutdown the player, hold the key and plug usb. If you have a menu key, that's a good candidate
14:01:32kugelJdGordon: 95% of the lockups or crashes in buflib I've debugged were caused by faulty client code and resuling memory corruptio
14:01:51kugelAnd I beleive I've fixed the other 5% :)
14:02:26zaytsevpamaury: mine seems to have the reset button that they suggest to press on aquabeat 1, but I will try menu, stop and others when it's charged. dmesg doesn't report VID & PID and I don't seem to be able to find them with lshw though. sorry for a stupid question, but what I should be looking at?
14:02:50JdGordonkugel: same theme works fine on clip+ (8mb vs 2mb build)
14:03:37pamauryzaytsev: lsusb or dmesg, but if dmesg doesn't report device insertion then it means it wasn't the right conbinaison
14:03:37wodzzaytsev: lsusb
14:05:43kugelJdGordon: perhaps some OOM? that seem loads a font, those can be large no?
14:06:09kugelthe audio buffer now reserves quite a bit so the shrink callback there can fail
14:07:01zaytsevin usb msc mode I suspect it's 1271:0211. I will try in service mode when it's charged and report back. thanks a lot for your help so far.
14:08:47zaytsevyes, very much looks like Foxda Technology Industrial
14:09:01JdGordonthis is not good.... crashes loading the failsafe skin for the wps
14:13:00JdGordonhmm, crap, ok might be my bad, i was loading a broken wps
14:13:03JdGordonit shouldnt crash though
14:15:08JdGordondont try loading a 71kb wps :p
14:16:47JdGordonespecially when the pluginbuffer is only 64KB...
14:16:57JdGordoni.e dont do ctrl-s in chrome on a .wps!
14:18:11 Part TomColler
14:18:45wodzJdGordon: Maybe we should exercise skin engine with purposely broken wpses? This technique revealed something like 1k bugs in mozilla codebase some time ago :-)
14:19:57 Quit zu (Quit: Lost terminal)
14:20:08JdGordonhaha, no need for that. The parser is very good and will reject broken skins. th eproblem is trying to load a file that is bigger than the pluginbuffer into the plugin buffer and then giving the skin engine negative buffersize to deal with :)
14:22:32 Join einhirn [0] (
14:23:00 Join bluebrother [0] (
14:23:00 Quit bluebrother (Changing host)
14:23:00 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
14:23:57kugelJdGordon: the one on the tracker is only 0.6MB
14:25:02 Quit fs-bluebot (Ping timeout: 248 seconds)
14:25:31JdGordonI know... do ctrl-s on it and look at the resulting file
14:26:09 Join fs-bluebot [0] (
14:26:23 Join zu [0] (
14:26:49 Quit bluebrother^ (Ping timeout: 276 seconds)
14:26:55kugelJdGordon: 633bytes
14:28:43kugel(using firefox)
14:29:08JdGordongot chrome installed?
14:29:15JdGordonit puts the whole css in there!
14:32:37 Join funman [0] (~fun@rockbox/developer/funman)
14:35:08 Join fatline [0] (
14:35:26 Quit casainho (Ping timeout: 252 seconds)
14:43:34CIA-22New commit by jdgordon (r30998): OOPS! Don't overwrite the trailing \0 from the skin text with the first alloc (should fix FS #12388)
14:44:12 Quit zu (*.net *.split)
14:44:13 Quit mikroflops (*.net *.split)
14:44:13 Quit ej0rge (*.net *.split)
14:44:14 Quit preglow (*.net *.split)
14:45:14wodzJdGordon: isn't it wrong indented?
14:45:44CIA-22r30998 build result: 2 errors, 0 warnings (jdgordon committed)
14:45:45kugelalso, why does it read line by line and not the whole thing?
14:46:29JdGordonwodz: ?
14:46:37JdGordonkugel: because thats very old code
14:46:42JdGordonfeel free to fix it
14:47:18JdGordonbut \r will break the parser, so windows edited skins will need fixing
14:47:20JdGordonor handling
14:47:24kugelI expect it would speed up loading quite a bit
14:47:35 Join hilbert [0] (
14:47:44JdGordonI doubt it highly
14:47:54JdGordonanyway, im going to bed, feel free to experiment
14:48:02kugelwell, read_line() does byte-by-byte internally
14:48:21pamaurythe filesystem code does buffering
14:48:24 Join zu [0] (
14:48:24 Join mikroflops [0] (
14:48:24 Join ej0rge [0] (~alhaz@
14:48:24 Join preglow [0] (thomj@rockbox/developer/preglow)
14:48:48kugelpamaury: our code too?
14:49:09JdGordondoing a single read there would be nice to trim the code though
14:49:15pamaurysure, at least I'm pretty damn sure, it would be disaster otherwise
14:49:34kugelI'm not so sure
14:49:44pamauryyes, just checked
14:49:56wodzJdGordon: it looks like start++; is shifted one char right
14:50:31kugelwodz: the raw diff looks alright
14:50:42wodzah ok
14:51:35wodzcrappy chrome :/
14:51:54kugelhow about using a browser that's not broken guys :)
14:53:09wodzspeed up firefox and I'll be in
14:59:40***Saving seen data "./dancer.seen"
15:02:56 Quit wodz (Quit: Leaving)
15:03:15CIA-22New commit by desowin (r30999): Add myself to COMMITTERS
15:03:37kugeldesowin: welcome \o/
15:04:21*funman looks for something to commit
15:04:42desowinhuge commit on the way :P
15:05:11funmanenjoy svn revision numbers while you can!
15:05:26CIA-22r30999 build result: 2 errors, 0 warnings (desowin committed)
15:07:10pamaurydesowin: please avoid errors on first commit ;)
15:08:03CIA-22New commit by desowin (r31000): Sandisk Sansa Connect port (FS #12363) ...
15:08:20desowinpamaury: it's JdGordon fault!
15:08:37 Quit fatline (Quit: This computer has gone to sleep)
15:10:55CIA-22r31000 build result: 2 errors, 9 warnings (desowin committed)
15:10:59desowincan I get tracker rights as well?
15:11:04desowinuuuu, warnings
15:11:40pamauryit's a bit ugly that we have yet another full sd driver, we really should factor the common parts: init and read/write
15:12:04desowinI didn't touch rbclient! it's not my warning
15:12:38desowinI mean, I didn't trouch iriver.c and gigabeat.c
15:13:31funmandesowin: different compilers being used
15:15:00desowinfunman: but is it something I should worry about now? (the iriver.c and gigabeats.c warnings)
15:15:32funmanno, the warnings were present already
15:16:03funmanwell the code causing them. the warning themselves just show or not, depending on particular of compiler and libc being used to build the simulator
15:19:05desowinpamaury: yes, I completely agree with you
15:20:23pamauryI don't remember why there were some objection, probably because some drivers have quirks here and there or because of clock settings but we can't keep duplicating code like this
15:21:08 Join Thra11 [0] (~thrall@
15:22:48funmani just think nobody did it
15:24:08pamauryI'll try to do something when implement the fuze+ sd driver
15:28:28 Join preglow_ [0] (
15:28:48 Join ej0rge_ [0] (~alhaz@
15:28:55 Join zu_ [0] (
15:29:12openstandardsquick question if you were to get a mp3 player would it be the sansa clip, sansa fuze or something else?
15:30:06desowinnew target!
15:31:10funmanopenstandards: are you asking if rockbox work fine on these models?
15:32:19openstandardswell i've checked the website and i've seen they are compatible but which would is considered the best for rockbox
15:33:38 Quit zu (*.net *.split)
15:33:39 Quit mikroflops (*.net *.split)
15:33:39 Quit ej0rge (*.net *.split)
15:33:39 Quit preglow (*.net *.split)
15:34:12funmanopenstandards: look at it could help you to make your choice
15:34:21 Join mikroflops [0] (
15:35:11openstandardsdidn't see that link i'll have a look now, thank you very much
15:36:10 Quit GeekShadow (Read error: Connection reset by peer)
15:39:26pamauryis it up to date ?
15:44:32 Quit dv_ (Read error: Connection reset by peer)
15:47:55desowinpamaury: fuze+ and hifimanhm801 has the same target_id number in configure
15:48:00 Join dv_ [0] (
15:48:26pamauryreally ? that's not my fault !
15:49:11desowinyes, they both have 80, I needed to sync my patch (to 81) as the stuff got commited during my initial works
15:49:27 Quit dv_ (Read error: Connection reset by peer)
15:49:48pamauryhuh, in my repo hifimanhm801 has 79
15:50:15pamauryoh no you're right
15:53:12 Join dv_ [0] (
15:55:19pamauryshould I ask mortalis to fix it or should I fix it now ? It will break the bootloader
15:57:28 Join Topy [0] (
15:58:01 Join fatline [0] (
15:58:13pamauryI think it's better if mortalis fix it
16:00:42 Quit T44 (Ping timeout: 253 seconds)
16:01:23 Join TheLemonMan [0] (
16:14:48 Join WalkGood [0] (~4@unaffiliated/walkgood)
16:24:58 Join balintx_ [0] (
16:26:13 Quit balintx (Read error: Connection reset by peer)
16:31:40 Quit hilbert (Read error: Connection reset by peer)
16:44:00GodEater_does anyone else understand what the heck Lambda is on about with regard to removing stuff to get an iOS port going? He's left me utterly bewildered.
16:46:09pamauryI didn't quite get it to. I thought port rockbox to iOS would mean *adding* stuff
16:47:00Zagoryeah, hence my question
16:47:26 Join dreamlayers [0] (
16:47:27 Quit dreamlayers (Changing host)
16:47:27 Join dreamlayers [0] (~dreamlaye@rockbox/developer/dreamlayers)
16:47:34 Quit dreamlayers (Client Quit)
16:49:09GodEater_also, I wanted to ask him "For god's sake whhhhhhhhhhhhy?" but didn't have the heart.
16:49:39funmanwhen rockbox will be reduced to int main(void){return 0;} a port to iOS will be easier
16:50:07 Join dreamlayers [0] (
16:50:07 Quit dreamlayers (Changing host)
16:50:07 Join dreamlayers [0] (~dreamlaye@rockbox/developer/dreamlayers)
16:51:35 Quit pamaury (Remote host closed the connection)
16:59:42***Saving seen data "./dancer.seen"
16:59:46 Part Zagor
17:01:13 Quit factor (Read error: Connection reset by peer)
17:01:31 Join mortalis [0] (~mortalis@
17:03:10desowincould the desowin/zsitool/"> be included in rockbox utils directory? (being the only author I can relicense it to GPL v2 or v3 without any problems if that's required)
17:03:30 Join n1s [0] (
17:03:30 Quit n1s (Changing host)
17:03:30 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:06:46funmandesowin: what does it do?
17:07:07CIA-22New commit by mortalis (r31001): fix target id for HiFiMAN HM-801
17:07:11funmanas long as it is free i think you can choose your own license for tools
17:09:36CIA-22r31001 build result: 2 errors, 0 warnings (mortalis committed)
17:12:14desowinfunman: upload the binary to Sansa Connect when in aither recovery mode or usbautoboot mode
17:18:57 Join factor [0] (~factor@
17:21:41*dreamlayers is trying to figure out an odd problem: if ata_transfer_sectors contains "if (!write) cpucache_invalidate()", then tempbuf_sort crashes when committing the database.
17:35:06 Quit Thra11 (Read error: Connection reset by peer)
17:36:02 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
17:50:31 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
17:52:28 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
17:54:55 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
18:02:11 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:03:51pamaurymortalis: ping
18:06:12mortalispamaury: pong
18:06:22 Join Keripo [0] (
18:06:36mortalispamaury: i've fixed target id already
18:06:47pamauryok perfect
18:14:23 Quit ezzieyguywuf (Disconnected by services)
18:16:00 Quit WalkGood ()
18:18:50 Join Thra11 [0] (~thrall@
18:20:01 Join WalkGood [0] (~4@unaffiliated/walkgood)
18:23:38pamauryI tried to contact Freescale to get some information about the stmp3700 (!=stmp3780 but have lot in common) and they won't except under NDA :(
18:23:41 Join wodz [0] (
18:24:06pamaurythat's definitely stupid since they give the stmp3780 manual without any condition !
18:24:21funmanpamaury: insist, insist and then insist
18:24:24funmanthen repeat
18:24:31wodzpamaury: Although I agree that sd code refactoring is needed this may be a bit tricky
18:24:43funmanit worked for AMS although because i pointed that they had already given us doc in the past
18:25:44funmani don't remember what was the problem with SD refactoring
18:25:55wodzpamaury: I am afraid it will end up as ifdef hell
18:26:03pamauryI already insisted a lot for the stfm1000 and didn't suceed (they pretended an IP problem)
18:26:06pamaurywodz: why ?
18:26:58pamauryfunman: I already pointed out they give away the stmp3780 manual but that's all, Freescale never gave us anything, at best they wanted to sign a NDA
18:26:58 Quit factor (Read error: Connection reset by peer)
18:27:22wodzpamaury: for example rk27xx has some serious problem with reporting tran state
18:27:50wodzs3c2440 driver is weird at best
18:28:09pamaurythe s3c2440 would need to be rewritten but it's not much used
18:28:28pamaurywhy do you need to wait for the tran state ?
18:28:34 Join ChickeNES-laptop [0] (~ChickeNES@
18:28:40wodzothers have different schemes for retrieving data from fifo/buffer
18:29:21 Join factor [0] (~factor@
18:29:31 Quit factor (Read error: Connection reset by peer)
18:29:55 Join ChickeNE_ [0] (
18:29:59 Quit Thra11 (Quit: kthxbai)
18:30:14wodzpamaury: generally you should be sure that the card is in right state before issuing data transfer commands
18:30:15 Join Thra11 [0] (~thrall@
18:30:26 Join factor [0] (~factor@
18:30:30wodzat least that is how I understand (incomplete) documentation
18:31:11pamauryPerhaps it's not the case for SD, but for MMC that's not a problem, after you complete init sequence, you can't escape tran mode by doing read/write only
18:31:23*pamaury has another try at convincing freescale guys
18:32:41 Join Bagder [0] (
18:32:42 Quit Bagder (Changing host)
18:32:42 Join Bagder [241] (~daniel@rockbox/developer/bagder)
18:32:58 Quit ChickeNES-laptop (Ping timeout: 258 seconds)
18:34:39wodzthe card should report as being in tran anyway :-)
18:35:26wodzsetting sd clock divider would be needed to factor out as target specific
18:35:57pamaurythe factored portion would be the init sequence basically, sending command would be driver specific of course
18:36:00pamaury*and data
18:36:41wodzbut switching sd clock is part of init
18:36:54pamauryyou can have a driver specific function to do that
18:37:06wodzisn't it that what I said?
18:37:10 Join sacarasc [0] (sacarasc@xmms2/bitch/sacarasc)
18:37:30pamauryyou said it like it was a blocker
18:37:59wodzit is a wording problem I guess
18:38:11wodzI have nothing agains
18:39:56 Part sacarasc ("Leaving")
18:43:23 Quit Thra11 (Ping timeout: 240 seconds)
18:43:50 Join Thra11 [0] (
18:50:28 Join thegeek [0] (
18:51:29 Join Poodlemastah [0] (
18:52:49 Join Pood [0] (5518f1cd@gateway/web/freenode/ip.
18:55:32 Quit Pood (Client Quit)
18:57:43 Join liar [0] (
18:59:45***Saving seen data "./dancer.seen"
19:00:28 Quit Keripo (Quit: Leaving.)
19:09:37amiconngevaerts, kugel: Regarding the 8MB recorder build, we have several options
19:09:59amiconnThe easiest would be to disable dircache and tc_ramcache
19:10:45amiconnAnother option (for a while) would be to upgrade gcc, which hopefully would produce a smaller binary (as we're using -Os)
19:11:58amiconnThird one would be to switch to using a bootloader .ajz/.mod on hwcodec and have main rockbox in /.rockbox like on all other targets
19:12:06dreamlayersI wonder if -ffunction-sections could free up enough space.
19:12:43amiconn-ffunction-sections usually *increases* binsize, unless there are quite a lot of functions to throw away
19:13:34amiconnOn SH anyway; it is due to the fact that -ffunction-sections aligns each function at section boundaries (16 bytes) instead of the usual 4 byte alignment
19:13:34dreamlayersOn ARM there doesn't seem to be a penalty because code is always aligned. Is that not the case on sh?
19:13:41dreamlayersoh, sorry
19:14:18amiconnSome time ago I checked whether it is possible to select the section alignment; 16 bytes is unnecessary and a waste
19:15:01amiconnI guess it's aligning to cache lines, but this only helps SH2 and higher, SH1 has no cache
19:16:30 Quit ChickeNE_ (Quit: Computer has gone to sleep.)
19:16:53dreamlayersgcc 4.6.2 with binutils that supports plugins, -flto, -fuse-linker-plugin and -fwhole-program decreases size, but I didn't create a working v2 Recorder binary yet, and the linking is *much* slower.
19:16:53 Join Horscht [0] (
19:16:53 Quit Horscht (Changing host)
19:16:53 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:18:05dreamlayersWhat would be needed for decreasing section alignment? Did you find an option or would binutils have to be patched?
19:18:23amiconnI didn't find an option (in 4.0.x)
19:19:28amiconnSwitching to a bootloader model would mean that we should change how booting from flash works. Since loading the uncompressed and unscrambled rockbox binary from disk is probably faster than decompressing rockbox.ucl from rom, I'd do away with ucl-from-rom altogether
19:19:57dreamlayersman sh-elf-ld shows some relevant options
19:21:02amiconnSo we'd essentially have our own flash loader ("bootbox+") which can boot either rockbox.archos (default) or an OF from .ajz/.mod
19:21:14 Quit liar (Read error: Connection timed out)
19:22:25Poodlemastah /msg *status help
19:22:31amiconnOn non-flashable archoses, bootbox would reside in / as ajbrec.ajz or archos.mod, respectively. Once it's flashed these files would be unnecessary
19:23:14amiconnI could test recv1, player, and the Ondios. Unfortunately I have no fm or v2
19:23:55dreamlayersI have a v2 I could use to test
19:24:22amiconnLinusN has an fm he doesn't use afaik
19:25:11dreamlayersFrom what I read, it seems they're the same hardware. One just has the additional radio module.
19:25:16amiconnSo the question is whether we want to go that route, and also whether the flash image should optionally (still) support ucl compressed rockbox and/or OF
19:25:45amiconnYes. There are even early v2s which still have the radio chip, so using rockbox on those makes them fms
19:26:25amiconnThere's one other difference - the model id in the .ajz, so you can't boot the OF of a v2 on an fm and vice versa
19:28:51 Join ChickeNES-laptop [0] (~ChickeNES@
19:29:01dreamlayersIf Rockbox is normally loaded from disk, removal of ucl support doesn't seem very beneficial. There's no need to save flash space then.
19:29:48*amiconn thinks there's a contradiction in that sentence
19:31:41amiconnIf there's no need to save space, why should we remove ucl support?
19:32:21dreamlayersI thought I said that.
19:32:38 Join metaphys [0] (
19:32:49amiconnErr, why *shouldn't* we remove it?
19:34:12amiconnIf we want an (optional) OF in flash (perhaps with automatic fallback to that OF if bootbox doesn't find rockbox) we might still need it. I need to check how large the varoious OF versions are, and whether they fit into the 256K ROM together with bootbox
19:34:15metaphysHey pamaury did you recieved the same mail as i did :D ?
19:35:34metaphyswell did you all recieved the same mail as I did? :D
19:37:00dreamlayers-ffunction-sections and -Wl,−−gc-sections saved 2048 bytes, and it still works on my v2 Recorder.
19:37:05metaphys"This might be a bit unorthodox, but my son Thomas (who is seven) insisted that I thank you for your work on the Rockbox port for the Fuze+. He got a Fuze+ for his birthday in June, and we have been eagerly awaiting the Rockbox port for it. He has ADHD, and we have found that listening to music makes it much easier for him to focus on his homework, which is part of the reason why he got a MP3 player for his birthday. But he wanted a por
19:37:07metaphystable game player as well, which is why Rockbox will be perfect for him. When I explained to him how Rockbox is developed and the port for the Fuze+ specifically, he was completely entranced. He and I thank you and all the other developers for your efforts, and want you to know that it is greatly appreciated." how nice is that !!! :D
19:38:14 Join japc [0] (~japc@
19:38:48 Join hilbert [0] (
19:41:51 Quit ChickeNES-laptop (Quit: Computer has gone to sleep.)
19:44:26 Quit WalkGood (Quit: CaminaBien ♪ ♫ ♪ ♫)
19:47:31 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:50:19 Quit japc (Ping timeout: 244 seconds)
19:52:40 Part metaphys
19:53:38 Join stripwax__ [0] (
19:58:22 Quit stripwax__ (Read error: Connection reset by peer)
20:00:17 Quit mortalis (Quit: KVIrc 4.1.1 Equilibrium
20:05:19 Join metaphys [0] (
20:05:26 Part metaphys
20:06:35gevaertsI assume neither a bootloader nor a new compiler are things that can realistically be done within a few days, so if we want the red gone from the build table we should probably go for a short term fix
20:08:44funmani.e. remove the recorder from the build network? :)
20:10:09gevaertsEither disable the 8MB recorder build for now or add an ifdef condition for it to disable dircache, yes
20:10:16gevaertsUnless someone has other ideas
20:10:36*gevaerts points out again that it's the *8MB* recorder build, not the regular
20:11:43amiconndreamlayers: Btw, if -ffunction-sections saves binsize for main builds, this is a sign of missing #ifdef's, or entirely unused functions which should be removed
20:11:54n1sdoes either of the two recorder users have the 8MB mod?
20:12:15gevaertsn1s: I believe there are at least three recorder users. Two is the ondio ;)
20:12:33n1sah, i mix them up
20:12:55amiconnWe don't know how many users there are, because most of the remaining archos users probably don't upgrade often
20:13:14*gevaerts nods
20:13:25n1sif they don't upgrade they won't be affected by whatever we do
20:13:49amiconnI didn't say they don't upgrade - they don't upgrade often. Probably just releases
20:13:50dreamlayersamiconn: So, you feel that every function correctly removed via -ffunction-sections is something that should be fixed via #ifdefs?
20:14:36n1sanyway, if we do something "short term" now it'll likely be permanent unless someone actually intends to work on a better solution
20:14:37amiconndreamlayers: Almost every, yes. Bootloaders are another matter
20:15:01gevaertsn1s: short term in this case buys us a few months at most
20:15:07amiconndreamlayers: Did you try gcc 4.6.x yet?
20:15:33 Quit wodz (Ping timeout: 260 seconds)
20:15:43n1sgevaerts: what happens in a few months?
20:16:03gevaertsn1s: the main recorder build is within a few kilobytes of the limit now
20:16:32n1sah yes
20:16:53dreamlayersamiconn: I've used gcc 4.6.2 to compile Rockbox without errors, but I didn't get any working binaries. It might have been partly due to binutils 2.21.1, which I used for linker plugin support.
20:17:13amiconnDefine non-working
20:17:38amiconnAlso, did you check whether the patch for jumptable padding is still necessary?
20:18:10amiconnAnd did you compare binsizes?
20:18:13dreamlayersI didn't even get to see the logo.
20:18:21amiconnSo it hangs?
20:20:18 Join japc [0] (
20:21:13dreamlayers4.6.2 increased v2 Recorder binsize unless paired with all the -flto related stuff. Then, the reduction wasn't huge, similar to -ffunction-sections.
20:21:16 Join wodz [0] (
20:22:19dreamlayersWhen I used -ffunction-sections, functions were word aligned.
20:23:51 Join Keripo [0] (
20:24:57amiconnWith 4.0.x?
20:27:03dreamlayersYeah, the standard Cygwin sh toolchain.
20:28:07 Join lorenzo92 [0] (
20:28:08amiconnHmm, weird. Where did I observe that 16 byte alignment then?
20:28:38 Quit lorenzo92 (Client Quit)
20:28:41amiconnAnyway, 4 byte alignment for each and every function still wastes a bit of space (1 byte per function on average)
20:33:57dreamlayersIt's better than wasting more space with unused functions, but I understand you prefer excluding them via the preprocessor instead.
20:34:36dreamlayersFor example, here's a list of functions excluded on the 5G iPod
20:36:43dreamlayers-ffunction-sections helps a bit, but not enough: firmware image is 206178 bytes while max size is 204800
20:41:24 Join bug2000 [0] (~bug@unaffiliated/bug2000)
20:41:47bug2000That "Branched" part got me confused on the website making me think it's out.
20:42:35bug2000[On the site that is. Noticing the 3.10 and the SVN ontop of it. Made me think 3.10 is out or being pushed except no other pages for what I've seen were editted to fit]
20:43:53bug2000"From now on, many RAM can be saved by allocating memory in correct (instead of worst-case) sizes " - Should it not be much?
20:44:07TheSevendefinitely much
20:45:35bug2000TheSeven, Will you change it? Since I am not going to register with real name to the Wiki on the next 30 seconds?
20:45:52TheSevenwhere exactly is that?
20:45:58bug2000Writing as you ask.
20:46:05bug2000 -> Major changes in version 3.10 -> 2011-08-30
20:47:03bug2000Thank you very much.
20:47:17*bug2000 thinks some of the stuff on the changelog was written in a rush.
20:47:43bug2000TheSeven, You can read the entire thing, I think there's stuff like `a` missing or stuff that could be better said.
20:47:58bug2000But I guess they don't bother as much as this many / much.
20:48:17bug2000[Which mostly bothers by just not sounding right]
20:53:06 Join caramou [0] (
20:54:06 Join promyloph [0] (~promy@unaffiliated/promyloph)
20:56:54promylophI’m getting “ROLO error: file too big” on my Sansa Clip+ when trying to load r31001 from r30720. rockbox.sansa is 545644 bytes large. Is this supposed to happen?
20:57:42n1sprobably not
20:59:28promylophWill rebooting help or brick the device?
20:59:48***Saving seen data "./dancer.seen"
20:59:48n1si'd say no :)
20:59:58gevaertsHow would that brick the thing?
21:00:29 Quit Keripo (Quit: Leaving.)
21:00:42*TheSeven wonders if he should try a final attempt at sorting out that ipod classic bootloader stuff
21:00:47n1sthat buffer is gotten with core_alloc_maximum which probably fails under some circumstances
21:01:10promylophNo idea, gevaerts. That’s why I’m asking. Better safe than sorry.
21:01:11n1sTheSeven: what stuff?
21:01:52TheSevenwell mostly figuring out the signing and the update protocol to finally build a rockbox bootloader that doesn't touch the boot flash
21:01:54n1spromyloph: i'd guess kugel is the one to look at that bug
21:02:24gevaertspromyloph: ROLO doesn't touch the bootloader or the filesystem
21:02:33n1sTheSeven: as in easier install?
21:03:15n1sthat would be nice
21:03:44TheSevennot only easier install, probably also finally dualboot
21:03:48promylophYep, gevaerts. But I’m not sure what happens when I reboot into r31001 now.
21:03:52n1salso nice
21:03:52TheSevenand if we're incredibly lucky lower power consumption
21:04:12TheSeven(in case some of our HW setup code is the culprit)
21:04:16n1spromyloph: rolo is separate from the bootloader
21:04:35promylophAh, ok. Thanks.
21:04:42n1sTheSeven: tripple nice!
21:04:50TheSevenpromyloph: you basically found a rolo bug in r30720
21:05:06promylophNot very impressive, huh? ;)
21:05:20TheSeventhe fact that this didn't boot doesn't mean that's anything wrong with r31001
21:05:29 Quit hilbert (Quit: Textual IRC Client:
21:05:32TheSevenand even if you install a broken build, you should be able to boot the OF and fix it from there
21:05:47promylophYeah, loading r31001 from itself works.
21:05:56promylophMust have been fixed in the meantime.
21:06:16TheSevenor just some intermittent problem
21:06:29TheSeventhe kind we like most :)
21:08:41*bug2000 thinks Rockboy is awesome. Letting you hide a gameboy disguised as a music player.
21:09:02TheSevenbug2000: try that with an ipod :)
21:09:25*bug2000 laughs out loudly.
21:09:27TheSevena clickwheel is kind of the least suited input device for gameboy games that was ever invented
21:09:32promylophTheSeven: I’ve seen the OF puts a lot of stuff in the flash (database files, a few hundred MB). Will that work even if flash is (almost) full?
21:10:21TheSevenbug2000: a gameboy hidden in an ipod case would definitely work better, if you just need it to be disguised :)
21:10:58TheSevenpromyloph: I've heard reports of trouble with this, but I don't have such a device myself, so I can't really answer that
21:10:59bug2000TheSeven, Yeah, I know. I tried playing on Sansa Fuze. A couple of times I got the wheel to move rather than clicking the correct buttons.
21:11:38bug2000TheSeven, It would be even better if I knew how to compile the bootloader for Rockbox so that Rockbox would run in `Hidden` mode.
21:12:23TheSevenhidden mode in terms of the OF being the default boot option?
21:12:35 Quit japc (Ping timeout: 240 seconds)
21:12:36bug2000TheSeven, Yeah.
21:12:42promylophOk, thanks everyone! Keep up the good work!
21:12:50 Quit promyloph (Quit: WeeChat 0.3.5)
21:12:53bug2000^ Ditto.
21:13:03TheSevenwell, that's probably a two-liner patch, but a rather risky one (it touches the code that's most likely to brick your device)
21:13:24bug2000TheSeven, I know. That's why I didn't even try it. Because I didn't even know where to start with it.
21:13:39bug2000Well I did copy the reposetory but noticed I can't figure where it is in the code.
21:14:00TheSevenin some assembler file
21:14:31bug2000That might be the reason. I only looked at the c files.
21:15:24TheSevenoh, it's even worse
21:15:27TheSevenit's somewhere in here:
21:16:06bug2000How do you even read that?!
21:16:27TheSevenyou just don't read that :)
21:16:39bug2000TheSeven, So how did someone write it in the first place?
21:16:50TheSevenit's a compiled version of this:
21:17:45bug2000That is by far more readable.
21:18:03bug2000I know which lines to edit and how thanks to you. But I fear of breaking too much.
21:18:16*bug2000 thinks comments are awesome.
21:18:39TheSevenyou should let someone who knows what he's doing do it for you, and test it on his own device first
21:20:09bug2000Anyway thanks for your time.
21:20:23bug2000Thanks for everyone in here [here or not here] for their time.
21:20:27bug2000Rockbox is really awesome.
21:20:37TheSeventhanks :)
21:22:29 Join liar [0] (
21:23:31 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
21:25:30 Join Xerion [0] (
21:29:57dreamlayersWhile looking at PP502x cache stuff, I learned how to use the per-core interrupt vectors in the cache. So far, I have code that works, except that loading via rolo from a patched copy causes an immediate crash.
21:34:00dreamlayersWho knows the most about PP502x?
21:39:58 Quit perrikwp_ (Read error: Connection reset by peer)
21:41:07 Join perrikwp [0] (
21:45:06 Join Transformer [0] (
21:47:12 Part Transformer
21:49:16zaytsevpamaury: I managed to connect my aquabeat 2 in firmware recovery mode... it's 0402:7103... you have to press < and hold and the insert the cable.
21:49:35zaytsevpamaury: dmesg says [ 5855.115975] scsi 9:0:0:0: Direct-Access USB 2.0 Boot Loader PQ: 0 ANSI: 0 CCS
21:53:17 Quit Stummi (Quit: Bye!)
22:02:17zaytsevvid is like acer labs.
22:06:29zaytsevand in the old aquabeat 1 (which is not what I have) firmware update package there are files called and this sounds familiar to me.
22:07:11 Join Keripo [0] (
22:14:04 Quit Keripo (Ping timeout: 260 seconds)
22:17:47 Quit liar (Read error: Connection timed out)
22:18:45 Quit benedikt93 (Quit: Bye ;))
22:18:50 Join saratoga [0] (9803c31c@gateway/web/freenode/ip.
22:18:57saratogazaytsev: open up the player and see what CPU it uses
22:20:09zaytsevsaratoga: that's a big of a problem. it's one of the few waterproof models, which I bought exactly for that reason =( I'd like to refrain from opening it, although I understand that it leaves little chance..
22:21:29saratogawell its some obscure mp3 player no one has heard of, so probably you'll have to do everything yourself anyway
22:21:38saratogamight as well get a few so you have spares when you break them
22:22:32saratogaany of the application people want to look at the new #12348?
22:22:42saratogaFS #12348
22:22:42fs-bluebot [New Port] Samsung YP-R0: the first patches (patches, unconfirmed)
22:30:24 Quit Thra11 (Ping timeout: 260 seconds)
22:32:37 Quit dreamlayers (Quit: dreamlayers)
22:33:24 Join Keripo [0] (
22:46:57 Quit n1s (Quit: Ex-Chat)
22:53:09 Join lovasoa [0] (~olojkine@
22:53:25 Quit Keripo (Quit: Leaving.)
22:54:48 Join Strife89 [0] (~Strife89@
22:55:55 Quit Xerion (Quit: )
22:56:06 Join Xerion [0] (
22:57:06 Join Keripo [0] (
22:59:52***Saving seen data "./dancer.seen"
23:00:37 Quit TheLemonMan (Quit: WeeChat 0.3.6)
23:02:13pamauryzaytsev: if it's a sb file there are two or three possibilities: stmp35xx, stmp36xx and stmp37xx. But as you said it's only for the v1, for the v2 you have no info and this particular vid:pid doesn't help. Since the recovery mode is a mass storage, can you read from it ?
23:04:53zaytsevpamaury: sudo dd if=/dev/sdb of=test-bin bs=1M count=1 -> dd: opening `/dev/sdb': No medium found :-(
23:06:05pamaurythat's unfortunate
23:10:04lovasoain apps/plugins/rockblox.c, on line 904, i think there is a file descriptor leak
23:11:15gevaertslovasoa: how?
23:15:01lovasoa*fd is opened, and never closed
23:15:52gevaertsIt's closed on line 903
23:16:30 Join metaphys [0] (
23:17:27 Join Thra11 [0] (~thrall@
23:17:49lovasoaMaybe we don't have the same line numbers
23:18:02lovasoaI just applied a big patch
23:18:53lovasoaI'm talking about void load_game(void)
23:18:53gevaertsAh, possible
23:19:03*gevaerts looks
23:19:23 Quit advcomp2019 (Read error: Connection reset by peer)
23:19:24gevaertshm, indeed
23:19:29gevaertsThat return looks wrong
23:19:59 Quit Keripo (Quit: Leaving.)
23:20:14 Join advcomp2019 [0] (
23:20:14 Quit advcomp2019 (Changing host)
23:20:14 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
23:22:27 Join Scromple [0] (
23:25:11CIA-22New commit by gevaerts (r31002): Fix file descriptor leak. Thanks to Ophir Lojkine for spotting.
23:25:31lovasoaThank you :)
23:26:02gevaertsThank you for finding it
23:27:32CIA-22r31002 build result: 2 errors, 0 warnings (gevaerts committed)
23:37:37 Quit advcomp2019 (Read error: Connection reset by peer)
23:38:01 Join advcomp2019 [0] (
23:38:01 Quit advcomp2019 (Changing host)
23:38:01 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
23:47:08 Quit advcomp2019 (Read error: Connection reset by peer)
23:47:35 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
23:52:38 Join low_light [0] (
23:53:00 Quit Strife89 (Ping timeout: 260 seconds)
23:54:26zaytsevpamaury: thanks anyway :-/ maybe when the firmware upgrade comes out I'll be able to investigate... I'm not rich enough to buy a couple of those for parts ATM
23:54:32low_lightpamaury: no luck with the rom dump
23:54:54pamaurylow_light: what happened ? reboot ?
23:55:09 Join robin0800 [0] (
23:55:48 Quit advcomp2019 (Read error: Connection reset by peer)
23:55:54low_lightpamaury: the recovery file loads ok, but the subsequent read reboots the player.
23:56:09pamauryit means the address is not good
23:56:11 Join advcomp2019 [0] (
23:56:11 Quit advcomp2019 (Changing host)
23:56:11 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
23:56:36pamaurycan you try a different combination ? like 0xffff0000 but only read 1024 bytes
23:56:42pamaurythen try 0xc0000000

Previous day | Next day