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

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

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

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

#rockbox log for 2010-04-07

00:01:05stripwaxsaratoga - I'm going to try a clean complete rebuild of tremor on my system; but I still think I'm getting some strange behaviour where an unpatched version works ok, but the patched version results in an app complaining about that LIKELY symbol at load link time, even when source tree doesn't contain the LIKELY symbol. very strange indeed.
00:01:45 Quit jgarvey (Quit: Leaving)
00:05:47 Quit robin0800 (Ping timeout: 240 seconds)
00:06:12 Join robin0800 [0] (~robin0800@
00:07:30 Quit JohannesSM64 (Ping timeout: 246 seconds)
00:12:29saratoga_labstripwax: once you're sure the benchmark results are correct, would you send a mail to the tremor list
00:12:37saratoga_labwith the results
00:12:38stripwaxwill do
00:13:41stripwaxsaratoga - is the patch supposed to have fft-ffmpeg.lo in it?
00:13:57saratoga_labstripwax: IIRC someone said that should be removed
00:14:03saratoga_labbut i haven't had a chance to remove it
00:15:08stripwaxI wonder if that's been contributing to my problems in some way
00:16:17 Quit robin0800 (Remote host closed the connection)
00:16:42 Join robin0800 [0] (~robin0800@
00:17:21stripwaxalso patch seems to make fft.h executable; probably wrong.
00:17:56stripwaxactually loads of files.
00:26:25 Join JohannesSM64 [0] (
00:28:22stripwaxsaratoga - really weird. mpd definitely does not like the patched version of tremor, but plays ok on the unpatched trunk version; clean rebuild. Unfortunately I'm not getting a meaningful error out of mpd any more so a bit tricky to see what's going on ..
00:28:33 Quit CGL (Quit: Saliendo)
00:30:20stripwaxah-ha. dmesg is telling me there's an alignment trap.
00:31:19saratoga_labcan you get an address out of it?
00:32:26 Quit dfkt (Ping timeout: 265 seconds)
00:32:49 Quit domonoky (Read error: Connection reset by peer)
00:35:44stripwaxyep, but not quite worked out how to make use of it in an appropriate way. Address is 0x000002e2 ... fwiw ....
00:37:48saratoga_labi guess its not clear what binary to objdump?
00:38:19stripwaxit's an LDR instruction of some sort however; so that's definitely an unaligned read. with a bit of luck I can coax gdb to do the next bit ..
00:38:25 Join yegor [0] (
00:41:22 Join xiainx [0] (
00:41:48saratoga_labcould the bitrev table somehow be stored unaligned?
00:41:48 Quit mt (Remote host closed the connection)
00:42:15saratoga_labno i guess not since its an integer number of bytes long and full of int16s
00:42:21saratoga_labsorry even number of bytes
00:43:12 Quit yegor (Client Quit)
00:43:20 Join yegord [0] (~yegor@2001:67c:7c:40d4:224:7eff:fedf:124d)
00:48:46 Join chrism [0] (
00:48:59 Quit xiainx (Ping timeout: 276 seconds)
00:49:08CIA-5New commit by 03Blue_Dude (r25503): Another small bookmark.c revision, no functional change, saves bin size
00:54:10 Quit evilnick (Quit: Leaving)
00:56:15 Part froggyman
01:03:37***Saving seen data "./dancer.seen"
01:04:12 Quit geertvdijk (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
01:07:14 Join planetbeing_ [0] (
01:07:20CIA-5New commit by 03torne (r25504): Fix typo in a bootchart message
01:12:35 Quit DataGhost (Ping timeout: 260 seconds)
01:15:29 Quit ender` (Quit: Man created gods. The opposite remains to be proved.-- Serge Gainsbourg)
01:22:41 Quit Luca_S (Quit: CGI:IRC)
01:28:15 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
01:28:26 Join Adubbb [0] (~Aldubuc@
01:28:27 Quit robin0800 (Ping timeout: 260 seconds)
01:28:51 Part toffe82
01:32:11 Quit komputes (Ping timeout: 260 seconds)
01:43:09stripwaxsaratoga - hm, I'm pretty much out of ideas. it looks like it could be any number of things. weird, though, that the example works, but an actual application does not. (and i've confirmed I get no alignment traps on stock trunk tremor)
01:43:15 Join aaron_ [0] (
01:43:45stripwaxov_read, surprisingly, has some dodgy code in it (casting a char* buffer to short*) - but that would appear to work (since trunk tremor has the same code)
01:44:13stripwaxI guess because the app that creates the buffer actually creates it aligned to short* boundaries (or better) in the first place
01:44:38stripwax(but that's a sloppy assumption on tremor's part imho!)
01:44:59 Quit stripwax (Quit:
01:45:08 Nick aaron_ is now known as oxymis (
01:46:14 Quit adnyxo (Ping timeout: 269 seconds)
01:47:18 Nick oxymis is now known as adnyxo (
01:52:39 Join komputes [0] (~komputes@ubuntu/member/komputes)
01:53:35 Join Sergio [0] (
01:53:50 Join krabador [0] (
01:55:34 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
01:57:14 Quit mikroflops (Ping timeout: 276 seconds)
01:57:28 Quit Surge97 (Ping timeout: 264 seconds)
01:58:53 Quit krabador (Quit: Sto andando via)
01:58:58 Join mikroflops [0] (
01:59:45 Join RadicalR [0] (
02:03:07 Quit mikroflops (Remote host closed the connection)
02:03:44 Join mikroflops [0] (
02:11:42 Join CaptainKewl [0] (
02:18:07 Quit mikroflops (Ping timeout: 252 seconds)
02:21:14 Join mikroflops [0] (
02:30:11 Join krazykit` [0] (
02:33:31 Quit krazykit (Ping timeout: 252 seconds)
02:37:08 Join fyrestorm [0] (
02:40:28 Nick fxb is now known as fxb__ (
02:40:49 Join n1s [0] (~n1s@rockbox/developer/n1s)
02:42:54 Join Boldfilter [0] (
03:02:01 Quit Adubbb (Ping timeout: 264 seconds)
03:02:14 Quit Adubb (Ping timeout: 260 seconds)
03:03:40***Saving seen data "./dancer.seen"
03:04:54 Join Adubb [0] (~Aldubuc@
03:07:59 Join Adubbb [0] (~aldubuc@
03:09:32 Join xiainx [0] (xiainx@wpa106034.Wireless.McGill.CA)
03:12:24 Quit komputes (Ping timeout: 265 seconds)
03:21:45 Quit n1s (Ping timeout: 240 seconds)
03:26:59 Quit adnyxo (Quit: Leaving)
03:36:38 Quit Adubb ()
03:37:44 Join MadRat- [0] (
03:40:43MadRat-I have a G5, 30GB iPod (color LCD, search feature, video playback, etc) and I just installed Rockbox. When I try to play MP4s encoded with the Nero encoder using the HE profile (v1 or v2) I get moment of playback, a pause then another moment of playback.
03:40:58MadRat-I didn't see anything in the documentation about it but I assume that's normal?
03:45:38 Join n1s [0] (~n1s@rockbox/developer/n1s)
03:48:41 Join Strife1989 [0] (
03:49:05MadRat-Oh yeah, I also noticed I can't play OGG files with more than 2 channels.
03:52:19saratoga_labMadRat-: yeah, the CPU on your player probably isn't fast enough for AAC-HE
03:52:28saratoga_labat least not without a lot more optimization
03:52:42n1sMadRat-: do multichannel oggs not play at all or only in 2 channels
03:54:21MadRat-n1s: That's pretty much what I figured. The OGG files will sit there for a bit then skip to the next file. I can try to DCC send an example if you like.
03:55:16n1swe have limited our decoder to 2 channels intentionally, but how it should handle multi channel files i don't know
03:55:28saratoga_labdoes Tremor handle multichannel vorbis?
03:55:46n1ssaratoga: i think it can but we limit it to 2 channels
03:55:47 Join fejfighter [0] (
03:56:04n1smainly so we can know how big the buffers are
03:58:16MadRat-n1s: No I was testing the limitations of Rockbox and playing multi channel OGG files doesn't seem to work.
03:58:45MadRat-I'm really impressed by the midi and mod playback
03:58:54n1sok, i don't know if that is intentional but i would guess so
03:59:14n1smaybe you could attach the file to a bug report?
03:59:30MadRat-Sure if that's possible. Where do I go to do that?
04:01:08 Quit planetbeing_ (Quit: planetbeing_)
04:01:20saratoga_labi don't have grep handy, but google doesn't show anything suggesting tremor supports 5.1
04:02:48n1ssaratoga_lab: i've gtg but i thought i have seen something suggesting it does support multichannel
04:07:13saratoga_labthe code seems like it should work with 5.1, at least skimming
04:07:23saratoga_labit maybe that the mallocs just fail due to lack of RAM
04:07:39MadRat-Could it be an iPod issue?
04:08:02MadRat-<−−- lousy programmer
04:11:33 Quit n1s (Ping timeout: 240 seconds)
04:16:40 Quit TheSeven (Disconnected by services)
04:16:53 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:17:04 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:26:08 Join froggyman [0] (~me@unaffiliated/froggyman)
04:32:01 Join Rob2223 [0] (
04:35:29 Quit Rob2222 (Ping timeout: 245 seconds)
04:37:43 Quit xiainx (Ping timeout: 240 seconds)
04:37:48 Part froggyman
04:38:30 Join S_a_i_n_t [0] (S_a_i_n_t@
04:44:58 Join xiainx [0] (
04:45:33 Join Barahir [0] (
04:47:45 Quit fejfighter (Ping timeout: 264 seconds)
04:49:36 Quit Barahir_ (Ping timeout: 276 seconds)
05:03:03 Join anewuser [0] (anewuser@unaffiliated/anewuser)
05:03:42***Saving seen data "./dancer.seen"
05:05:01 Quit Strife1989 (Quit: Logging off the desktop.)
05:15:48 Join CGL [0] (~CGL@
05:20:39 Quit Strife89 (Quit: Gone sleepin)
05:22:46 Quit RadicalR (Read error: Connection reset by peer)
05:27:25CIA-5New commit by 03ranma (r25505): Remove vim option line
05:39:30 Quit Horscht (Quit: Verlassend)
05:43:51CIA-5New commit by 03jethead71 (r25506): Gigabeat S (imx31): Begin voltage and frequency scaling code. For now, to avoid overdrive voltage, just lower core voltage to 1.35V since voltage ...
05:44:36 Quit anewuser (Quit: Like renoise + like music? 3 days to submit your entry!)
05:44:55 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
05:57:02 Quit jordan` (Read error: Connection reset by peer)
06:09:08 Join jordan` [0] (~jordan@
06:12:56 Quit CaptainKewl (Ping timeout: 264 seconds)
06:16:20 Join S_a_i_n_t_ [0] (S_a_i_n_t@
06:17:06 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
06:19:32 Quit S_a_i_n_t (Ping timeout: 264 seconds)
06:26:36 Quit togetic (Quit: WeeChat 0.3.0)
06:30:33 Part S_a_i_n_t_
06:31:21 Join S_a_i_n_t [0] (S_a_i_n_t@
06:36:49S_a_i_n_tIs there a list of things I can bind to "Hotkey" someplace? The only thing I can find is in the manual just defines what hotkey is, and not how to use it.
06:37:36S_a_i_n_tThe impression I get at the moment is that its just broken, as most of the things I've tried setting to the hotkey just end in a data abort.
06:37:50MadRat-S_a_i_n_t: I have no idea, but just out of curiosity what kind of player do you have? Mine is a gen 5 iPod
06:38:07S_a_i_n_tNano1/2 gen
06:38:19MadRat-how's your battery life?
06:38:46MadRat-heh I get about 10 or 12 hours out of mine
06:38:47S_a_i_n_tBetter than the OF can give me ;)
06:39:14MadRat-Sorry I can't personally help, I just installed last night.
06:40:36S_a_i_n_tNo worries, man. Someone will see it at some point.
06:41:47 Quit JohannesSM64 (Quit: WeeChat 0.3.2-dev)
06:43:33LloreanS_a_i_n_t: Have you posted a bug report on the data aborts?
06:43:58 Join JohannesSM64 [0] (
06:45:32S_a_i_n_tLlorean: Not yet, just found it then.
06:47:36S_a_i_n_tI wanted to clear up if I was doing something wrong...I did find the actual bit in the manual that explains how to set a hotkey just now, and it turns out I bluffed my way through it, and I was doing it right but so-far all the things I've tried assingning as "hotkey" end in data abort or just don't set.
06:47:54LloreanA data abort should never happen, even if you're doing something wrong.
06:48:33S_a_i_n_tSweet, I'll write down the exact address of the data abort and post a bus in Flyspray now then.
06:48:33LloreanIf you can reproduce one with a current build, you should check if there's a bug report already, and if not, file one with instructions how, even if "how" is doing something stupid or weird.
06:49:30S_a_i_n_tThat's what I wanted to clear up really, whether or not my "how" was correct, or stupid and weird.
06:50:53 Join MadRat [0] (
06:51:01 Quit MadRat- (Disconnected by services)
06:51:02 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
06:51:20 Nick MadRat is now known as MadRat- (
06:53:05S_a_i_n_thehe...just noticed when logging into Flyspray that if you mess up your password you get "error #7 paassword failed" ;P
06:54:07 Quit CGL (Quit: Saliendo)
06:55:19 Join lpereira [0] (
07:03:43***Saving seen data "./dancer.seen"
07:12:10 Quit Boldfilter (Quit: Boldfilter)
07:13:14 Nick fxb__ is now known as fxb (
07:23:25JdGordonS_a_i_n_t: are you getting data aborts with your build? the exact number isnt really useful unless you attach your file to the report
07:24:01S_a_i_n_tyep, and Oh...didn't realize.
07:24:18S_a_i_n_tWelcome back btw \0/
07:24:34JdGordonwe annoyingly dont archive that file with builds so data aborts are annoying to see in flyspray :/
07:25:13S_a_i_n_tBlargh...I see why now.
07:25:25 Quit Adubbb (Read error: Connection reset by peer)
07:26:20 Quit JohannesSM64 (Quit: WeeChat 0.3.2-dev)
07:26:41S_a_i_n_tIn the tree context menu, I just can't set a hotkey, and in the wps context everything fails with a data abort :/
07:27:20JdGordonI have no idea how hotkeys work
07:27:51 Join JohannesSM64 [0] (
07:28:43 Join Tux2 [0] (
07:28:55S_a_i_n_tNeither did I until last night/now...but it turns out I bluffed my way through it, and got it just doesn't seem to work.
07:29:20S_a_i_n_tI also think that menu+play is a horrible key combination to use...but that's just me I guess.
07:29:44JdGordonwell, ipods dont have spare buttons, so probably shouldnt include hotkey anyway
07:29:54S_a_i_n_tThe whole thing went in too quick IMO
07:30:05*JdGordon refuses to comment
07:30:24S_a_i_n_tNuff sain ;)
07:30:31S_a_i_n_tdammit *said
07:30:36 Quit xiainx (Ping timeout: 260 seconds)
07:35:09 Join xiainx [0] (
07:42:34 Join Adubb [0] (~aldubuc@
07:44:10 Quit JohannesSM64 (Quit: WeeChat 0.3.2-dev)
07:44:34 Join JohannesSM64 [0] (
07:45:22 Quit kramer3d (Quit: Leaving)
07:47:15CIA-5New commit by 03FlynDice (r25507): sd-as3525v2: Set SD FIFO TX/RX watermarks once in init_controller() and leave them alone after that. ...
07:48:31Tux2Hey guys, I just bought a Sansa Fuze v2 and I would love to help you guys, I just have one question: should I update the firmware on my Fuze before I upload the roxbox firmware?
07:48:59saratoga_labthe installation instructions for the Fuze involve updating the firmware
07:49:05saratoga_labjust read through them it should be clear
07:49:31 Quit xiainx (Ping timeout: 264 seconds)
07:49:46 Join mikroflops_ [0] (
07:50:04Tux2Ah, okay, I was reading the steps in the forum about if you want to test for the fuze and didn't see that. :) I'll go to the other page and read that. Thanks. :D
07:50:13FlynDiceTux2: You don't need to, just use the latest OF firmware when you run mkamsboot and that should update the OF for you
07:51:11saratoga_labwe have instructions on the forums?
07:53:13Tux2 Page 101
07:53:35Tux2Are these the instructions I should be following here:
07:54:01 Quit mikroflops (Ping timeout: 276 seconds)
07:54:24saratoga_labthats just some test build, you probably shouldn't be installing that now
07:54:49 Join xiainx [0] (
07:55:22Tux2Ah, I see.
07:55:48saratoga_labmade it more clear not to actually follow those
07:57:01 Quit arbingordon (Quit: `)
07:57:23FlynDiceTux2: That second link you posted was right
08:01:36FlynDiceTux2: The process is the same with all the AMS Sansas but since the Fuzev2 is still in the unusable category you'll have to build your own bootloader and main.
08:03:40pixelmaS_a_i_n_t: isn't Menu and Play the top and bottom keys on the clickwheel?
08:04:05 Quit xiainx (Ping timeout: 258 seconds)
08:04:11saratoga_labwe have binaries for the main build actually:
08:04:31saratoga_labbut we should probably think about putting up an official bootloader, I think all the issues with backlight are now fixed and it should be ready to go
08:04:55S_a_i_n_tpixelma: did I say menu+play? If I did, sorry...I meant select+play, but yes, menu and play are at the top and bottom respectively.
08:05:15S_a_i_n_tAha...I did say that...shit, sorry.
08:06:53S_a_i_n_tAhem... " "select+play" is an awful key combo for the hotkey, though I don't see one that would work any better."
08:08:13Tux2Okay, so, just to make sure, what exactly do I need to do? I know I have FlynDice and saratoga_lab throwing things at me, but, for you guys to test your builds on my device, what do I need to do?
08:08:32saratoga_labfollow the instructions
08:08:55Tux2Okay, reading yet again. :D
08:09:41 Join xiainx [0] (
08:11:02FlynDiceTux2: The install instructions are right after the unlikely heading.... "Installation: " ;-)
08:11:03Tux2Never dealt with cross compiling yet...
08:11:15 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
08:11:40 Join tanbo [0] (
08:12:15Tux2I would have to say that I am not an experienced C programmer... but that I do know enough to compile programs for myself...
08:13:00Tux2According to the instructions I assume that I have an ARM CPU in this Fuze, right?
08:13:16FlynDiceTux2: Correct
08:13:48Tux2Perfect. I'll get back with you guys when I get this thing compiled...
08:16:45MadRat-Does Rockbox have floor 1 playback for OGG files?
08:17:12saratoga_labMadRat-: yes, basically all vorbis files are floor1
08:17:46saratoga_labfloor0 hasn't been used in almost a decade
08:17:53MadRat-I have two OGG files that have 6 channels but won't play on my iPod G5
08:19:00saratoga_labare they different then the ones you mentioned before?
08:19:20MadRat-Would you like to see a mediainfo report?
08:19:27saratoga_labmy random guess would be they run out of memory and the codec quits when a malloc fails
08:19:47saratoga_labif you compile the sim and add some print statements you can check
08:20:29MadRat-I wouldn't know how to do that I'm afraid.
08:20:48saratoga_labare there multichannel vorbis sample files online
08:22:55MadRat-Do you have a ULR?
08:23:33Tux2should I do the target arch as arm, arm-eabi, or both when I run the
08:23:47 Join ender` [0] (
08:24:15 Quit lpereira (Quit: Leaving.)
08:24:32MadRat-saratoga_lab: Do you have the URL of a mitichannel vorbis sample?
08:25:03saratoga_labi asked you if you had one . . .
08:26:10MadRat-I'm sorry. I have two files I use for testing.
08:26:21CIA-5New commit by 03jethead71 (r25508): Gigabeat S: Extend the upper temperature range for battery charging to 50C: OK and 45C: Resume. Currently extended disk activity can cause it to ...
08:26:29MadRat-I can dcc send them or post them if that helps
08:26:41 Quit xiainx (Ping timeout: 240 seconds)
08:27:02saratoga_labi was going to suggest finding or making a link and then filing a bug report
08:27:36saratoga_labnot sure if this is even really a bug, but its worth documenting
08:28:05MadRat-saratoga_lab: I agree, I'm not sure if it's a bug or limitation
08:30:43 Join xiainx [0] (
08:32:35 Quit tanbo (Quit: CGI:IRC (EOF))
08:32:40amiconnjhMikeS: Hmm, what would I have to do now if I want the beast to run at 528MHz?
08:34:16jhMikeSamiconn: Remove the PMIC write or just set it to 1.600V.
08:35:25jhMikeSAccording to the specs, it shouldn't run above 1.470V (iirc) except for exceptional circumstances (overdrive).
08:35:47 Join mitk [0] (~mitk@
08:36:38jhMikeSAt least it's looking promising by lowering the voltage, perhaps 30%-50% more runtime guessing from 120 minute samples, comparing from similar voltage difference on old benches.
08:37:58jhMikeSWe'll see though. I'll try to do a proper bench today with SVN.
08:39:38 Quit amiconn (Disconnected by services)
08:39:40 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
08:40:02 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
08:40:14 Quit pixelma (Disconnected by services)
08:40:16 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
08:40:36 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
08:42:34 Quit xiainx (Ping timeout: 264 seconds)
08:43:08 Quit Tux2 (Remote host closed the connection)
08:46:25 Join xiainx [0] (
08:46:44 Join flydutch [0] (
08:48:44MadRat-Here's the bug report:
08:49:39 Join Tux2 [0] (
08:53:19 Quit xiainx (Ping timeout: 246 seconds)
08:56:50 Join Bagder [0] (~daniel@rockbox/developer/bagder)
08:59:24 Join xiainx [0] (
09:03:45***Saving seen data "./dancer.seen"
09:05:14 Join petur [0] (~petur@rockbox/developer/petur)
09:07:01 Join lpereira [0] (
09:07:06 Join LinusN [0] (~linus@rockbox/developer/LinusN)
09:16:39 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
09:17:17 Join n1s [0] (~n1s@rockbox/developer/n1s)
09:17:53 Join xavieran [0] (
09:19:45pixelmaS_a_i_n_t: you could try a build before the hotkey addition. The same key combo was used then for a shortcut to the playlist viewer in the WPS, might be interesting (potentially be pointless as well though) to know whether it was broken back then already
09:19:51Tux2Just to let you guys know, your instructions don't work with the newest firmware update... It gives me this error: [ERR] Original firmware unknown, please try an other version. Tested Fuze versions are : 2.01.17, 2.02.26
09:21:12Tux2I downloaded firmware version 2.02.26 and re-did the mkamsboot and it worked.
09:21:23 Join tanbo [0] (
09:22:07tmzt_somebody has to test and add the md5/sha1 I think
09:22:09pixelmait's "fun" how the addition with the hotkey functions show that this (playlist viewer shortcut) was not well tested and probably unknown before. Shows something about not documenting things...
09:22:17 Quit xiainx (Ping timeout: 258 seconds)
09:22:43Tux2Ah, got it tmzt_
09:24:12Tux2Okay, here goes nothing... putting on the bootloader...
09:26:17JdGordonpixelma: it is very unlikely that the button combo causes the abort
09:26:26JdGordon*much* more likely that new code is broken
09:27:24Tux2Well, firmware upgrade went well... the device still boots!
09:28:18 Join xiainx [0] (
09:28:32 Quit mikroflops_ (Ping timeout: 276 seconds)
09:30:04pixelmaJdGordon: probably - and it is broken (or not working) on other targets as well but has been there before and only a very few people knew about it...
09:30:20 Join mikroflops [0] (
09:32:56JdGordonmenu+play does squat on my mini2g (pre hotkey patch)
09:33:00S_a_i_n_tpixelma: the default hotkey for the WPS works fine, you just can't set any new ones...
09:33:10JdGordonoh, select+play
09:33:13JdGordonthat works
09:33:19 Quit xiainx (Ping timeout: 264 seconds)
09:33:41S_a_i_n_ttrying to do so in the WPS ends in data abort, and trying to do so in the tree ends in....well, nothing.
09:34:10JdGordonopen up your file and find the closest number to the abort number
09:34:31pixelmasome targets have the hotkey only in the WPS, some in tree and WPS and some have neither
09:34:40S_a_i_n_twhere do I find said file?
09:34:50JdGordonin the build directory
09:35:58 Quit mikroflops (Remote host closed the connection)
09:37:10 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
09:38:18 Join mikroflops [0] (
09:38:34 Join xiainx [0] (
09:38:39 Quit kramer3d (Quit: Leaving)
09:38:57tanbohi, when i type make,the console display cc apps/action.c /home/Administrator/rockbox/apps/action.c:26;18: error: lang.h:No such file or directory make:***[/home/Administrator/rockbox/build/apps/action.o] Error 1 how can i manage this?
09:39:31JdGordonmake clean and try again
09:39:48tanbothank you. i`ll try
09:43:07Tux2Okay, just loaded rockbox on my fuze... :) :) I guess I am ready to test now.
09:43:35S_a_i_n_tHmmmm, not too sure what I'm supposed to be looking at. Tried searching for the address listed in the data abort, couldn't find it. The file is here:, the data abort address is: 0000000A (0), version is: 25486
09:43:57JdGordonhaha then something is totally fubar
09:44:10S_a_i_n_tWhy's that?
09:44:27JdGordonfunction pointer probably broken i tihnk
09:45:22 Quit xiainx (Ping timeout: 252 seconds)
09:48:21JdGordondo you ever see the "hotkey not set" splash?
09:49:00S_a_i_n_tIn the tree *something* flashes while it fails...but its too damn quick to see.
09:49:18S_a_i_n_tIn the wps, it just aborts immediately trying to set it.
09:49:28 Join xiainx [0] (
09:52:23*JdGordon suspects the settings
09:53:28JdGordonS_a_i_n_t: put this "splashf(HZ*5, "%d", this_hotkey);" in onplay.c line 1263 please
09:54:02S_a_i_n_tWill do.
09:55:00JdGordonyeah, the settings are a bit messed up
09:55:14tanboi just want to program A Hello world
09:55:45JdGordontanbo: still not working?
09:56:01JdGordonyou did run ../tools/configure right?
09:56:09tanbocan i have some guide? sorry i can not find on the wiki... it works!! i am so happy :-)
09:59:46tanbothank you! i have compile Rockbox complete now following the guide. now how can i write my first hello world program and compile and run it?
09:59:59*JdGordon wonders how much testing hotkeys got
10:00:06JdGordonI cant believe the settings actually work
10:00:13S_a_i_n_tsweeet F**K all...
10:00:28S_a_i_n_the was SOOOO damn pushy about getting it into SVN
10:00:47S_a_i_n_tmurdering keymaps for targets he's never even seen/heard of.
10:00:51*S_a_i_n_t grumbles...
10:03:38 Quit mikroflops (Ping timeout: 246 seconds)
10:03:55LloreanS_a_i_n_t: To be fair, I pushed him to try to implement keymaps for other targets, but I did keep trying to be clear that I meant targets that had a button free, like other targets with a recording button.
10:04:47S_a_i_n_tYes, I know..even I picked up on "tagets should have/need a free botton" and I'm a retard.
10:04:55S_a_i_n_tI should have been more clear.
10:06:01 Join m3dlg [0] (~m3dlg@
10:06:37JdGordonS_a_i_n_t: what was the flyspray number? sec.
10:07:16*JdGordon is sick of clicking "Save now" then finding it then goign back to drafts :p
10:07:44S_a_i_n_tJdGordon: FS #11181
10:07:58S_a_i_n_t*not* battery/charging (see comment) ;P
10:09:58saratoga_labtanbo: if you just want to write stuff to the screen, take a look at the plugins
10:10:03saratoga_labon the wiki there should be some info
10:12:10tanboi know .i just do not know how to do .according to the wiki,i should make a project and cd that dir and run ../tools/configure to make am i right?
10:12:33saratoga_labyes thats how you begin to compile
10:14:12tanboand it means that i can not just write a hello.c and call gcc hello.c to compile and link . Am I right?
10:15:28 Join mikroflops [0] (
10:15:30 Quit xiainx (Ping timeout: 240 seconds)
10:15:55 Join xiainx [0] (
10:18:37 Quit n1s (Ping timeout: 258 seconds)
10:19:09 Quit dockimble (Ping timeout: 264 seconds)
10:20:45linuxstbtanbo: Plugins are in apps/plugins/ and are compiled as part of a full Rockbox compile. See apps/plugins/hello_world.c for an example... You need to add new plugins to apps/plugins/SOURCES and apps/plugins/CATEGORIES
10:21:11 Quit xiainx (Ping timeout: 276 seconds)
10:22:32 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:24:50tanboThank you! I understand now...I just do not want to compile all rockbox each time....though it do not take long time. I know the correct way now~~ thank you all ~~ I am going to try now~~Good luck for me~~ :-)
10:25:37pamauryMy question will probaly sound stupid but how can I get the current audio volume ? In sound.h/audiohw.h, I see that I can get min/max/default and set current volume but can't get current one !
10:26:17linuxstbglobal_settings.volume ?
10:26:49 Quit m3dlg (Ping timeout: 240 seconds)
10:27:12niekieAfter building RockBox from source, what's the proper way to prepare for another rebuild (make clean seems to remove too much)
10:27:25linuxstbpamaury: Or do you need it in firmware/ code?
10:27:39pamauryit would be better in firmware/
10:27:52linuxstbniekie: You mean preparing for a different build (i.e. different target) ?
10:28:05S_a_i_n_tniekie: do you mean another build for a different target?
10:28:10S_a_i_n_tif not, do nothing.
10:28:17S_a_i_n_tjust build in the same directory
10:28:24niekieHm :)
10:28:39linuxstbpamaury: Then you probably can't. I can't think of a reason why the firmware code would need to know.
10:28:41niekieI didn't mean for a different target though. :)
10:28:56pamaurylinuxstb: for ubs audio
10:29:08linuxstbpamaury: I mean a _current_ reason.
10:29:15S_a_i_n_tSweet, then just build in the same dir, only the updated files will be built, and your compile will be a LOT faster. ;)
10:29:33 Join xiainx [0] (
10:30:27linuxstbpamaury: Then I think you would need to add it to the audio driver. i.e. save the value when "set_volume" (or whatver it's claled) is called, and write a "get_volume" function to match.
10:31:03niekieAlso, I asked this yesterday, but got no answer. Though from what I can gather, there currently isn't a proper way of doing translation for the Database, right? (e.g. the tags which say Composer, Artist, Album . . . in the menu)
10:31:06pamauryTell if I'm wrong but that probably means modifying all existing driver ?
10:31:56pamauryah no, hopefully, there are wrappers in sound.c so it seems feasable
10:32:03S_a_i_n_tniekie: I guess you could write your own tagnavi file...
10:32:10amiconnWhy does usb audio need the current volume?
10:32:12S_a_i_n_tHave you looked at that option?
10:32:25niekieS_a_i_n_t: yeah, though I kinda consider that a nasty hack :P
10:32:40S_a_i_n_tOnly way to do it though AFAIK
10:33:00niekieIt also makes language switching harder.
10:33:14niekieWould need to upload a file again each time to change language.
10:33:15linuxstbpamaury: Yes, I was hoping there was a single wrapper function...
10:33:25niekieBut yeah, guess it's the only way currently.
10:33:29pamauryamiconn: strictly speaking it doesn't need it but I feel a usb audio device should at least expose volume and mute control...
10:33:40S_a_i_n_tWell, could keep two/three/four builds on the player...ut thats even nastier ;P
10:34:01linuxstbpamaury: How does that work though? You need to be able to tell the host the current volume?
10:34:11Lloreanpamaury: Can USB Audio plus HID be combines, so you could control the host computer volume / mute?
10:35:46pamauryLlorean: HID could work while usb audio is on but iirc there is a special kind of interaction that is possible between HID and usb audio, something to tell the host that there are hid control on the device that control, say, the volume. But I don't remember the details
10:36:17pamaurylinuxstb: I need to be able to tell min/max/current and set current value, to claim support for volume control
10:37:58pamauryAnd by the way, it's a bit clumsy to have hid control on the device to change volume on host to change volume on device...
10:38:40pamauryI will implement getters in sound.c and see what I can get
10:38:55LloreanIt seems to me that most sound cards don't have an independent volume control though.
10:39:02LloreanYou control their volume through the host OS
10:39:09LloreanPlus, possibly, volume on the speakers.
10:40:09amiconnpamaury: Isn't that independent from the volume in rockbox? What direction are we talking about?
10:41:32 Quit n1s (Ping timeout: 260 seconds)
10:41:32linuxstbLlorean: What do you mean by "independent". The host OS controls the volume on the soundcard... USB Audio is just a USB soundcard IIUC.
10:41:34pamauryI don't get it amiconn, I just want to set the volume on the device... Any audio card allows to set the volume of its output, so the OS is not obliged to do software mixing...
10:42:12linuxstbpamaury: What units do you expose the volume in? Is it just an arbitrary integer range?
10:42:26pamauryarbitrary as far as I can tell
10:42:35 Quit Rob2223 (Read error: Connection reset by peer)
10:42:42Lloreanlinuxstb: I guess I misunderstood, it sounded like it would be able to set volume on the device.
10:42:49LloreanRather than having the host set the volume
10:43:08 Join Rob2222 [0] (
10:43:22pamauryAh no, it's in dB
10:44:13linuxstbCan it handle fractional dBs?
10:44:43pamauryoh yes, the unit in 1/256 dB...
10:44:51 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:44:53linuxstbI _think_ our volume is stored as 1/10 dB
10:45:20pamauryI think so, looking at the headers
10:46:38pamauryI will do it for the volume and if I manage to make it work, I'll do the same for other controls
10:47:27 Quit Tux2 (Quit: Leaving)
10:48:16linuxstbpamaury: Is the audio going through the dsp layer? On some devices, we have software tone controls.
10:49:12pamauryI know nothing about how audio works :) I current play audio using pcm_play_data, so tell me...
10:50:05linuxstbpamaury: I'm not sure of the exact flow, but DSP code is in apps/
10:50:26pamauryI think you are right
10:50:42linuxstbAnd yes, thinking about it, pcm_play_data() is the lowest-level - i.e. it sends it straight to the DAC.
10:50:55pamauryI know there are several approaches to usb audio. I see two of them:
10:51:08pamaury- export low level interface: usb audio cap=hardware cap
10:51:25pamaury- export high level interface: usb audio cap=everything, do it though dsp, resampling, ...
10:52:49linuxstbWell, the high-level approach is obviously the most flexible. You could then simply disable DSP/resampling/etc and get the equivalent of the low-level interface if you wanted.
10:53:24tanboI can see my hello world now~~ Yeah!
10:53:35 Join Luca_S [0] (
10:53:47pamauryTrue. But, given that usb audio is probably implement only on OSes like Linux, Windows and MacOS, which usually run on quite powerful machine which are already probably doing so audio decoding, I guess the host can do the software mixing that the device doesn't have
10:53:51linuxstbYou may also want to (for example) let the user record the usb audio being streamed to the device. That would be easier if it's apps/ code.
10:53:51amiconnpamaury: What I'm asking is whether you're doing audio in or audio out (as seen from the host os)
10:54:35amiconnIf you're doing audio out, aren't you just setting the volume in rockbox itself?
10:54:50pamauryamiconn: current I doing audio out (host->device) but I want to do everything at the end. I would like to be able to stream out to line-out (and perhaps to a file) and stream in from lin-in/FM/file
10:56:43pamauryMy current approach is (as you probably understood), a low level approach. I think the driver should only expose hardware capability, and host fills the gap
10:56:57 Quit xiainx (Ping timeout: 258 seconds)
10:56:58 Join BdN3504 [0] (
10:57:55pamauryBut I'm open to suggestions of course
10:58:36Luca_SFuzeV2: Data abort at 3003A624 FSR 0x8 (domain 0 fault 8) address 0xE59FF098. I was skipping a couple of songs in a playlist of about 400 items.
10:59:41Luca_Splaying from internal memory, frequently reproducible by skipping songs.
10:59:45 Join kugel [0] (~kugel@rockbox/developer/kugel)
11:00:04 Quit Topy44 (Ping timeout: 276 seconds)
11:00:59 Join xiainx [0] (
11:01:14pamauryamiconn, linuxstb, Llorean: I not sure that's a good argument but usb audio is a driver so it's implemented in firmware/. Using the DSP and all that would require a complete access to apps/. Of course if you tell me there is an easy way to make usb audio appear as a some kind of buffering source then the problem is different
11:02:53linuxstbpamaury: Maybe it should be implemented as a kind of line-in, and be available to select from the apps/ code. So to use your DAP as a USB audio device you would need (at least currently), to go into the recording screen and select that input.
11:03:23 Quit Kitar|st ()
11:03:47***Saving seen data "./dancer.seen"
11:04:09 Join JdGordon_ [0] (~jd@rockbox/developer/JdGordon)
11:04:18 Quit JdGordon_ (Client Quit)
11:04:45tanboExcuse me! Make zip not run, the message is sh:zip command not found.What should i do?
11:05:11pamaurylinuxstb: yes that's what I mean, but does the recording screen does playback ? (I never use it)
11:05:18topiktanbo: install zip ?
11:05:48Lloreanpamaury: You can monitor what you're recording.
11:06:53pamauryBut then it would be difficult to export volume/mute/... controls
11:07:02BdN3504yesterday i replaced the hard drive in an 20gigabyte ipod 4th gen
11:07:07linuxstbpamaury: Why?
11:07:24linuxstbpamaury: Because they can also be changed in Rockbox itself at the same time?
11:07:36BdN3504when i first plugged it into the usb port of a win xp computer the charging sign appeared
11:07:36pamaurybecause it doesn't make sense, if you're recording to a file for example
11:07:52BdN3504then i installed itunes and restored it
11:08:04linuxstbI would expect recording to a file to dump the PCM data unmodified.
11:08:26Lloreanlinuxstb: Doesn't recording let you adjust monitor volume *and* input gain?
11:08:37LloreanIt seems like it would complicate use somewhat.
11:08:47BdN3504after i disconnected it from the and plugged the device into the wall charger, the screen now only shows a white background
11:08:56linuxstbLlorean: Yes. But this is the equivalent to an S/PDIF input, where I doubt you can adjust the gain.
11:09:05BdN3504and some bars in a very light grey but no font
11:09:16BdN3504the piezo loudspeaker is working
11:09:46BdN3504is this a known problem? and will installing rockbox on that 4th gen ipod fix it?
11:10:05tanbotopik:Thank you.
11:10:06Lloreanlinuxstb: Are you sure you can't? I'd imagine you generally *shouldn't* but it seems like a feature to be able to, right?
11:10:25LloreanBdN3504: If you've got a broken player without Rockbox installed, this isn't really the place to be asking about fixing it.
11:10:37linuxstbBdN3504: No, if the Apple firmware doesn't work, it's unlilkely Rockbox will. Does your ipod even appear as a disk when connected to your computer?
11:10:56BdN3504linuxstb: yes it gets recognized
11:10:57linuxstbLlorean: No, I'm just guessing. I've never tried S/PDIF recording.
11:11:02 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
11:11:26S_a_i_n_tJdGordon: Applied you change to onplay.c as sugested, same data abort - same address.
11:11:36BdN3504Llorean: thanks for the heads up. i guess i'll have to replace the screen then, too.
11:11:41Lloreanlinuxstb: I tried it, but that was long, long ago and I can't remember how it behaved. :)
11:11:44linuxstbBdN3504: Well, it does sound like a hardware issue, which as Llorean says isn't really on-topic here. The only think Rockbox would help you with is that it has a voice UI...
11:12:04pamauryAll this seems complicated. I propose the following: first implement a low level interface, not too intrusive (my code probably already modifies 20+ files but only in firmware/), stream out only for know, and then we'll see what we do. Having a high-level interface is probably interesting and doesn't cost so much because the code is already there but it requires at least a partial rework of apps/ or recording code
11:12:21kugelpamaury: wouldn't the high level approach have a higher probability to make usb audio work the same on all (sw usb) targets?
11:13:02 Join Topy44 [0] (
11:13:10linuxstbpamaury: I agree that a simple first implementation is a good idea, to concentrate on getting the basics working reliably. After that, it depends if anyone is interested in improving it...
11:13:31pamaurykugel: yes, the current code highly depends on the hardware, for example the host can only stream at a frequency that the hardware can playback, but it's simpler
11:13:56linuxstbAnd also, a PC is likely to resample better than Rockbox
11:14:04pamauryThat's my opinion
11:14:39 Quit BdN3504 (Quit: CGI:IRC (EOF))
11:15:06 Join Kitar|st [0] (
11:15:25*pamaury goes back to usb audio. current goal=make it work with windows
11:16:02kugelwe should still be able to tell the pc the hw caps in the apps/ variant
11:16:33LloreanIt's too bad the only player with optical out is hw-usb.
11:17:05kugelcalling firmware from apps is fine, the other way around is not, that's why an apps/ variant is always more flexible which can do everything firmware can but possibly more
11:18:16pamauryyes but then apps/ must know the feature exist in firmware/. And the current code doesn't have a general way of handling audio source/sink, no ?
11:18:29pamaury* yes ?
11:18:47kugelno idea, but the current code is the current code not the future code :p
11:18:54saratoga_labcould you just write the basic firmware stuff and then have a plugin that handles the USB and audio stuff?
11:19:41 Join JdGordon_ [0] (~jd@rockbox/developer/JdGordon)
11:19:44kugelyea as a plugin would be nice, we always wanted the usb drivers as plugins (that's why the hid gsoc project failed unfortunately)
11:19:54 Quit JdGordon_ (Client Quit)
11:20:00 Join JdGordon_ [0] (~jd@
11:20:00 Quit JdGordon_ (Changing host)
11:20:00 Join JdGordon_ [0] (~jd@rockbox/developer/JdGordon)
11:24:21pamauryMy answer would be, yes it's possible, but not so simple :) I think it's possible to have an interface for say, audio in and audio out. The audio out would call a callback each time there is data (or preferably would copy data somewhere) and the audio in would do the converse. But the reality is also that there are lots of timing issues, but everything can be done with enough time and motivation
11:25:29*S_a_i_n_t suggests hotkey only be applied to builds with a free key on both the wps and tree
11:25:51S_a_i_n_tpotential for "longplay" shutdown on the iPods is just fricki' stupid
11:27:21S_a_i_n_tWere builds even offered for testing before commit?
11:28:02kugelplease, don't bitch now
11:28:42S_a_i_n_tAs opposed to bitching *before* I knew it didn't work?
11:28:43kugelis move is perfectly understandable. I did the same just without a discussion.
11:29:39kugelS_a_i_n_t: opposed? you guys bitched before and after
11:30:27S_a_i_n_tApparently with good reason.
11:30:42kugelthe choices were: make untested(and thus sometimes stupid) keymaps to keep the crowd quite or don't.
11:31:02kugelyou guys obviously made him do the first
11:31:18gevaertsS_a_i_n_t: how about a bug report?
11:32:00Lloreankugel: Actually, I pretty clearly said several times I thought he should do keymaps on players that, like the Sansa, had a recording button or similar free key.
11:32:21kugelso instead of gruming (remember it wasn't his idea to work on a keymap for so many targets, so it's not entirely his fault) around you should go and fix it
11:33:13LloreanThe problem also wouldn't have happened if he'd actually provided testing builds in the testing build forum and gotten feedback on the keymaps before committing rather than rushing to commit it.
11:33:32kugelit's not obvious if a player actually has a free button without spending multiple dozen hours to go through all targets
11:34:16LloreanAnd yet he did that.
11:34:32LloreanThen he continued beyond that and assigned it to combos on other targets, which is where the mistakes seem to have happened.
11:35:17LloreanIt also is apparently quite buggy, since it apparently consistently causes data aborts for some users.
11:35:39Luca_Sdoes anybody know why GPIOD "flickers" between 9F and BF when playing music from internal sd on fuzev2? if i read correctly, it's the third bit from the left, but it's not described on the amshardwaremappings page
11:35:44kugelobviously he just re-used the playback viewer combo in some places
11:36:43kugelworking on rockbox shouldn't be a pain in the arse, it should be fun in the hope the feature is useful to others. I'm fairly certain in the end it wasn't fun for him anymore
11:37:13 Quit xiainx (Ping timeout: 276 seconds)
11:37:14S_a_i_n_tWell, thats because his commit only had his target in mind.
11:37:19Lloreankugel: Working on Rockbox also means considering the effect your actions have on thousands of users.
11:37:45LloreanIf it's not fun to make it work right, don't do it, rather than refusing to put in the necessary effort and committing something half assed.
11:38:06S_a_i_n_tIt was a case of, I want this in my build for my target and unless anyone says otherwise...I'm committing.
11:38:09LloreanIN this case though, it's definitely not wholly his fault - the playlist viewer apparently chose some bad combinations in the first place.
11:38:31LloreanBut JdGordon's already mentioned the other problems with the code in an email to the list.
11:39:18 Join perfectdrug [0] (~marko@
11:39:44kugelthat argument is good way to keep developers away from our project
11:40:14LloreanWhat, "we have users, so try to make sure your code works well before committing"?
11:41:02 Quit n1s (Ping timeout: 258 seconds)
11:41:08kugel"well" doesn't mean "on all of or two dozen targets"
11:41:41gevaertsA bad choice on one or two targets is to be expected I think
11:41:59 Quit tanbo (Quit: CGI:IRC (EOF))
11:42:03kugelif a feature needs so many work individually on targets because they hit target specific areas, then it should be a collaboration instead of forcing 1 guy to do all the work
11:42:21 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:42:31Lloreankugel: Yes, it should. Which means the one guy shouldn't say "I'm going to commit it now, and refuse to spend any time gathering feedback first"
11:42:44kugelhe *did* gather feedback
11:42:48saratoga_labi prefer to commit early and often
11:42:49LloreanNo, he really didn't.
11:42:50saratoga_labusers be damned
11:42:53 Join xiainx [0] (
11:42:57 Quit mikroflops (Ping timeout: 258 seconds)
11:43:10kugelthe bad keymaps on some targets are the result of the feedback
11:43:15LloreanHe didn't, for example, say "I need keymaps for Target 1, Target 2, Target 3, and Target 4 in Screen 1 and Screen 2, can anyone with these players tell me a good free button or combo?"
11:43:40LloreanIt's a really simple question which almost anyone with the player could've answered in one or two words and probably helped head something like this off.
11:43:43kugeltarget 1 ... target 30 rather
11:44:10LloreanThere's so much overlap in targets that, for example, iPod 4G and newer would all be the same.
11:44:27LloreanAs well, targets with a recording button wouldn't be in that list since the keymap was obvious for those.
11:44:48 Join mikroflops [0] (
11:44:58kugel"I'd also like more targets to be added. I'd appreciate any feedback." sounded like asking people to name free buttons to me
11:45:20Lloreankugel: Specific questions _always_ work better.
11:45:22kugelparticularly because he avoided to add the targets himself
11:48:25LloreanAs it stands though, a lot of features get added with little or no followup on fixing problems with them from the original developer. For example theme behavior just gets buggier and buggier. There's constant reports of the screen not updating right after the backlight turns on, bug reports about scrolling lines continuing to scroll with when the USB screen comes up, etc.
11:48:33 Quit scorche (Ping timeout: 252 seconds)
11:48:48LloreanYet the people who've introduced these bugs are usually more interested in the next feature to catch their fancy rather than finishing off the last thing they did.
11:49:06 Quit BHSPitMonkey (Remote host closed the connection)
11:49:46kugelwe're in a development phase, not a bug fixing phase, so I can understand the situation (that doesn't mean I agree with)
11:49:51 Quit xiainx (Ping timeout: 258 seconds)
11:50:04LloreanWhat do you mean "not a bug fixing phase"
11:50:19LloreanIs someone just supposed to say "no new features for the next year, bug fixing time?"
11:50:41pamauryIf I can say something, now that's it's committed, the most important point is to follow the bugs and fix them and not let it get buggier and buggier
11:51:13pamauryAnd preferrably, it should be done by him because he know the code
11:51:58kugelLlorean: we're doing that once in a while, e.g. before releases. not for a whole year, but still
11:52:11LloreanOne week of bug fixing every three or four months is *not* enough
11:52:38LloreanMost of the time the feature-happy devs just wander off and do something else for that period, then come back with commits after the branch/freeze ends.
11:53:04 Join scorche [0] (~scorche@rockbox/administrator/scorche)
11:53:08LloreanBug fixing will almost never be fun, but if people aren't held accountable for the bugs they introduce (if from listening to bad advice, or being in a hurry, or for whatever reason, their fault or not) then the count of bugs is just going to increase.
11:55:05 Join xiainx [0] (
11:58:04 Quit kugel (Ping timeout: 246 seconds)
11:59:45 Quit saratoga_lab (Quit: Page closed)
12:00:40 Quit n1s (Ping timeout: 240 seconds)
12:03:08 Quit JohannesSM64 (Quit: WeeChat 0.3.2-dev)
12:07:19 Quit xiainx (Ping timeout: 268 seconds)
12:08:01 Join n1s [0] (~n1s@rockbox/developer/n1s)
12:10:50 Join xiainx [0] (
12:14:51 Join moos [0] (moos@rockbox/staff/moos)
12:15:57 Join mikroflops_ [0] (
12:17:24 Quit Zagor (Ping timeout: 276 seconds)
12:18:11 Join watto [0] (~watto@
12:18:51 Join mikroflo1s_ [0] (
12:19:54 Quit mikroflops (Ping timeout: 252 seconds)
12:21:05 Join geertvdijk [0] (
12:22:45 Quit mikroflops_ (Ping timeout: 264 seconds)
12:23:15 Quit Topy44 (Ping timeout: 276 seconds)
12:28:01 Join archivator [0] (~archivato@
12:33:08 Join Topy44 [0] (
12:36:10 Quit perfectdrug (Ping timeout: 240 seconds)
12:40:10 Quit xiainx (Ping timeout: 240 seconds)
12:40:39 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:43:36amiconnlinuxstb, Llorean: S/PDIF in doesn't let you adjust input gain, because it wouldn't make sense
12:44:49 Join xiainx [0] (
12:45:17amiconnInput capabilities can be defined per target in config/*.h
12:46:59JdGordonS_a_i_n_t: odd.. ok so its crashing even beofre there which is odd
12:47:27S_a_i_n_t"before there"?
12:47:41 Join emrecelikten [0] (~c18cf902@gateway/web/freenode/x-wjlweoiystpjstum)
12:47:45JdGordonbefore that splash
12:47:48JdGordonor did you see the splash?
12:47:51S_a_i_n_tI compiled with your dice.
12:48:00JdGordonso yes
12:48:01S_a_i_n_tOh, no splash at all.
12:49:16 Quit n1s (Ping timeout: 258 seconds)
12:50:29S_a_i_n_tIt seems odd (to me, with little to no understanding of the code) that trying to define a hotkey in the wps context menu causes a data abort, but the tree context menu just does.....nothing.
12:51:01LloreanS_a_i_n_t: Is playback occurring while you do it in the tree?
12:51:01S_a_i_n_t"nothing" being not only not set a hotkey, but literally nothing.
12:51:23S_a_i_n_tLlorean: both yes, and no.
12:51:29S_a_i_n_tie, same result either or.
12:55:48pixelmaLlorean, kugel: the reuse of the playback viewer shortcut in itself is ok, I think. The fact that this turns out to have problems on some targets shows that the playlist viewer shortcut mapping isn't working and because it wasn't documented, not a lot people even knew about it and so didn't try combos and test on their target...
12:56:19Lloreanpixelma: I agree with that - it's the playlist viewer shortcut's fault.
12:56:40S_a_i_n_tthe combo works fine to get to the playlistviewer in the WPS, just shit for setting a hotkey apparently ;)
12:57:01pixelmaeven if the data abort has other reasons
12:57:11pixelmaS_a_i_n_t: setting the hotkey?
12:57:57S_a_i_n_tyes, its "supposed" to be the key combo to set a hotkey.
12:58:40 Join adnyxo [0] (
12:59:00 Quit lpereira (Ping timeout: 246 seconds)
12:59:01S_a_i_n_tAnd it most likely is, it just doesn't seem to work.
12:59:15 Join n1s [0] (~n1s@rockbox/developer/n1s)
12:59:20 Quit xiainx (Ping timeout: 264 seconds)
12:59:41pixelmahow does that work? I haven't read the documentation yet but
13:00:30S_a_i_n_tselect a context menu item, press select+play on it, you *should* get a menu asking if you want to set it as hotkey
13:01:09pixelmaI would have expected it to be in the usual context menu
13:01:18pixelma+an item
13:01:36S_a_i_n_tsame here...the implementation is well...shit IMO
13:02:14kugelS_a_i_n_t: what happens instead?
13:02:17S_a_i_n_tselect "hotkey" from the context menu, get "what item would you like to define as hotkey?" (or similar). select it, done.
13:02:22S_a_i_n_tseems simpler to me.
13:02:42Lloreanpixelma: You mean, "call up the context menu on an item in the context menu, then select 'make hotkey'" kinda thing?
13:02:48S_a_i_n_tkugel: in the WPS a data abort, in the tree: nothing
13:02:55pixelmaS_a_i_n_t: no, I mean on the item... what Llorean said
13:03:21S_a_i_n_tOh, yes..I see, easier still.
13:03:50***Saving seen data "./dancer.seen"
13:05:07 Join xiainx [0] (
13:09:01 Part MadRat-
13:10:20 Join arbingordon [0] (~w@unaffiliated/arbingordon)
13:14:26 Join wodz [0] (
13:27:58 Quit xiainx (Ping timeout: 258 seconds)
13:32:54wodzI am reading firmware/sound.c and I am wondering why MAS35xx codecs are treated so differently? Souldn't lowlevel function be moved to firmware/drivers/audio/mas35xx.c ?
13:33:50LloreanThey decode in hardware, rather than having software codecs to decode the audio.
13:33:59 Join xiainx [0] (
13:34:51 Join lpereira [0] (
13:34:52linuxstbwodz: They use a completely different playback engine - "hwcodec" vs "swcodec"
13:35:35 Quit mikroflo1s_ (Ping timeout: 260 seconds)
13:36:37wodzI know but there is no particular difference between wmcodec_write() and mas_codec_writereg(). The former do not appear in sound.c. Setting sound settings like loudness, super_bass etc. is lowlevel stuff
13:36:51*n1s spots potential beast battery time improvement, yay, great work jhMikeS
13:37:09linuxstbwodz: Hmm, maybe that's simply something no-one has cleaned up.
13:38:21amiconnwodz: Btw, fs #11153 looks fine to me
13:38:28 Join mikroflops [0] (
13:38:43wodzamiconn: cool
13:38:52amiconnThe original idea was that different daps using the same SoC might have different memory layout, but it turned out that this is usually not the case
13:39:14amiconn(as you can see from the other SoCs like the various PP)
13:40:45n1samiconn: could you look at i am pretty certain gcc is buggy since it should do a sign extension afaiu
13:40:56n1sbefore the compare
13:40:59 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:41:09n1sor am i missing something?
13:41:59 Quit JdGordon_ (Quit: Bye)
13:42:15wodzcomming back to mas35xx code in sound.c - static int tenthdb2reg(int db) is defined in sound.c but all others equivalents like tenthdb2master() are defined in lowlevel drivers
13:43:11wodzIs it a matter of historical reasons (no-one has cleaned up) or is it by design?
13:43:13 Quit mikroflops (Ping timeout: 265 seconds)
13:43:53 Join mikroflops [0] (
13:44:19 Join einhirn [0] (
13:51:19 Quit TillW (Read error: Connection reset by peer)
13:54:12 Quit archivator (Ping timeout: 248 seconds)
13:56:00amiconnn1s: That looks buggy, but it doesn't look like this is rockbox code?
13:56:27amiconnRockbox is compiled using -fomit-frame-pointer, or at least it should be
13:57:18n1samiconn: yeah, i just isolated the saturate function and compiled it stand alone, thanks for looking, I will post a gcc bug
13:58:42wodzWhy there are mas.h and mas35xx.h header files? Souldn't this be merged into one?
13:59:25 Join archivator [0] (~archivato@
13:59:45n1swodz: lots of that is very old code from before swcodec was introduced, so i imagine a lot of it isn't in the now odeal place etc, i think noone is against cleaning up
14:01:48wodzOk. I'll do my best to clean this up.
14:09:10 Join TillW [0] (
14:14:10pixelmadon't forget and hesitate asking for testers if you don't have any such device. I volunteer testing with the OndioFM I got (only one of the two different MAS chips though)
14:16:28wodzI will post my work as a FS patch
14:18:34 Quit mikroflops (Ping timeout: 258 seconds)
14:22:55JdGordondoes audio_peek_track() have some odd side effect which could cause the wrong track to start playing if it is called too early in buffering?
14:23:29JdGordonor playlist_peek()?
14:24:18 Join mikroflops [0] (
14:28:31 Join Schmogel [0] (
14:30:41emreceliktenThere goes my SoC proposal. Couldn't find much to fill though.
14:31:09 Quit xiainx (Ping timeout: 240 seconds)
14:35:29 Join xiainx [0] (
14:38:34TillWhey, i was just filling out my application to rockbox for GSoC, and i had written out a bunch of stuff. are you looking for just the major questions answered in short and concise, or looking for those questions answered in paragraphs?
14:39:05gevaertsTillW: you're better off with too much information than too little
14:41:37TorneHmm. Reading sector 1 on the broken Toshiba 1024 byte sector drive just returns -6 and then everything works fine afterward
14:41:40TorneThis looks like a good sign
14:41:56TillWso, if my paragraphs answer all the questions, i should just stick with them then
14:43:37Torneis it just the archoses that have the dumb USB bridge that can't do LBA48? is it *all* the archoses?
14:45:27gevaertsTillW: we do appreciate following the template if that's what you mean
14:45:59GodEaterTillW: yes, please supply answers to the questions in the right order. We'd rather not have to go backwards and forwards through your submission to find them all.
14:46:24TorneAh. Problem is that it retries and ends up taking five seconds.
14:46:31TorneThat's probably not desirable ;)
14:46:41TillWwell it is following the template, i'll just rearrange things to more closely follow the order of questions
14:46:44pixelmaTorne: definitely not the Ondios which are flash based
14:47:00Tornepixelma: ok, only the ones which use STORAGE_ATA, i mean
14:47:14TillWalso, if i submit it in the next few hours, would someone be able to take a quick look over it and say if i should change some things, or be more or les detailed in places?
14:47:22TorneI am going to add LBA48 support to all ATA targets, because the code is not very big and it removes one obstacle to using upgraded drives
14:47:42Tornebut there's no point adding it when the usb bridge chip can't do it
14:47:45Torneor at least, it might be misleading
14:47:57Tornepeople still need to follow the Special Magic to use those devices with a larger disk ;)
14:48:11GodEaterTillW: we'll no doubt check it over once it's in, and you'll receive feedback when we have some for you
14:48:22GodEaterTillW: we have a queue of them we're working through right now
14:48:25pixelmaTorne: then I think the HD ones (of the currently supported Archoses) are the same in this regard. At least they all use 2.5'' disks
14:48:45TillWalright, that's fair
14:48:46n1sTorne: all archoses have similar usb bridges with the same limitation afaik but i don't know about the bridges other targets use
14:49:28Tornehm. is there a config.h flag for the usb bridge thing?
14:50:21Torneaha, it's !HAVE_USBSTACK I think
14:50:35Tornehm, maybe not
14:50:59 Join froggyman [0] (~187b533e@gateway/web/flash/
14:51:44amiconnTorne: It's the disk-based archoses. The ISD300 doesn't support LBA48, and the even older ISD200 doesn't either
14:52:13Torneamiconn: right, so it's safe to turn it on for all the other STORAGE_ATA targets?
14:52:20amiconnFor the Ondios, things are different, as there is no LBA48 for MMCs...
14:52:33Tornei assume they all at least have the hypothetical possibility of a disk upgrade
14:52:36Torneor CF upgrade
14:52:47TorneYah, MMC isn't using ata.c though
14:53:12Torneit looks like reading sector 1 is a reasonable test for the broken drive, also
14:53:33Torneit retries for 5 seconds, which I need to do something about, but it does eventually return -6 on the 80GB 5.5G disk
14:53:35amiconnYou can use harddisks >128GiB in an archos, but then you need more than just HAVE_LBA48 and you have to be careful wrt partitioning
14:53:51Torneyeah, so I was gonna leave it off for those since users would need to do a bunch of stuff themselves anyway
14:54:02Torneplus archos has rombox so the delta is best avoided :)
14:54:31amiconnArchos has no rombox anymore execpt the player :\
14:54:38Tornedidn't know
14:54:46amiconnRockbox grew too large
14:54:49Tornebut anyway, yeah. best not to turn it on since it might make people fail to realise the problem
14:54:59Torneon everything *else* it should Just Work, i hope
14:55:20 Quit xiainx (Ping timeout: 276 seconds)
14:55:32amiconnThe flash rom is 256KB. 28KB are taken by the flash loader and (compressed) bootbox. So exceeding 228KB uncompressed means no more rombox
14:57:01 Join jgarvey [0] (
14:57:13amiconnRombox has been gone on the recorders and Ondio FM for a long time already. For the Ondio SP it happened not long ago
14:57:59amiconnWe don't have a config flag that tells whether the usb bridge is lba48 capable. Perhaps we should introduce one
14:58:40Tornei was just gonna put HAVE_LBA48 in all the configs for the relevant targets..
14:59:19amiconnhmm, ok
14:59:44 Join xiainx [0] (
15:00:32TorneIt might be better to infer it, though, I guess
15:00:36Torneanywya, meeting time, back later.
15:01:48 Quit jgarvey (Client Quit)
15:03:52***Saving seen data "./dancer.seen"
15:04:28 Part lpereira
15:04:36 Quit antil33t (Read error: Connection reset by peer)
15:04:44 Join antil33t [0] (
15:08:06 Quit geertvdijk (Ping timeout: 246 seconds)
15:14:29 Join geertvdijk [0] (
15:16:38 Quit archivator (Quit: Leaving)
15:16:52 Quit xiainx (Ping timeout: 248 seconds)
15:19:10 Join Barahir_ [0] (
15:22:09 Quit Barahir (Ping timeout: 240 seconds)
15:22:14 Join xiainx [0] (
15:23:31 Join evilnick_B [0] (~0c140464@rockbox/staff/evilnick)
15:24:25 Quit Barahir_ (Ping timeout: 276 seconds)
15:32:05 Quit TheSeven (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
15:32:58 Join JdGordon_ [0] (~jd@
15:32:58 Quit JdGordon_ (Changing host)
15:32:58 Join JdGordon_ [0] (~jd@rockbox/developer/JdGordon)
15:33:20 Quit JdGordon (Ping timeout: 246 seconds)
15:35:00 Quit xiainx (Ping timeout: 248 seconds)
15:38:14 Join Blue_Dude [0] (
15:38:47 Quit TillW (Remote host closed the connection)
15:42:29 Join xiainx [0] (
15:49:20 Join JdGordon [0] (~7bf38c1f@gateway/web/freenode/x-dfepwobrmsvphnoh)
15:52:37 Quit froggyman (Remote host closed the connection)
15:54:29 Quit mitk (Quit: Leaving)
15:55:41JdGordonBlue_Dude: did you see my email?
15:56:30 Join TillW [0] (
15:56:57Blue_DudeYeah, I'm taking a harder look at the settings struct to make sure I didn't cause any problems. I found a couple of things already that might cause problems. But it may not need a complete redo either. I'm still checking.
15:59:19JdGordoneither way, the config line should be human undertsnadable
15:59:30JdGordonbut I dont beleive it works at all as it is in svn
16:00:51Blue_DudeDoes anyone have a target handy to check this on?
16:02:17pamauryBlue_Dude: check what ? I have a e200, I can check something quicky if it's no too long
16:02:54Blue_DudeNo data abort on the e200. That's my DAP too. It probably needs a Nano.
16:04:10 Join JdGordon1 [0] (~jonno@
16:04:14 Quit JdGordon_ (Quit: Bye)
16:05:37Blue_DudeJdGordon: the problem with a full rework in settings_list.c is that this was never intended to be a menu item, just a placeholder for persistent variables. I suppose there's nothing really wrong with putting them in an INT macro, but most of the values won't ever be referenced.
16:05:50 Quit JdGordon (Ping timeout: 248 seconds)
16:06:39JdGordon1that list is how the config.cfg file is handled, its got (almost) nothing to do with the menus
16:06:53JdGordon1it should be human readable anyway
16:07:54Blue_DudeHm. Drat. OK.
16:13:19Blue_DudeWould there be any problem with using NULL as callbacks in the CHOICE settings as long as the config strings are preset?
16:13:25pixelmaI find it weird (sounding at least) that you have to press the hotkey or respective combo on the menu item to set it
16:14:01JdGordon1Blue_Dude: callback is only needed if something needs to know when it has changed
16:14:02JdGordon1so no
16:14:10JdGordon1but I think CUSTOM is the better way to go
16:14:17 Join evilnick_ [0] (
16:14:21 Quit Luca_S (Quit: CGI:IRC (Ping timeout))
16:14:40Blue_Dudepixelma: What do you have in mind?
16:14:54 Quit emrecelikten (Ping timeout: 248 seconds)
16:14:57JdGordon1something like how the quickscreen items are set maybe
16:15:33 Join Luca_S [0] (
16:16:30pixelmaput a "set as hotkey item" or somesuch into the context menu
16:17:10JdGordon1context menu in a context menu isnt very nice :)
16:17:51 Quit evilnick (Ping timeout: 265 seconds)
16:17:58S_a_i_n_tseems nic"er"
16:18:20pixelmaI thought it's on other menu items too? me at least, and pixelma apparently ;)
16:18:48Blue_Dudepixelma: So you want a context menu on a context menu item?
16:18:54 Quit xiainx (Ping timeout: 268 seconds)
16:19:18Blue_DudeAs opposed to a keypress, or key-combo?
16:19:21S_a_i_n_tIt doesn;t even need to be a menu, just an option.
16:19:32S_a_i_n_tit doersn't necessarily need "yes/no"
16:20:48Blue_DudeThe hotkey code is a mod of the context menu itself. Now we're talking possibly recursive context menu calls...
16:20:50pixelmaBlue_Dude: I wasn't aware that you could only set things that are in some other context menu as a hotkey function. Even then, that would be the way I'd expect in Rockbox settings for these things
16:21:12JdGordon1a menu option to select the action would be best
16:21:27pixelmaor that - if it is restricted anyways
16:21:54pixelma(I mean you only have the choise of a handful things to put on the hotkey)
16:22:15Blue_Dudepixelma: The hotkey was just a way to bring a specified context menu item out for easier access, vs. drilling down into the context menu all the time. It was never intended to run arbitrary functions.
16:22:20JdGordon1I dont particularly care about how it is set, just that the setting handling is fixed
16:23:41Blue_DudeIf a setting is not able to be hotkey'ed, the hotkey keypress in the menu is ignored. It's only when the item is acceptable that the confirmation dialog pops up.
16:25:56pixelmaBlue_Dude: how many things are there you could put on the hotkey?
16:26:12 Quit n1s (Ping timeout: 258 seconds)
16:27:28Blue_Dudepixelma: There's no specific limit, but how many context menu items are really one-button useful? I coded the one's that I most wanted to see, but it's easy enough to add more.
16:28:06 Quit JdGordon1 (Quit: Leaving.)
16:28:10 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
16:36:58Blue_DudeJdGordon: Yes, I think I see what you're talking about re: CUSTOM.
16:37:31 Join n1s [0] (~n1s@rockbox/developer/n1s)
16:37:31 Join TopyMobile_ [0] (
16:39:47 Quit JdGordon (Ping timeout: 246 seconds)
16:39:55 Join toffe82 [0] (~chatzilla@
16:40:09 Part LinusN
16:40:25 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
16:41:28 Join JdGordon1 [0] (
16:41:45 Quit TopyMobile (Ping timeout: 276 seconds)
16:41:50Blue_DudeJdGordon: looking at the quickscreen CUSTOM code, it seems as though you had to duplicate functionality. Is that going to be required for any CUSTOM setting?
16:43:25 Nick fxb is now known as fxb__ (
16:44:09pixelmaare there some only available through the WPS context menu? Things like playlist, sound settings etc. are in the main menu too and maybe are easier to set from there. It's just that the way setting seems to work now sounds very very different to anything else
16:44:55 Quit JdGordon (Ping timeout: 265 seconds)
16:45:52 Quit JdGordon1 (Ping timeout: 260 seconds)
16:46:33 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
16:49:32Blue_Dudepixelma: It's really not so much to change settings but to make it easier to access the associated context menu. I found I spent a lot of time negotiating the context menus looking for the same item, and wanted to be able to select that item without so much looking and scrolling.
16:50:24Blue_Dudepixelma: Also, the settable menu items are noted in the manual. If they're not settable, current behavior is unchanged: the keypress is ignored.
16:51:32 Join Xerion [0] (
16:52:04 Join stoffel [0] (
16:54:25S_a_i_n_tnoted in the manual where exactly?
16:55:36 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
16:55:40pixelmaBlue_Dude: but once set (through a different mechanism than now) you can access this one context menu item the same as now (if your target supports it)
16:56:24Blue_DudeWhich one?
16:56:26 Quit mikroflops (Ping timeout: 240 seconds)
16:56:36 Quit Sergio (Ping timeout: 260 seconds)
16:57:32 Quit TillW (Remote host closed the connection)
16:58:07S_a_i_n_tIf hotkey settable items are indeed listed in the manual, I'm having some difficulty finding them. I expected to find them listed under Browsing and Playing/Hotkeys.
16:58:47 Quit n1s (Ping timeout: 258 seconds)
16:59:11Blue_DudeThey are noted under the context menu items.
16:59:21pixelmae.g. through the context menu on a menu item - not the context menu itself _if_ all are accessible through both
17:00:08Blue_Dudepixelma: Sorry, I still don't follow you.
17:00:41 Quit gevaerts (Read error: Operation timed out)
17:01:32 Join mikroflops [0] (
17:01:55 Quit evilnick_ (Read error: Connection reset by peer)
17:02:43 Join evilnick_ [0] (
17:03:32S_a_i_n_tpixelma: this may clear things up, perhaps we were expecting more usability than it has? I only count 4 "hotkeyable" context menu items
17:03:56***Saving seen data "./dancer.seen"
17:03:57 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
17:04:17 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
17:04:37S_a_i_n_tErrrr...sorry, 3
17:05:11S_a_i_n_tDelete, Delete Dir, and Open With, but apparently Pitchscreen is one also.
17:05:39Blue_DudeIn the WPS: View Playlist, Show Track Info, Pitch, Open With, Delete.
17:06:20 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:06:21Blue_DudeIn the File Browser: Open With, Delete (also Delete Dir), Playlist Insert
17:06:38Blue_DudeWhat else would you like?
17:07:04S_a_i_n_tAny context menu item. IMO
17:07:09pixelmaonly plain "Insert"?
17:07:10Blue_DudeWithin the exsiting context menus, yes.
17:07:38Blue_DudeInsert into Playlist. I misspoke.
17:08:02 Join Sergio [0] (
17:08:47pixelmayes, but there are different ways to Insert - Insert (plain) and Insert Next, Last, Shuffled, Last Shuffled, same with Queue
17:08:52Blue_Dude"Any" can be done, but not all of them make much sense. OTOH, each item only costs about 20 bytes.
17:09:50Blue_DudeI went with only plain Insert, but others can be added easily enough.
17:10:26pixelmaok, I didn't think of things like "Delete" or "Delete Directory" etc. which are really only available through the context menu on files
17:11:30Blue_Dudepixelma: What I really wanted was a Delete key, but that was just shut down as being too narrow. Someone else wanted an Insert key. So why not make it do what you want?
17:11:30 Quit JdGordon (Ping timeout: 265 seconds)
17:11:46Blue_DudeThat was the concept anyway.
17:13:19 Join TillW [0] (
17:17:39 Join fyre^OS [0] (
17:21:49 Quit guymann (Ping timeout: 246 seconds)
17:23:43 Join guymann [0] (
17:28:40 Quit Bagder (Quit: It is time to say moo)
17:33:53 Join Boldfilter [0] (
17:35:36Tornegah. i can't think of a non-horrible way to selectively stop ata_transfer_sectors from retrying
17:35:53Torne(also, is this even a good idea? not sure.)
17:36:27Tornetrying to test for needing large sector emulation by reading sector 1. it works fine, but it takes 5 seconds because the driver just keeps resetting the bus and trying again until the timeout
17:38:11 Join panni_ [0] (
17:38:43 Join liar [0] (
17:39:00 Quit Zagor (Quit: Leaving)
17:41:05wodzTorne: can't the timeout be shorter than?
17:41:29Tornewodz: the timeout has to be a reasonable time in general, because the disk might be being shaken or similar
17:41:59Tornehm, i guess I could have it check the error register and not retry for errors which are "definately" fatal and won't get any better
17:42:20Torneif the disk says the sector you are reading is out of bounds then retrying isn't going to help.
17:44:24n1smeh, that supposed gcc bug is not a gcc bug but me missing something as usual, too bad i reported it before figuring it out...
17:45:21wodzn1s: what have You missed?
17:45:37 Quit Adubb (Quit: "The mind and body must be subjected to extreme stimulus, by means of drugs and music." - HST)
17:46:08n1ssince the destination was an address register, movea is used which does sign extend, only gcc doesn't print out move it uses move for everything and i guess as does the translation
17:46:29n1sdoesn't print out movea
17:46:38n1sthis is seriously braindead
17:47:09wodzthat makes sense
17:47:32wodzbut this IS a bug anyway
17:48:58wodzor it does not matter if You feed m68k-elf-as with move or movea (it chooses properly based on destination reg)?
17:50:26kugelIIRC it doesn't matter
17:50:47kugeln1s: does objdump also not show movea?
17:51:18n1syeah, i
17:51:18n1sthink it has to
17:51:43n1skugel: it does
17:52:44wodzok so it is not a bug from generated binary point of view but generated assembly file is at least missleading
17:52:45n1sthe programmers manual says " MOVEA is
17:52:47n1sused when the destination is an address register
17:52:59wodzI would call this a bug still
17:53:12n1sthat it is :)
17:54:03n1smaybe, but it's a pretty unimportant one
17:54:06 Join mitk [0] (
17:54:24 Quit pamaury (Quit: Quitte)
17:54:28wodzbut annoing
17:56:31wodzbtw. the way gdb shows dissasembly of coldfire binary makes me crazy :-) I hacked it to at least show numbers in hexadecimal
18:00:04 Join merbanan [0] (
18:06:09 Join archivator [0] (~archivato@
18:07:08 Quit geertvdijk (Ping timeout: 240 seconds)
18:07:49 Quit petur (Quit: *plop*)
18:07:52n1sah, never used that
18:08:38Torneaha, there we go. no retries on IDNF, and now it only takes 40ms to probe sector 1
18:11:26S_a_i_n_tOooops, a little "Hotkey" updata, it was a little naughty not to check first before I reported the bug I know...but I had a couple of patches applied. I built current SVN from a clean checkout and the data aborts are still present.
18:11:52S_a_i_n_t*phew* ;)
18:12:50S_a_i_n_tWell, not *phew* it's broken, but *phew* I didn't fuck up by reporting a bug introduced by a patch.
18:15:23 Quit n1s (Quit: Lämnar)
18:15:37pixelmaand I first wanted to ask you if you tried an unmodified SVN version a few hours ago
18:16:05S_a_i_n_twell, now 'ya know ;)
18:16:55 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:18:07Torneanyone familiar with ATA able to take a quick look at ?
18:18:20Torneamiconn: you still around? :)
18:19:10 Join Barahir_ [0] (
18:20:04 Quit Topy44 (Ping timeout: 276 seconds)
18:20:19 Quit kugel (Ping timeout: 264 seconds)
18:25:11 Join Lear [0] (chatzilla@rockbox/developer/lear)
18:27:22 Quit Sergio ()
18:30:36 Join Topy44 [0] (
18:32:19 Join Strife89 [0] (~michael@
18:32:31Strife89Found a bug in the text viewer.
18:33:00Strife89Apparently, if it can't locate the previously set font, it gives a divide by zero error.
18:33:17Strife89Crashing Rockbox, of course.
18:33:31Blue_DudeS_a_i_n_t: Before I send out this svn, try something out for me: in settings_list.c, there are 4 hotkey config lines ending in 0),
18:34:08Strife89Could just be a corrupted font file, though.
18:34:13Blue_DudePlease replace the "0)," with "1, NULL)," and see if that solves the problem.
18:34:27S_a_i_n_twill do.
18:34:56Blue_DudeIt won't fix the config output but if it solves the data abort that would help.
18:36:16 Quit moos (Ping timeout: 264 seconds)
18:39:41 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
18:40:38S_a_i_n_tBlue_Dude: Building it now, 'gonna be AFK for ~15-ish, will report success/fail re: data abort then.
18:40:46Torneamiconn: Hm.. you said the ISD300 doesn't do LBA48.. the h300 appears to use that as well..
18:40:49Blue_Dudek, thx
18:41:08Torneamiconn: so maybe it's not just the archoses?
18:45:15 Quit Luca_S (Quit: CGI:IRC (EOF))
18:45:15*bluebrother gets annoyed by a dev being too trigger happy about filing bugs (and not checking the svn logs first). grmbl.
18:45:49*Strife89 tries to reproduce the text viewer bug.
18:47:30Strife89Still happens if I manually set and then remove a font.
18:48:01Strife89The text viewer returns a "font load failed",
18:48:10Strife89then Rockbox crashes with a Divide by zero.
18:48:46 Quit TillW (Remote host closed the connection)
18:48:48 Join PatrickB [0] (
18:49:38Strife89I don't see any bug report for this, so shall I file it?
18:50:12 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
18:51:00 Quit flydutch (Quit: /* empty */)
18:53:06 Quit PatrickB (Ping timeout: 252 seconds)
18:53:22 Join xiainx [0] (
18:53:47 Quit antil33t (Read error: Connection reset by peer)
18:53:53 Join antil33t [0] (
18:55:04Strife89Task added.
18:55:18 Quit Strife89 (Quit: Changing buildings.)
18:56:45 Join PatrickB [0] (
19:00:57 Join tomers [0] (
19:02:30Blue_DudeCan someone check out FS #11178 for possible commitment? Thanks,
19:03:58***Saving seen data "./dancer.seen"
19:04:32CIA-5New commit by 03tomers (r25509): Text viewer: Documentation and code cosmetics (no functional changes)
19:06:34CIA-5New commit by 03tomers (r25510): Text viewer: Fix possible memory corruption
19:08:01bluebrothertomers: I would really appreciate discussing possible issues before filing bug reports. Discussing things is usually less work.
19:11:11CIA-5New commit by 03tomers (r25511): Text viewer: Fix limit calculation in col_limit() which maximizes overlapping in wide mode
19:12:22CIA-5New commit by 03tomers (r25512): Text viewer: Make global variables static
19:15:26CIA-5New commit by 03tomers (r25513): Text viewer: Use exact width for lcd_fillrect (no actuall change)
19:16:47 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:18:30 Join fml [0] (
19:19:34fmlBlue_Dude: hello. I have a couple of questions/notes re hotkeys. May I shoot?
19:19:54Blue_Dudefml: Sure, I've got a few more minutes.
19:19:57CIA-5New commit by 03tomers (r25514): Text viewer: Replace macros with inline functions (which are about to be extended in the following commits)
19:20:47 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
19:20:49fmlBlue_Dude: ok. (1) Why do you need two settings for each hotkey (wps/tree)? IMO it could be stored in just one. The desc_id can be inferred.
19:21:25fmlBlue_Dude: (2) In the "disply hotkey" function, the words "WPS" and "Tree" are not localizable
19:21:50Blue_Dudefml: I needed to keep the variable for the info screen. I could set up a lookup table, but this way is easier.
19:22:01fmlBlue_Dude: (3) In the manual, where you write "Hotkeyable", it would be good to put a reference to the hotkey section IMO
19:22:18CIA-5New commit by 03tomers (r25515): Text viewer: Fix RTL languages and diacritic characters support
19:22:44fmlBlue_Dude: but you already have the lookup table, the onewith items
19:22:52Blue_Dudefml: 2: true enough. The info screen was a last minute addon. I could make a localizable text for each one...
19:23:24fmltomers: hello.Why do you read size - 1 bytes only? To not to lose the last byte which is overwritten by 0?
19:24:10Blue_Dudefml: can't read the table from the settings menu (it's static). I could make it un-static and extern it. Messy though.
19:24:37Blue_DudeOr I could move the info screen to onplay, but it doesn't really belong there.
19:25:11Blue_Dudefml: 3) Probably should. I'm not a very good tech writer.
19:25:35PatrickBCould anyone please add me to the WikiUsersGroup?
19:25:53bluebrotherPatrickB: what is your wiki name?
19:26:08PatrickBsame as here :PatrickB
19:26:15tomersbluebrother: ping
19:26:23bluebrotherPatrickB: meep
19:26:51bluebrotherPatrickB: please read the registration page carefully. Your wiki name is _not_ PatrickB.
19:27:03bluebrothertomers: pong
19:27:19fmlBlue_Dude: I'd create some (non-static) functions "id->name" and "name->id" in the hotkey "module" and use them in th settings code
19:27:21CIA-5New commit by 03tomers (r25516): Text viewer: Fix red
19:28:23Blue_Dudefml: Works for me. Costs some bin, though the config page is cleaner.
19:28:40fmlBlue_Dude: what works?
19:29:04Blue_DudeThe id->name funcs.
19:29:31Blue_DudeI'll take a look. Shouldn't be messy.
19:29:43PatrickBblubrother, it says here that that's my name:
19:29:43PatrickB+ on my own wikipage: PatrickB">
19:30:19TheSevenwell, if your last name is really just "B", poor guy, that might confuse people a bit :-P
19:31:13gevaertsPatrickB: he means that the registration page says to use your real full name
19:32:17PatrickBI understood that you can pick this as the wikiname i.e. you don't have to use your full name like in the registration. I registered with my full name of course but then as the wikiname I picked that shorter version
19:32:37 Nick YPSY is now known as Ypsy (
19:32:51S_a_i_n_tBlue_Dude: Data abort...
19:32:55bluebrotherPatrickB:"Use your real name as WikiName in the form FirstnameLastname."
19:33:11bluebrotherI don't see what's unclear about that
19:33:14Blue_DudeYeesh. Alright. Back the drawing board.
19:33:40Blue_DudeI'll take fml's advice and get rid of a couple of variables and go from there.
19:33:43S_a_i_n_tsame address as before: 0000000A (o)
19:33:55S_a_i_n_t*(0) rather
19:34:20Blue_DudeY'know, that sound's like a null pointer to a function.
19:35:26Blue_DudeOK, got to go. Back later.
19:35:49 Quit leavittx (Remote host closed the connection)
19:35:49 Quit leavittx_ (Read error: Connection reset by peer)
19:35:50 Quit leavittx__ (Read error: Connection reset by peer)
19:35:53amiconnTorne: It *is* just the archoses. The H300 (and H1x0, and...) don't use the ISD300, but the Cypress CY7C68310. It's a direct successor of the ISD300, and pin compatible, but unlike its predecessor it does support LBA48
19:36:00 Quit Blue_Dude (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
19:36:50 Join napfsuelze [0] (
19:37:00napfsuelzeCould anyone please add me to the WikiUsersGroup?
19:37:02 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:37:34bluebrothernapfsuelze: what wikiname?
19:37:55 Quit napfsuelze (Client Quit)
19:39:29bluebrotherwell, not this way.
19:39:46PatrickBbluebrother could you try my fullname then : PatrickBistelmann
19:39:47PatrickBsorry for the trouble
19:40:04 Quit fml (Quit: CGI:IRC)
19:40:16 Join fml [0] (
19:41:44bluebrotherPatrickB: you should really sign up with that name before someone else does :)
19:43:14PatrickBsee I'm not the only one who got confused apparently because here on the list are more people having only the first letter of their last names :
19:43:14PatrickBI'll try to re-register then...
19:44:43bluebrotherPatrickB: WikiUsers is everyone that registered on the wiki. Not everyone has gotten write permissions though :)
19:45:02bluebrotherthat list has to get cleaned up manually, and it's done every now and then by the admins.
19:46:02PatrickBi did register again, it should work now: PatrickBistelmann
19:46:35bluebrotherB4gder: just spotted a wiki page System/UserRegistration that looks quite different than UserRegistration.
19:46:50bluebrotherPatrickB: you should already be able to edit pages now, I've just added you
19:47:27 Quit linuxstb (Read error: Operation timed out)
19:48:15 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
19:52:00S_a_i_n_tIs it just me, or does USB transfer speed in RB seem to have taken a bit of a dive?
19:52:57 Join jgarvey [0] (
19:53:00gevaertsS_a_i_n_t: did this start around r24333?
19:54:28 Quit fml (Quit: CGI:IRC)
19:54:55 Join ale__ [0] (
19:55:03 Nick ale__ is now known as fml (
19:55:08PatrickBthanks bluebrother
19:55:25tomersbluebrother: ok
19:55:33fmltomers: hello. Have you seen my question about the read size?
19:56:15tomersbluebrother: I would like to fix the Hebrew detection in rbutil. I'm not sure I understand how to solve it
19:57:13fmltomers: another question/request: could you please comment the function is_diacritic? What do the paramsmean? And: why is the diacritic.c in the drivers directory? Shouldn't it be in common?
19:57:22 Quit PatrickB (Quit: Miranda IM! Smaller, Faster, Easier.
19:57:24 Join Strife89 [0] (~michael@
19:58:57 Quit stoffel (Remote host closed the connection)
20:02:14archivatorbluebrother, domonoky: Any pressing issues with rbutil I can help with?
20:03:02*fml thinks he's in tomers' ignore list
20:03:14domonokyarchivator: i think there is nothing really pressing, but there are ofcourse many things to improve :-)
20:04:22domonokyif you want to continue working with the TTS system,for example it could need some way for voice corrections (like its done in
20:04:37tomersfml: If I read into buffer of size X bytes, and call read(size=X), then read might read the whole buffer, i.e. return X. Then the code writes buffer[X] which is wrong, since buffer's last cell is buffer[X-1]
20:05:19 Join DataGhost [0] (
20:05:19 Quit DataGhost (Changing host)
20:05:19 Join DataGhost [0] (~dataghost@unaffiliated/dataghost)
20:05:46domonokythen there is the still unfinished opensapi thing in the tracker and maybe some UI improvements needed (like the ability to select multiple dirs for talkfile creation).
20:06:02fmltomers: but you could just do buf[read-1]=0. But this could overwrite the last read byte. Is *this* the rationale?
20:06:49CIA-5New commit by 03tomers (r25517): FS #11185 - Text viewer returns div by 0 if font is missing, renamed or corrupted
20:06:54archivatordomonoky: I see RockboxUtilityDevelopment lists "Talkfile generation for database" as a Milestone 3 entry. Is there really support in core rockbox for that or is it yet to be written?
20:07:51domonokyarchivator: that would need support in the rockbox first i think..
20:09:03archivatorHmm, is it really possible (not that I'm too keen to work on it..), seeing how the database strings are highly customizable? I thought this was one of the main reasons we want a real TTS engine in core..
20:10:26CIA-5New commit by 03bluebrother (r25518): Fix layout special case for hebrew failing for system language (FS #11180). ...
20:10:40bluebrothertomers: just fixed it.
20:11:05domonokyarchivator: i think there was some prove of concept patch in the tracker. just making all strings in the db speakable is easy. the problems start if strings are concatenated like <Artist><Album>
20:11:08tomersbluebrother: I would like to fix the Hebrew detection in rbutil
20:11:23domonokyarchivator: ofcourse a real tts would be much better.
20:11:35bluebrothertomers: btw, have you checked the comment I've posted to FS #11163?
20:11:46*bluebrother points to r25518
20:11:51tomersbluebrother: any idea how to do it? I understand the reason for the bug that you described, but I'm not sure I know the right solution
20:12:11*bluebrother feels that something is lagging pretty much right now
20:12:21fmlBlue_dude: andcouldyou please comment the hotkey settings in settings.h? I.e. what are they for, what the possible values are etc.
20:13:14bluebrothertomers: see r25518
20:13:28domonokyarchivator: so there are many parts in rbutil which could need improvement, it just depends on what you want to try :-)
20:13:49archivatordomonoky: hopefully, come autumn, we'll have that TTS engine :) Anyway, I think I'll try to make sense of the opensapi stuff.
20:14:04CIA-5New commit by 03tomers (r25519): Text viewer: Fix yellow
20:14:09wodzI have a question: in archosplayer.h I have CONFIG_CODEC MAS3507D and HAVE_DAC3550A. Functions related to DAC3550A are in dac.c and are used exclusively by functions related to MAS3507D. Is there any other reason than historicaly to separate the two?
20:14:16archivatordomonoky: I was really just asking if you guys need help with rbutil, with the new release coming up and everything.
20:14:50domonokytaking a look at opensapi could also help for the parellising, there would be no need for the sapi script anymore if we have it in rbutil.
20:15:16tomersbluebrother: thanks!
20:15:29bluebrotherarchivator: more eyes looking at the rbutil code is always a good thing
20:15:36domonokyarchivator: ah, for the release i think there is nothing urgent. Ofcourse testing for bugs is always good before releases.
20:15:38archivatordomonoky: I'm not sure I understand how it works - it's a server that connects to SAPI and a client that uses a brand-new protocol to voice things?
20:16:02bluebrotherI don't see any critical issues right now, but there's always enough work :)
20:16:22domonokyarchivator: yes, the important thing is that it also works under linux/mac with wine.
20:17:12archivatordomonoky: So what's the showstopper? Anything in particular bothering you about that v4 patch?
20:18:46domonokyi am not sure, its a long time since i looked at it last time. probably only some cleanup/bug hunting needed and the decicion if we really want it in rbutil built-in, because its pretty big.
20:19:23pixelmawodz: the MAS is not the DAC, maybe it was expected that one or the other would also be used in relation to another decoding chip or DAC respectively.
20:19:37pixelmawhich didn't happen
20:20:26archivatordomonoky: I hope those 1.3MB (total) .exes are not stripped. 'Cause otherwise that's a lot of space.
20:20:45domonokyarchivator: also if i remember correctly there where some server startup problems when i last looked.
20:20:49 Join Stephen__ [0] (~S@
20:21:54archivatordomonoky: those *could* have been fixed with my memory leaks patch - if you recall, we had some serious startup issues with festival as well. They seem to be resolved now.
20:22:11domonokytrue, someone should try it :-)
20:22:48 Part tomers
20:23:23domonokyarchivator: if i remember correctly its already upx-ed, its a Tcl program, thats why its so big.
20:23:30archivatordomonoky: btw, do we *really* want the server as a built-in resource? That seems kinda excessive..
20:23:32pixelmawodz: this is just my guess though
20:23:40archivatordomonoky: Then we should look into rewriting it.
20:23:53wodzpixelma: I see now - just reading MAS35xx datasheets to understand what I am touching...
20:24:10archivatordomonoky: upx'd Tcl app, packed into a Qt binary, all upx-d again. Ugly doesn't quite cut it.
20:25:33domonokyarchivator: jup, its not pretty. But if we dont include it, we dont have built-in sapi support on windows anymore (unless we keep both sapi interfaces).
20:26:15Torneamiconn: ah, ok
20:26:38domonokyarchivator: if you are motivated to rewrite it sure, i tried several times to use the SAPI5 api directly from C++ and always failed at that damn win32 api. :-)
20:27:33bluebrotherwhy not put it on a server and download it on request? Is it a good idea to always ship that code even if it's not used? Might get quite soms size
20:27:43bluebrotherdomonoky: w32api is ... well, interesting :)
20:28:15domonokybluebrother: putting it on the server and automatically downloading it might be a good idea.
20:28:40domonokywe could even extend that to all TTS engines and make them Qt plugins :-)
20:29:04 Join stripwax [0] (
20:29:49archivatordomonoky: I was thinking, that tcl server seems to rely on a single Tcl package. Perhaps we could could against it and then use QtScript to run the server? It's not exactly pretty, either, though. As for Win32, I have absolutely no experience with it so don't count on me there :)
20:30:56bluebrotherdomonoky: well, wouldn't make much sense for e.g. the Mac one as it's only in the Mac binary (and the code shouldn't make a major difference).
20:31:04domonokyarchivator: the opensapi should stay a external exe, or you wont be able to run it under wine (with native rbutil).
20:31:14bluebrotherand with Qt plugins we also get a problem: static binaries can't load plugins
20:32:04archivatordomonoky: ah, true, I forgot about that.
20:32:17archivatorbluebrother: they can't?
20:33:35bluebrotherarchivator: no. That's the reason for the Q_IMPORT_PLUGIN in main.cpp
20:34:09bluebrotheri.e. you can use plugins, but you have to compile them statically into the binary. But then it doesn't give any benefit having it as plugin
20:34:17 Quit mitk (Quit: Leaving)
20:34:59 Part watto
20:35:47amiconnwodz: The MAS3507D utilizes an external DAC, the DAC3550
20:35:50archivatorbluebrother: yeah, I saw it. Doesn't explain the technical details, though.
20:35:52bluebrotherdomonoky: what do you think about FS #11182?
20:36:05amiconnThe MAS3587F and MAS3539F have a built-in DAC
20:36:14 Part fml
20:36:18amiconn(actually both a DAC and an ADC)
20:37:07 Quit toffe82 (Ping timeout: 240 seconds)
20:37:18domonokyno bug :-)
20:37:33CIA-5New commit by 03tomers (r25520): rbutil: Hebrew translation
20:38:34domonokythe "disabled Targets" checkbox should only be visible in svn builds, so its fine if such targets arent saved.
20:39:21bluebrotherwell, it's also present in releases. At least right now as I haven't found a good solution how to toggle that. svn builds use the same script.
20:41:35domonokyoh, i think we should hide that in releases. so you should modify your script to add a extra define if it makes releases :-)
20:44:03 Quit Strife89 (Quit: Leaving the building.)
20:45:14CIA-5New commit by 03tomers (r25521): Add comment to is_diacritic()
20:47:43 Join PatrickB [0] (
20:49:40PatrickBso I read the wiki introductions because I want to make a page for a player(a more experienced user recommended this to me on the forum). but I can't find where I have to create the page/topic. I looked around but no idea. please help me.
20:51:31bluebrotherPatrickB: just navigate to a page that doesn't exist.
20:51:58archivatordomonoky: are we really interested in getting rid of the vbs for SAPI? We should really pick - either opensapi, or a new class to use the native API. We can't have both for obvious redundancy reasons.
20:52:35bluebrotherthe easiest (IMO) way is to simply put the page name in the URL. You could also use the "Go" box at the upper right in the wiki, type the (new) page name, and then you'll get a page that allows you creating that new page
20:52:58PatrickBthank you again bluebrother...
20:53:54bluebrotherin the latter case you can just leave the entry boxes on the page as they are (assuming your page name is correct ;-) and press the "create this topic" button
20:54:01bluebrotherPatrickB: you're welcome
20:56:25 Join moos [0] (moos@rockbox/staff/moos)
20:57:18 Join Luca_S [0] (
20:57:20 Quit Xerion (Quit: )
20:57:25 Join toffe82 [0] (~chatzilla@
20:58:22wodzarrgh... firmware/mp3_playback.c is terrible
20:58:50wodzor horrible rather
20:59:26archivatordomonoky: also, for transformations, I think we should steal wxWidgets' idea and create a SpeechTransformer chain-of-command architecture that has the individual transformers added/removed by the user (with sensible defaults).
20:59:57pixelmawodz: that's what they say about software codec playback too ;)
21:00:50wodzpixelma: I mean intermix of lowlevel, hardware specyfic stuff and more abstract...
21:01:15 Quit xavieran (Ping timeout: 276 seconds)
21:01:45 Join xavieran [0] (
21:03:59***Saving seen data "./dancer.seen"
21:06:07 Join captainewkl [0] (~2669ecc2@gateway/web/freenode/x-oawftrvwfuxckltj)
21:07:27domonokyarchivator: dont know how that works, but sounds good :-)
21:08:06archivatordomonoky: well, their thing is far too complicated (it can convert from html to XML or SSML) but I think I have a basic idea of how it should be structured.
21:08:36archivatorHowever, I can't figure out if the language setting is used for anything other than downloading the list of strings. seems to take care of several language-specific options. The transformer should be aware of the language, the engine, *and* the voice selected. Not too easy, given that we don't have a TTSSettings struct.
21:10:08 Quit toffe82 (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
21:12:57bluebrotherarchivator: optimally and the transformer would both use a separate file holding the predefined transformations.
21:13:07bluebrotherare the transformations identical between voices?
21:13:13archivatorbluebrother: no
21:13:22bluebrothertoo bad :(
21:13:36archivatorit's actually quite sophisticated and I don't think we should try to describe it in an external file
21:14:07bluebrotherwhy not? I don't like duplicating such lists ...
21:14:40archivatorWell, some of the transformations are only triggered if the language is German, the engine is SAPI, and the vendor of the selected voice is AT&T Labs.
21:15:11bluebrothercouldn't we handle this in a similar way to target-specific strings in voice files?
21:15:20archivatorHow is that handled?
21:16:20bluebrotherit's some kind of tag-based list.
21:17:24bluebrothersee for example
21:18:09bluebrotherif we go for a more xml-compatible syntax we could even use the xml parsers that are already around, like for No idea about visual basic for that, though.
21:20:36archivatorbluebrother: I thought you meant something else. well, yes, we could indeed use XML-like syntax to describe the transformations but I'm not too fond of integrating an entire parser into rbutil
21:22:23domonokyarchivator: i think there is QXml :-)
21:23:41archivatorwell, I'm not sure the .lang files are well-formed XML any way
21:23:45domonokyarchivator: language and engine is available in TalkGenerator.cpp, the voice (if any) could maybe be queryd from the tts.
21:23:47bluebrotherthere is (it's QtXml when speaking about Qt modules ;-) but not needing to include that would be a good thing too.
21:23:50 Quit stripwax (Quit:
21:24:08bluebrotherarchivator: no, the .lang files are not. There was quite some discussion about this when langv2 was introduced
21:24:35 Quit captainewkl (Quit: Page closed)
21:24:49 Quit Lear (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
21:24:53bluebrotherbut my point is simply to have an external file that holds transformation strings with some additional markup to note which setup they are to be used for
21:25:38bluebrotherthis could be as simple as a "engine:voice:transformation" format
21:25:44archivatorWell, we don't need xml in any case. We only need to find the start and end of a transformation, the criteria when it should be applied and the actual strings. That should be doable with basic string parsing.
21:26:22archivatorbluebrother: I follow. I was thinking of a better-looking format for the criteria but we're basically on the same page.
21:26:56archivatorWe would still need to expose the vendor in ttssapi.cpp - QUERY\tVENDOR\r\n seems to be the script command in question.
21:27:15archivatorNot sure if I should expose this in the GUI as well - similar to how festival handles the voice descriptions.
21:28:04bluebrotherwell, for such a simple file format we could implement a table view a user can edit. Maybe only show the tranformations for the currently selected engine / vendor pair
21:28:19 Part PatrickB ("I'm a happy Miranda IM user! Get it here:")
21:28:46bluebrotheror allow specifying an external file with additional transformations. Might be useful to get users submit the transformations for their languages
21:29:07archivatorbluebrother: I was thinking of having an online database, actually
21:29:08bluebrotherwe could even go as far as hosting user transformations similar to the theme website :)
21:29:19archivatorYes, my point exactly.
21:29:42bluebrotherit _might_ be a good idea to consider giving the user the option to override specific transformations
21:29:55bluebrotherthough I'm not really sure about that.
21:30:34CIA-5New commit by 03alle (r25522): Set SVN keywords; small addition to the comment.
21:30:55archivatorHow about we start simple - a set of predefined transformations, in an external file to be embedded as a resource.
21:31:08domonokyusersubmitted voice corrections surely need some supervision, or someone will make yes to no :-)
21:31:34bluebrotherarchivator: exactly what I thought :)
21:31:40domonokystarting simple is always good :-)
21:32:00bluebrotherdomonoky: well, we could allow the user to select the transformation he wants to use. Pretty similar to the theme installation, only that he can only select one
21:33:21domonokyyou mean a complete set of voice transformations or the single replacements ?
21:33:41bluebrotherI'm thinking of sets
21:35:29 Join Strife89 [0] (~michael@
21:36:19CIA-5New commit by 03alle (r25523): Set SVN keywords
21:37:40 Join fml [0] (
21:38:13fmlWhy is the function is_diacritic declared as returning an int (and not bool)?
21:38:38fmlAnd why is it in "drivers" and not in "common"?
21:39:38 Join drostie [0] (
21:41:45fmlThat same function already has a bool in its signature. I could understand it if we didn't want "non standard" types in common/drivers.
21:43:13gevaertsfml: bool is perfectly standard as far as I know :)
21:44:34gevaertsI agree about drivers/ being definitely the wrong place for it
21:45:54 Quit kugel (Ping timeout: 260 seconds)
21:46:36fmlgevaerts: I think the bool as the return type would more clearlyconvey what the function does.
21:46:46gevaertsit would, yes
21:46:52*bluebrother agrees with fml
21:47:12 Quit archivator (Quit: "No message")
21:48:52fmlSo would someone change that? I can commit but can't make a verifying build.
21:49:43fmlgevaerts: is bool really a standard type? In C? I thought it must be #define'd.
21:50:10gevaertsfml: it's #defined, but the spec says how it must be defined
21:51:23fmlgevaerts: so it's not part of the language but part of ... what?
21:51:48gevaertsthe standard headers
21:51:57bluebrotherfml: bool is part of C99.
21:52:22bluebrotherunfortunately, most C compilers do only support C90. Which includes the Microsoft compiler
21:52:37fmlOK. Now I'm a C99 expert! :-)
21:53:20bluebrotherthat created the interesting effect that some code I was trying to compile in VS did compile in cpp mode but not in c mode (it was parts of the MTP dll for beastpatcher)
21:53:20gevaertsfml: if you include stdbool.h, you get bool, which is defined as _Bool, which *is* a type
21:54:14gevaertsIt also defines, true, false, and __bool_true_false_are_defined
21:55:36fmlShould I file a bug report about diacritic.c being in drivers so that it doesn't get forgotten?
21:58:03bluebrotherfml: wouldn't be wrong, unless you're planning to do it yourself shortly :)
21:58:07 Join kugel [0] (~kugel@rockbox/developer/kugel)
21:58:43fmlNo, I'm not planning it because I can't verify the change (have no buildenv)
21:59:28CIA-5New commit by 03jethead71 (r25524): Gigabeat S: Add some sanity checks for a strange charging anomaly that I have personally witnessed twice−− no, I don't have photos or a YT video but ...
21:59:34bluebrotherthen I'd add a task for it. We agreed that it needs changing :)
22:00:10 Quit xiainx (Quit: Good Bye)
22:00:37 Quit n17ikh (Remote host closed the connection)
22:01:25CIA-5New commit by 03torne (r25525): Make ATA code not bother to retry requests that return IDNF (specified sector not valid). ...
22:02:07 Quit Strife89 (Quit: Changing buildings again.)
22:02:35kugelTorne: are all ata-target.h files the same?
22:02:38fmlbluebrother: FS #11187 is at your disposal ;-)
22:02:48Tornekugel: no
22:03:03Tornekugel: they're pretty similar, admittedly :)
22:03:22bluebrotherfml: I won't fix it, I'm a rbutil guy ;-)
22:03:33Tornekugel: but there's some byte order thing that means taht one of the cpus ahs different values, and the base registers are at different addresses, etc
22:03:50 Join xiainx [0] (
22:04:19kugelbut these error values are the same on all?
22:05:23Tornekugel: it's always the same bit, it's from the ATA spec, but the coldfire version has them in the top byte of a halfword for some reason
22:05:41kugeljust thought the common parts could be in a common file ;)
22:05:48fmlbluebrother: I hope tomers reads logs then.
22:05:50kugelata.h for example
22:05:50TorneYeah, the common parts are really not very big
22:06:03TorneSomeone could probaly tidy it up but I just wanted one more bit ;)
22:06:18bluebrotherfml: me too
22:06:41Tornehow on earth did my commit add 400KB on vibe 500
22:06:57Tornethis is nonsensical :)
22:07:12gevaertsTorne: the usual build system bug
22:08:32gevaertsIf you read perl fluently, you might want to have a look
22:08:43Torneer no
22:08:50Tornei read perl only at gunpoint
22:09:25kugelnobody reads perl, only some write in it :)
22:10:42pixelmait's an Ondo vx747 build in the Vibe500 zip
22:10:49pixelmaOnda too
22:12:10 Join amr [0] (~quassel@
22:19:40fmlbluebrother: are you also a manual guy? Then you could try FS #11188
22:22:29 Quit Topy44 (Ping timeout: 276 seconds)
22:23:25 Join Adubb [0] (~aldubuc@
22:24:02amrjust want to know, how to repoen a closed task on flyspray ? thanks
22:24:36gevaertsamr: ask here :)
22:25:59amrdo I need an extra permission ?
22:26:14 Join toffe82 [0] (~chatzilla@
22:27:12bluebrotherfml: I did manual work in the past, now rather rarely.
22:27:57fmlI've already done the work, you could just check it
22:28:44fmllike "make manual"
22:29:37bluebrotherfml: will try :)
22:31:03amror should I create a new task again ?
22:31:53 Quit TopyMobile_ (Ping timeout: 246 seconds)
22:32:18gevaertsamr: which task is it?
22:32:26fmlamr: just tell us what task you're talking about and why do you want to reopen it
22:32:54 Quit merbanan (Ping timeout: 246 seconds)
22:33:37amror alterntatively this:
22:34:11gevaertsWell, I think we should only reopen one :)
22:34:29gevaertsSo that still happens?
22:34:56amryeah, I'm going to explain the case more in the comment
22:35:18gevaertsOK, I reopened FS #11185
22:35:46kugelfml: " And why is it in "drivers" and not in "common"?" -> because there's no clear distinction between them
22:35:56amrok thanks
22:35:56kugelat least no commonly known one :P
22:36:15jhMikeSthose builds are sure quick now, which I'm not complaining about. kinda nice to have it nearly done before I even get over the table to check it. :)
22:36:17gevaertskugel: well, can you give one reason for "drivers"? :)
22:36:36kugelI guess part of the raaa gsoc project will be to clean that up, so that linking OS libraries is easier
22:36:45bluebrotherfml: FS #11188 builds fine
22:37:55fmlbluebrother: does it also look decently (I mean the PDF)?
22:39:14jhMikeSwhat's the deal with the monster-sized delta anomalies? (~400K) ??
22:39:16kugelgevaerts: well, it contains high-level drivers, but also a lot of other stuff
22:39:49gevaertsjhMikeS: a bug in the build system. Every now and then it mixes up targets
22:40:01pixelmajhMikeS: wrong builds in the zip
22:40:03gevaertsIf you read perl, please help :)
22:40:05fml kugel: IMO diacritics.c is of the same type as ctype
22:40:47kugelfml: except that ctype is ansi c, diactritics is rockbox specific
22:40:51kugelbut generally, yes
22:40:59jhMikeSso long as it all works out in the end! :P
22:41:01gevaertskugel: I'd argue that it *only* has drivers, apart from diacritics.c
22:41:21CIA-5New commit by 03tomers (r25526): FS #11187 - diacritic.c is in 'drivers' but it does not belong there ...
22:41:44bluebrotherimo it's ok:
22:41:57 Quit Luca_S (Quit: CGI:IRC (EOF))
22:42:26kugelit's also a mix of high level drivers and shared low-level ones, a bit confusing
22:42:40kugelsome high level drivers are not in drivers/, too
22:43:27gevaertsOh, things can certainly be improved
22:43:50fmlbluebrother: and the link points to the right place? If yes, could you commit it?
22:45:21bluebrotherfml: haven't checked the links in detail, I trusted you ;-)
22:45:34*pixelma wonders when her theme stopped working correctly on an M3 sim (the monochrome bitmaps don't show up correctly)
22:45:35 Join TopyMobile_ [0] (
22:46:40fmlbluebrother: I copied the reference code from anotherplace in the hope that it's correct. But I can't check it here. Can you just click it and see?
22:48:42 Part fml ("bye")
22:48:43 Quit bmbl (Quit: Bye!)
22:53:12CIA-5New commit by 03bluebrother (r25527): Insert links to the explanations of hotkey where it's mentioned. ...
22:53:34 Join Adubbb [0] (~Aldubuc@
22:54:04 Quit Adubb (Quit: "The mind and body must be subjected to extreme stimulus, by means of drugs and music." - HST)
22:54:20 Quit fyre^OS (Quit: Ur skills' fireproof like a wooden panel -- U got feds talking leet on your IRC channel!)
23:04:01***Saving seen data "./dancer.seen"
23:10:14wodzHere is my attempt to cleanup codebase a bit. I did my best to separate mas35xx lowlevel stuff from sound.c. FS #11189. Any comment welcome
23:10:35 Quit GeekShadow (Quit: The cake is a lie !)
23:14:31pixelmahmm... shouldn't the "mandelbrot" screenshots in the manual also be renamed to "ss-fractals-something"
23:16:04 Quit liar (Quit: bye)
23:16:32pixelmaalthough they still show the Mandelbrot set
23:21:10wodzpixelma: could You try my patch on Your Ondio?
23:23:17 Quit TopyMobile_ (Ping timeout: 264 seconds)
23:24:20kugelwodz: ah, I like these kind of patches
23:29:56 Join TopyMobile [0] (
23:30:48 Quit soap (Quit: soap)
23:31:14pixelmawodz: sure, might take a bit (building and then unzipping to the Ondio takes ages). And I'm a bit tired so results more likely tomorrow
23:31:34 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
23:32:25wodzok, time to sleep, bye
23:32:32 Quit wodz (Quit: Leaving)
23:33:15 Join n17ikh [0] (
23:35:31 Quit parafin (Quit: So long and thanks for all the fish)
23:35:32 Join soap [0] (~soap@rockbox/staff/soap)
23:35:36 Quit TopyMobile (Ping timeout: 240 seconds)
23:35:51 Join parafin [0] (
23:36:03 Join Topy44 [0] (
23:39:43 Quit Zarggg (Read error: Connection reset by peer)
23:42:49 Join Zarggg [0] (
23:45:17 Join TopyMobile [0] (
23:45:57 Quit soap (Quit: soap)
23:46:58 Quit Topy44 (Ping timeout: 276 seconds)
23:49:53 Join Topy44 [0] (
23:52:02 Join Topy [0] (
23:53:33 Quit Topy44 (Client Quit)
23:55:08 Join rd1381 [0] (~rd1381@
23:56:16 Join anewuser [0] (anewuser@unaffiliated/anewuser)
23:56:17rd1381is it me or rockbox file transfer spped is usb1?
23:56:32Torneit's you
23:56:55Tornewe may indeed be slower than the original firmware on some players, but it's not anything to do with usb1/2
23:56:56rd1381i try with rockbox and spped is low
23:57:07rd1381i try with ipod firmware and spped is high
23:57:23Torneyes, on some ipod models our disk access runs slowly for some reason
23:57:30Tornewhich ipod?
23:57:34rd1381with rockbox it maxes out at - mB/s
23:57:54rd1381ipod 5.5 30gb
23:58:05rd1381ipod video 30gb
23:58:06gevaertsWell, usb1 would max out at about one megabyte per second
23:58:20rd1381so its not usb 1 but very low
23:58:34TorneOur implementation is just not very fast on some models

Previous day | Next day