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

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

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

#rockbox log for 2009-06-16

00:00:58 Join robin0800 [0] (
00:03:08 Quit jordan` (Read error: 60 (Operation timed out))
00:03:23 Join gregzx_ [0] (
00:04:15 Join jordan` [0] (i=gromit@
00:04:57Unhelpfulis there now, or is there likely to be soon, a bigendian armv6 target? :)
00:06:02funmanUnhelpful: if you have a armv6 target you can configure it to be big endian
00:07:39Unhelpfulfunman: i'm on my third version of ARM asm 4-point IDCT. it doesn't pay to load two 16-bit values together on ARMv5, not at this size, anyway, but it looks like that will pay off all the way down to 2-point on ARMv6. as soon as i start loading multiple values into one register, though, i need to be endian-aware - unless it's not going to matter. :)
00:09:01amiconnJust write it for little endian atm. Leave a note that this is LE only, and needs fixing should a BE ARM target be included
00:09:07funmanendianess can be configured in the system control coprocessor, so i perhaps if such ports are made we could switch them to little endian
00:09:33amiconnI guess that many ARM asm blocks silently assume LE
00:10:11Unhelpfulwell, SWAR was too slow on ARMv5. interestingly, the IDCT transpose is faster *in pure C* at 8x8.
00:10:40funmanthen use gcc -S and optimise that :P
00:10:56Unhelpfuli've made it optional, i'll enable it for ARMv6 if it's a benefit on 4x4.
00:11:17amiconnIf gcc manages to be faster than hand-coded asm, you're doing something wrong...
00:11:52Mikachuthey removed some asm from glibc where gcc did better from the c code
00:14:25Unhelpfulamiconn: no. i meant that with the pure C code, i've worked up an option to transpose coefficients during unpack, have the vertical IDCT read in rows from that and output columns to a second buffer, and then the horizontal IDCT reads and writes rows.
00:14:51Unhelpfuland that is faster, in C, than essentially the same code without the transposes, in C.
00:14:59kugelshould the fixed bugs list in ReleaseNotes33 only contain bugs that appeared in 3.3 or in current builds inbetween too?
00:15:22Mikachuyou mean 3.2?
00:15:36BagderI think it should mention bugs fixed in 3.3 that were present in 3.2
00:15:40Unhelpfulmemory locality issues, i would assume, as i doubt that gcc is doing anything terribly clever.
00:15:51*gevaerts still has to document the full USB situation in ReleaseNotes33...
00:16:05kugelyes, 3.2 sorry
00:16:07gevaertsBagder: what's the exact release planning?
00:16:27BagderI don't have any exact planning, I play by ear!
00:16:41Bagderbut thursday we said before
00:16:43BagderI think
00:16:45kugelso no bugs that came up between the releases?
00:17:12 Join HellDragon [0] (
00:17:34gevaertsok. I wasn't sure if it was wednesday or thursday.
00:17:46*gevaerts actually wasn't even sure what *today* was...
00:17:50kugelI think that would a) make the bug fixed list bigger (which looks nicely :P ) and b) also appropriate since quite a few people use current builds after a few weeks or months (particularly this time due to usb)
00:19:26gevaertskugel: I think we should try to get to a system where we can expect that people who don't use the release follow things more closely, so the don't really need that list. I know we're not there at this moment, but...
00:20:01 Quit gregzx (Read error: 110 (Connection timed out))
00:20:33kugelI think that's no reason to not have fixed bugs in release notes
00:20:43bluebrothermaking the list of fixed issues isn't necessarily a good thing −− release users are less likely to read it if its longer.
00:21:09JdGordon|they will read the fixed list
00:21:19 Quit matsl (Read error: 110 (Connection timed out))
00:21:23gevaertsthe list is meant for release users. They don't care about bugs that were present for two days in trunk and that they never saw
00:21:47kugelwe should probably remove fixed bugs in rbutil though if we don't plan on paralell releases (also, rbutil has its own release notes)
00:22:03gevaertshm, indeed
00:22:33kugelgevaerts: I wouldn't sign that statement
00:23:01kugelpeople that want stable builds aren't automatically those which don't care about developement
00:23:26bluebrotherhow's the state of this "numerical sorting" thing? Unless I've missed something there hasn't anything changed since 3.2, thus the same arguments apply and it should get disabled in 3.3 (again).
00:24:08bluebrotherkugel: if those care about development they can follow the svn activity.
00:24:18gevaertskugel: people who care about development shouldn't need the release notes to keep up to date
00:24:23kugelwhich is not exactly an overfew about fixed bugs
00:24:55kugelbluebrother: I asked on the -dev list. Nobody raised his voice against enabling it
00:25:16gevaertsWhy do we need a huge list of fixed bugs?
00:25:54bluebrotherhow many user are only using release builds, not following svn activity and _still_ interested in the intermediate bugs? As much as the morse-code section? Or the hwcodec ones?
00:26:19Mikachumy guess is there are some bugs reported against the svn builds that are also present in 3.2
00:26:43gevaertstrue. In an ideal world we should list those
00:26:51gevaertsI'm not sure if it's practical though
00:26:54bluebrotherkugel: can't find that email. Was that in the last couple of days?
00:27:02kugela few weeks ago
00:27:47gevaertsSo yes, if you're fairly confident that a bug was in 3.2 even while it's not been reported against 3.2, by all means add it to the release notes
00:27:48kugelMay 27th
00:28:19kugelok, so only 3.2 bugs
00:30:22kugelI think we should leave it enabled
00:30:25 Quit ender` (" A man without religion is like a fish without a bicycle. -- Vique's Law")
00:31:01 Part Jax184
00:31:12Unhelpfulhrm. do we have a document somewhere with cycle counts for add16?
00:32:03saratogathats an arm5 instruction?
00:33:22kugelUnhelpful: are you going to keep a equivalent C version ?
00:33:37amiconnThat's a requirement
00:33:46kugelit's always nice to have that, although IDCT is above many heads anyway
00:34:08amiconnWell, unless he wants to write versions for all target architectures, plus all potential sim architectures
00:34:49amiconnUnhelpful: Timing on armv6 is quite complex, as it depends on when you're going to use the result (pipelining)
00:35:06kugelwell, I mean equivalent in the means of optimizations which come in during the asm'ifying
00:35:21kugelotherwise he could just keep the current c version
00:35:24amiconnGet the arm1136jf-s technical reference manual from
00:35:43amiconnSome optimisations are impossible to express as C code
00:36:05kugelsure, as possible of course
00:36:17 Quit PaulJam_ (".")
00:36:39 Join stripwax [0] (
00:36:40amiconnThat's one reason why asm is often faster than C. It allows to exploit all the special instructions which don't translate to C
00:37:01CIA-37New commit by saratoga (r21304): Remove unused variable and assignment.
00:37:47amiconnOne very simple example is if you want to test all bits of a word sequentially. In C you need to test the lsb, then right-shift
00:38:20kugelno rotate in c, yea, sad sometimes
00:38:35Mikachuhow would rotate be an optimization over shift?
00:38:42amiconnIn asm, you can just right-shift on most processors, as the shifted-out bit automatically ends up in the carry flag (or equivalent)
00:39:05 Quit jgarvey ("Leaving")
00:39:08 Join webtaz [0] (
00:39:16Mikachuand gcc doesn't do that optimization? :)
00:39:26kugelthat too, I worked that way in a recent 8086 asm course
00:40:10webtazdo i need to compile a new bootloader in order to use a new build?
00:40:13 Join leachim6 [0] (
00:40:18 Quit bertrik (Read error: 113 (No route to host))
00:40:18saratogausually not
00:40:18webtazon the fuze*
00:40:23amiconnSame thing works for lsb and left-shift (if you need the opposite bit order)
00:40:35leachim6I've got the latest rockbox installed...and I found this awesome plugin for for lyrics ->
00:40:36amiconnErr, msb
00:40:40leachim6is that a trivial or nontrivial isntall?
00:40:57saratogaleachim6: check the guide to patching on the wiki
00:41:06leachim6I knew you'd say that
00:41:11leachim6I don't wanna know how to do it
00:41:17leachim6if it's nontrivial I'm not even gonna attempt it...
00:41:25leachim6because I have like 5 minutes to work now
00:41:30 Quit mcuelenaere ()
00:41:33saratogaits going to take longer then that
00:41:49saratogaunless you've already got the tools installed
00:42:04***Saving seen data "./dancer.seen"
00:42:12 Quit bluebrother ("leaving")
00:42:14webtazand now how compiling works
00:42:56leachim6oh crap...
00:43:04leachim6I need to have the special arm compilers and such...
00:43:13leachim6I have patch and gcc and all that jazz...
00:43:44kugelsaratoga: thanks for fixing the red btw. I obviously wasn't awake enough :)
00:43:59leachim6so I'm gonna need the entire rockbox tree...
00:44:03saratogano problem
00:44:04leachim6...and the entire toolchain
00:44:17leachim6...once I've got it all compiled...can I flash it via rbutil ?
00:44:18saratogaleachim6: you don't need to tell us these things, we already know
00:44:20krazykitleachim6, you don't really need to think aloud in the channel
00:44:30leachim6sorry about that...
00:44:38leachim6they were kind of questions though...
00:44:45webtazi know it is not supported, but if i try new builds on my fuze (which works withmy olderbuild) it trys to start but hangs after 7 lines(loadign firmware; length; checksum; model name; loading rockbox.sansa sum; executing)
00:44:54webtazany idea why?
00:45:12Mikachuleachim6: it's been 5 minutes now
00:45:28funmanwebtaz: yes, because it's unsupported ..
00:45:36leachim6since I entered the channel?
00:45:45Mikachusince you said "i have 5 minutes"
00:45:53funmansome people have put builds available on the web so you shoud get support from them
00:45:54saratogahas the fuze bootloader changed?
00:45:56webtazI'm sure this is not the reason.... just an answer...
00:46:10evilnick_homeleachim6: By all means ask questions, but as this channel is logged for the developers it'd be best if you could keep from the internal monologue!
00:46:34funmansome people needed to update the bootloader when data cache was enabed
00:47:02webtazok, so i try compiling a new one
00:47:40leachim6evilnick_home, sorry :/ my bad...
00:48:07krazykitleachim6, you don't "flash" a new build, but i believe rbutil can install from a provided
00:48:19krazykitor you could just unzip it yourself
00:48:25leachim6that's true...I forgot...
00:48:29leachim6sorry :(
00:48:37leachim6I'm being stupid today...and rather annoying I gather...
00:52:53leachim6I searched the wiki for how to apply a patch ...
00:52:57leachim6and I got nothing
00:55:19 Quit funman ("leaving")
00:55:31saratogaI searched for "patch" and it came up
00:55:40kugelgevaerts: the list should be complete now
00:55:42leachim6ohh...I search for "how to apply a patch"
00:55:53stripwaxsaratoga - so, when you mentioned that the mdct in lib could be optimised to take advantage of 'middle symmetry' - what did you have in mind?
00:56:05*stripwax says hello by the way - hi!
00:57:18saratogaleachim6: in general when using a search engine, you should not ask it questions, but rather type a list of keywords
00:57:27saratogaits just a computer it won't understand english
00:57:52leachim6...I know just has been a long day...I'm normally a relatively proficient user
00:57:55stripwax(context: I think you said we're doing twice as many multiplies as we need to)
00:58:02leachim6...and am usually good at formulating queries ...
00:58:05saratogastripwax: the MDCT returns 2N samples, but really you only need N since half are duplicated
00:58:38saratogaso we could get rid of half the samples and save about 1 fixed point mul, a few loads/stores, and 16KB of IRAM on all MDCT targets
00:58:41stripwaxsaratoga - yep, and I think the mdct lib shifts the N into the first half of the output array.
00:59:13saratogai think the speed up would be fairly small, but at least we'd save some IRAM
00:59:59leachim6I never realized there were so many cool patches for rockbox...what patches do you guys use ?
01:00:04saratogai think it would be really easy to add this to our MDCT, since we could just remove the last couple lines that duplicate the entries
01:00:14saratogabut changing the individual codecs i harder
01:00:14 Join lilltiger [0] (n=lilltige@
01:00:28stripwaxdo we not need 2N-size input, and operate in-place in IRAM? not sure how we'd save IRAM in that case
01:00:44stripwaxAgree, don't need the lines that duplicate entries, codecs surely not using those (I'd hope)
01:00:50saratogaI don't know about vorbis, but WMA at least does not operate in place
01:01:05saratogai spent a long time trying to make it do so but gave up, without realizing this symmetry existed
01:01:08lilltigerhi, i read that rockbox now partly works on the Fuze, but does anyone know how well it works?
01:01:11stripwaxthen yep, definitely
01:01:24saratogalet me look at vorbis
01:01:31stripwaxvorbis operates inplace
01:02:40stripwaxif there's enough IRAM we now actually have two iram buffers and flip between them (rather than memcpy anything)
01:03:05stripwaxwhich is actually super wasteful but gives a bit of a speedup
01:03:46soapPlease excuse me if this has already been said - but does anyone else thing the LUA mention in the 3.3 Release notes should be clarified? If I read that line w/o prior knowledge I would assume the Rockbox interface can be LUA scripted...
01:04:32saratogastripwax: the imdct is in place or the windowing + imdct are in place?
01:04:47kugelsoap: I don't disagree
01:05:02kugelthe whole page is work-in-progress still
01:05:17 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
01:05:26Mikachujust say plugins can be written in lua?
01:05:27stripwaxthe imdct is in place. the windowing adds (a+b) to either a or b depending on which would result in fewer multiplies
01:05:59stripwaxat least I'm pretty sure the mdct is in place (I think we call with the same arg for input and output buffer)
01:06:12safetydansoap: it could be clearer. Maybe "Rockbox now supports plugins written using the LUA language."
01:06:39safetydanYeah I know they're not strictly plugins, but it's sort of the same concept
01:06:49saratogastripwax: the imdct is always in place
01:06:53stripwaxsaratoga- sorry, the windowing is in place; the overlap doesn't use a third buffer
01:06:57leachim6"rock" is the rockbox terminology for plugin correct?
01:07:03leachim6a "rock" and a plugin are the same thing
01:07:08saratogathird buffer? does it use two?
01:07:19saratogarock is the file extension used for plugin files
01:07:54stripwaxsaratoga - for lapping, we have A (in iram if possible) which is previous frame, and B (always in iram), and we add either A to B or B to A.
01:08:04leachim6saratoga, so that's a yes then...
01:08:06soapit's a fine matter of semantics. Would saying "Addition of a LUA interpreter plugin" be better than how it is phrased now?
01:08:09stripwaxat this point, both A and B have already been windowed
01:08:21saratogastripwax: if I understand you correctly, thats what WMA does now
01:08:50stripwaxthought you said wma doesn't do mdct in place?
01:08:56saratogabut I think it should be possible to have one buffer that holds the previous block and holds the iMDCT, though it'll need to be 2N samples big
01:08:56kugelfunman: (for the logs) backlight on GPIO on c200v2? The backlight works exactly like on the other targets IIRC (powering the dcdc)
01:09:14stripwaxyou'll also need at least N additional working space
01:09:32saratogaright now WMA uses an 2N sample buffer to hold the previous and present blocks, and another N sample buffer for the iMDCT
01:09:53stripwaxactually that's what vorbis used to do- 2N iram buffer plus N buffer for previous frame
01:10:19 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
01:10:25stripwaxwe only recently increased the vorbis iram usage [if enough iram available]
01:10:41stripwax(that commit of mine :-)
01:10:46saratogahow much memory does it use now for N sample blocks?
01:10:58saratogaincluding iMDCT working space, windowing, and previous frame?
01:11:33stripwaxdepending on bitrate, either 2N + 2N in iram, or 2N in iram plus N in dram. I think.
01:11:51stripwaxdepending on bitrate (which influences N) and iram availability that is
01:11:53saratogayeah so thats basically what WMA does with the added optimization to avoid memcpy
01:12:04saratogaN is independent of bitrate actually
01:12:22saratogaor at least it should be unless vorbis encoders are very strange
01:12:31saratogausually its determined by the sample rate
01:12:51stripwaxI *thought* N could depend on bitrate but sample rate sounds more likely
01:13:30stripwaxVorbis encoders might *also* be very strange :)
01:13:44saratogausually you pick how long you want your long and short blocks in milliseconds and then adjust the window sizes to achieve that at each sample rate
01:14:10leachim6how long does it usually take to run
01:14:18leachim6I have a 3.0ghz p4
01:14:20saratogaminutes to hours depending on your CPU
01:14:29stripwaxmakes sense. so I'm still trying to work out which multiply we could save but I'll sleep on that ..
01:14:31leachim6speak of the devil...done
01:14:44saratogastripwax: easy
01:15:04saratogayou save one in windowing, since right now we window both halves of the symetric data
01:15:07stripwaxAs for getting rid of the last loop to reorder the results - I think that's what the lomem branch for Tremor does
01:15:56stripwaxsaratoga - I don't think we do (for vorbis at least). We only window the first N samples not all 2N of the current frame
01:16:15kugelsaratoga: didn't you work on backporting tremor code to upstream?
01:16:18kugelsaratoga: didn't you work on backporting tremor code to upstream?
01:16:18stripwaxor do you mean we only need to window one buffer and not the other?
01:16:35saratogastripwax: you window half of each block then window the other half on the next pass
01:16:43saratogathe forward and reverse multiply loops in tremor
01:16:59saratogakugel: yeah thats on my todo list still, just been busy with other things
01:17:05stripwaxsaratoga - right, but with a (potentially) different shaped window.
01:17:36saratogai think this trick will still work even if different windows are used, though you'll need one extra load if you use two windows
01:17:48saratogalet me pastebin something
01:18:51 Quit tessarakt ("Client exiting")
01:18:55saratogadoes vorbis use two different windows?
01:19:03stripwaxyep, short and long.
01:19:14saratogathey're still the same window
01:19:22saratogai thought you meant two windows like AAC with Sin and KBD
01:19:43stripwaxah, no. same window function but different size. anyway, I gotta go sleep now, so I'll check logs in the a.m.
01:19:52saratogahold on i'll give you code showing how its handled
01:19:53stripwaxdifferent size => different window factors
01:19:56 Quit stripwax ("")
01:20:53saratogaanyone care if I fix tabs in tremor?
01:22:18saratogahmm it does look like lowmem tremor is already using this optimization
01:22:20kugelI thought we don't touch imported code
01:22:34Unhelpfulamiconn: i'd be surprised if add16 without the saturate option is slower than add, although it might be. a partitioned add or subtract amounts to cutting the borrow or carry logic between certain bits.
01:22:54saratogain this case we broke tabs by converting parts of some files to use 4 spaces and others to use 2 spaces
01:23:06saratogai think its worth making them consistent
01:23:23leachim6ok...I got all the stuff setup...the environment is in place...I've got my patchfile in the right spot and included it in SOURCES...
01:23:25leachim6now what?
01:23:31saratogaxiph is kind of stupid with this lowmem tremor crap, they should have just fixed normal mem tremor
01:23:33leachim6it just says "Build rockbox again" in the readme
01:24:00saratogawhat step are you at
01:24:07 Quit amiconn (Nick collision from services.)
01:24:09 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
01:24:12leachim6all I need to do is compile
01:24:14 Quit pixelma (Nick collision from services.)
01:24:14 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
01:24:15leachim6everything is in place
01:24:25Unhelpfulkugel: the asm is using *exactly* the same math. there are definitely parts i'm doing in ASM that aren't really possible in C, like doing two adds at once (unless you use vector extensions), using the arm MULxy instructions that let you select either the top or bottom of each input register for a 16x16->32 multiply. many things can not be done in C. only the math is the same.
01:24:27saratoganot everything or you wouldn't be asking
01:24:31 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
01:24:34 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
01:24:39leachim6saratoga, well I just need to know what command I need to run from where
01:24:55saratogaso what step are you at?
01:25:04leachim6...the...building? step?
01:26:27leachim6I'm in the root of the "Rockbox" folder I checked out of svn..and I moved my patch file into place in the hierarchy into apps/sncviewer.c and added it to SOURCES per the documentation
01:26:42leachim6the documentation says to "Build Rockbox" and i wanna know what command I run from where to build rockbox
01:26:43saratogahave you patched the source yet?
01:26:51leachim6haha is that what you were asking?
01:27:05leachim6it's not really a patch doesn't patch's actually just a standalone c file
01:27:31saratogathen you need to compile
01:27:36saratogafollow the instructions I linked you before
01:28:00leachim6would you mind linking me again? I seem to have lost it
01:28:01 Quit robin0800 (Remote closed the connection)
01:28:06saratogacheck the logs
01:28:39leachim6...yeah I know...but a lot of links were flying around here
01:28:47leachim6you mean the WorkingWithPatches link? because there is no help there...
01:29:13Mikachuyou are allowed to click the links on that page
01:29:27saratogai'm stepping out good luck
01:30:05 Quit Thundercloud (Remote closed the connection)
01:30:21webtaz <- try this
01:31:10 Join barrywardell [0] (
01:31:29webtazi get now (with new bootloader and new build) the error "no .rockbox directory" and "installation incomplete"
01:32:10webtazi'm sure the .rockbox directory is on the player (root)
01:32:22webtazwhat could be the reason for this?
01:34:22webtazrb then apparently can't access the filesystem (weird filenames in the file-menu, no plugins ...)
01:37:48leachim6ok I got it, thanks for you guys' help
01:39:15 Join goffa_ [0] (n=goffa@ I tried to compile the simulator...and I get this error ->
01:41:32leachim6this is with the bleeding edge svn
01:43:32kugellooking at pondlife's fix
01:43:44kugelit looks like the timestrech buffer is always allocated now?
01:45:21Mikachuleachim6: run this command: file /media/data/rockbox/devenvironment/apps/plugins/sncviewer.c
01:47:21 Join intrados_ [0] (
01:47:34kugelit should covered with global_settings.timestrech_enabled, shouldn't it?
01:50:33 Nick fxb is now known as fxb__ (
01:51:07 Quit goffa__ (Read error: 110 (Connection timed out))
01:52:25 Join webtaz_ [0] (
01:52:41 Quit webtaz_ (Client Quit)
01:55:41 Join kitas [0] (n=5b27783a@gateway/web/cgi-irc/
01:56:16 Join midgey|w [0] (i=836b004d@gateway/web/ajax/
01:56:40 Quit barrywardell ()
01:59:08kitaswebtaz: the internal file system on some AMS sansas seems to be broken since mmu and cache are enabled
02:00:57kitasall before r21228 should be functional, but slow
02:02:41 Quit fyrestorm (Read error: 104 (Connection reset by peer))
02:05:06 Join tempnick [0] (
02:06:11 Quit kitas ("CGI:IRC (Ping timeout)")
02:06:18 Join webtaz_ [0] (
02:07:34 Quit leachim6 (Read error: 104 (Connection reset by peer))
02:08:17 Quit intrados_ (SendQ exceeded)
02:08:21webtaz_new compiled builds produce the error: "no .rockbox directory" and "installation incomplete"
02:08:30webtaz_why is it like that?
02:09:10 Join kitas [0] (n=5b27768f@gateway/web/cgi-irc/
02:10:02 Quit gregzx_ ("ChatZilla 0.9.84 [Firefox 3.0.11/2009060215]")
02:10:22kitassee forum
02:11:06 Quit webtaz (Read error: 110 (Connection timed out))
02:11:11kitasfirmware can't read the files in .rockbox directory
02:12:13 Join webtaz [0] (
02:14:05 Quit intrados (Connection timed out)
02:16:57kugeltimestrech is broken again :(
02:18:27 Quit tvelocity[a] (Read error: 113 (No route to host))
02:18:38 Quit linuxstb (Read error: 113 (No route to host))
02:19:06 Part toffe82_
02:21:34 Join webtaz__ [0] (
02:21:37 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
02:22:01Unhelpfulhrm, sadd16 is supposed to have a result latency of 1. how can this *possibly* be slower...
02:24:14Unhelpfulis ldmia perhaps slower for loading only two registers than doing two loads?
02:29:33 Quit webtaz_ (Read error: 110 (Connection timed out))
02:29:37saratogais anyone familar with the text editor? Yoshihisa Uchida's patches should probably be commited
02:29:38 Join yesssss [0] (
02:31:20yessssshey, can I get wiki write permissions? theres an error in the LinuxCompilingGuide page
02:31:33 Join webtaz_ [0] (
02:31:35 Quit webtaz_ (Client Quit)
02:32:18saratogasure whats your name
02:32:32yesssssMeyer Merenstein
02:33:31 Quit midgey|w (" ajax IRC Client")
02:34:08yessssscool thanks : D
02:35:36 Quit Lss (Read error: 104 (Connection reset by peer))
02:36:00 Quit yesssss ("leaving")
02:39:00 Quit webtaz (Read error: 110 (Connection timed out))
02:42:05***Saving seen data "./dancer.seen"
02:47:42kugelsaratoga: I thought about looking at it further too
02:47:52kugelsome seem to be useful
02:49:12saratogai hate to let good patches rot
02:49:15saratogaparticularly for plugins
02:49:34saratogai'd favor giving people like him who are interested in maintaining plugins svn access
02:49:36 Quit webtaz__ (Read error: 110 (Connection timed out))
02:50:35JdGordon|which patches are you talking about?
02:51:45kugelsaratoga: if he at least wouldn't need google translator or something :/
02:52:25saratogasome combination of 9855, 9892, 9893, 9898, 9902, and 9990
02:52:31saratogayeah the language thing is a problem
02:53:26 Quit JdGordon| (" ajax IRC Client")
02:54:30 Quit Ridayah (Read error: 60 (Operation timed out))
02:54:56kugelalthough he's quite patient and willing to edit patches in a way we could like them
02:56:34saratogaas long as hes asking before commiting things to core I don't see much problem with him working on plugins
02:59:12 Quit efyx__ (Remote closed the connection)
02:59:15saratogahmm maybe that optimization i suggested to stripwax won't save as much memory as I thought
02:59:53CIA-37New commit by kugel (r21305): Get rid of tdspeed_enabled in struct dsp_config and use the global_settings member instead. That one needs to be checked in tdspeed_init() as well, ...
03:01:53kugelthe svn server is slooooow
03:06:03 Join Ridayah [0] (
03:08:42Unhelpfulamiconn: if i understand this correctly, as long as i have the operands ready one instruction before it, and can do other things for two instructions after it, SMLAxy with an appropriate constant loaded is faster than SXTH + ADD. insane.
03:09:44 Quit perrikwp (" ajax IRC Client")
03:10:21 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
03:20:11 Quit kitas ("CGI:IRC")
03:23:22kugelinteresting arm binsize increase again
03:26:54 Join barrywardell [0] (
03:27:39 Quit kugel (Remote closed the connection)
03:37:11 Quit perfectdrug ()
03:41:29 Nick tempnick is now known as intrados (
03:45:20 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
03:52:36 Join Blue_Dude [0] (
03:52:58Blue_DudeHi. I just posted a bug fix for FS #10335. Please take a look.
03:58:23 Join fdinel [0] (
04:00:23 Quit notlistening (Remote closed the connection)
04:10:48 Join cool_walking__ [0] (i=cb3b81c3@gateway/web/ajax/
04:13:01 Quit cool_walking_ (" ajax IRC Client")
04:17:37 Nick cool_walking__ is now known as cool_walking_ (i=cb3b81c3@gateway/web/ajax/
04:24:31 Quit Blue_Dude ("ChatZilla 0.9.84 [Firefox 3.0.11/2009060215]")
04:34:34 Join leachim6 [0] (
04:34:41leachim6do any of you guys have lyrics on your Rockbox ?
04:34:47leachim6if so, which patch do you use for it
04:37:49 Join Lss [0] (
04:42:09***Saving seen data "./dancer.seen"
04:54:13 Quit perrikwp (" ajax IRC Client")
04:54:31 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
05:13:24Unhelpfulplease be more patient. many of the developers are scattered across europe, and it's fairly late in the US, and it's quite possible that nobody paying attention at this very moment can offer a useful answer to your question. i certainly can't, i don't use lyrics.
05:31:22safetydanleachim6: might do what you want but I don't use lyrics myself
05:31:24JdGordonwe are gplv2 or later yeah?
05:32:03safetydanJdGordon: yes I think that was the decision
05:47:43 Quit n17ikh (Connection timed out)
05:48:34 Quit Horscht ("Verlassend")
05:54:36 Quit barrywardell ()
05:57:00 Join barrywardell [0] (
06:01:40*JdGordon prepares a *big* plugins fix commit :)
06:08:03*JdGordon is almost upset that jewels is losing its custom menu...
06:08:10JdGordonit was the only custom menu which was nice
06:09:11LloreanEventually it should be possible to re-create it with list viewports anyway, right?
06:12:18JdGordonforget eventusally... it could be done now
06:13:33 Join Chesteta [0] (
06:14:21 Quit Chesteta ()
06:14:32 Join Chesteta [0] (
06:14:32Unhelpfulyou could use that to get the playback control submenu to match as well, couldn't you?
06:18:15 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:23:12JdGordonyou can do anything! :D
06:23:16JdGordonok... here we go...
06:25:25CIA-37New commit by jdgordon (r21306): Accept FS #10094 by Teruaki Kawashima: ...
06:25:32*JdGordon expects to need to do a follow up commit
06:25:52JdGordonoh bugger.. presed enter too soon
06:31:21 Quit intrados (Read error: 54 (Connection reset by peer))
06:34:27JdGordonhmm.... does it make more sense to use ACTION_STD_CANCEL or ACTION_STD_OK to get rid of a splash?
06:35:09JdGordonand... WTF? green build!
06:35:49JdGordonprobably cancel which would make it almost impossible to remove a splash accidently
06:35:52 Join intrados [0] (
06:41:07LloreanThe problem with getting rid of splashes, in general, is they also go away on their own
06:41:24LloreanIt makes it really easy to accidentally click or cancel something else if you don't press the button to get rid of them soon enough
06:41:27JdGordonnot soon enough though
06:41:51JdGordonI thought we had this discussion already and the outcome was it would be ok
06:42:13***Saving seen data "./dancer.seen"
06:42:14LloreanI wasn't here for that one, I guess.
06:42:33LloreanIs one second really that long?
06:43:24JdGordonno, but its annoying... also the fact that there is no set time for splashes, the other purpose if this is to have them consistantly at say 3s or something
06:44:53JdGordon817 splashes need to be fixed though which is why i dont want to do that :p
06:45:08LloreanI don't know why there needs to be a set time for splashes.
06:45:20LloreanConsistency with what?
06:45:26JdGordon1s is too slow if you arnt expecting the splash
06:45:36LloreanMaking everything the same for the sake of making everything the same is kinda pointless.
06:45:59JdGordonthe same can be said about consistant keymaps between targets
06:46:09LloreanWhy not make all splashes with a timeout require cancelling instead, then?
06:46:30JdGordonbecause that means you have to press a button... and people wont know to do that
06:46:33LloreanJdGordon: That's not "for the sake of consistency", that's "to ease transition between use of the targets, and allow easier development for targets not owned by the developer"
06:46:55LloreanIf a splash pops up and doesn't go away, the very first thing they'll do is press a button
06:47:06LloreanIf you make *every* button cancel the splash, it's as simple as that
06:47:10JdGordononly the cancel button will get rid of it
06:47:17LloreanSo don't limit it to the cancel button.
06:47:25JdGordonallowing any button will mean accidental dismissals
06:47:56LloreanAnd having timed splashes will mean accidental triggering of other things - something worse than mere accidental dismissals.
06:48:20LloreanI'm sorry, a question mark on its own gives me no clue which part of that you didn't understand.
06:48:33JdGordonwell, there is no smiley for a blank face...
06:48:42LloreanA blank face wouldn't either.
06:48:50JdGordoni have no idea what that last statement was about
06:49:01LloreanIf splashes go away on their own *and* a button makes them go away, if someone presses the button too late, they'll cause something else to happen instead.
06:49:10LloreanFor example, the cancel button may also stop playback if used in the list.
06:49:39JdGordonwell.. here is the fun.. with a tiny bit of extra code I can make that impossible
06:49:48LloreanHow do you make that impossible?
06:49:56JdGordonby not quiting untill the button is relelased
06:50:04Lloreanyou missed the point
06:50:15LloreanI'm talking about if the button is pressed right *after* the splash vanishes
06:50:21LloreanIf they're too slow
06:50:33 Join nibbler_ [0] (
06:51:43JdGordonwell, then tough
06:52:02LloreanWell then tough to your people who might accidentally dismiss the splash without a timeout
06:52:40LloreanIt's better to accidentally dismiss a splash than to accidentally do something else unexpected.
06:53:32JdGordonthe window between the button dismissing the splash and being used int he next screen is soo bloody small its not worth worrying about.. AND there is nothing that can be done for it
06:53:38 Join killan_ [0] (
06:54:18LloreanThere's plenty that can be done for it - for example not having splashes time out
06:54:31LloreanThere's nothing that can be done for it that *you* like is what you meant.
06:55:03LloreanIf you want to make all splashes have the same timeout, you're going to have to pick a longer time that's suitable for the splash with the most text.
06:55:13LloreanThat means most people are going to be annoyed by splashes that used to be shorter.
06:55:22 Quit killan (Read error: 104 (Connection reset by peer))
06:55:39LloreanYou already said, if you pick a single button, people won't know how to cancel.
06:55:39JdGordonit turns out no was to that comment also.. but mostly to my yes
06:55:56LloreanSo the majority of people are just going to be stuck with longer splashes.
06:56:01LloreanBy your own reasoning, even.
06:56:49JdGordonno, because if people read the manual it will say that splashes can be dismissed... however if what you want to happen happens, then they will wonder why the thing has locked up, because we know they dont read the manual
06:57:35LloreanWhat I want to happen is any key dismissing the splash
06:57:46LloreanSo, I doubt they'll think it's locked up.
06:57:57JdGordonwhich *will* lead to accidentl dismissals
06:58:04LloreanWhich is better than accidental other things
06:58:10JdGordonit wont happen
06:58:20LloreanYou're crazy if you think it'll never happen
06:58:37JdGordonok, it will happen... but not enough to worry about....
06:58:39LloreanYou'll do it yourself sometimes unless the splash time is set to so long that it's almost impossible for it to go away before you get to the button.
06:59:28LloreanIt'll happen occasionally to everyone who knows they can cancel the splashes, because they'll be just a bit too slow sometimes and not think "oh, I should just let it vanish at this point"
06:59:34JdGordonso long? 3s is long enough for voice and to be able to read.... there is never (iirc) a splash which is triggered by something other than the user, so its not like you will look at the dap and realise a splash is showing
07:00:00JdGordonits either trigger splash and cancel immediatly, or read the thing and have it dismiss by itself
07:00:26Llorean3s is a long, long time for splashes like "Cancelled" that users will see all the time
07:00:40LloreanEspecially since most of those won't have read the manual and know they can cancel it (as you yourself have pointed out)
07:00:50 Quit barrywardell ()
07:00:57JdGordonthen they should read the manual
07:01:04JdGordonthats their punishment :)
07:01:20LloreanYes, that's great - let's make the UI annoying for the majority of our users so you and a couple other people can cancel the occasional splash.
07:01:39JdGordonsounds good to me
07:01:43LloreanWhich splash is it that annoyed you enough, frequently enough, to make you want this?
07:02:06JdGordonthats irrelevant
07:02:11LloreanNo, it's really not
07:02:18LloreanThis sounds a lot like a solution in search of a problem.
07:02:29JdGordon"nothing to resume" and "canceled"
07:02:31LloreanDo we actually have a problem being solved, or is this something you think is a "good idea" and don't care about the consequences?
07:03:02LloreanBoth of those are due to user error.
07:03:11JdGordonno they arnt
07:03:13LloreanStop trying to resume when there's nothing, and don't go into settings you don't want to change, and you'll never see them
07:03:24JdGordonhow do you know there is nothing?
07:03:27LloreanThat's *your* punishment, to use your phrasing
07:03:35LloreanWell me, I actually know when my playlist ends.
07:03:41LloreanMusic stops on its own, etc.
07:03:51JdGordonand when you start up with the start screen not set to the wps?
07:04:00LloreanYou basically want to "punish" our users (your words) to avoid the consequences of your own actions?
07:04:10LloreanMy start screen's been the list for years.
07:04:14LloreanWhat does that have to do with it?
07:04:33JdGordonthere is no way to know if there is a playlist on boot or not
07:04:45LloreanI have a little something called "memory"
07:04:52JdGordonwell good for you
07:04:58LloreanIf my playlist didn't end last time I used my player, then there's a playlist on boot.
07:05:25LloreanAnother solution would be to have the "Resume Playback" option say something else if there's no playlist
07:05:46LloreanOf course, that wouldn't have the added benefit of "punishing" tens of thousands of users.
07:05:52 Quit timc (Remote closed the connection)
07:06:41 Quit Chesteta ()
07:07:38 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:07:48LloreanIt could probably even simply say "Nothing to resume" re-using the string.
07:09:36JdGordonnot unless you really want to punish everyone and have a slight delay every time the menu gets redrawn
07:10:02 Join jfc^2 [0] (
07:10:12*JdGordon double checks that though
07:11:26JdGordonhmm.. that could be done actually
07:11:32LloreanEither way, my point is that if the problem is not knowing whether the playlist is empty or not on boot, the solution isn't "make tens of thousands of users have to wait extra time for the most common splashes to exit"
07:11:51LloreanIt's "find a way to show the playlist status"
07:11:57LloreanAt least, in my opinion
07:12:25JdGordonI will accept that yes thats a problem, but also I am still against not being able to dismiss splashes when you know they are coming
07:13:31JdGordon:) doesnt work anyway
07:13:44LloreanMaybe give splashes a priority? "Things you must know" "Things you should know" "Low priority"
07:13:52LloreanThen let the user pick which priority level to accept splashes from
07:14:00*JdGordon slaps self
07:14:17LloreanSo you can turn off all but essential splashes (shutting down, this setting needs a reboot, etc)
07:14:47LloreanNever having them show is even better than being able to cancel them, I'd think.
07:14:55CIA-37New commit by jdgordon (r21307): For those who dont have the memory of a <Something that has really good memory> show in the main menu that there is nothing to resume to stop the ...
07:15:15LloreanAnd we could probably add a third tier, "Extra Splashes" with some accessibility related information (Charging started, Charging ended, etc) that they'd like.
07:15:18LloreanEr, fourth tier
07:15:22JdGordonim pretty sure that would fall under the category of settings bloat
07:15:38LloreanNot if we need the option for extra accessibility information anyway
07:16:06JdGordonthose arnt timed splashes anyway
07:16:23Llorean"This setting needs a reboot" is.
07:16:38LloreanBut my point was more that if you set it to "Things you must know" basic informational ones like "Cancelled" or "Nothing to resume" wouldn't even display.
07:17:09JdGordonthat really is over kill imo
07:18:02LloreanAnd intentionally "punishing" thousands of users so that you don't have to see cancelled for an extra second every now and then isn't overkill?
07:19:25JdGordonthey will learn quickly
07:19:35LloreanAs you said, most of them never see the manual
07:19:42LloreanThey'll just assume we changed the timing on it.
07:19:51LloreanWe'll probably get bug reports that splashes take longer than they used to.
07:20:04JdGordonwhich we'll close as by design/rtfm
07:20:10JdGordonwhich we expect users to do anyway
07:20:34LloreanI love this new "screw the users" attitude. Really I do. Why are you so attached to the idea of punishing them, when it's clear there's plenty of alternatives that don't require it?
07:21:45JdGordonwanna show me where on the site it says "rockbox, putting users first?" its always been about what we the devs want
07:22:14LloreanIt's never been about screwing over the users because the dev can't be bothered to choose the right setting, though.
07:22:57LloreanI mean with your latest commit, now the only splash that really seems to be an issue is "cancelled". I'd say accept your own punishment and sit through it if you choose the wrong setting, and then move before backing out of it.
07:23:07LloreanI mean, 99% of the time even if you choose the wrong setting, you don't see cancelled.
07:26:08 Quit jfc (Read error: 110 (Connection timed out))
07:30:19JdGordon1) its not *only* those.. but they were the annoying ones... and 2) why is cancelling a setting a stupid thing to do? it doesnt have to be a mistake
07:30:35LloreanIt's always a mistake of some degree.
07:30:42LloreanIf you didn't want to change it, you didn't need to enter in the first place.
07:31:05 Nick fxb__ is now known as fxb (
07:31:18JdGordonyou might want to just look at the aviable options, or decide while your there you want to not change
07:31:31JdGordonor change it for some time during the track and cancel quickly
07:32:23LloreanAnd it's possible to just re-select the same value to not see "Cancelled" too, isn't it?
07:32:49CIA-37New commit by jdgordon (r21308): fix mysterious yellow which was one commit too slow
07:33:03JdGordonnot if you dont know what it origionally was
07:33:16LloreanAh yes, because you weren't looking at the screen when you entered the setting I guess?
07:33:21JdGordonthe *point* is that splashes are annoying if you expect them
07:33:35LloreanIf you expect them, you can avoid them anyway
07:34:15JdGordonand how would you avoid the "restart to enable" splash without editing the config on the conputer?
07:34:42LloreanConsidering I'm going to be restarting my player, an extra second or two doesn't bother me in the slightest there.
07:34:59JdGordoni might not care about restarting right now though
07:35:05LloreanThen don't change the setting right now.
07:35:10LloreanBam, avoided the splash
07:35:38JdGordonbah, no you havnt
07:35:59LloreanI didn't make thousands of users wait extra time to save me one second, either, though.
07:38:01LloreanThere's a difference between doing things the way devs feel they should be done, and choosing to screw over users for insanely trivial benefits.
07:38:48LloreanSo far the "Reboot" splash is the only one you've mentioned it's physically impossible to avoid through good planning anyway.
07:39:06LloreanAnd you still haven't explained what would be the benefit of this "consistency" longer splashes would add.
07:40:13JdGordon1) easier/more maintanble code, 2) voiceing, 3) blind users
07:40:56LloreanHow so for 2 and 3?
07:41:04LloreanNo matter what we set the splash time to, slower voices can be longer than it.
07:41:59LinusNthe reboot splash is just a band-aid fix anyway, before we fix the real issue with reallocating buffers
07:42:30LloreanSo the reboot splash has a "real" fix as well, eventually.
07:42:44LloreanAnd the fix for "nothing to resume" has already been committed.
07:42:52Llorean"Cancelled" is darn easy to avoid in the first place.
07:43:00LloreanWe're on a roll. Any other problematic splashes?
07:43:26JdGordonyay arguing specifics instead of the idea
07:43:41JdGordononce again.. its the splashes in general.. not only those specific ones which are annoying
07:43:54LloreanWell, the idea was supposed to solve a problem. In that case, it's pretty important to identify what the problem is, and that's what the specifics are about.
07:44:17LloreanAs I said up at the beginning, it really seems like a solution searching for the problem.
07:44:27LloreanAnd you yourself have described it as "punishing" our users.
07:44:35LloreanIf we're going to do that, we should have a damn good reason.
07:47:35LloreanThat forum poll is pretty pointless, since the very people for whom it will be a problem (those that don't know about it) are the same people who don't see the poll
07:48:59 Nick fxb is now known as fxb__ (
07:49:55LloreanNot to mention you've gone quite out of your way to try to be as far from impartial on the matter as possible in your posting.
07:50:19JdGordonI never said i was impartial
07:50:37JdGordonit was gong to be worse, but i thought i better tone it down
07:51:01amiconnJdGordon: Conditional (i.e. timing dependent) reaction on buttons would be very bad. One reason to not add splash cancelling
07:51:28JdGordonI *hugely* disagree that it will be a problem
07:51:39JdGordonsplashes only happen after you have triggered them....
07:51:48LloreanJdGordon: Then don't trigger them and they won't be annoying!@
07:52:58LloreanJdGordon: But if you plan to try to use the poll to justify something, you should make it as impartial as possible so people like me won't point out later that you intentionally created a biased poll, so all results are meaningless due to intentional skewing
07:53:33JdGordonif my previous Prime Minister can swing the vote with the question, then who am I to say thats not right?
07:53:50*Llorean thinks splashes that may be dismissed are only a good idea if there's no timeout, as that's when accidents are least likely, and the accidents that can occur are least harmful
07:54:23JdGordonand they will be missed more often
07:55:13 Join AndyIL [0] (i=AndyI@
07:55:20LloreanMissing a splash is less harmful than pressing the cancel button at the wrong time.
07:55:46LloreanIt also means you never have to worry about whether there's enough time to voice it.
07:56:35LloreanMaybe you should take discussion of it to -dev instead of a random poll in the forums?
07:57:21JdGordonbut you are all about the users... not the devs
07:58:07LloreanAs I've pointed out, the user group that I fear your idea would harm couldn't possibly respond to the poll anyway
07:58:27LloreanBesides, as you're the one claiming it's about what devs want, shouldn't you be thrilled to take it to the -dev list?
07:58:50*JdGordon just likes arguing
07:58:58JdGordonI'm pretty sure this got the ok already anyway
07:59:32*JdGordon doesnt really want to admit that the wps text commit was enough for me to not care about this anymore
08:04:37 Join timc [0] (n=aoeu@
08:06:53 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:08:07 Quit AndyI (Read error: 110 (Connection timed out))
08:13:50 Join Rob2222 [0] (
08:18:10 Join ender` [0] (
08:20:20 Join Dauron [0] (
08:20:42 Quit _Auron_ (Nick collision from services.)
08:20:46 Nick Dauron is now known as _Auron_ (
08:21:41JdGordonnow .eu is awake.... I wrote up a while ago and wasnt sure if it should be posted ot the dev-ml or not... im going to bed now so If someone wants to take it and send it be my guest
08:23:11 Quit perrikwp (" ajax IRC Client")
08:23:43 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
08:31:31 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:32:05 Join ender1 [0] (
08:32:19 Quit Rob2223 (Read error: 110 (Connection timed out))
08:35:16 Join petur [50] (n=petur@rockbox/developer/petur)
08:39:14 Quit safetydan ("Leaving.")
08:42:17***Saving seen data "./dancer.seen"
08:48:29 Join bertrik [0] (
08:54:39 Quit ender` (Read error: 110 (Connection timed out))
08:59:33 Join matsl [0] (
09:10:07CIA-37New commit by pondlife (r21309): Renamed routines and changed comments for clarity.
09:22:09 Join flydutch [0] (
09:26:11 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:26:40 Join kugel [0] (n=kugel@rockbox/developer/kugel)
09:27:41kugelpondlife: I wasn't sure if "global_settings.timestretch_enabled = enable;" is needed. I added it for the case it's called outside the setting (i.e. without the setting as parameter)
09:30:12 Join Thundercloud [0] (
09:33:52 Quit kugel (Remote closed the connection)
09:43:30 Quit bertrik (Read error: 113 (No route to host))
09:44:21pondlifekugel: Ideally there would be no need for the parameter, but settings.c wants it.
09:44:56pondlifeOops, missed him
09:49:34 Quit Thundercloud (Remote closed the connection)
09:59:13 Quit DataGhost (Nick collision from services.)
09:59:19 Join DataGhost [0] (
10:00:24 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
10:00:45 Quit BHSPitMonkey ("Ex-Chat")
10:01:54 Join ap0_ [0] (
10:04:14 Join funman [0] (n=fun@rockbox/developer/funman)
10:07:23 Quit framo4_ ("muuh")
10:09:07 Join x72 [0] (
10:09:13 Join fyrestorm [0] (
10:09:37 Quit cool_walking_ (" ajax IRC Client")
10:12:05x72hi! I try to build the bootloader for ipod-video; it builds just fine, but it's size is about 750MiB, so ipodpatcher complains that image is too big. any thoughts?
10:12:35x72build system is gentoo-x86; gcc 4.3.3
10:13:25LloreanIt should be less than 1MB I believe, and that's not the GCC you used to compile it (or shouldn't be) since the iPod is ARM
10:15:04x72it's the cross-arm-elf gcc actually
10:15:22x72and the rockbox itself works
10:15:24gevaertstry the toolchain built by
10:15:52x72so it's the toolchain issue, you think?
10:16:18 Join kugel [0] (n=kugel@rockbox/developer/kugel)
10:16:22gevaertsif you're building from unmodified sources, I don't see what else it could be
10:16:52pondlifekugel: ping
10:16:54x72well, yes, will try that, thank you!
10:17:01kugelpondlife: pong
10:17:41pondlifeI can't repro your timestretch issue at all on the sim. :/
10:17:59kugelI can repro it everytime
10:18:18pondlifeTo clarify, you start playback, then go into the pitchscreen, press MODE twice, then press RIGHT to speed up?
10:19:17pondlife"sdl_audio_callback: No Data" is classic underrun, not a bug. But it shoudln't skip any track - might just take a while to get playing again though.
10:19:49pondlifeI'm also not using EQ or crossfeed or similar..
10:22:55 Quit timc (Read error: 110 (Connection timed out))
10:23:02 Quit nibbler_ (Read error: 110 (Connection timed out))
10:24:03 Join einhirn [0] (
10:27:00kugelpondlife: also happens with empty config.cfg
10:27:31kugelempty except for timestrech that is
10:28:02pondlifeAh, ok
10:28:12pondlifeSo did I get the recipe correct?
10:28:17kugelpondlife: "pitchscreen, press MODE twice, then press RIGHT to speed up" (right is speed only on my targets)
10:28:23kugelso yes
10:28:51pondlifeBut it's ok if you press MODE, MODE, DOWN ?
10:29:08pondlifei.e. increase speed and decrease pitch
10:29:19kugeldown does nothing here
10:29:57kugelyes, doing that works fine. and after doing that, pressing select to reset, then RIGHT to speed up only works too
10:31:51pondlifeMost odd
10:32:33kugelmaybe something is used initialized
10:33:20pondlifeYes, but the only difference might be in a call to sound_set_pitch()
10:35:46gevaertskugel: FS #10200
10:36:08kugelah thanks
10:36:55kugelgevaerts: I think that's solved by using tolower too. isn't " " between Z and a in ASCII?
10:37:23pondlifekugel: If you press RIGHT repeatedly, what happens on the screen - i.e. does the displaysd speed value increase while pitch remains at 100%?
10:37:27Mikachuspace is around 0x20 iirc
10:37:36gevaertsit *is* 0x20 :)
10:37:45kugelpondlife: nothing happens but the displayed speed increases
10:38:18kugelgevaerts: that's definitely weird then
10:39:09gevaertskugel: the good news is that it includes a real example, so it can be tested :)
10:39:32pondlifekugel: Try pressing SELECT (or whatever the reset key is) before you press RIGHT? i.e. MODE MODE SELECT RIGHT
10:40:27gevaertsAnyway, that's a pretty common naming scheme for classical music, so I think the default sort order should handle it
10:40:41kugelgevaerts: yes sure
10:40:45kugelpondlife: no success
10:41:00pondlifeYou mean that still shows the problem?
10:41:13 Quit r0b- (Read error: 110 (Connection timed out))
10:41:24pondlifeWell, that's even odder as it calls sound_set_pitch()
10:41:25 Join r0b- [0] (
10:41:43pondlifeWhat music format are you using?
10:42:21***Saving seen data "./dancer.seen"
10:47:14 Join fml [0] (n=4fd3c61a@gateway/web/cgi-irc/
10:48:26fmlHrm... I'm a bit disappointed by the new main menu entry "Nothing to resume". I see the rationale behind it but it's a weird title for a menu entry. I can't say what would be better though.
10:48:59pondlifeJust hide the option?
10:50:17kugelpondlife: mp3
10:50:30pondlifeSame here...
10:51:58FrankTM : last comment
10:52:38 Quit r0b- (Read error: 104 (Connection reset by peer))
10:52:50 Join efyx [0] (
10:52:59 Join r0b- [0] (
10:53:03pondlifeFrankTM: Do you have both dircache AND load database to RAM enabled?
10:53:27pondlifeIf not, try enabling both and rebooting
10:53:36FrankTMi have no idea
10:53:49FrankTMwhere do I find that options
10:54:00 Join PaulJam [0] (
10:54:55pondlifeSettings > General Settings > Database > Load To RAM
10:55:16FrankTMI do not have that option :o
10:55:19pondlifeand Settings > General Settings > System > Disk > Directory Cache
10:55:30pondlifeFrankTM: Which device?
10:55:36FrankTMSansa Fuze
10:55:48FrankTMno Disk under System aswell
10:56:57pondlifeAh, I don't know then...
10:57:54pixelmacould you state the device with your bug report as well? I think it's not unreasonable to think that ports in progress (especially reading that there are disk access /file system issues)... I don't say it's impossible that it's a general Rockbox bug but I'd be careful reporting this
10:58:17kugelgevaerts: funnily enough, xfce4-terminal doesn't handle those as well
10:58:22FrankTMpixelma: good point :)
10:58:37FrankTMi can't edit a comment in flyspray, can i?
10:59:32kugelgevaerts: that's without -v. ls -v fails too
10:59:47kugelnot as bad as without -v at least
11:01:38PaulJamHi, does anybody know what SVN revision was used to buld the official H300 bootloader (v5)?
11:02:02LloreanIt should be tagged in svn
11:02:16LloreanLooks like 8341
11:02:46 Quit fml ("CGI:IRC (EOF)")
11:02:56pondlifeAh, the H300 bootloader question... ;)
11:03:01PaulJamthank you
11:03:01LloreanPaulJam: What are you going to do with it?
11:03:04kugelFrankTM: sansa fuze is unsupported.....
11:03:29kugeldatabase problems can come from a variety of problems on unsupported targets
11:03:36FrankTMkugel: i'm just trying to help squash the bug?
11:03:57kugelit does not help though if we cannot make sure the target itself works
11:04:23kugeland it currently even has filesystem problems
11:04:28pondlifeAm I right to think that database update relies on dircache and/or load to RAM in order to delete entries (moved files etc.) or is that out of date?
11:04:42kugelI think that's true, yes
11:04:52kugel(for deleting at least)
11:07:56PaulJamLlorean: since the more recent bootloaders (for example the one on the CFModGuide page) don't work reliably on my H300 and i would like to use the RTC alarm feature, i thought it might be possible to just take the old working bootloader and just remove the check for the ON/PLAY button. I'm not sure if this would work. (and yes, i know there is the risk of bricking.)
11:09:22funmani can only transfer red component of pixels to the c200v2 graphic ram ..
11:09:51LloreanPaulJam: You might want to check, I'm not *entirely* sure we used the same compiler version back then.
11:14:35PaulJamLlorean: yes, i was planning to first build an unmodified bootloader and compare it to the official files and only proceed when the checksums match. how can i find out which compiler was used (i think it was v3.4.5 but i'm not sure)?
11:14:52linuxstbYou could try the history of the CrossCompiler wiki page
11:15:45linuxstbI don't know if LinusN can remember - he built all those bootloaders.
11:17:13LloreanIt definitely wasn't 3.4.5
11:17:16LloreanI'm thinking it was 3.4.3
11:17:43LinusNi don't remember :-(
11:18:40*linuxstb wouldn't either...
11:19:07 Join biengo [0] (
11:19:14 Join safetydan [0] (n=Dan@rockbox/developer/safetydan)
11:19:52pondlifePaulJam: What goes wrong for you on the later bootloaders?
11:20:33 Nick AlexP_ is now known as AlexP (n=alex@rockbox/staff/AlexP)
11:20:46pondlifeI have an H300 that works ok with recent bootloaders, but am happy to help test any fixes you can come up with.
11:24:43PaulJampondlife: on some players the bootloader crashes sometimes while booting (i don't remember what the errormessage was). LinusN and petur were looking into this issue during last years devcon. they probably can tell you more details.
11:25:05pondlifeIs that what you see on yours though?
11:25:33pondlifeI don't think petur and LinusN had enough time to make any progress.
11:25:54peturtoo bad Linus won't be here this year, he could pin it down to a certain commit but not what caused it
11:26:06gevaertskugel: ls -v does indeed handle them badly. Plain ls has no problems here though, probably because of LANG=POSIX here
11:26:27pondlifepetur: Don't suppose you know the revision that broke it?
11:26:46pondlife(That info ought to be added to the wiki, IMHO.)
11:26:50peturmaybe I wrote it down at home, else ask Linus
11:27:14pondlifeLinusN: pingy?
11:27:48*gevaerts hopes that there will be things at devcon where the conclusion *won't* be "Ask Linus" ;)
11:28:14*pondlife wishes he had the time and money to attend
11:29:43*LinusN too
11:32:27 Join timc [0] (n=aoeu@
11:35:04 Quit perrikwp (" ajax IRC Client")
11:35:21 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
11:37:09kugelyea, LANG=POSIX works better
11:37:28gevaertslocale also specifies sorting order
11:37:51gevaertsTry with just LC_COLLATE=POSIX, that should work as well
11:38:28kugelwhy could this result in such a weird sorting?
11:54:30kugelpondlife: can you repro the sim skips at least?
11:55:17pondlifeYes, but that's expected - same happens quite often when it's buffering anyway here.
11:55:43pondlifeI use −−root n: where n: is a slow networked drive :)
12:04:04kugelpondlife: I thought skipping is not expected
12:04:12kugelalso, I never see those while playback
12:04:53pondlifeNot skipping, gapping/pausing
12:05:07pondlifei.e. stop/start playback, but with nothing missed out
12:08:33pondlifekugel: It only happns when I have album art (on normal playback)
12:09:12pondlifeBasically there's something that's not yielding enough on the sim.
12:10:27kugelI never experience that
12:10:47pondlifeIt's probably unrelated to what you're seeing though
12:11:13kugelI doubt that
12:11:32kugelI don't get the skipping if I use speed-corrected pitch before
12:12:19safetydankugel: regarding FS #10335, I think the reporter mentioned it was the last 1000 or so samples... so that's like what, 1/20 of a second?
12:13:19kugelpondlife: that skipping is really only when I use my recipe
12:14:44 Join robin0800 [0] (
12:15:28kugelsafetydan: seems safe then. Put it should probably be in the if (playing) block
12:17:19kugelsafetydan: also it would be great if you could say something to the recent beep/accessiblity discussion on the ml
12:17:47kugelpondlife: maybe you try it without that slow network drive?
12:17:54pondlifeYep, maybe
12:18:02pondlifeNot got time at the moment though
12:18:12 Quit robin0800 (Remote closed the connection)
12:18:35 Join robin0800 [0] (
12:21:04pondlifekugel: Has your problem only been with the "mpegplayer-fix"?
12:21:18pondlifeOr was the original FS #8894 commit also broken in this way?
12:21:30*kugel goes trying
12:22:45gevaertsFS #10335 is at least a year old. amiconn saw it while trying to load software on his zx spectrum
12:24:17safetydangevaerts: man, that took me a while to parse... was amiconn trying to playback spectrum tapes on rockbox?
12:24:35pixelmamaybe there's a duplicate? I also wonder if this is the reason the voicing of the "end of playlist" splash is broken on swcodec and would be fixed with the patch
12:24:40gevaertssafetydan: yes. You connect a spectrum to your DAP, and you have your tapes as flac on the DAP
12:24:49funmanmy c200v2 seems to accept r(5)b(5)g(6) instead of r(5)g(6)b(5) Oo
12:25:06kugelpondlife: no skipping in sim, didn't try target yet
12:25:24pondlifeBest if you stick to sim for now
12:25:56kugelI already looked if the mpegplayer commit could've caused it, but I can't see how
12:26:13pixelmasafetydan: just using the DAP for playing the tapes back (and there's info in the end "noise")
12:27:06kugelpondlife: there seems to be another bug
12:27:35pondlifeSorry, I don't have time to help now
12:27:47pondlifeGotta go
12:27:50 Part pondlife
12:30:11safetydankugel: am I missing something? Is there a reason you think I should contribute to the bleeping thread?
12:31:22kugelisn't it your patch?
12:31:47kugelalso, I thought the topic could touch you..
12:32:36pixelmakugel: are you confusing him with sdoyon maybe?
12:35:20safetydanquite possibly
12:35:30safetydanI haven't contributed anything for a long time :)
12:35:52 Quit safetydan ("Leaving.")
12:37:09kugelI guess I did
12:37:37*kugel goes in his corner
12:38:13*scorche|sh should have brought the stomping boots
12:39:12gevaertsscorche|sh: surely there must be a good stomping-boots-shop in Amsterdam!
12:40:56 Join perfectdrug [0] (
12:41:24 Quit perrikwp (" ajax IRC Client")
12:42:24***Saving seen data "./dancer.seen"
12:47:41 Quit bmbl (Connection timed out)
12:48:44 Quit kugel (Remote closed the connection)
13:04:44 Quit Lss (Read error: 104 (Connection reset by peer))
13:12:47 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
13:14:51 Join TheSphinX^ [0] (
13:21:08 Quit bubsy (Read error: 60 (Operation timed out))
13:21:27 Join wark [0] (
13:21:36 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:41:51CIA-37New commit by mcuelenaere (r21310): Onda VX747: fix backlight driver so it doesn't 'flicker' upon init
13:42:47FrankTMhmm. would that be on device start?
13:50:26 Join bubsy [0] (i=Bubsy@
13:52:40 Quit markun (Read error: 60 (Operation timed out))
13:57:20funman /away lol
13:57:27funman>< sorry ..
13:57:52gevaertsfunman: boring away message!
14:03:49 Quit PaulJam (".")
14:05:32 Quit perfectdrug (Remote closed the connection)
14:06:56 Join PaulJam [0] (
14:07:18 Join perfectdrug [0] (
14:10:26 Join Horschtl [0] (i=3ee19662@gateway/web/ajax/
14:11:10Horschtlis the pegbox plugin a completely new game or a port/remake of an existing one?
14:12:03pixelmait's ported from a TI calclator game, as far as I know. Should be in the header of the c file
14:13:11pixelmahmm, not there. I wonder where I read it
14:13:49HorschtlI am asking because I realy like it. 15 Levels is just too little but I am not good at *creating* challenging puzzles. Só I was hoping that it would be a port of which we could try to get a few more levels :D
14:15:40pixelma - and I also looked for more levels ;) If I didn't miss something, there were not more in the original game
14:17:02 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:17:11Horschtl <- I also found the original game...
14:19:34Horschtlwhich realy contains only 15 levels :(
14:21:31 Join barrywardell [0] (
14:24:05 Quit PaulJam (Read error: 60 (Operation timed out))
14:24:37 Join PaulJam [0] (
14:27:35 Quit antil33t (Read error: 104 (Connection reset by peer))
14:42:25***Saving seen data "./dancer.seen"
14:48:13 Quit intrados (Remote closed the connection)
14:51:15 Join PaulJam__ [0] (
14:52:23 Quit PaulJam (Read error: 113 (No route to host))
14:52:43 Join akur [0] (
14:54:32 Part akur
14:55:08 Join evilnick_7 [0] (i=0c140464@gateway/web/ajax/
14:55:33 Join intrados [0] (
15:01:32 Join akur [0] (
15:01:49 Quit barrywardell ()
15:03:57akuri have been reading the wiki page about the lang file
15:04:13akurand that table seems a bit outdated...
15:05:05 Quit evilnick_7 (" ajax IRC Client")
15:05:18 Join Rondom [0] (
15:07:23 Quit at0m ("it is now safe to power off your machines.")
15:10:52 Quit Horschtl (" ajax IRC Client")
15:13:43 Join evilnick_7 [0] (i=0c140464@gateway/web/ajax/
15:14:26 Quit akur (Read error: 60 (Operation timed out))
15:18:47funmanthe storage works on c200v2 (i can load rockbox) but rockbox shows the logo, and then fastly i only see the backlight without anything on screen (backlight timeout still works though). And I still don't know how the colors are packed in one pixel
15:19:14 Quit ap0_ (Client Quit)
15:26:02 Quit shodanX (Read error: 104 (Connection reset by peer))
15:29:22 Part LinusN
15:32:01 Join akur [0] (
15:36:25 Quit parafin (Read error: 60 (Operation timed out))
15:40:11 Join shodanX [0] (
15:41:14 Quit PaulJam__ (Read error: 113 (No route to host))
15:42:23 Join PaulJam [0] (
15:48:18 Quit perfectdrug ()
15:48:34 Join perfectdrug [0] (
15:48:59 Join parafin [0] (
15:49:24 Join fml [0] (n=4fd3c61a@gateway/web/cgi-irc/
15:51:31 Nick fxb__ is now known as fxb (
15:51:31fmlI'd like to pick up my earlier statement today about the "Nothing to resume" as a menu entry. Wouldn't it be better to have "Resume playback (Nothing to resume)" in that case? The drawback is that the title is more lengthy, but as a menu entry it would do more sense than just "Nothing to resume" IMHO since the latter doesn't tell what action is behind the menu entry.
15:54:38 Join avacore^ [0] (
15:55:27 Part akur
16:00:20 Quit robin0800 (Remote closed the connection)
16:01:55 Quit avacore^ (Read error: 60 (Operation timed out))
16:02:19 Quit PaulJam (Nick collision from services.)
16:02:22 Join PaulJam_ [0] (
16:04:23 Join Lss [0] (
16:05:23 Join avacore^ [0] (
16:07:13 Quit avacore (Read error: 110 (Connection timed out))
16:07:58 Join at0m [0] (n=at0m@
16:13:56leachim6does anyone here have lyrics on their rockbox?
16:19:13evilnick_7I used the sncviewer briefly a few years ago
16:19:52linuxstbfml: I definitely think a menu option called "Nothing to resume" which can be used to display a splash saying "Nothing to resume" is very odd...
16:21:06leachim6is there a way to view non-synced lyrics from ID3 tags
16:21:12leachim6I just don't see the need for synced lyrics...
16:21:21*linuxstb wonders what happened to discussing feature changes prior to commit...
16:21:59leachim6is there a patch I can use to view noncynced lyrics?
16:22:37linuxstbHave you searched the patch tracker for "lyrics" ? There's at least one there.
16:28:29pixelmalinuxstb: wasn't the plan to get rid of that splash instead? But I agree "Nothing to resume" sounds odd in a menu
16:31:05linuxstbI don't have a problem with splashes, but I guess I'm more patient than most people...
16:31:26fmllinuxstb: this phrase as an addition to "Resume playback" would be OK as a mean to avoid the splash that would pop up after selecting the entry. But these words alone are hard to understand.
16:31:32 Quit biengo (Remote closed the connection)
16:31:44leachim6I keep getting "dlopen(simdisk/_temp_codec0.dll): simdisk/_temp_codec0.dll: invalid ELF header" on the simulator when I try to play an MP3
16:32:34FrankTMleachim6: did you do make fullinstall?
16:33:02FrankTMhehe.. that is what i did wrong last week :P
16:33:15leachim6also ... I did search the patch tracker for Lyrics ... but I only found the Synced Lyrics plugin
16:33:24leachim6and it's a PitA to get synced lyrics files for all of your music
16:33:35leachim6especially when you're using the DB from the iPod
16:33:40linuxstbfml: Your suggestion just feels awkward to me. I wouldn't say the current (prior to today's change) is perfect (displaying a menu option that doesn't work), but it does the job.
16:33:57pixelmamake fullinstall is only needed for the fonts, usually make install should do it
16:33:58leachim6is there a way on linux to convert iPod Playlists to work in rockbox?
16:34:07leachim6I didn't run install either
16:34:09leachim6just make
16:34:37FrankTMlinuxstb: isn't it possible to hide that menuoption, when there's no current playlist
16:34:52leachim6so I compiled the syncronized lyrics viewer
16:35:00leachim6how do I use it? haha
16:35:03*FrankTM pets leachim6
16:35:15linuxstbFrankTM: Yes, but that has the downside of making it harder for blind navigation (which includes sighted users using the DAP without looking at the screen).
16:35:18evilnick_7leachim6: You can sync your own lyrics within that plugin...
16:35:44pixelmaleachim6: could you please stop hitting the enter key so often (especially before and after ...)?
16:36:16FrankTMlinuxstb: i tend to forget about those everything i suggest something
16:36:32leachim6pixelma, yeah sorry...I do that a lot...and it can get really annoying...sorry about that...very annoying habit ...
16:37:06FrankTMyou also should read more
16:37:54fmlFrankTM: that would make it obscure for sighted users as well as it would change the menu for no clear reason.
16:38:50FrankTMi agree with the first part of that :)
16:42:27***Saving seen data "./dancer.seen"
16:42:53FrankTMleachim6: my point is, things you ask are probably described in the manual or in the bug report ;)
16:45:16leachim6FrankTM, except one of my questions...where can I find a plugin to dispaly unsynced lyrics from ID3 ...
16:45:31 Quit matsl (Read error: 110 (Connection timed out))
16:45:33leachim6the only one on the patch tracker is synced lyrics
16:45:51FrankTMit probably does not exists then
16:46:24 Join Blue_Dude [0] (
16:46:39leachim6I saw something on the mailing list...but there was never any code...
16:46:46 Join jgarvey [0] (
16:46:48evilnick_7leachim6: IIRC, it displays either, if the lyrics file isn't synced then the user can scroll down as they please.
16:47:15Blue_DudeLast night I posted a bug fix for FS #10335. Comments are appreciated.
16:48:14leachim6evilnick_7, oh if only I could get it installed...I wish there was a simpler system for managing plugins
16:49:24evilnick_7If you *only* want to try that out, then get the revision of code mentioned in the patch and then it will work, but you'll probably have an older version of Rockbox.
16:50:11evilnick_7Each patch is patched against one revision of the code, the problems with updating them only happens if you use a revision that is *after* that revision
16:54:41*linuxstb wonders if he's alone in not liking JdGordon's change to the main menu
16:57:29fmllinuxstb: you very probably are not
16:58:25leachim6what patches do you guys use?
16:59:24linuxstbleachim6: I expect most people here use none. You may want to look at the unsupported builds forum to see which ones users seem to like.
17:00:00 Quit fml ("CGI:IRC (EOF)")
17:00:05 Join robin0800 [0] (
17:00:35leachim6in all guys don't seem to be typical users :)
17:00:56linuxstbleachim6: Yes, that's why I pointed you there...
17:01:13linuxstbBut I'm not sure if there is such a thing as a typical user.
17:03:09leachim6haha...that's true...
17:04:08leachim6anyone who'd hack their Mp3 player's firmware probably isn't typical...
17:04:21FrankTMhavent checked the patches. rockbox pretty much does what i missed in the original firmware ;)
17:05:07leachim6not me...I miss lyrics ... :(
17:05:54 Quit Zagor ("Don't panic")
17:07:16*amiconn uses several patches... all of them only exist locally
17:07:32leachim6amiconn, what do your patches do?
17:08:44 Quit robin0800 (Remote closed the connection)
17:14:43 Quit r0b- (Read error: 60 (Operation timed out))
17:15:23 Join toffe82 [0] (n=chatzill@
17:21:28leachim6where can I get themes from? only has 18 themes for iPod video
17:21:51Blue_DudeCurrently I'm juggling three different uncommitted patches. Is there a convenient way to generate a diff without having to trim them all out?
17:22:45linuxstbWhen you do "svn diff", you can specify directories and files to include - but that obviously depends on how the patches interact. Others will tell you to use git...
17:23:16Blue_DudeSo I can specify exactly which files I want to diff?
17:23:19PaulJam_leachim6: you may find additional themes in the wiki (WpsGallery) that havent yet been transferred to the new themes site.
17:23:48linuxstbBlue_Dude: Yes.
17:24:06Blue_DudeThat helps. I'll have to look up the syntax on that.
17:24:21linuxstbSimply "svn diff file1 file2 dir1 etc etc"
17:25:22 Join BryanJacobs [0] (i=809743f3@gateway/web/ajax/
17:25:56Blue_DudeI just tried it. That's extremely helpful. Thanks.
17:27:15 Quit intrados (Remote closed the connection)
17:34:21 Join intrados [0] (
17:34:26leachim6is there an app for linux that puts the albumart from all my tracks in rockbox format?
17:35:22 Quit HellDragon (Client Quit)
17:35:24linuxstbThere isn't really a "Rockbox format" - they just need to be .jpg or .bmp files. See the AlbumArt wiki page.
17:35:37leachim6yeah that's what I mean...for lack of a better term
17:36:06leachim6what I really mean is extract the jpegs from the id3 and put them in ~/.rockbox/albumart/AlbumName-100x100.bmp
17:36:12leachim6is there a script for that?
17:36:31linuxstbRead that wiki page ;)
17:37:57leachim6all the tools are windows only :(
17:38:01leachim6I don't have a windows machine atm
17:38:11linuxstbNo they're not...
17:39:08leachim6mb I'm reading the wrong section
17:43:23 Join Sajber^ [0] (
17:44:18 Quit Sajber^ (Client Quit)
17:45:01 Join Sajber^ [0] (
17:46:34 Quit Sajber^ (Client Quit)
17:46:49 Join Sajber^ [0] (
17:47:31 Quit funman ("leaving")
17:50:31 Join gregzx [0] (
17:56:00 Join r0b- [0] (
17:58:30 Join HellDragon [0] (
18:01:21 Quit HellDragon (Read error: 54 (Connection reset by peer))
18:07:38 Join avacore [0] (
18:13:46 Join JdGordon| [0] (i=836b0053@rockbox/developer/JdGordon)
18:17:31 Quit Blue_Dude ("ChatZilla 0.9.84 [Firefox 3.0.11/2009060215]")
18:19:34 Quit avacore^ (Read error: 110 (Connection timed out))
18:21:41CIA-37New commit by mcuelenaere (r21311): Onda VX747: add ability to exit keyboard input using physical keys
18:23:02 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
18:23:04 Join HellDragon [0] (
18:27:32 Join kugel [0] (i=kugel@rockbox/developer/kugel)
18:28:33pondlifekugel: New patch up at ..
18:28:54pondlifeI've still not been able to repro on sim, but this is tidier
18:29:12pondlifeIt now allows you to run timestretch in mpegplayer ;)
18:29:36pondlife(Although I guess a bit of UI would be nice in there, so you don't have to preset it.)
18:31:13 Quit Erant (Read error: 60 (Operation timed out))
18:33:07 Join Erant [0] (
18:38:36 Join bertrik [0] (
18:39:20 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:40:12 Quit petur ("work->sports")
18:42:30***Saving seen data "./dancer.seen"
18:42:51 Join perfectdrug_ [0] (
18:52:27 Quit bmbl (Connection timed out)
18:56:51 Quit perfectdrug (Read error: 110 (Connection timed out))
18:57:23 Join kugel_ [0] (i=kugel@
19:04:51CIA-37New commit by mcuelenaere (r21312): Touchscreen targets: add calibration screen + rewrite calibration driver (FS #10295)
19:05:07 Quit kugel (Read error: 113 (No route to host))
19:26:07 Join robin0800 [0] (
19:26:24 Quit r0b- (Read error: 110 (Connection timed out))
19:26:31 Join r0b- [0] (
19:26:55 Join tessarakt [0] (
19:31:35 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:32:11 Quit robin0800 ("Konversation terminated!")
19:32:47 Quit Rondom ("Ex-Chat")
19:33:25 Join Genesis [0] (
19:33:35bertrikhi Genesis
19:36:23 Join funman [0] (i=56428f5e@gateway/web/ajax/
19:39:17 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:40:27 Quit funman (Client Quit)
19:50:54HorschtI was wondering: why does the sudoku plugin notsave the scratchpad upon exit?
19:52:09JdGordon|there is a patch on the tracker for that iirc
19:54:39Horschtwhy would a patch of such importance not be in SVN? ;)
19:55:04Horschtalso: you guys impress me with your knowledge what's in flyspray....
19:55:04JdGordon|because noone has been bothered taking the time to commit it?
19:55:34*JdGordon| hopes Horscht doesnt figure out we just say "there is a patch for that already" to everything and hope they dont actually go checking
19:55:43gevaertsThere are much more important things to commit!
19:55:45*gevaerts hides
19:56:06amiconnBecause the standard format used by sudoku.rock has no means for storing the scratchpad
19:57:11 Join midgey|w [0] (i=836b0052@gateway/web/ajax/
19:57:12Horschtdoes it also save the current process as an .ss file?
20:00:44 Quit TheSphinX^ ("XChat@Linux")
20:03:18 Join Thundercloud [0] (
20:03:41BryanJacobsis rockbox going to have a BugDay some time where developers swat some of the flies building up?
20:04:19BryanJacobsI mean, just the "patches" portion of flyspray has 450 entries
20:04:35pixelmaamiconn: I'm not sure Rockbox is (still) using the "standard" format. I remember last time this discussion came up, generated Rockbox files couldn't be loaded in Simple Sudoku (if I remember correctly)
20:05:08evilnick_7BryanJacobs: Isn't that the point of the feature freeze?
20:05:37 Quit HellDragon (Client Quit)
20:05:47 Quit n1s (Read error: 104 (Connection reset by peer))
20:05:48 Quit Sajber^ (Read error: 60 (Operation timed out))
20:06:06 Join Sajber^ [0] (
20:06:10BryanJacobsevilnick_7: but in all the time I've been watching Rockbox, flyspray has never really been cleaned up
20:06:16JdGordon|BryanJacobs: we usually use the freeze for that, and even occasionally have a clean up week to go with it, but really its everyone's own time so they use it how they want... if you want to go smash some bugs, go for it :)
20:06:36BryanJacobsJdGordon: how about committing the "power off menu option" patch, then?
20:06:44BryanJacobsI've been hand-applying that one for ages
20:07:08evilnick_7How does that offer better functionality than is currently done?
20:07:19JdGordon|and now you'll see why patches dont get applied... there is almost never unanimous support for them, which means either pissing people off, or not commiting
20:07:40BryanJacobsevilnick_7: you can turn off your DAP immediately - for airplanes and such
20:07:46JdGordon|as can be witnessed in te irc logs and the dev ml...
20:08:19BryanJacobsI guess I just feel that there must be some consensus eventually reached
20:08:20 Quit flydutch ("/* empty */")
20:08:35BryanJacobsnot unanimous, but there has to be some way to decide either "we're taking this one" or "close the ticket"
20:08:36gevaertsof course, some people try to force the issue by calling other people names, which doesn't help
20:08:38 Join n1s [0] (n=n1s@rockbox/developer/n1s)
20:09:03 Join biengo [0] (
20:09:08pixelmaBryanJacobs: isn't pressing the "off" button even quicker?
20:09:15BryanJacobspixelma: what off button?
20:09:18BryanJacobsI have an iPod
20:09:27evilnick_7BryanJacobs: How much actual time is saved with that patch? Maybe one second. And how often would a user need to use it to have that extra 1 second? I can't see the aeroplane argument holding much water to be honest.
20:09:28BryanJacobsI don't think it HAS an off button
20:09:29midgey|whold play for ~5 sec
20:09:29*JdGordon| would be in favour of shutdown in the menu for the ipods
20:09:36pixelmaBryanJacobs: long Play it is on those, I believe
20:09:41BryanJacobsthat works?!
20:09:43BryanJacobswell, blimey
20:10:00linuxstbBryanJacobs: We have a manual ;)
20:10:08*BryanJacobs remembers that not working in Rockbox way back when
20:10:17*BryanJacobs apparently has become too set in his ways
20:10:44BryanJacobsso then, if we have that, why not close FS #6733 with "just press the power button"?
20:11:03JdGordon|because noone can be bothered
20:11:22BryanJacobsah. OK then.
20:11:22pondlifeAny c200/e200 users fancy testing the timestretch patch on FS #10341? I'd like to commit it soonish.
20:11:29*BryanJacobs goes back to playing with buffering code
20:11:36linuxstbIt's also that no single person speaks for Rockbox
20:11:50evilnick_7BryanJacobs: Out of interest, how had you been turning off your iPod?! :)
20:12:09JdGordon|headphone removal and idle shutdown ftw
20:12:16BryanJacobsevilnick_7: using the menu option
20:12:33evilnick_7BryanJacobs: Touche :)
20:12:34BryanJacobsI'm also still not sure that holding play turns it off - I think it might just sleep instead of shutdown
20:12:51gevaertsthere's no sleep support in rockbox
20:12:52evilnick_7It's DEFINITELY shutdown (from within Rockbox)
20:13:10BryanJacobsgevaerts: I meant display-off instead of hard shutdown
20:13:24BryanJacobslike where you don't get the splash screen when you press a button
20:13:38pixelmasee evilnick_7's reply
20:13:56BryanJacobsI'm just glad now I have an easy way to turn it off
20:14:02BryanJacobsand boy do I feel silly
20:14:42JdGordon|thats ok.... we wont hold it against you or anything...... :D
20:14:56gevaertsDoes anyone happen to have a somewhat working working beast bootloader?
20:14:56*linuxstb makes some notes for the mid-term evaluation...
20:15:03*BryanJacobs cringes
20:15:09gevaertssomehat recent...
20:18:10 Join funman [0] (n=fun@rockbox/developer/funman)
20:18:12AlexPgevaerts: I do, not ure from when
20:18:19AlexPgevaerts: It says v1
20:18:53gevaertsmine is from just after charging worked I think, and the last time I tried building one it didn't work
20:19:18AlexPOn disk it says April IIRC, but not sure when it is really from
20:20:07AlexPgevaerts: I think it is still at if you want it
20:20:24*AlexP adds an http:// somewhere
20:28:00 Join HellDragon [0] (
20:29:28obomidgey|w: yes, tis a pain
20:30:10oboI hadn't bothered doing the configure work until it had some actual code worth running :)
20:30:12midgey|wfrom -community: i've been messing around with the view as well
20:30:51midgey|wright now it's basically a copy of the e200 bootloader but signed correctly
20:31:04midgey|was in, signed for the view
20:31:15 Quit leachim6 (Read error: 104 (Connection reset by peer))
20:31:22 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
20:31:30obomidgey|w: yeah, I've just been doing that with a bash script for my testing so far
20:31:36midgey|wi have dummy files in the target tree too
20:31:55 Quit n1s ("Lmnar")
20:31:58 Quit funman (Read error: 113 (No route to host))
20:32:34 Join funman [0] (n=fun@rockbox/developer/funman)
20:32:52obomidgey|w: that was actually my task after submission for gsoc, except I used a completely fake google/dummy target
20:33:23midgey|wis there anything that needs to be done to run the code you put in the wiki?
20:33:36midgey|wi added GPIO_SET_BITWISE(GPIOD_OUTPUT_VAL, 1<<8); to main-pp and nothing seemed to happen
20:34:26obomidgey|w: I commented out a lot of the existing functions - lcd_init, but left kernel_init there so I would have sleep functions
20:35:03midgey|wah i see
20:35:18midgey|wwell maybe i'll mess around with that later
20:35:30oboI'm then running all my code before the adc_init line, with a reboot function before it gets to that point
20:35:46midgey|wthat's likely my problem
20:36:02CIA-37New commit by mcuelenaere (r21313): Make some asm() statements volatile
20:37:09midgey|wyou said you thought you found the lcd init code, any luck?
20:38:17oboI've found some lcd init code, the section which sets up the GPIOs and touches two of the LCD registers, but there must be more to setup the lcd controller itself somewhere
20:39:09obobut since I'm running my code after the OF bootloader all that code has already been run.. so I doubt it would have any effect if it was run again?
20:40:51kugel_pondlife: just add the pitchscreen to the plugin api
20:41:02kugel_I cannot test right now, but will do later
20:42:34***Saving seen data "./dancer.seen"
20:42:36funmanbertrik: i have lcd output on c200v2 although the colors are wrong. rockbox loads from SD and quickly shows a "white screen" after showing the logo (only backlight on, no pixels visible)
20:43:24pondlifekugel_: It's not that I aimed to add timestretch to mpegplayer, more that I was surprised ir worked so well (on sim at least). I really just want to know if that patch fixes your particular problem.
20:43:41bertrikfunman, how does the white screen look, is it a kind of uneven fade towards the white screen?
20:44:36bertrikthe e200v1 does that when booting to the OF, I think it's caused by lack of screen refresh by the LCD controller
20:45:24CIA-37New commit by bluebrother (r21314): Clean up several debug messages.
20:48:41 Join funman_ [0] (
20:48:50funman_no it's not white, it more looks like all pixels are black
20:48:55bluebrotherdomonoky: thoughts about this systrace try?
20:49:25funman_i tried to add delays in the lcd driver, disable the data cache, without effect
20:49:39 Quit funman (Read error: 104 (Connection reset by peer))
20:49:51funman_if someone is going to devcon and can come in dijon on the way i can give you the c200v2
20:50:40funman_but then i'll have to do bugfixing for other targets, so in fact no :P
20:50:43bertrikI think no french people are planning to come to devcon
20:51:33 Join Lear [0] (i=chatzill@rockbox/developer/lear)
20:53:06bertrikfunman_, I hope to hack a bit on the e200v2 radio problem with domonoky's e200v2
20:53:32domonokybluebrother: first look at the code looks fine. lets try it :-)
20:53:55bertrikI'm quite convinced now that it's some kind of radio initialisation problem, not a communication problem with the radio
20:54:21bluebrotherI'm a bit unsure if we should try to get it into the next rbutil release. OTOH it might be really helpful when users have problems.
20:54:32bertrikthe giant delays that people use in the communication with the radio must have some other side-effect
20:55:50domonokybluebrother: this debugbuffer string could get very long.. perhaps there is a better way to store the history ?
20:56:16 Quit intrados (Read error: 104 (Connection reset by peer))
20:56:22bluebrotherif you have a better idea let me know :)
20:56:30 Join intrados_ [0] (
20:56:58bluebrotherI wasn't too fond of that QString myself but haven't had a better idea.
20:57:46 Quit intrados_ (SendQ exceeded)
20:58:47 Quit saratoga (" ajax IRC Client")
20:59:47 Quit jordan` (Read error: 110 (Connection timed out))
21:00:30domonokybluebrother: maybe use a QStringList and remove entrys from the beginning, when it gets "too long"
21:00:46 Quit BryanJacobs (" ajax IRC Client")
21:01:15bluebrotherthat would be a possibility but (a) what is "too long" and (b) how likely is it to remove important entries from the beginning?
21:01:37bluebrotheralso, how likely is that list to become too big on a standard PC?
21:02:39domonokydepends on how much we log.. i will now try it and see how long a "normal" log is.
21:03:09bluebrotherjust installed a build, pdf manual and fonts. Saved log is ~8k
21:03:49 Join intrados_ [0] (
21:05:10 Quit funman_ (Read error: 60 (Operation timed out))
21:05:21 Join jordan` [0] (i=gromit@
21:06:02*domonoky has to rebuild rbutil to get all qDebug() messages :-)
21:06:10bluebrotherhehe :)
21:06:27 Join Shadz [0] (
21:06:29 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
21:08:37domonokybluebrother: but it looks very usefull. if the log doesnt get too long, we should put it into svn before release.
21:09:35 Join funman [0] (n=fun@rockbox/developer/funman)
21:10:08domonokywe could even open the log window automatically when the progresslogger gets a error. (but that open it too often. i think aborting a install also counts as error)
21:10:38bluebrotherI would rather not do that but leave it to the user to use the "Troubleshoot" menu entry
21:11:11bluebrotherIMO a log of 50k should be safe on current machines. rbutil itself has a 10MB binary, so that's not much.
21:11:34domonokyyes, even 1MB of log is ok.
21:12:48bluebrotherthough a normal use shouldn't get that big
21:13:25domonokycan we make that log window automatically refresh ?
21:15:09bluebrothernot sure. At least not easily.
21:16:12domonokyit could either check for new text every second, or we somehow have to notice it on new entrys.
21:16:53bluebrotherbut is that needed at all? My idea was to have this so we can instruct users to go to that dialog, save its output and post it somewhere if we need to debug things
21:17:44bluebrotherwe could replace the qDebug() calls with something different later that is a complete singleton class. That would allow auto-refresh and stuff easily.
21:18:24domonokythats also oki. but then you can just create the windows on the stack and exec() it, then you dont need the OnClose Signal etc.
21:18:25bluebrotherpolling just for auto-updating seems a bit overkill to me.
21:19:05bluebrothertrue, but not requiring to close is isn't bad either :)
21:21:52domonokyits a bit halfway at moment. if it can stay open, it should either auto-update or be a completly independent window.
21:23:03 Quit Thundercloud (Remote closed the connection)
21:24:56 Quit JdGordon| (" ajax IRC Client")
21:25:48 Quit Zambezi (Remote closed the connection)
21:26:20bluebrotherwell, fortunately the latter is rather easy.
21:26:24domonokybluebrother: you cant click the refresh button while a progress logger is open..
21:26:35domonokyat least at moment :-)
21:27:03 Join Rondom [0] (
21:27:52bluebrotherhmm, indeed that doesn't work even when making the window top-leve.
21:29:12 Quit LambdaCalculus37 (" ajax IRC Client")
21:29:22bluebrothermaking that window modal is also an option, but that makes the refresh button somewhat superfluous.
21:34:37domonokymaking it autoupdate is also not too hard. implement a small QThread which just waits on a static QSemaphore and then calls refresh. the debug function then just flushes the semaphore when it adds a text.
21:35:50bluebrotherhmm. I'm not sure if it would be better to replace the whole qDebug() stuff with our own tracing at some later point.
21:35:53domonokyor instead of the QThread, just start a function which does the same with: QtConcurrent::run()
21:37:27bluebrotherhmm. Maybe its best to wait with that after the next release and polish it a bit more first.
21:37:28domonokyusing qDebug() benfits from Qt internal qDebug messages..
21:37:37 Join stripwax [0] (
21:38:29bluebrotherinternal debug messages? I haven't seen any. Besides, my idea was to have a program flow trace, not any Qt internal stuff. We can still use qDebug() for that later.
21:39:22domonokyah, i thought more of a logging mechanism instead of program flow.
21:40:31domonokymy log is 186KB: build+autodetection+voicefile. so it can grow quite big, but it shouldnt be really a problem.
21:40:35bluebrotherno. If we really need detailed logging we can use qDebug() in conjunction with a debug build. But my idea was to easily show what's happening −− f.e. it would have been interesting for the 404 issues.
21:41:36domonokymy log is cut-off at then end ? does QString have a size limitation ? :-)
21:42:04bluebrotheranother option would of course be to write it to the disk immediately, but I'd like to avoid writing to the disk more than needed.
21:44:10 Join tvelocity [0] (
21:44:50bluebrotherhmm. It's weird to see "encoding <foo> to <foo>.mp3" when building the voice file for a non-archos target.
21:49:18domonokybluebrother: unfortunatly thats from history. the buildsystem does it the same. (i think voicefont wants it this way)
21:49:51bluebrothercan't find anything about a size limit for QString. Maybe some of the debug output isn't using qDebug() (like voicefont)?
21:50:39 Quit biengo (Remote closed the connection)
21:50:58Horschtwhat is commit 21307 suposed to be doing, by the way?
21:51:05domonokymy first log is cut off in the middle of the word "encoding" ... now trying to repoduce it.
21:51:12Horscht"show in the main menu that there is nothing to resume to stop the annoying "nothing to resume" splash"
21:52:00Horschtresume playback is still in the main menu, even if the "current playlist" does not have any entries nor is there a song to resume...
21:52:08bluebrotherHorscht: you might be interested in that thread:
21:53:12 Quit PaulJam_ (".")
21:57:26Horschtso... it was reverted? I always thought that would show up in the commit section as well
21:59:04 Join Zambezi [0] (
21:59:24bluebrotherit wasn't reverted yet. Reverting was proposed.
22:01:31Horschtin that case: it is not working anyways.
22:02:21Horschtnow it says "Resume Playback". Pressing play or selecting that option brings a small "Loading" message, then nothing happens
22:02:24 Join gregzx_ [0] (
22:03:07 Quit gregzx (Nick collision from services.)
22:03:09 Nick gregzx_ is now known as gregzx (
22:07:22LloreanHorscht: You've got something else wrong if that's happening.
22:07:25Horschtoh, wait... I actualy had to create a small playlists and let that one finish playing...
22:08:23Horschtso, basicaly what I had was: an empty playlist making rockbox think it has something to resume when it actualy didn't. So pressing play would not do anything except for the small "Loading" dialog
22:08:32 Quit HBK (Read error: 60 (Operation timed out))
22:08:52evilnick_7Was it a playlist that had come to the end of the last song?
22:10:11 Quit funman ("leaving")
22:12:14Horschtno, it was a playlist i manualy removed all songs from
22:12:43Horschta playlist that had come to an end actualy showed me "Nothing to Resume" in the mainmenu
22:13:26Horschtbut in my first case, there was *nothing* that actualy told me that there was nothing to resume.
22:13:34 Quit wark (Read error: 104 (Connection reset by peer))
22:14:34 Join HBK [0] (
22:14:41 Quit stripwax ("")
22:15:42HorschtIf I was a blind user I would not have known...
22:20:18 Join stripwax [0] (
22:22:35 Quit dmb (Read error: 113 (No route to host))
22:22:38 Join JdGordon| [0] (i=836b0052@rockbox/developer/JdGordon)
22:23:02LearHorscht: I think that behavior is the same as before the change...
22:24:26*Horscht builds 21306
22:24:41 Join tapiocapudding [0] (n=chatzill@
22:26:49LearYep, the old code would only show the splash if resume index is -1, otherwise try to resume. The new code shows "Nothing to resume" directly in the menu if resume index < 0...
22:27:36tapiocapuddingOK, a couple of questions: I've been trying to change some of the system strings, and the guide in the wiki recommends setting up a "build environment." I did some google searches but still can't come up with any other info on this. What is a build environment, and how do I go about setting it up?
22:28:08LearCheck the developer docs on the wiki.
22:28:42 Quit stripwax ("")
22:28:46 Join funman [0] (n=fun@rockbox/developer/funman)
22:29:46Mikachufwiw i think it's a bit weird that the playlist is erased when the last track is finished
22:29:59Mikachuevery other player just starts the same playlist from the beginning if you press play again
22:30:14 Join stripwax [0] (
22:30:35 Join Thundercloud [0] (
22:31:35JdGordon|Mikachu: if we did it like that, we would need some other way to erase the playlist so you could add new tracks without having the old ones played (assumign you didnt want them)
22:31:47LloreanJdGordon|: Uhm, you erase the playlist by stopping
22:31:50LloreanWhich can still be resumed
22:31:56LloreanThat has nothing to do with end-of-playlist behaviour
22:32:03 Join raps [0] (i=raps@unaffiliated/raps)
22:32:32LloreanMikachu's suggestion wouldn't interfere with the ability to insert into a new playlist at all, really, since that just depends on being in the "stopped" state (which still has a populated playlist until you insert something new)
22:33:29LloreanIn fact, we still keep track of what was in the previous playlist when playback ends for end of playlist (you can view it in the playlist viewer I believe)
22:33:55LloreanI think the "Nothing to Resume" mostly dates back to when we didn't keep track of what was in the playlist after it ended, now that we do it kinda makes sense to allow "Resume" to re-start it.
22:36:03Mikachuyeah as long as it's stopped, inserting something removes the old playlist, even while playing there's the "play next" entry which queues the currently playing and inserts, so it is deleted when it finishes
22:36:32LloreanMikachu: I'd say post a patch for your idea, and get feedback
22:36:57 Join Harryy [0] (i=Harry@botters/harryy)
22:37:43gevaertshm, that would also make the "Nothing to Resume" debate mostly go away. It would become pretty rare...
22:37:54 Join PaulJam [0] (
22:38:38 Quit HBK (Read error: 54 (Connection reset by peer))
22:39:58 Part Harryy
22:40:15 Quit tchan (Read error: 104 (Connection reset by peer))
22:41:45MikachuLlorean: while i haven't looked at the playlist code in two years, thinking about it gives me shivers
22:41:54Mikachuis it better now? :)
22:42:17 Quit kugel_ (Read error: 113 (No route to host))
22:42:28gevaertsJdGordon|: shut up! Don't discourage people!
22:42:36***Saving seen data "./dancer.seen"
22:42:37 Join HBK [0] (
22:42:44JdGordon|but yeahm, what you want to do should be pretty straight forward
22:43:53 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
22:44:11Horschtalso, reagrding the issue I had: I can't seem to reproduce such a playlist. I must have broken something the first time around when I was deleting entries from the playlist. Now I am not able to realy delete the last entry from the playlist...
22:44:39Horschtdisregard my previous ramblings then
22:47:27 Quit stripwax ("")
22:51:22 Quit midgey|w (" ajax IRC Client")
22:52:00 Quit funman (Read error: 60 (Operation timed out))
22:55:17 Join Ubuntuxer [0] (
22:55:18pixelmawouldn't Mikachu's idea remove the possibility to start a new playlist by e.g. "Insert shuffled" on a folder, or am I misunderstanding?
22:55:35Mikachupixelma: if you are in 'stopped' it will already start a new playlist
22:55:50Mikachuor do you always skip to the last track when you want to make a new playlist?
22:55:52 Quit HBK ()
22:55:54Mikachuthat would be a horrible ui :)
22:56:15 Join funman [0] (n=fun@rockbox/developer/funman)
22:56:52Mikachuyou are listening to album A, after 3 songs you decide to listen to album B instead, what do you do?
22:57:40pixelmaok, so that's only for the case you press the "Resume" button or the menu entry?
22:58:17 Quit x72 ()
22:58:22pixelmaalright, there was my misunderstanding
22:58:46pixelmaI thought it would also have an influence on creating a new playlist
22:59:14Mikachuthat must be handled somewhere else since inserting in stop mode already works even though resume_index isn't -1
23:00:23pixelmasounds ok then (as I imagine it now)... and I'm somehow reminded of the c200 keymap patch... :\
23:01:10Mikachuheh, running make reconf also prints the "you should run 'make reconf'." warning
23:04:35 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
23:06:41Mikachuthat seemed to be easy enough
23:06:42 Quit Lear ("ChatZilla 0.9.84 [Firefox 3.5/20090612072440]")
23:07:00 Part n00b81 ("Leaving")
23:10:54 Join grimman [0] (
23:11:11bertrikfunman, did anyone ever look at the built-in bootloader of the as3525?
23:11:27grimmanHallo guys. I manually initiated a database update on my iPod, and it burned through the battery like candy and after a charge it goes at it again. What gives?
23:11:28 Join dmb [0] (n=dmb@unaffiliated/dmb)
23:11:42funmanbertrik: daniel_at had dumped the content of the ROM when our bootloader was not finished
23:12:13bertrikwe don't really need to know about it, but I'm still curious
23:12:51bertrikI noticed that some ams sansas seem to have some kind of marker at a pin of the flash chip, not just the e200v2 ones
23:13:50grimmanAlright. Nobody knows, fair enough. However, does anyone know if I can abort a database update?
23:14:39linuxstbgrimman: You need to wait more than 2 minutes for people to answer questions...
23:15:13grimmanI waited almost 2.5 minutes. :P
23:15:26grimmanBesides, if anyone DID know, they're unlikely to type that slow in a nerd channel.
23:15:55Mikachudo what you want with it;a=commitdiff;h=015f563982dd4b0e "Resume last playlist even if we reached the end."
23:16:00AlexPgrimman: People aren't constantly staring at the channel just hoping for a question to answer
23:16:05linuxstbBut anyway, what do you mean by "burned through the battery like candy" ?
23:16:17AlexPgrimman: Also, people are also in the middle of other conversations
23:16:44AlexPlinuxstb: I guess quickly?
23:16:52Horschtwouldn't clearing the settings help?
23:17:01JdGordon|Mikachu: I guarwentee nothing will happen if you dont put it on the tracker
23:17:02bertrikfunman, for example, both the clip and the fuze have those markings
23:17:08grimmanAlso, if anyone here has any control over the changes page, "Rockbox now supports scripting using the LUA language" is slightly wrong. The scripting language is called Lua. It's not an acronym. :P
23:17:15grimmanAlexP: Please calm down.
23:17:16MikachuJdGordon|: yeah i figured you'd say that :)
23:17:23 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
23:17:31bertrikfunman, do you where to get that dump or how to contact daniel_at?
23:17:33Horschti.e. shut down ipod, start ipod, as soon as the backlight turns on showing the apple logo: flip the hold switch grimman
23:17:47grimmanlinuxstb: Pretty much that I started the update and like 30 minutes later the battery was completely drained.
23:18:00AlexPgrimman: What? I'm not remotely uncalm - I was just trying to explain to you why 2 minutes is a slightly short time to expect an answer.
23:18:11grimmanAlexP: In a rude manner, no less.
23:18:23Mikachugrimman: you're being extremely oversensitive
23:18:29grimmanHorscht: Won't that just slap me into the Apple firmware?
23:18:31AlexPgrimman: Excuse me? Care to point out where I was rude?
23:18:39grimmanMikachu: Not at all.
23:18:43grimmanAlexP: The mannerism.
23:18:55funmanbertrik: i'll give you his jabber id in PM, but i don't know if he's active on rockbox anymore
23:18:56Horschtno, if you flip the switch when the backlight has turned on, you will clear rockbox settings
23:18:58linuxstbgrimman: How many tracks do you have on your device? 30 minutes seems a long time...
23:19:07grimmanlinuxstb: Seems a very long time.
23:19:21grimmanEspecially since it already had a database since before I added a couple of tracks yesterday.
23:19:35grimmanFigured I'd update at work -> keep listening. iPod did not agree.
23:19:35Horschtlinuxstb, 30 minutes is pretty much what I always got when building the database on my ipod
23:19:46grimmanHorscht: It starts from scratch?
23:19:48AlexPHorscht: Really? Which ipod?
23:19:51HorschtThat's why I asked for a PC based tool to do that
23:20:09HorschtAlexP, 80GB Ipod Video 5.5G with ~ 54GB Mp3s on it
23:20:23MikachuHorscht: the rockbox simulator?
23:20:35AlexPHorscht: That's a bit mad :)
23:21:03Horschtgranted it got quite a bit faster with dircache enabled, but it's still pretty slow on my ipod
23:21:09AlexPThat's very slow
23:21:10Mikachuwhy would you buy an 80GB mp3 player if not to put 80GB of mp3s on it?
23:21:18 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
23:21:21JdGordon|I tihnk there is a PC app to generate the db files
23:21:27JdGordon|either in the wiki or tracker....
23:21:27 Part n00b81 ("Leaving")
23:21:31grimmanMikachu: Because FLAC exists! :P
23:21:39HorschtJdGordon| i know
23:21:49AlexPMikachu: Indeed so, but there is something else at play - I have a 120 GB (albeit beast) which is full and takes nothing like that
23:21:51Horschtthat's what I am now using to build my database ;)
23:22:06Mikachuwell, i have a 2GB ipod nano so i obviously have no experience in the matter :)
23:22:18HorschtMikachu, because a 40GB Device wouldn't hold 54GB
23:22:36JdGordon|it would if you use light weight bits :)
23:22:44Horschtand it was a lot easier to find a used 80GB Ipod than a 60GB one
23:23:10JdGordon|which reminds me to buy a back plate for his 5.5g
23:23:16JdGordon|+ /me
23:23:48HorschtAlexP, as I said. I mentioned this issue quite often before. Truly, it takes only ~25 -30 minutes with dircache *disabled*
23:23:48Mikachuthat wouldn't work with /me ;)
23:24:00Horschtwith dircache enabled it "only" takes ~10 Minutes
23:24:25Horschtwhich was still too long for my taste
23:24:35AlexPHorscht: I believe you :)
23:24:46MikachuJdGordon|: want me to assign it to someone in particular?
23:25:30LloreanDircache shouldn't significantly alter database generation times, should it?
23:25:38HorschtAnd suffice to say: I would have never thought that dircache would affect the database in any way. I Always assumed from the description it would only affect people using the filebrowser
23:25:44grimmanSo... does it resume a database build where if stopped if I reboot it (or, in this case, it runs out of juice)?
23:25:52grimmanOr does it start from scratch?
23:27:08JdGordon|Mikachu: nup
23:27:15funmanbertrik: what do you mean by 'marking' ?
23:27:16Horschtthanks to FS 9371 I no longer have to build the database on my ipod
23:27:17JdGordon|Llorean: yeah, I was tihning the same...
23:27:21 Quit Shadz (Read error: 60 (Operation timed out))
23:27:37JdGordon|it would stop the directory listing coming from the disk though
23:27:46Horschtgrimman, i am not entirely sure what you are asking...
23:28:00LloreanJdGordon|: Yeah, but since the disk is spinning nearly constantly, and that should be a very small percentage of time used anyway, it shouldn't change it from 25-30 down to 10
23:28:15bertrikfunman, some white stuff also used to write numbers on the PCB
23:28:19grimmanIf an update is in progress and is interrupted for some reason.
23:28:21grimmanWith me?
23:28:29grimmanAnd I start the iPod again.
23:28:47grimmanIt starts the process again. I can feel it, the entire device is vibrating.
23:29:08grimmanBut does it start from zero or does it pick up where it left off?
23:29:23Horschtthat's a good question. I have never tried.
23:29:28bertriknear pin 9 of the flash chip, see for example
23:29:40grimmanI sort of need to know. Either that or how to stop it from restarting the process.
23:29:43Horschtdoesn't it show you a progress status message or something?
23:29:52funmanbertrik: have you looked at ONIF (Nand Flash specification) to see which pin it is ?
23:29:58grimmanNo. It does it in the background, and it hogs _all_ the CPU. So I can
23:30:02MikachuJdGordon|, pixelma: 10343
23:30:06AlexPHorscht: There is one in the debug screen
23:30:07grimman't listen to music.
23:30:15HorschtI remember seing a status message telling me "XXXXX files found" counting up
23:30:56funmanbertrik: ah i can see it clearly
23:31:06grimmanRockbox autostarting the database build seems to have rendered my iPod completely useless anyway. Boot device -> wait for battery to drain. Doesn't even take very long.
23:31:09Horschtthen again, it has been months since I used it on my Ipod thanks to the PC app, which only takes ~ 1 Minute building the database on my PC
23:31:09grimmanSuper awesome. :<
23:31:37JdGordon|then plug into the computer and edit the config.cfg so it doesnt start in the db
23:31:38bertrikit seems pin 9 is marked, which is CE1#
23:31:49Horschtgrimman, i think you can delete all the tagcahce files
23:31:52JdGordon|and anyway.. cant you back out of the db build screen?
23:32:13grimmanHorscht: I'm going to have to play with it some more tomorrow.
23:32:40AlexPHorscht: If the PC is ~25x faster I guess that isn't too suprising
23:32:52grimmanIt's either that or sleep, and I value my ability to stay awake at work. ;)
23:32:55 Quit tapiocapudding ("ChatZilla 0.9.84 [Firefox 3.0.9/2009040821]")
23:33:31Horschtas in, connect the ipod to your pc, browse to .rockbox and delete all "database*" files except database.ignore
23:33:47grimmanWill give it a whirl real quick then.
23:33:48HorschtAlexP, I didn't say it's surprising, I am just saying it's nice ;)
23:34:00grimmanIs that supposed to murder the autostarting db building?
23:34:25HorschtI assume it deletes all the database process and the whole initiation.
23:34:58grimmanBah... I'll just delete the entire directory and start fresh.
23:35:00AlexPHorscht: Yes :) - I more meant that if the PC takes a minute I'm not surprised that the player takes 25 minutes
23:35:18Horschtwell, I didn't time it....
23:35:43Horschtbut my PC clock never passes 2 minutes or more, so I said 1 minute
23:38:01Horschtin fact I have build myself a batch script that syncs my Ipod to my local music folder, resizes the coverart to 200x200, builds a database and creates a playist of all the songs :D
23:39:44 Join chandoo [0] (
23:40:50 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
23:40:56 Quit Ubuntuxer ("Leaving.")
23:41:33 Quit funman ("g'night")
23:41:40Horschtonly issue I have with it: the File structure must be exactly the same on the sourcedrive and destination drive, else I can't resize the album art properly (haven't found a way to "strip" paths in cmd)
23:42:58 Join robin0800 [0] (
23:43:41 Join saratoga [0] (i=9803c6dd@rockbox/developer/saratoga)
23:46:27saratogaits surprising that the PC is so much faster since I would expect tag parsing to be disk limited
23:46:39 Part pondlife
23:47:38Horschtwell, no idea realy
23:48:29Lloreansaratoga: Maybe database generation still needs a lot of optimizing?
23:48:37 Quit bluebrother (Read error: 60 (Operation timed out))
23:48:57Horschtanyways, I am rushing off to bed now.
23:50:20saratogai just tried on my sansa and did 2500 songs in less then a minute
23:50:44Horschton my ipod it actualy starts realy fast
23:50:53 Quit robin0800 ("Leaving")
23:50:57Horschtpretty much the same speed as you just mentioned
23:51:34Horschtbut as the filecount gets higher it *realy* slows to a crawl, adding only a few (~6) songs per 1 or 2 seconds
23:51:51 Join robin0800 [0] (
23:52:19HorschtI think at the end it's only 3-4 songs...
23:53:00Horschtbut now seriously: good night :)
23:53:45saratogacheck whats with those songs
23:55:15Lloreansaratoga: There have been other reports of database slowing as it gets to later files

Previous day | Next day