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 2009-01-07

00:05:22 Quit _Auron_ ("Infinity repeatedly denies rumours of plotting with zero to bring down the Universe.")
00:06:57 Join AndyI [0] (i=AndyI@
00:08:41 Join jeffronius [0] (n=kvirc@
00:10:08 Join nernie [0] (n=chatzill@
00:10:45 Nick fxb is now known as fxb__ (
00:11:44nernieI have an iPod Classic (80GB) and would like to install Rockbox on it. Would I use the iPod Video 60/80GB installer?
00:12:08Bagdernernie: nope, the classic is not supported by rockbox
00:12:47nernieBagder: Doh, ok. Thank you very much!
00:12:57 Join ameyer [0] (
00:13:19 Quit Thundercloud (Remote closed the connection)
00:13:30 Part nernie
00:14:20 Quit blahrus ("Ex-Chat")
00:19:09 Quit timc ("Leaving")
00:21:37 Quit bertrik (Remote closed the connection)
00:22:20 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
00:26:26 Join _Auron_ [0] (
00:26:29 Join n17ikh|Lappy [0] (
00:36:22 Quit bluebrother ("leaving")
00:41:35 Join domonoky1 [0] (
00:41:35 Quit domonoky (Read error: 104 (Connection reset by peer))
00:41:49 Join MarcGuay [0] (
00:43:10 Quit MethoS- (Remote closed the connection)
00:43:13MarcGuayRemove the warning about mingwm10.dll?
00:44:38gevaertskugel: you have multivolume issues related to RaaA?
00:46:43kugelgevaerts: well, it doesn't have mutlivolume defined
00:47:15kugeland I get errors that some structs in fat.c don't have drive/volume element
00:47:32*jhMikeS pongs the pingers
00:47:47gevaertsWhy does it even have fat.c?
00:48:33kugeljhMikeS: I updated, I think it's good now
00:48:52kugelgevaerts: I don't know, shouldn't it?
00:49:02kugelprobably not if you say that ;)
00:49:40gevaertskugel: if you're using files from the host system, which I thought is about half of the point of it, you're not going to handle FAT at all
00:50:00kugelhm, makes sense
00:50:25***Saving seen data "./dancer.seen"
00:50:27kugelanyway, I enabled those for the raaa, and still had filesystem access w/o problems
00:50:59kugelbut ok, thanks for the info, then it's building it for some reason
00:51:22linuxstbMarcGuay: Maybe better to say something like "or upgrade to v1.0.9" (for now) ?
00:51:42gevaertsI think it's probably a good idea to make sure it doesn't compile fat.c and hunt down all places wanting to use it
00:51:52linuxstbv1.0.9 doesn't appear for me on the download server I'm connecting to though...
00:56:32jhMikeSBACKLIGHT_FADING_TARGET == not controllable? :)
01:00:40linuxstbdomonoky1: Looking at the RockboxUtility wiki page (specifically the OS X bug for Sansa/Ipod), is there a reason rbutil installs the bootloader first? The manual install instructions generally say to install Rockbox first, then the bootloader, so the bootloader always has something to run when it boots.
01:01:43domonoky1linuxstb: no real reason i think.
01:02:10 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:07:56kugelshouldn't it?
01:08:06kugelI understood it from what you were saying
01:09:23kugelyou said your define is basically only needed for the setting
01:09:32jhMikeSimho, target-specific implementations imply no particular setting type. The beast is just one example of such an implementation that happens to use bool.
01:10:27kugelwell, bool is the only one that works for _TARGET anyway
01:10:54kugelif you use int, you'll likely have different int settings for different target specific backlight fading
01:11:10jhMikeSYou mean it wont work if some target selects INT and TARGET?
01:12:11kugelI have my doubts, since I imagine that different target specific implementations cannot use the same int settings (kinda by definition)
01:12:52MarcGuaylinuxstb: the iAudio7 has FIRMWARE_OFFSET_FILE_CRC as 0 and FIRMWARE_OFFSET_FILE_DATA as 8... This seems off from the TC file format...?
01:12:55jhMikeSI was thinking the setting type #define should only select the strings and settings code and be independent.
01:13:00kugelin that case _TARGET isn't sufficient
01:13:10linuxstbdomonoky1: Another question - does rbutil have 7zip support?
01:14:47domonoky1linuxstb: i dont think so i uses zlib
01:15:27linuxstbdomonoky1: Ah, OK. It's just that the original firmware for the Beast is available in .cab format inside a .exe inside a .iso inside a .zip from Toshiba's website. 7zip can hande all those formats...
01:16:11kugeljhMikeS: the strings are dependant on the settings so that doesn't quite work too (at least not in the way I thought it to work)
01:17:35jhMikeSyou mean the list in settings_list.c? That can be variable.
01:17:46domonoky1linuxstb: if really needed, i dont think it would be much of a problem to replace zlib with a 7z lib in rbutil. :-)
01:18:11linuxstbIt would make normal installs slightly faster...
01:18:44*linuxstb realises the .7z files aren't actually available...
01:19:51kugeljhMikeS: so you would do "#if defined(HAVE_BACKLIGHT_FADING_INT_SETTING) && defined(MY_TARGET)"?
01:20:45kugelanyway, I thought apps/ should only see INT_SETTING and BOOL_SETTING, but those aren't sufficient if you plan to have TARGET be usable for both
01:22:45jhMikeSI was thinking about simply turning the cfg value list into a single #define for the int setting.
01:23:54jhMikeSOf course we can just not worry about it until it comes up. I doubt it's on the horizon at all.
01:25:33kugelthat was my impression too
01:27:00kugeland a target specific fading is going to be directly in the driver, and thus probably not configurable
01:28:37Lloreanlinuxstb: If we used .7z as our main archive format for it instead of .zip (as in, not just what the util downloaded) we'd at least force MacOS users to explicitly download an archive handler that wouldn't be problematic like the OS-native one is. I'm not sure if forcing them to acquire software is good, but forcing them not to use bad software might be.
01:29:59jhMikeSEven Gigabeat F would be a bool since it's technically closest to SW fading.
01:31:14 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
01:33:01saratogalinuxstb: for what its worth, I don't think the 7zip lib can extract the nk.bin, looking at it I think its 7zip only, and that the 7zip Windows GUI probably uses some other libraries to handle extracting exe files
01:33:25saratogabut that was only from a quick glance at the linux library source so i may have missed something
01:37:37ameyerhrm, switching to .7z to make mac users to download 7zip would also force linux and windows users to download 7zip
01:38:00Lloreanameyer: Well, most people *should* be using the utility anyway
01:38:04ameyeralthough you guys might be able to include the binary with rbutil or something along those lines
01:38:10LloreanSo, in general, we won't be forcing anyone actually following the directions to download anything new.
01:38:44ameyerthat works, and making hackers download 7zip isn't that big of an additional barrier to entry
01:39:38 Quit Seed (Read error: 145 (Connection timed out))
01:39:56saratogai think recovering the nk.bin makes more sense, since it can be done more or less automatically whenever we correct the broken firmware partition
01:40:00Unhelpfulthe linux 7z also does iso and whatever exe archiving it uses - i extracted with it.
01:40:05 Quit DerDome ("Leaving.")
01:40:10 Join Dauron [0] (
01:40:10saratogathe command line too?
01:40:18 Quit _Auron_ (Read error: 54 (Connection reset by peer))
01:40:27 Nick Dauron is now known as _Auron_ (
01:40:29Unhelpfulyes, that's all i tried
01:40:31 Join Seed [0] (
01:40:37saratogasorry mean tool, good to know
01:42:56 Quit efyx_ (Remote closed the connection)
01:44:37 Part toffe82
01:45:41 Quit itcheg (" ajax IRC Client")
01:46:11 Part domonoky1
01:46:28 Quit culture (Read error: 60 (Operation timed out))
01:48:00 Join fdinel [0] (
01:48:27linuxstbsaratoga: Yes, I agree using your idea to recover the nk.bin is a nicer solution - it has the advantage of also making sure we keep the identical version nk.bin to the one the user was previously running.
01:49:03linuxstbsaratoga: I'm not sure if it's a good idea, but writing a Rockbox plugin to do that recovery would be relatively easy.
01:51:53 Quit JdGordon|zzz (Read error: 54 (Connection reset by peer))
01:52:48 Quit Seed (Read error: 60 (Operation timed out))
01:53:05 Join Seed [0] (
01:53:32 Join JdGordon|zzz [0] (
01:53:45linuxstbsaratoga: I'm not sure what you mean by "whenever we correct the broken firmware partition" though...
01:54:03*linuxstb needs to sleep and resigns himself to never being online the same time as saratoga...
01:55:06 Join culture [0] (
01:57:19kerwoodare pictures of a "real" v2 Sansa C250 still needed?
01:57:19kerwoodI made some scans of mine because I couldn't find them anywhere on the net, but I don't know if that's because they already exist somewhere, or because they're not needed.
01:58:24 Quit ameyer ("leaving")
02:04:17 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121623]")
02:04:39MarcGuaykerwood: Have you checked the c200v2 wiki page at
02:04:41 Join itcheg [0] (i=62db4767@gateway/web/ajax/
02:07:08 Join AndyIL [0] (i=AndyI@
02:08:42 Quit itcheg (Client Quit)
02:10:50 Quit AndyIL (Read error: 54 (Connection reset by peer))
02:11:49 Quit AndyI (Read error: 104 (Connection reset by peer))
02:13:14kerwoodMarcGuay: yes, which is what made me wonder...
02:13:33MarcGuayIf there are none on that page it can't hurt to add some.
02:17:45 Part XavierGr
02:17:51 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
02:18:54 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
02:19:47kerwoodIn that event I should like wiki write permssion, name: MarshallJose
02:20:18JdGordonanyone know how to build rbutil release on the mac? is it just svn up + make?
02:21:01 Quit culture (Read error: 110 (Connection timed out))
02:25:16saratogakerwood: added
02:30:10saratogalinuxstb: the Beast comes with an invalid firmware partition that bothers the linux driver
02:30:36saratogathe windows one doesn't seem to mind however, so technically its only needed if the user intends to use linux
02:30:47saratogaactually the data partition may have it too
02:31:16saratogasince i couldn't seem to mount it in windows, but i'm not the best one to to ask about that since i've intentionally not changed my partitions at all on the beast
02:32:00 Quit saratoga ("CGI:IRC (EOF)")
02:42:07 Join AndyI [0] (i=AndyI@
02:46:55 Join Darksair [0] (n=user@
02:50:27***Saving seen data "./dancer.seen"
02:54:21 Quit MarcGuay ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
02:55:03 Join robin0800 [0] (
03:00:04 Join Dauron [0] (
03:05:02 Quit _Auron_ (Read error: 60 (Operation timed out))
03:05:02 Nick Dauron is now known as _Auron_ (
03:07:58 Quit AndyI (Read error: 104 (Connection reset by peer))
03:11:33 Quit Darksair ("Use the Force, Luke!")
03:12:14 Join toffe82 [0] (n=chatzill@
03:22:51 Quit PaulJam (".")
03:23:23 Quit robin0800 (Remote closed the connection)
03:33:58 Join Darksair [0] (n=user@
03:41:21 Join Willwolfe [0] (
03:46:34 Quit blithe ("Lost terminal")
03:46:45 Join blithe [0] (
03:57:21 Join itcheg [0] (i=62db4767@gateway/web/ajax/
04:01:12 Join robin0800 [0] (
04:03:12 Quit robin0800 (Client Quit)
04:03:29 Join robin0800 [0] (
04:05:06 Join blkhawk- [0] (
04:12:50 Quit tvelocity (Remote closed the connection)
04:15:19 Join Barahir_ [0] (
04:20:23 Join sbhsu [0] (
04:21:00 Quit blkhawk (Read error: 110 (Connection timed out))
04:21:04 Nick blkhawk- is now known as blkhawk (
04:28:50 Quit robin0800 ("Leaving")
04:29:12 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
04:29:15 Join robin0800 [0] (
04:30:52 Quit itcheg (" ajax IRC Client")
04:32:21 Quit Barahir (Read error: 113 (No route to host))
04:35:56 Quit miepchen^schlaf (Connection timed out)
04:36:56 Join itcheg [0] (i=62db4767@gateway/web/ajax/
04:39:52 Quit robin0800 (Remote closed the connection)
04:40:10 Join robin0800 [0] (
04:41:25 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
04:41:45 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
04:47:21 Quit kadoban (Read error: 113 (No route to host))
04:50:32***Saving seen data "./dancer.seen"
04:56:14 Nick Bensawsome is now known as BensJammin (n=Bensawso@unaffiliated/bensawsome)
04:59:53 Quit robin0800 (Read error: 60 (Operation timed out))
05:03:47 Join plutonian [0] (
05:04:30 Join AndyI [0] (i=AndyI@
05:04:44 Nick BensJammin is now known as Bensawsome (n=Bensawso@unaffiliated/bensawsome)
05:05:36 Quit plutonian (Client Quit)
05:05:49 Quit Seed ("cu, Andre")
05:10:54 Quit itcheg (" ajax IRC Client")
05:26:50 Quit jfc (Read error: 104 (Connection reset by peer))
05:27:18 Join jfc [0] (
05:27:27 Join Rob2223 [0] (
05:27:36 Join midkay_ [0] (
05:27:38 Quit Rob2222 (Read error: 104 (Connection reset by peer))
05:28:05 Join toffe82_ [0] (n=chatzill@
05:36:07 Quit Willwolfe ("ChatZilla 0.9.84 [Firefox 3.0.1/2008070208]")
05:38:41 Quit Nibbler (Remote closed the connection)
05:38:46 Join fyrestorm [0] (
05:43:09 Quit toffe82 (Read error: 110 (Connection timed out))
05:44:20 Part toffe82_
05:44:41 Quit midkay (Read error: 110 (Connection timed out))
05:57:03 Quit amiconn (Nick collision from services.)
05:57:06 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
06:01:14 Join wmw1900 [0] (
06:01:50 Nick wmw1900 is now known as Walter (
06:02:13 Quit Walter (Client Quit)
06:08:02 Quit aurix_lexico ("Leaving.")
06:14:53 Quit pixelma (Read error: 110 (Connection timed out))
06:17:48 Join gartral1 [0] (n=Gartral@
06:20:53gartral1r19702m has a bug: the backlight turns off, then, back on
06:24:53 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:41:26Unhelpfulsaratoga: if i recall correctly, the data partition is normal, but that doesn't matter, since linux refuses to expose any of the partitions if it doesn't like the table
06:45:06*lucent reads backlog
06:48:53 Part gartral1
06:50:36***Saving seen data "./dancer.seen"
06:59:11 Join midkay [0] (n=midkay@rockbox/developer/midkay)
07:02:03 Quit midkay_ (Read error: 60 (Operation timed out))
07:18:07 Quit jhulst (Remote closed the connection)
07:26:29 Join kadoban [0] (
07:59:09Angesdo many people break their players when testing the early versions of rockbox?
07:59:36scorcheit depends on the device
08:00:04 Quit BHSPitLappy (Remote closed the connection)
08:01:37JdGordonregardless of player I'd still say the count of dead players because of rockbox is pretty damn small
08:02:55Angeslucent sent me rockbox for the sansa fuze, with so many warnings it's scary!!
08:05:06scorcherockbox is not yet deemed ready for public consumption for that device yet...while there may be little chance of it permanently doing something harmful to your device, you should be aware that the port is mainly for developers to...well...develop at this stage and that the port is unsupported at this point
08:06:07AngesI want to help with the tests, not just using it
08:09:16 Join GodEater_ [0] (i=c2cbc962@gateway/web/ajax/
08:11:35 Join Willwolfe [0] (
08:18:56 Join kugel [0] (n=chatzill@unaffiliated/kugel)
08:24:52 Quit JdGordon (Remote closed the connection)
08:26:57 Quit Acksaw (Connection reset by peer)
08:27:10 Join Acksaw [0] (
08:28:13 Join bertrik [0] (
08:30:08 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
08:32:08 Join Rob2222 [0] (
08:32:31 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:33:00 Part _Auron_
08:35:15 Quit BigBambi (Read error: 113 (No route to host))
08:43:16 Join timc [0] (n=aoeu@
08:46:20JdGordonanyone wanna test a quick patch which fixes up the debug menu wierdness? (and changes the simplelist handling a bit which is what actually needs testing)
08:46:26ZagorUnhelpful: I bet you won't be surprised by me saying I don't think 576 bytes saving is worth the hassle.
08:46:55UnhelpfulZagor: no, i suppose i'm not shocked. :)
08:49:09 Quit Anges ("Leaving.")
08:49:33Unhelpfulwould using #define open rb->open be more acceptable? it leaves a good deal less mess in the core file than API() does.
08:50:05Unhelpfuli just don't really like increasing the core binsize, without any new features, on some of the targets with the least memory
08:50:23Zagoras amiconn said #define open will break on sims, since they already do that
08:50:27Unhelpfulwithout any new *core* features, sorry.
08:50:36 Quit Rob2223 (Read error: 110 (Connection timed out))
08:50:39***Saving seen data "./dancer.seen"
08:51:21Unhelpfulhrm. but, plugins on sim are surely still expected to use the plugin_api?
08:52:14Zagorindeed. I need to look at that.
08:52:20 Join LinusN [0] (n=linus@gateway/web/cgi-irc/
08:52:58Unhelpfulif that is indeed the case, since such a macro would only happen when the file is compiled as a plugin, it should still be safe.
08:57:40Unhelpfuli could "just try it"... but i should probably ask somebody who knows how the sim works ;)
08:58:05Zagorit ought to work. open() is redefined in firmware/include/file.h, which is never included in plugins
08:58:23Zagoryeah, try it
08:58:54Unhelpfuleven if so, as long as plugins are still supposed to call via the API, i can undef it, first.
08:59:18Unhelpfulyeah, i'll save the hack-job for scaling support in-core to a branch, in case we really need to go that way
08:59:46Zagorsim plugins are definitely supposed to go via the api
09:00:24 Join pixelma [0] (
09:00:45Unhelpfulgood deal... i think the plugin-only x_init function and plugin_api pointer should just go into the wrapper file, as well...
09:01:12Unhelpfulare the archos targets the ones where we have non-square pixels?
09:01:31ZagorI don't know
09:03:02amiconnUnhelpful: Yes, although the iAudio remotes also have (slightly) non-square pixels
09:03:13amiconnThat fact is ignored by rockbox atm
09:03:34 Join Anges [0] (
09:03:44Unhelpfulamiconn: there's no scaler output for remotes at present, though a format plugin could provide it.
09:04:13amiconnWell, the M3 uses the remote as its main display
09:04:19Unhelpfulright :/
09:04:59Unhelpfulthen i already have *somewhere* that i need to do AR compensation. do we have accurate PAR values for the displays?
09:05:13 Quit Willwolfe ("ChatZilla 0.9.84 [Firefox 3.0.1/2008070208]")
09:05:27amiconnUnhelpful: open() and friends are prefixed with 'sim_' in the sims. Check the PREFIX() business in plugin.[ch]
09:06:04Unhelpfulaspect ratio, sorry. and Pixel aspect ratio.
09:06:06Zagoramiconn: hmm, why are we doing that?
09:07:32amiconnBecause plain open() etc are the host's functions which we have to call. Hence our own functions (which are wrappers for the host functions) need to be named differently
09:08:01Zagoramiconn: but the plugin functions are just pointers in a struct. why do they have to be renamed?
09:08:01amiconnUnhelpful: Archos pixels are taller than wide, w=0.8*h
09:08:43Unhelpfuli *think* that if we just apply an appropriate multiplier to each of the input dimensions before calculating the output dimensions, it will be fine
09:09:45Unhelpfulso, on sim, the call should be rb->sim_open?
09:10:04amiconnThe iAudio remote pixles are wider than tall, but only a little bit. I measured AR some time but didn't document it :\
09:10:09pixelmaUnhelpful: Archos display pixels are 80% of their height wide - example: the chessboard in th plugin is 80x64 to compensate that and looks square on the display
09:10:22*pixelma too slow
09:12:45amiconnZagor: I don't remember the reason, but there was one. Needs a little digging into history
09:13:06Unhelpfuli think i can make the math work, as long as i have the actual pixel dimensions. the only issue i sees is the constraints, and i think that pretty much needs to be left up to the plugin developer - it seems more sane to ask them to use a non-square region to get a square image on archos than to ask them to deal with scaled bitmaps on archos actually being larger or smaller than the requested size.
09:13:08Zagorhmm, it was done by bagder nearly four years ago
09:13:29 Join wjx [0] (n=debc0a28@gateway/web/cgi-irc/
09:14:09amiconnUnhelpful: Does the scaler keep aspect, or is that something the caller needs to do?
09:14:55 Join ender` [0] (
09:14:59Unhelpfulamiconn: the scaler keeps aspect if the caller asks it to. basically, a flag determines whether the specified rectangle is the desired size of the output, or a constraint for that size.
09:15:35 Join btbr [0] (n=79eb3100@gateway/web/cgi-irc/
09:16:45amiconnSo I don't see the problem. If it is asked to keep aspect, it can take a non-square pixel aspect into account
09:17:20amiconnAnd if it's not, there's nothing the scaler can do and the caller needs to handle it
09:17:36Unhelpfulwell, the question is whether the constraining rectangle is in display pixels. i think it pretty much has to be, for things to be sane.
09:18:55 Quit wjx ("CGI:IRC (Ping timeout)")
09:19:12 Join wjx [0] (n=debc0a28@gateway/web/cgi-irc/
09:19:29wjxI'm in
09:19:55wjxsay something,Wrong!
09:20:30btbryou idiot
09:21:08wjxHey,guys,dit M6 progect alive?
09:22:50Unhelpfulthe info about ports-in-progress is generally in the wiki or an appropriated forum thread. you should look there instead of asking for status updates here.
09:23:28Unhelpfulespecially since developers work on what interests them, and on hardware they actually have... so there may very well be nobody in here who knows anything about a particular device at the moment.
09:23:33wjxI know it
09:24:43wjxIt's a great job for rockbox members
09:25:30wjxYour name is so intersting
09:30:40 Nick Barahir_ is now known as Barahir (
09:33:38JdGordonhmm... now resizing AA is in, should I remove the AA size in the info screen?
09:34:33Unhelpfulwell, we do still do the check for "prescaled" AA... something else that it might be a good idea to remove, at this point.
09:37:38Zagoramiconn: as far as I can see the only reason is that we have poor #include control in plugins. this should be fixed.
09:44:10ZagorI especially "like" that codecs.h have prefix around some of the file functions, but not others. open, but not close etc.
09:44:21JdGordondoes anyone know what the difference between gwps.c and gwps-common.c is?
09:47:38 Quit btbr ("CGI:IRC (Ping timeout)")
09:49:20 Join Thundercloud [0] (
09:50:05 Quit wjx ("CGI:IRC (EOF)")
09:51:47 Quit axionix (Read error: 104 (Connection reset by peer))
09:53:32Unhelpfulso, button_get(true) basically no longer works for waiting for a button press... but it looks like it should be enough to check the return value and loop?
09:54:37Unhelpfulthe greylib scaler test plugin needs fixed for this... and i think ppmviewer, as well, since that's what i borrowed the method from.
09:55:07 Join axionix [0] (
09:55:51JdGordonUnhelpful: no.. it was changed back to how it should work...
09:55:58JdGordonso it will block properly again
09:56:00Unhelpfuloh! i missed that, then :)
09:57:39Zagoreek. what is the purpose of intptr_t?
09:58:04UnhelpfulZagor: it's an integer the same size as a pointer.
09:58:06JdGordon64bit builds...
09:58:37JdGordonis using linked lists much slower than a simple array?
09:59:03ZagorI see what it is, but when is it needed?
09:59:04UnhelpfulJdGordon: depends on the operation being performed.
09:59:37UnhelpfulZagor: when you need to do operations other than simple addition/subtraction on pointers, or at least that's the main use i know of
09:59:54 Quit bertrik ("Leaving")
09:59:55JdGordonUnhelpful: replacing a for loop with lots in the middle with a LL
10:00:04Unhelpfuliterating a LL is going to be pretty fast.
10:01:01Unhelpfulinsert/delete are of course much faster than an array. looking up an arbitrary value is much slower, and you'll probably want some more advanced structure instead of a "pure" LL
10:02:12JdGordonnot really... it would only be used with iterating and searching (one direction), adding will always be at the end... the reason would be to use that instead of a static array so the count isnt limited
10:02:15kugel"insert/delete are of course much faster than an array" how that?
10:02:35Zagorkugel: if you need to insert in the middle
10:02:41JdGordonfixing pointers instead of memcpy/memcpy
10:02:48Unhelpfulkugel: if you insert, not merely modify, a value in the middle of an array, you have to move the other values.
10:03:47Unhelpfulin a LL, you just make a new node, change the next pointer for the node it's being inserted after, and set its next pointer to the node that belongs after it.
10:04:23 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
10:04:50UnhelpfulJdGordon: if you're never doing inserts in the middle, why not an array that you expand the allocation of? dynamic size is going to require *some* kind of dynamic allocation...
10:05:36Zagorwhy is queue_event using intptr_t rather than long for data? surely not many systems have a smaller long than int*
10:06:01Zagorit seems this is the cause of most other uses of intptr_t
10:06:14JdGordonis intptr_t causing problems?
10:06:34Zagornew types are bad
10:07:17JdGordonUnhelpful: thinking about mem mamangement in the wps... considering using a LL for the tokens instead of a static buffer... its size is unknown at the start so a LL makes sense.. but im thinking maybe its not worth the effort
10:07:49Zagorand besides, yes it does cause problems since files need to be included for its' use.
10:08:13Zagorand that causes #include ballooning
10:08:15Unhelpfuli guess i'm just saying, if you can allocate space for new LL nodes, you can allocate space for a larger array. and the array will use less memory.
10:09:25Zagorr11818 jhmikes. "Convert queues to use intptr_t for event data and return values as most of the time pointer are not passed and it should make some things a bit cleaner."
10:09:38Zagoris it too late to say "I disagree?" ;-)
10:10:39ZagorI agree void* was not ideal. but long would be better than a new type.
10:11:14JdGordonwould a simple intptr_t -> long patch work with 64bit sims?
10:11:25JdGordonand by work, i mean not cause annoying warnings
10:11:48ZagorJdGordon: since everything is casted to intptr_t today, I would think so
10:12:13 Nick Bagderr is now known as B4gder (n=daniel@gateway/web/cgi-irc/
10:22:50kugelUnhelpful: oh yea, I wasn't thinking of inserting in the middle (I had queue in mind)
10:31:00 Quit Thundercloud (Remote closed the connection)
10:31:16 Quit XavierGr ()
10:31:44 Quit GodEater_ (" ajax IRC Client")
10:33:20 Quit kachna (Read error: 113 (No route to host))
10:34:05 Join GodEater_ [0] (i=c2cbc962@rockbox/staff/GodEater)
10:36:45Unhelpfulok, API is gone :)
10:37:00ZagorUnhelpful: sim prefix removal patch coming up soon
10:37:34Unhelpfulok, i've got other things to do here, so i'll hold on to this cleanup for now
10:41:55JdGordonwhere shold the .map be? I cant seem to find it
10:43:54ZagorJdGordon: for the core bin? it should be in the same dir as .bin and .elf
10:44:06Zagori.e. the root
10:44:27JdGordonis it not generated for the sim maybe?
10:44:34Zagoran, no
10:50:40***Saving seen data "./dancer.seen"
10:54:25JdGordonif the wps is the while playing screen... what should the actual drawing code be called? theme_engine or something?
10:55:02JdGordonor keep that as wps and use now_playing or something for the actual screen logic
11:05:41JdGordonhave you all forgotten the name is the most important part???!
11:06:24Zagorwps is rivaling playback for hot-potato-ness :-)
11:08:04ZagorI suppose if the drawing code is used globally, something with "theme" would be appropriate
11:09:15B4gderI'd prefer still calling the actual screen wps as the S is for screen
11:09:29amiconnZagor: Why remove the prefixes? The method is proven to be reliable
11:09:52Zagoramiconn: reliable? why should the code become less reliable without the prefix?
11:10:33ZagorI removed them because they don't belong there. the prefix is a simulator implementation issue, not an plugin api issue.
11:11:36Zagoras it was, it had even infected the implementation of plugins
11:12:53ZagorB4gder: who runs the build server? it is throwing odd warnings: "htobe16" redefined
11:12:56amiconnNo, but open is mapped to sim_open by a macro for the sims, and depending on how you write the rb->open call, it may or may not get replaced by sim_open, depending on what headers are (recursively) included
11:13:08Zagoramiconn: not any more
11:13:14*scorche assumes obo
11:13:18amiconnI remember having run into this issue in the past. Just having the prefix everywhere avoids it.
11:13:24*B4gder hands scorche 10 points
11:14:17Zagoramiconn: that was a kludge. my commit is a fix.
11:14:21lucentoh um, could I bug someone to sponsor a commit on FS #9647 ?
11:16:43Zagorlucent: rasher has grabbed it. bug him :-)
11:16:54kadobanwhen i open and close a menu in a plugin, the screen gets offset downward by the size of the status bar (even after redraws). is that known about, or should i add a bug or something?
11:17:34lucentZagor: his last comment on the task was a question about freeze
11:17:39lucentI don't know the answer
11:17:41lucentdo you?
11:18:03Zagorlucent: well the freeze is over ...
11:18:12 Join merbzt [0] (n=benlar@
11:18:17 Quit kugel (Remote closed the connection)
11:18:52JdGordonkadoban: what revision nummber? I hopefully fix that one
11:19:29kadobanJdGordon: one today. i think it was 19702. i can check in the most recent one
11:19:46JdGordoncheck the most recent
11:19:51JdGordoni tihnk my last commit fixed that
11:20:03kadobanJdGordon: oh nice, i'll check then
11:20:55 Quit plus_M (Remote closed the connection)
11:21:12lucentrasher: ping w/re FS #9647
11:21:50 Quit blithe (Remote closed the connection)
11:22:18 Join plus_M [0] (
11:26:24 Join PaulJam [0] (
11:27:03 Join blithe [0] (
11:27:13kadobanJdGordon: yeah, it does seem to be fixed in the most recent. good job :)
11:31:05 Join kachna [0] (
11:32:20PaulJamhi, are there any known problems when compiling the uisim under cygwin? i get some warnings (warning: -ffunction-sectio
11:32:20PaulJamns may affect debugging on some targets) and when it compiles doom some errors. (i did make veryclean and ran configure before compiling.)
11:40:39 Join culture [0] (
11:52:40 Join flydutch [0] (
11:56:41oboZagor: that box is running libc6-dev v 2.9 - I don't know yet if that change is from upstream or a distro patch (kubuntu)
11:58:25 Join moos [0] (i=Mustapha@rockbox/staff/moos)
12:00:27jhMikeSZagor: How is intptr_t a new type? It was there from the start, isn't exactly an obscure one and ensures it's correct for integral and pointer data.
12:02:49 Join kugel [0] (n=chatzill@unaffiliated/kugel)
12:09:14ZagorjhMikeS: what does it do that a long doesn't?
12:14:47jhMikeSAs long as long can always fit a pointer in every case, nothing. Otherwise it's trouble for sure.
12:16:13 Quit JdGordon (Remote closed the connection)
12:16:18Zagoras I said earlier, I know of no systems where sizeof(int*) > sizeof(long)
12:17:11Zagorsure, it's part of C99. but I dislike the use if it for the same reason I dislike using uint32_t where it isn't strictly necessary
12:17:48B4gderwin64 does have 32bit longs but 64bit pointers...
12:17:55jhMikeSActually, in Windows 64-bit a pointer is bigger than a long
12:18:07*jhMikeS got beaten
12:18:22Zagorwindows saves the day again... :-)
12:18:35Zagorok, I give in
12:19:42Zagorhow do you declare a "system" int on win64?
12:19:55Zagorlong long?
12:20:05 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
12:20:26B4gdera "system" int is still int, isn't it?
12:20:29 Join MTee [0] (n=mtee@
12:20:54ZagorB4gder: so int is larger than long? isn't that forbidden?
12:20:56 Join JdGordon_ [0] (n=jonno@rockbox/developer/JdGordon)
12:21:18B4gderZagor: no int is 32 bit
12:21:20MTeeHi all
12:21:26B4gderbut so it is on most *nixes on 64bit as well
12:21:52B4gderI don't get what you mean with "system" int
12:21:53 Quit JdGordon_ (Client Quit)
12:21:56kadobanB4gder: i believe he means, how do you declare a 64 bit integer? long long should work, right?
12:22:12ZagorB4gder: I mean register width
12:22:13B4gderyes, long long should work afaik
12:22:52B4gderthey use the type jhMikeS mentioned
12:23:00B4gderthey as in Microsoft
12:23:16B4gder"The _PTR suffix means that the integer has the same size as a native pointer."
12:23:54kadobanthat seems odd. if i saw those, i'd assume they were actual pointers
12:24:15B4gderI agree, it's a weird name
12:24:37Zagorin microsoft land, the type is shown in the variable name rather than the type name ;-)
12:24:57kadobanoh yeah, i forgot. hungarian notation...that stuff sucks
12:24:58*Zagor wanders off to #rockbox-community
12:25:25 Join BdN3504 [0] (n=55b23f9e@gateway/web/cgi-irc/
12:26:25 Quit JdGordon (Remote closed the connection)
12:27:57 Quit gevaerts (Nick collision from services.)
12:28:12 Quit fyrestorm (
12:28:12 Quit Darksair (
12:28:12 Quit jeffronius (
12:28:12 Quit Anges (
12:28:12 Quit sbhsu (
12:28:12 Quit sarixe (
12:28:12 Quit GodEater (
12:28:12 Quit rasher (
12:28:12 Quit PaulJam (
12:28:12 Quit markun (
12:28:12 Quit SUSaiyan (
12:28:12 Quit crashd (
12:28:12 Quit ch4os (
12:28:12 Quit goffa (
12:28:12 Quit courtc (
12:28:12 Quit larstobi (
12:28:12 Quit GodEater_ (
12:28:12 Quit AndyI (
12:28:12 Quit japc (
12:28:12 Quit scorche (
12:28:12 Quit CaptainKwel (
12:28:12 Quit jon-kha (
12:28:12 Quit Slasheri (
12:28:12 Quit desowin_ (
12:28:12 Quit J-23 (
12:28:12 Quit parafin (
12:28:12 Quit freqmod_qu (
12:28:12 Quit BdN3504 (
12:28:12 Quit pixelma (
12:28:12 Quit amiconn (
12:28:12 Quit n17ikh|Lappy (
12:28:12 Quit Bagder (
12:28:12 Quit amigan_ (
12:28:12 Quit write__erase (
12:28:12 Quit unstable (
12:28:12 Quit crwl (
12:28:12 Quit lastebil (
12:28:12 Quit Hadaka (
12:28:12 Quit BlakeJohnson86 (
12:28:12 Quit moos (
12:28:12 Quit culture (
12:28:12 Quit jhMikeS (
12:28:12 Quit trisiak (
12:28:12 Quit thegeek (
12:28:12 Quit flux (
12:28:12 Quit loswillios (
12:28:12 Quit fred_2 (
12:28:12 Quit pabs (
12:28:12 Quit soap (
12:28:12 Quit Neovanglist (
12:28:12 Quit TMM (
12:28:12 Quit reacocard (
12:28:12 Quit timc (
12:28:12 Quit midkay (
12:28:12 Quit obo (
12:28:12 Quit Beta2K (
12:28:12 Quit merbzt (
12:28:12 Quit Zagor (
12:28:12 Quit Rob2222 (
12:28:12 Quit jfc (
12:28:12 Quit blkhawk (
12:28:26 Quit HBK (
12:28:26 Quit HellDragon (
12:28:26 Quit linuxstb (
12:28:26 Quit alexbobp (
12:28:26 Quit feisar (
12:28:26 Quit plus_M (
12:28:26 Quit axionix (
12:28:26 Quit ender` (
12:28:26 Quit Acksaw (
12:28:26 Quit kerwood (
12:28:26 Quit ajb` (
12:28:26 Quit dionoea (
12:28:26 Quit martian67 (
12:28:26 Quit Unhelpful (
12:28:26 Quit synergist (
12:28:26 Quit balou (
12:28:26 Quit suom1 (
12:28:26 Quit LinusN (
12:28:26 Quit lostlogic (
12:28:26 Quit liiwi (
12:28:26 Quit Dieterbe (
12:28:26 Quit shodanX (
12:28:26 Quit preglow (
12:28:26 Quit DaCapn (
12:28:26 Quit Zambezi (
12:28:26 Quit tim__b (
12:28:26 Quit Kopfgeldjaeger (
12:28:26 Quit Tuplanolla (
12:28:26 Quit agaffney (
12:28:26 Quit kachna (
12:28:26 Quit advcomp2019 (
12:28:26 Quit Lss (
12:28:26 Quit lucent (
12:28:26 Quit esthar (
12:28:26 Quit DataGhost (
12:28:26 Quit B4gder (
12:28:26 Quit Xerion (
12:28:26 Quit shadearg (
12:28:26 Quit avacore (
12:28:26 Quit ChanServ (
12:28:26 Quit Barahir (
12:28:26 Quit Bensawsome (
12:28:26 Quit gromit`` (
12:28:26 Quit troy___ (
12:28:26 Quit scorche|sh (
12:28:26 Quit tchan (
12:28:26 Quit fxb__ (
12:28:26 Quit killan (
12:28:26 Quit Galois (
12:28:26 Quit romain (
12:28:26 Quit offset (
12:28:26 Quit hobbs (
12:28:26 Quit flydutch (
12:28:26 Quit kadoban (
12:28:26 Quit Slack_ (
12:28:26 Quit yosafbridge (
12:28:26 Quit Llorean (
12:28:26 Quit krazykit (
12:28:27 Quit tmzt (
12:28:27 Quit Tristan (
12:30:12NJoinChanServ [0] (ChanServ@services.)
12:30:12 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
12:30:12NJoinBdN3504 [0] (n=55b23f9e@gateway/web/cgi-irc/
12:30:12NJoinmoos [0] (i=Mustapha@rockbox/staff/moos)
12:30:12NJoinflydutch [0] (
12:30:12NJoinculture [0] (
12:30:12NJoinkachna [0] (
12:30:12NJoinPaulJam [0] (
12:30:12NJoinplus_M [0] (
12:30:12NJoinmerbzt [0] (n=benlar@
12:30:12NJoinGodEater_ [0] (i=c2cbc962@rockbox/staff/GodEater)
12:30:12NJoinB4gder [0] (n=daniel@gateway/web/cgi-irc/
12:30:12NJoinaxionix [0] (
12:30:12NJoinender` [0] (
12:30:12NJoinAnges [0] (
12:30:12 Join pixelma [0] (n=pixelma@rockbox/staff/pixelma)
12:30:12 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
12:30:12NJointimc [0] (n=aoeu@
12:30:12NJoinZagor [0] (n=bjorn@rockbox/developer/Zagor)
12:30:12NJoinRob2222 [0] (
12:30:12NJoinAcksaw [0] (
12:30:12NJoinkadoban [0] (
12:30:12NJoinmidkay [0] (n=midkay@rockbox/developer/midkay)
12:30:12NJoinamiconn [50] (n=jens@rockbox/developer/amiconn)
12:30:12NJoinfyrestorm [0] (
12:30:12NJoinjfc [0] (
12:30:12NJoinAndyI [0] (i=AndyI@
12:30:12NJoinadvcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
12:30:12NJoinsbhsu [0] (
12:30:12NJoinBarahir [0] (
12:30:12NJoinblkhawk [0] (
12:30:12NJoinDarksair [0] (n=user@
12:30:12NJoinn17ikh|Lappy [0] (
12:30:12NJoinjeffronius [0] (n=kvirc@
12:30:12NJoinLss [0] (
12:30:12NJoinjapc [0] (
12:30:12NJoinjhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
12:30:12NJoinSlack_ [0] (
12:30:12NJoinparafin [0] (
12:30:12 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
12:30:12NJoinBagder [241] (n=daniel@rockbox/developer/bagder)
12:30:12NJoinmarkun [50] (n=markun@rockbox/developer/markun)
12:30:12NJoinscorche [0] (i=Blah@rockbox/administrator/scorche)
12:30:12NJoinlucent [0] (
12:30:12NJoinkerwood [0] (
12:30:12NJoingromit`` [0] (
12:30:12NJoinCaptainKwel [0] (
12:30:12NJoinyosafbridge [0] (
12:30:12Mode"#rockbox +o ChanServ " by
12:30:12NJointroy___ [0] (n=toppy@
12:30:12NJoinesthar [0] (
12:30:12NJoinsarixe [0] (
12:30:12NJoinjon-kha [0] (
12:30:12NJoinHBK [0] (
12:30:12NJoinGodEater [0] (n=ge@rockbox/staff/GodEater)
12:30:12NJoinagaffney [0] (n=agaffney@gentoo/developer/agaffney)
12:30:12NJoinNeovanglist [0] (i=Neovangl@
12:30:12NJoinSlasheri [0] (i=miipekk@rockbox/developer/Slasheri)
12:30:12NJoinlarstobi [0] (n=larstobi@
12:30:12NJoindesowin_ [0] (n=desowin@
12:30:12NJoinfred_2 [0] (
12:30:12NJoinlastebil [0] (
12:30:12NJoinflux [0] (
12:30:12NJoincrashd [0] (
12:30:12NJoinamigan_ [0] (
12:30:12NJoinfreqmod_qu [0] (i=quassel@2001:700:300:1430:213:d3ff:fee9:5ed0)
12:30:12NJoinrasher [50] (n=rasher@rockbox/developer/rasher)
12:30:12NJoincourtc [0] (n=court@unaffiliated/courtc)
12:30:12NJointrisiak [0] (
12:30:12NJoinloswillios [0] (n=jan@unaffiliated/loswillios)
12:30:12NJoinHadaka [0] (
12:30:12NJoinpabs [0] (
12:30:12NJoingoffa [0] (n=goffa@
12:30:12NJoinBlakeJohnson86 [0] (
12:30:12NJoinreacocard [0] (i=reacocar@
12:30:12NJoinch4os [0] (n=ch4os@gentoo/user/ch4os)
12:30:12NJoinwrite__erase [0] (
12:30:12NJoinsoap [50] (n=soap@rockbox/staff/soap)
12:30:12NJoinJ-23 [0] (
12:30:12NJointhegeek [0] (
12:30:12NJoinunstable [0] (i=unstable@tor/regular/sid)
12:30:12NJoinSUSaiyan [0] (
12:30:12NJoinTMM [0] (
12:30:12NJoincrwl [0] (
12:30:12NJoinDataGhost [0] (i=dataghos@unaffiliated/dataghost)
12:30:12NJoinajb` [0] (
12:30:12 Join obo [0] (n=obo@rockbox/developer/obo)
12:30:12 Join Llorean [0] (n=DarkkOne@rockbox/administrator/Llorean)
12:30:12 Join HellDragon [0] (n=jd@Wikipedia/HellDragon)
12:30:12NJoinkrazykit [0] (
12:30:12NJoinBeta2K [0] (
12:30:12NJoinDieterbe [0] (
12:30:12 Join dionoea [0] (n=dionoea@videolan/developer/dionoea)
12:30:12NJointmzt [0] (n=tmzt@
12:30:12NJoinXerion [0] (
12:30:12NJoinmartian67 [0] (i=lol3izer@about/linux/regular/martian67)
12:30:12NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
12:30:12NJoinalexbobp [0] (n=alex@
12:30:12NJoinfeisar [0] (
12:30:12NJoinlinuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
12:30:12NJoinUnhelpful [0] (n=Militant@rockbox/developer/Unhelpful)
12:30:12NJoinsynergist [0] (
12:30:12NJointim__b [0] (
12:30:12NJoinshodanX [0] (
12:30:12NJoinDaCapn [0] (
12:30:12NJoinpreglow [0] (
12:30:12NJoinliiwi [0] (
12:30:12NJoinlostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
12:30:12NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
12:30:12NJoinbalou [0] (
12:30:12NJoinsuom1 [0] (i=markus@2001:4c40:1:0:0:0:0:11)
12:30:12NJoinKopfgeldjaeger [0] (
12:30:12NJoinZambezi [0] (
12:30:12NJoinshadearg [0] (
12:30:12 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
12:30:12NJoinoffset [0] (
12:30:12NJoinavacore [0] (
12:30:12NJoinfxb__ [0] (
12:30:12NJoinkillan [0] (
12:30:12NJoinGalois [0] (
12:30:12NJoinromain [0] (
12:30:12NJoinTristan [0] (
12:30:12NJoinhobbs [0] (n=nandrew@p3m/member/hobbs)
12:31:05 Quit japc (Read error: 145 (Connection timed out))
12:31:18pixelma%mo is not in official Rockbox
12:31:57CtcpVersion from freenode-connect!freenode@freenode/bot/connect
12:32:01 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
12:33:45 Nick fxb__ is now known as fxb (
12:34:13 Quit JdGordon (Client Quit)
12:35:23 Join JdGordon [0] (
12:39:34 Quit JdGordon (Remote closed the connection)
12:40:58BdN3504hm, ok then i'll just use the %mv tag...
12:41:00 Join JdGordon [0] (
12:43:06ajb`Ok, it seems the m4a crash address I'm seeing is in mad_main_data. Any ideas why?
12:43:43 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
12:43:52ajb`8 bytes into it to be precise
12:46:09 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
12:48:22kadobanajb`: stack corruption?
12:49:51ajb`I assume init_mad is the entry point of the codec when an m4a gets queded up to play
12:50:44***Saving seen data "./dancer.seen"
12:50:47 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
12:53:29Zagorajb`: no, codec_main is the entry. init_mad() is called from there.
12:54:29*ajb` tests the stack corruption theory
12:55:12ajb`the ipod video is a duel core machine right? Core 0 being menu core 1 decode?
12:59:15ajb`Well adding 128 bytes of space into IBSS moved the undefined instruction up 128 bytes which would seem to confirm the stack corruption idea
12:59:54amiconninit_mad sounds like a function in a codec that isn't used at all for m4a
13:01:25B4gderI think you checked the wrong .map
13:01:47ajb`actually I take that back, the crash address has moved but the map makes the address somewhere different
13:02:01ajb`40018928 is in:
13:02:03ajb` .ibss 0x0000000040018924 0x4 /home/alex/src/rockbox/
13:02:03ajb` 0x0000000040018928 . = ALIGN (0x4)
13:02:03ajb` 0x0000000040018928 iend = .
13:03:16 Quit JdGordon (Read error: 54 (Connection reset by peer))
13:04:25Zagorajb`: shouldn't you rather be looking in
13:04:43domonokycan someone please move the new rbutil mac build from"> to the download server ?
13:05:19ajb`Zagor: I looked for maps that contained the address and that's what it pointed to...
13:05:24*ajb` tries v3
13:05:53Zagorajb`: all codec maps use the same addresses. but only symbol start addresses are listed.
13:05:57 Join herefornow [0] (n=45df5a8b@gateway/web/cgi-irc/
13:06:50ajb`Ok so this crash is 40017fA4
13:07:37ajb`Doing a grep of the .maps for 40017 gives me:
13:07:59Zagorajb`: you can't grep. all codecs use the same memory.
13:08:06ajb`, (not that) and (no idea what that is)
13:08:21ajb`Zagor: ahh, I misunderstood
13:08:41Zagoryou simply have to look in the .map of the codec you are running.
13:08:45ajb`Zagor: So to translate from the reported address to where in the codec I do what?
13:08:56ajb`For m4a that will be?
13:09:03 Quit herefornow (Client Quit)
13:11:13ajb`Zagor: The end of is:
13:11:19ajb` 0x0000000040016818 iend = .
13:12:15ajb`the ibss section of filtbank.c
13:12:17 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
13:12:34Zagorwhat does the crash message say?
13:12:46 Join tvelocity [0] (
13:13:01ajb`Undefined Instruction at 40017FA4 (0)
13:13:32ajb`(see FS #9745 for other crashes)
13:14:37Zagoryou've established that there is no code there so that crash message is to be expected. now the question is why does it call there.
13:14:51ajb`I notice is crashes before the mp3 before it finishes. I assume the mp3 codec is out of the way by this point?
13:15:05Zagorsee if the sim crashes too
13:15:15ajb`Zagor: it doesn't
13:15:34ajb`I could really do with a backtrace on the device!
13:15:36*ajb` wishes
13:15:42Zagorthen it's time to explore logf over usb-serial :-)
13:17:06Zagorthere's a huge karma bonus to be collected for he who creates a gdb stub for usb-over-serial. nudge nudge ;-)
13:17:36ajb`Zagor: any pointers?
13:18:39ajb`Currently pluging in USB resets the ipod to flash firmware
13:18:48Lloreanajb`: Just a question, as I've just dropped in. Are you using an up-to-date bootloader?
13:18:56 Join n1s [0] (n=nils@rockbox/developer/n1s)
13:19:09pixelmacould the crash be in metadata? (just a suggestion...)
13:19:31 Quit MTee ("Java user signed off")
13:19:44LloreanI seem to recall addresses like that showing up with using recent builds and older bootloaders on PP targets.
13:20:16ajb`Llorean: define upto date? I haven't touched the bootcode since rbutil set up rockbox for me about a month or so ago
13:21:23Lloreanajb`: That ought to be up to date. And it's just "the bootloader". Using other terms can cause confusion if you talk with other people about it.
13:22:23ajb`Llorean: noted.
13:22:38ajb`Llorean: can I check easily?
13:22:52ajb`The boot partition is special and hidden isn't it?
13:22:52Zagorajb`: if you're interested in getting your paws muddy, is one place to start digging
13:23:46Zagorthere's something at too. I haven't checked the status of either, other than seen that both are rather old.
13:23:50ajb`Zagor: do we already have working serial for the USB on iPod or does that need doing as well?
13:24:09pixelmalinuxstb: can you tell me more about the something still not right re. beast svg/pdf? Does it "only" not look right or did you have some other trouble? I'm just curious because I could imagine two possible problems and just want to know which thing is the most important to avoid...
13:24:22Lloreanajb`: you'd need a bootloader from over a year ago for it to be the problem I was thinking of.
13:24:25 Join TheSphinX^ [0] (
13:24:25ajb`Getting logging first might be preferable to a full GDB stub for this bug
13:24:42Zagorajb`: we have usb-serial working. I'm not sure what the status is on ipod specifically.
13:24:51Lloreanajb`: Does the same thing happen if you use our precompiled builds?
13:25:01LloreanIt's very strange that you're the only person experiencing this, even with the same files.
13:25:20n1spixelma: what do you think about adding reset holes and battery switches to the labels in the target images in the manual?
13:25:26ajb`Llorean: not had the ipod that long :-) So I guess that's out. FWIW I did have working m4a's when I first tested the my other halfs un-DRMed iTunes music. It has broken since
13:25:37*ajb` considers a straight bisect
13:26:44pixelman1s: wouldn't be very hard and could be helpful for the Gigabeat F/X installation instructions, I think
13:27:02pixelmamaybe there are other things like that
13:27:04 Quit AndyI (Read error: 145 (Connection timed out))
13:27:15 Join AndyI [0] (i=AndyI@
13:27:27n1spixelma: many iriver users have never seen the hole until someone tells them about it :)
13:28:54pixelmathere's even a reset hole on the Iaudios but it also has a hard power off... the only time I saw someone needed to use it was when he was running the OF
13:29:09Lloreanajb`: About using our precompiled binaries?
13:29:29 Quit thegeek (Read error: 104 (Connection reset by peer))
13:29:50 Join thegeek [0] (
13:29:51pixelman1s: but they shouldn't need to know about it ;\
13:32:43n1spixelma: right :)
13:32:56 Join JdGordon_ [0] (
13:33:13n1sbut for the gigabeast for example the batteryswittch is needed to escape the OF
13:37:34ajb`Llorean: Not very helpful if I want to keep developing. Besides I'd like to fix the bug for other people :-)
13:37:42Lloreanajb`: That doesn't answer my question.
13:37:55 Quit bmbl (Read error: 110 (Connection timed out))
13:38:03LloreanIf the official binary works, there may be no bug other than some quirk of your build environment.
13:38:22LloreanIt will help us determine where to look further.
13:38:25ajb`Llorean: You mean for testing. Well other people have tried but failed to reproduce on other players. I'll give the official ipod build a go npw
13:38:47LloreanThe very first thing you do when you encounter a bug like that from a self-compiled build is verify whether it's in the one we provide.
13:39:29LloreanIt gives a much better idea of where to look. :)
13:40:34ajb`Llorean: you're right of course. I'd just gotten so used to working from the src tree!
13:41:51pixelman1s: ok, at least adding battery switches for the two Gigabeats would be very useful then
13:42:28n1spixelma: yes, although i see no harm in adding the reset hole either
13:42:53ajb`Llorean: Yep, works OK with the official build
13:43:16ajb`Compiler issue?
13:43:22Lloreann1s, pixelma: I really think it's probably best to show every button/switch/whatever, also so people don't discover it by chance and wonder what it does.
13:44:15 Quit GodEater (Read error: 60 (Operation timed out))
13:44:17n1sajb`: that is nasty business, could be just about anything,
13:44:31n1sBad ram for example, or overheating cpu
13:44:37Lloreanajb`: You used you said, but you also have some stuff around for Palm development which is also ARM?
13:45:16ajb`Llorean: No palm stuff. Just what I built with rockboxdev
13:45:29pixelmaLlorean: agreed. And it would give me the chance to do the small fix in the M5 graphic I deemed not important enough yet...
13:45:33ajb`I may build a fresh gcc from scratch and see what that does
13:45:39Lloreanpixelma: Sounds good.
13:47:10 Quit TheSphinX^ ("XChat@Linux")
13:48:35 Quit JdGordon (Read error: 110 (Connection timed out))
13:49:49 Nick JdGordon_ is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
13:50:41ajb`I'm off to grab lunch. I may take objdump to this sucker later
13:51:13B4gderthat's the spirit!
13:51:23B4gdernot the lunch part though -)
13:51:32*B4gder adds a :
13:56:11 Quit BdN3504 ("CGI:IRC (Ping timeout)")
14:01:39 Quit JdGordon (Remote closed the connection)
14:04:23 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
14:09:21 Quit kugel (Remote closed the connection)
14:17:25linuxstbpixelma: I couldn't get the PDF showing the image at the right size. If I increased the size from what it is now (which is too small), then it would appear cropped.
14:19:02B4gderZagor: care to svn up the web site to get my update?
14:19:56B4gderimage messed up...
14:20:45linuxstbZagor, B4gder: There is still an (almost) empty cell where the gigabeat-s used to be in the current build table - - does something still need to be run there?
14:21:38 Join BdN3504 [0] (n=55b208dc@gateway/web/cgi-irc/
14:21:47Zagorthat could be our "mystery target" :-)
14:21:59Zagorno, the url gives it away
14:22:40B4gderI'll remove it
14:22:50BdN3504hey i got another question; concerning the buffer for the wps
14:23:02pixelmalinuxstb: hmm, then it's something else than what I thought. How did you create the pdf?
14:23:23BdN3504i don't understand why i get a image too large for buffer error, when i'm using images that are only 39 K in size
14:24:18BdN3504i get this error message from the wps parser: too large for buffer: 74008
14:24:54*B4gder blames linuxstb's revert...
14:25:14B4gderI'll fix it
14:25:25BdN3504but the file is only 39k big. it consists of eleven parts, and is loaded using bitmapstrips.
14:26:35BdN3504the funny thing is, i have already created other wps's which use bigger files and the parser doesnot give out an error message with these... i'm using the sansa e200 btw
14:27:10B4gderZagor: svn up the file now and we'll get the rockbox logo where the beast image should be
14:27:12domonokyB4gder: coud you please move the new rbutil mac build to the download server ?">
14:27:40ZagorB4gder: ok
14:27:50kadobandoes rockbox load uhm, bitmaps which support run length encoding and all that? the file size might not necessarily have much to do with the actual size of the buffer needed for the data, if so
14:28:18B4gderdomonoky: transfer in progress
14:28:19 Join LambdaCalculus37 [0] (i=44a04303@gateway/web/ajax/
14:28:19JdGordonBdN3504: is the image wider than the lcd?
14:28:20 Quit ender` (" Python: executable pseudocode. Perl: executable line noise.")
14:28:44BdN3504if i remove other images it works perfectly, only not all of them together
14:29:09JdGordonyou're using too many then
14:29:14B4gdernow the beast manual has appeared nicely on the manuals page
14:30:24JdGordonwhats up with the M3 manual? just a placeholder?
14:30:41LambdaCalculus37B4gder: Is it almost time to break out the good beer and celebrate a new port release? :)
14:30:59BdN3504but i'm only using five images, which total in 92.052 Byte
14:31:09B4gderJdGordon: I should just remove it from that table until there's something to show...
14:31:43B4gderLambdaCalculus37: hah, we already did once and then took it back, so we can soon do it a second time! ;-)
14:31:47LambdaCalculus37JdGordon: IIRC it had something to do with the fact that the M3 has no LCD on the device itself, just on the remote. I think amiconn or pixelma mentioned it.
14:32:12LambdaCalculus37B4gder: Any excuse to have a good beer. :)
14:32:22B4gderthat's the case for M3, but since we have no manual for it, it won't have to be listed on the manuals page
14:32:23pixelmathe LCD is not the issue - the importance of remote and main target buttons is
14:32:44LambdaCalculus37pixelma: Ahh, okay. I was a little confused as to what the real issue was.
14:32:48B4gderpossibly it serves to clearly show that there is no manual
14:33:33JdGordonBdN3504: hmm... yeah you're definetly right... the e200 can hold up to 232Kb of images
14:33:51JdGordonyou havnt gone over the image count have you? i.e max 56?
14:34:02pixelmaBdN3504: are there images loaded more than once?
14:34:26BdN3504no only once, image count let me see
14:36:59BdN3504i got 23 individual tokens, plus one background image plus the backdrop
14:37:23 Join robin0800 [0] (
14:39:20linuxstbpixelma: I used "save as" in inkscape to save as .eps, then "ps2pdf" to convert to pdf.
14:39:56pixelmaBdN3504: background image + backdrop?
14:40:39pixelmalinuxstb: weird, no idea then
14:40:55BdN3504yes, i got an individual backgraound for the wps and a backdrop for the menu
14:41:34JdGordonwas the 92K total including that background image?
14:41:53BdN3504yes, but without the backdrop
14:41:54JdGordonhmm... even if its not.. you're still well under
14:42:33pixelmaI thought you were talking about the WPS only. A backdrop in the WPS shouldn't use the image buffer if it's loaded with the %X tag
14:42:36 Quit thegeek (Read error: 104 (Connection reset by peer))
14:42:37BdN3504130.852 Bytes with everything
14:42:52 Join thegeek [0] (
14:43:23 Quit CaptainKwel (Remote closed the connection)
14:43:52JdGordonBdN3504: pastebin the output from rockboxui −−debugwps
14:43:58JdGordonand the wps
14:44:13BdN3504well, it's been a while since i've done my last wps. back then i always had individual background for the wps and backdrop for the menu. i loaded the background using %X
14:44:32BdN3504how do i pastebin?
14:44:54LambdaCalculus37BdN3504: Copy your stuff here:
14:45:05LambdaCalculus37Then submit and copy the link it gives you back and paste it here.
14:49:02BdN3504first is parser, second wps
14:49:23*GodEater_ sees a "Failed to load image"
14:50:39JdGordonthat output isnt really verbose enough :/
14:50:45***Saving seen data "./dancer.seen"
14:51:08*JdGordon has no idea
14:51:41 Join japc [0] (n=japc@
14:51:54BdN3504does it matter if i load a backdrop? do the wps and the menu share the same imagebuffer?
14:53:04JdGordonumm... they might... but thats not the issue
14:54:28 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
14:55:18BdN3504maybe 11 tokens is too much for bitmap strips... or wait, you asked me if one image was larger thatn the screen, right? well the volume image is 638px high... but it's using bitmap strips, so the size used is 58x58.
14:55:46BdN3504the thing i don't get is that when i comment out only one single image, everything works fine...
14:55:50JdGordonthe limitation is on the width... it can be taller than the screen.. not wider
14:56:25JdGordoncomment out the %xl|F line.. does it work?
14:57:26BdN3504errm, which line?
14:57:39JdGordonthe 3rd line of the wps
14:58:10BdN3504but that is used in the %mv tag
14:58:27JdGordonit will fail to load once the %xdP lines.. but if it gets that far its good enough
15:01:08n1sjhMikeS: the reason i pinged you yesterday was to ask what (if any) effect the battery capacity setting has on charging on the beast to put it in the manual
15:01:33pixelma638x58 pixel is quite large already, I mean amount of data wise, almost the same as a complete background image for an e200
15:01:33 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
15:02:22BdN3504so shall i try reducing the size.... which dimensions would be reasonable ?
15:02:35 Nick fxb is now known as fxb__ (
15:02:41n1sBdN3504: also make sure you save them as 16 bit and not 24 bit bmps
15:03:49pixelman1s: that isn't really needed, it would only reduce the space that's needed on the disk
15:03:57BdN3504erm they are already 8 bit
15:04:04pixelma(or flash memory)
15:05:52n1spixelma: oh, it buffers as 16 in any case?
15:06:36pixelmathought so
15:08:05n1sthinking about it that makes good sense so i think you are right :)
15:08:43 Quit Zambezi (Remote closed the connection)
15:10:29BdN3504this might take a little while
15:10:49pixelmaso if the image is 8bit, the size you stated won't be the real size it needs to buffer. I'm slowly getting confused though, thinking of mono images (I believe they are an exception but), maybe someone else who is more confident in this should answer...
15:11:32JdGordonBdN3504: convert that image up to 16bit... see how big it is then
15:11:50JdGordonI'm assuming that its stored as 16bit in rockbox anyway
15:12:56pixelmait should be double the size, no?
15:14:52BdN3504well it was 39k but i already deleted it
15:14:57 Quit sarixe ("Connection reset by the motherfucking peer")
15:15:02BdN3504i got an image that works now
15:15:14BdN3504but it only uses six tokens
15:15:28BdN3504so the animation doesn't look that smooth anymore...
15:16:07BdN3504and the funny thing is, the image is also 40.424 Bytes in size
15:16:13BdN3504i don't get it
15:16:23 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:17:01n1sit's not the filesize that's important but rather how much memory it needs to buffer, and apparently any (colour?) bmp is buffered as 16 bit
15:20:06BdN3504i don't understand how one file with approximately the same size and colour depths works and the other does not
15:22:00kadobanare you going by file size or width * height * color depth?
15:23:35BdN3504one was 58*638 the other 58*348
15:24:46 Join Schmogel [0] (
15:27:08n1sBdN3504: well, it's _possible_ it's a bug
15:34:50pixelmahow is that approximately the same size?
15:35:51 Join EspeonEefi [0] (
15:37:50BdN3504well they are both about 38Kb
15:38:07gevaertsSize on disk is irrelevant
15:38:20 Quit blithe (Remote closed the connection)
15:38:30 Join blithe [0] (
15:39:33BdN3504ok so here is what i found out: 16 bit colour bmp 58*638 does not work;
15:39:53BdN350416 bit black & white 58*638 does not work
15:40:20BdN35041 bit monochrome 58*638 does work
15:42:37 Join tyfoo [0] (
15:46:02 Quit kachna (Read error: 60 (Operation timed out))
15:48:57 Join evilnick [0] (i=0c140464@gateway/web/ajax/
15:58:34 Quit EspeonEefi ("さよなら")
15:58:36 Join blahrus [0] (n=blahrus@
16:08:48*B4gder gets a synopsys data sheet viewed on their site
16:08:50B4gder2 pages
16:08:59B4gder_that_ is not what I call a data sheet
16:10:22 Quit BdN3504 ("CGI:IRC")
16:10:42B4gderit looks like a product brief
16:11:27Zagorit seems every company has a different definition of "the document you want"
16:11:38Zagoruser manual, reference manual, data book etc.
16:19:28 Join MethoS- [0] (
16:25:33 Join kerwood_wk [0] (n=80f48719@gateway/web/cgi-irc/
16:28:16B4gderI mailed AMS now to see if they can help
16:31:48 Quit evilnick (" ajax IRC Client")
16:32:53 Quit kerwood_wk ("CGI:IRC (EOF)")
16:34:08linuxstbB4gder: There's an image in SVN for the beast now (I committed last night)
16:34:08 Quit jhMikeS (Nick collision from services.)
16:34:14 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
16:34:27B4gderyes, it just has to be converted to one of them tiny ones
16:34:33linuxstbIt was...
16:46:01LambdaCalculus37The blind description for the beast needs a little fixing up.
16:46:40LambdaCalculus37The bottom of the beast is actually, from left to right: battery switch, USB port, dock connector.
16:49:35 Quit suom1 (Read error: 104 (Connection reset by peer))
16:50:47***Saving seen data "./dancer.seen"
16:52:49 Join kachna [0] (
16:52:59 Quit Llorean (Read error: 104 (Connection reset by peer))
16:58:09 Join suom1 [0] (
17:00:25 Join toffe82 [0] (
17:01:46 Quit Zagor ("Client exiting")
17:01:50 Join haithun [0] (n=4dbb5e1b@gateway/web/cgi-irc/
17:03:48ajb`Anyone know the magic to get objdump to dump a codec file? Is it just a raw binary?
17:04:29GodEater_why would you want to dump it? The source code is available for all our codecs...
17:04:53ajb`GodEater_: I'm seeing what the compiler has screwed up vs the official binary build
17:05:50GodEater_oh I see
17:06:10*GodEater_ doesn't recall the command line required for objdump as he uses it very infrequently.
17:07:18 Quit haithun ("CGI:IRC (Ping timeout)")
17:07:37 Join haithun [0] (n=4dbb5e1b@gateway/web/cgi-irc/
17:10:14 Quit haithun (Client Quit)
17:14:48 Join haithun [0] (n=4dbb5e1b@gateway/web/cgi-irc/
17:15:32 Quit haithun (Client Quit)
17:18:21 Join haithun [0] (n=4dbb5e1b@gateway/web/cgi-irc/
17:19:04 Join evilnick [0] (i=0c140464@gateway/web/ajax/
17:19:10ajb`Is the 5.5G iPod a LE and BE ARM?
17:19:36B4gderall current Rockbox ports on ARM are little endian
17:20:00gevaertsajb`: I'd objdump the .elf file if it's available
17:21:38ajb`gevaerts: The elf filea aren't packaged in the "official" builds. Just the final binaries
17:21:55ajb`gevaerts: If anyone can send me the generated files that would be great
17:22:14 Quit haithun (Client Quit)
17:23:31 Join {phoenix} [0] (
17:24:56*ajb` is almost there with objdump -d −−target binary -m arm -EL official.r19705/codecs/aac.codec
17:25:39B4gderthe "official" elf files aren't kept anywhere
17:25:54GodEater_ajb`: how is your build different from the "official" one ?
17:26:06GodEater_All you need to do is have the right tool chain set up, and then you'd be able to build it too
17:27:10ajb`GodEater_: Except my build breaks on m4a files and we narrowed it down to build. I want to compare the code my rockboxdev compiler spat out vs the official build
17:27:38GodEater_ajb`: the "official" build is done on a collection of random people's servers though
17:27:59GodEater_so one svn rev might be from one person's PC, and the next from someone completely different
17:28:48ajb`Well the "official" build works with m4a's mine doesn't. Until I can do a side by side dump of the codec we have no idea why
17:29:05*ajb` begins to suspect the .codec files aren't just raw binaries
17:29:07GodEater_and you're using all the recommended versions of all the build tools ?
17:30:07ajb`GodEater_: I built my compiler with
17:30:23ajb`GodEater_: And if there is a compiler bug I'd like to report it
17:31:17GodEater_hahaha - good luck
17:31:21GodEater_the gcc folks hate us
17:31:35GodEater_just ask amiconn ;)
17:32:22GodEater_ajb`: what build platform are you on?
17:32:28ajb`Well looking at the hexdump there is only one byte difference
17:32:28B4gdermaybe if he doesn't say he built rockbox with it! ;-)
17:33:11ajb`GodEater_: 16:06 alex@danny/x86_64 [] >arm-elf-gcc −−version
17:33:11ajb`arm-elf-gcc (GCC) 4.0.3
17:33:35*ajb` looks to see how the codec files are generated
17:37:16linuxstbajb`: You're compiling on a 64-bit Linux host?
17:39:41 Part B4gder
17:40:05ajb`linuxstb: yep
17:40:39ajb`linuxstb: 6:31 alex@danny/x86_64 [bins] >uname -m
17:41:55ajb`Assuming the offset of 0x750 in the hexdump is the same as the map that says the invalidate_icache code is slightly different
17:42:17ajb`Unfortunatly I can't see the instructions as objdump segs (I assume on the .codec header)
17:42:46ajb`16:31 alex@danny/x86_64 [bins] >diff -b my.dump off.dump
17:42:46ajb`< 00000750 c0 5e ea f3 18 18 01 40 18 18 01 40 18 68 01 40 |.^.....@...@.h.@|
17:42:46DBUGSent KICK ajb` to server
17:42:47Kick(#rockbox ajb` :No flooding!) by logbot!n=bjst@gateway/web/cgi-irc/
17:42:59 Join ajb` [0] (
17:43:10ajb`oops sorry, I should pastebin this stuff
17:48:54 Quit Slack_ (Connection timed out)
17:48:58 Join miepchen^schlaf [0] (
17:52:01ajb`Found it. Even objdump can't dissasmble this instruction
17:52:26 Nick fxb__ is now known as fxb (
17:52:54ajb`is codec_start a shared function?
17:54:40linuxstbajb`: If you want, I can do an ipod video build, and give you some bin/elf/map etc files to compare.
17:55:28linuxstbWhat memory size are you compiling for?
17:55:57ajb`80 GB 5.5G iPod 64Mb
17:56:03ajb`linuxstb: But I have found the problem
17:56:27ajb`linuxstb: Although it does beg the question what was the address being reported by the UEI handler?
18:03:52*ajb` waits for his iPod to finish commiting the db
18:06:16 Quit miepchen^schlaf ()
18:06:20 Join karashata [0] (n=karashat@
18:07:07ajb`Ok, that's confirmed. Copying the official aac.codec over my compiled one and I can play m4a's again. Don't know why my compiler is breaking it
18:08:43 Join ender` [0] (
18:15:44*ajb` calls it a day for today
18:16:33 Join miepchen^schla [0] (
18:17:23 Quit evilnick (" ajax IRC Client")
18:24:28 Join TheSphinX^ [0] (
18:24:35 Join kugel [0] (n=chatzill@unaffiliated/kugel)
18:26:37 Join fredddy [0] (
18:28:59 Quit tvelocity (Remote closed the connection)
18:29:20 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
18:43:38 Join Slack_ [0] (
18:46:54 Join bertrik [0] (
18:50:51***Saving seen data "./dancer.seen"
18:56:10 Quit obo (Remote closed the connection)
19:01:02 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:02:31 Join Llorean [0] (
19:09:07 Join obo [0] (
19:13:54 Join tessarakt [0] (
19:15:59amiconnBagder: ping
19:17:00amiconnajb`: You're building on linux-amd64, and you're using the official arm-elf-gcc 4.0.3, built using
19:18:25 Quit Darksair ("People who are zhuangbility want to show their niubility but only reflect their shability.")
19:19:45 Join Thundercloud [0] (
19:19:48Bagderamiconn: yes?
19:20:05amiconnThe two AMS Sansas are still missing from the binsize table
19:20:24Bagderyes, early ports don't tend to be as useful there
19:20:48Bagderyou want them added?
19:20:49amiconnHmm. They're the only target builds which are in the build table but not the binsize table
19:21:19domonokycould we also add the m200v4 main build ?
19:21:40Bagderdomonoky: sure!
19:23:19amiconnAnd if we don't want early ports there, the Cowon D2 and the creative ZENs shouldn't be there either.
19:23:37amiconnBut imho the two table should be kept in sync, if only to reduce confusion
19:23:39Bagderwell, its not about "what we want"
19:23:45Bagderit's about what's useful to anyone
19:25:48*amiconn thinks that having some lowmem swcodec ports there could be useful
19:26:34BagderI've now added m200v4 normal + sim to the build table, and clip, fuze and m200v4 to the delta table
19:27:01*domonoky hopes that the m200v4 sim builds :-)
19:27:24Bagderat next commit we'll see! :-)
19:27:25*amiconn wonders whether we'll support all m200 versions at some point ;\\
19:29:06BagderI doubt that
19:29:07 Join jgarvey [0] (
19:29:27LambdaCalculus37amiconn: Depends on when Mr. Someone decides to finish off the ports to the TCC m200s.
19:29:30bertrikm200v4-sim doesn't build here
19:29:45Bagderis it close to build or shall I remove it again?
19:29:53amiconnLambdaCalculus37: v1..v3 are all TCC?
19:29:58*domonoky will try to fix.
19:30:16domonokyits probably not much missing.
19:30:25Bagderamiconn: v1 and v2 are at least
19:30:26bertrikI think it's close, but have no details
19:31:14bertriksome button defines are undeclared
19:31:29domonokywe need a picture..
19:32:14LambdaCalculus37amiconn: Yes.
19:32:32 Join BigBambi [0] (n=alex@
19:32:51LambdaCalculus37Bagder: v3 is the HARP model; that's the one I have and it's TCC based as well. It should be, anyway, it did belong to shotofadds.
19:33:12amiconnAre the differences between v1..v3 known?
19:33:15 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
19:34:06LambdaCalculus37amiconn: v1 and v2 have NAND flash, and the v3 has an SD interface bridge, IIRC. preglow may know a little more about that.
19:34:27amiconnHmm, so the v3 should be a bit easier..
19:36:11bertrikooh, nice green deltas
19:36:20BagderSummer of Code 2009 will happen
19:36:25 Quit HellDragon (Read error: 104 (Connection reset by peer))
19:37:11Bagder"a bit smaller (targeting ~1000 students this year) so we
19:37:11Bagderwill likely target fewer organizations (~150). "
19:37:31bertrikhow many was it last year?
19:38:32 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:38:40bluebrother1125 organization? wow.
19:38:52ajb`amiconn: yes built with the rockboxdev script
19:39:14Bagderthe number of orgs isn't stated and I can't bother to count them myself ;-)
19:39:51amiconnajb`: That *is* weird then. Some build boxes used for the current builds are also running 64 bit linux (e.g. mine).
19:40:33Bagderwe should make sure someone else's 64bit build is tested against ajb's problem
19:40:59BagderI mean, it could be 64bit-related or it is ajb`-specific
19:41:03amiconnajb`: What's your target?
19:42:37ajb`5.5G 80Gb iPod Video
19:42:44 Nick miepchen^schla is now known as miepchen^schlaf (
19:43:18ajb`amiconn: It only gets a few bytes different from the official build but it's a crucial few bytes
19:43:29amiconnYeah, I've read it
19:43:54amiconnI could provide a build built on my debian-amd64 box (the same that is providing a few of the official builds)
19:43:59 Join GodEater [0] (n=ge@rockbox/staff/GodEater)
19:46:21ajb`amiconn: sure. If you could just bung me the aac codec elf file I can do a better compare
19:47:01ajb`amiconn: whats arm-elf-gcc −−version report on your box?
19:48:51linuxstbajb`: I've just made a 64MB ipod video build on my 32-bit Linux - do you want anything from it?
19:49:52ajb`linuxstb: apps/codecs/aac.elf
19:50:13linuxstbajb`:"> - contains aac.elf, aac.codec and
19:50:29 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
19:51:26amiconnajb`:"> contains .codec, .elf and .map, built on debian-amd64
19:52:02*ajb` compares
19:52:31 Quit BigBambi (Remote closed the connection)
19:53:27 Quit TheSphinX^ ("XChat@Linux")
19:53:29saratogai've been meaning to suggest for a while that the buffering from flash patch eventually be commited, and a special lowmem build be configured [sansa or whatever with all but 1-2MB of RAM disabled] so that we can keep an eye on low mem swcodec
19:54:04saratogait'd be nice to know when people are commiting code that implicitly limits playback to >X MB of RAM
19:54:27linuxstbWhy (mis)-configure an existing target?
19:54:53linuxstbIf it's just the build table you're interested in.
19:55:20saratogaso we notice when things break
19:55:49linuxstbI mean, why not just add a real lowmem target?
19:56:07saratogai figured a stable target would be easier to compare to
19:56:14domonokywe have real lowmem targets now in the table :-)
19:56:15 Quit japc (Read error: 145 (Connection timed out))
19:56:17amiconnImho the flash buffering patch isn'tz ready for commit, and it isn't really needed either
19:56:18saratogasince the AMS targets are already quite buggy on their own
19:56:46saratogathough i suppose if they become stable there'd be no sense in it
19:56:58linuxstbsaratoga: I agree, it's useful for testing, but you said you were talking about the build table.
19:58:37saratogalinuxstb: you think it'd only be useful if people actually ran the build, rather then just checking if it compiled?
19:59:37ajb`linuxstb: amiconn: I can't dissasemble the same bits on yours
20:00:08ajb`but it is a single bit difference
20:00:19*ajb` needs to know how to hand decode AMR
20:00:44 Quit markun (Read error: 110 (Connection timed out))
20:02:59 Join BigBambi [0] (
20:03:31Bagderoo fancy url
20:03:45ajb`so what ever is breaking my compiler seems to have confused my objdump too
20:03:46linuxstbsaratoga: Yes. I'm not sure what would show in the build table for a 2MB e200v1 that wouldn't show for a 2MB e200v2 or Clip.
20:04:23linuxstbI guess remove "2MB e200v2" from that sentence... ;)
20:07:42 Quit Bagder ("*plopp*")
20:07:50 Join Zagor [242] (
20:08:04ZagorI rather think the buffering from flash patch should be rejected
20:09:29 Quit karashata ("G'bye everyone!")
20:10:13kugelZagor: btw: I couldn't reproduce the hissing noice wiht the watermark patch on my e200v1, seems to be another ams sansa issue
20:10:32saratogaZagor: how different is it then the current buffering system with a small buffer size?
20:12:36 Join Chronon [0] (
20:13:01 Join petur [50] (n=petur@rockbox/developer/petur)
20:13:49ChrononI just found a case of GPL violation on our wiki:
20:14:12ChrononIt makes use of bmp2rb but isn't licensed in a GPL compatible way.
20:15:17linuxstbChronon: How do you know it's using bmp2rb?
20:15:26ChrononIt claims so on the wiki page
20:16:11amiconnajb`: Weird. These constants are in the constant pool of codec_start, and I wonder how gcc managed to put a wrong constant there for you
20:16:19 Quit soap (Read error: 104 (Connection reset by peer))
20:16:34ChrononSorry. . . it claims it in the text file in the .zip: "This Windows application makes it possible to change logos contained in Rockbox binary files.
20:16:36ChrononA compiled version of the Rockbox tool "bmp2rb" is used for the bitmaps conversions."
20:16:46Zagorsaratoga: well it differs massively, since buffering_flash.c is custom code to handle this particular case and ignore all others (such as MoB). and I don't see how it would gain any runtime either. memcpy isn't a big battery waster.
20:17:55linuxstbChronon: Does that .exe run directly, or is it an installer? (I'm on Linux, so can't run it)
20:18:14ChrononI guess it made its way to our wiki from this misticriver thread:
20:18:50ajb`amiconn: so it's not even code?
20:19:25ajb`amiconn: pop{r4, pc} is a return then?
20:19:46Chrononlinuxstb: it runs directly.
20:20:08*ajb` wonders which constant it is then
20:20:27 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:20:30amiconn0x03ea5ec0 is plugin_end_addr. Somehow your gcc manages to put 0xf3ea5ec0 there
20:20:39linuxstbamiconn, ajb`: Could this not be a binutils (as/ld?) issue, rather than gcc?
20:21:48amiconnAh no, it's plugin_bss_start. But in case of aac.codec, both are equal
20:21:58linuxstbChronon: "A compiled version of the Rockbox tool "bmp2rb" is used for the bitmaps conversions." isn't very clear, but it does sound like they've incorporated the code...
20:22:01amiconnThat just means aac.codec has no .bss
20:22:37amiconn(probably everything in .ibss for performance)
20:23:12ajb`No it would be gcc that generates it surely?
20:23:24amiconnajb`: objdump isn't really clever and tries to disassemble everything if it is located in a code segment, even if it is a constant pool
20:23:52ChrononI don't know of any way to contact the author for clarfication.
20:24:49ajb`ok that solves the objdump mystery, now as to why the address is wrong
20:24:55amiconnajb`: It could very well be caused by the linker
20:25:46 Part jon-kha ("[IRSSI] the Cadillac of all clients")
20:26:11amiconnarm-elf-gcc -v ==> gcc-Version 4.0.3, arm-elf-ld -v ==> GNU ld version 2.16.1
20:26:39 Join jon-kha [0] (
20:27:54ajb`arm-elf-ld −−version -> 2.18
20:28:43linuxstbChronon: I guess the first thing would be to post to that thread.
20:30:18linuxstbChronon: But hosting the file on our wiki seems questionable anyway, regardless of the license issues.
20:30:29ajb`19:30 alex@danny/x86_64 [] >which arm-elf-gcc
20:30:35ajb`19:30 alex@danny/x86_64 [] >which arm-elf-ld
20:31:27Chrononlinuxstb: I can edit the wiki to remove the download link for now. I don't have a misticriver account. Maybe someone who already has an account can pursue this?
20:32:53linuxstbI would wait until others comment - e.g. Llorean who last edited that page.
20:34:00amiconnajb`: You have a cross-linker in /usr/bin? Do you know what package it belongs to? What distro are you suing?
20:35:05ajb`amiconn: Gentoo, I suspect it's because I have a multilib binutils or some such
20:35:38Chrononlinuxstb: Sounds good.
20:36:25ajb`amiconn: Unfortunatly a clean build with path tweaked to skip those still gives the wrong code
20:36:37 Join rico [0] (
20:37:36 Join nibbler [0] (
20:38:00linuxstbajb`: Did you re-run configure?
20:38:17ajb`linuxstb: just doing so now to be sure
20:38:28*linuxstb has a feeling the generated Makefile has the full paths
20:38:33gevaertsIt has
20:38:34bertrikand a ccache -C wouldn't probably hurt either
20:38:59amiconnThat should fix it. configure puts the full paths to the tools into the Makefile, for a bit of speedup
20:39:54*ajb` taps fingures
20:40:34linuxstbajb`: Do you have the other v2.18 arm binutils in /usr/bin as well? e.g. ?
20:40:50ajb`linuxstb: yep, the whole lot
20:40:58linuxstbOK, so it seems 2.18 is to be avoided...
20:41:15ajb`Mystery solved.... broken by a later binutils
20:41:24amiconnWell, someone would need to find out what's going wrong in 2.18
20:44:02*ajb` tries 2.19
20:45:32 Join evilnick [0] (i=0c140464@gateway/web/ajax/
20:45:46 Join DerDome [0] (
20:46:03ajb`amiconn: Whatever it was it seems fixed in 2.19
20:48:16ajb`What should I do with FS #9745, I guess it needs closing now?
20:49:10*ajb` realises other half will be home soon and needs feeding.
20:49:23ajb`Be back later!
20:50:53***Saving seen data "./dancer.seen"
20:51:18 Quit saratoga ("CGI:IRC (EOF)")
20:56:07 Quit jhulst (Read error: 60 (Operation timed out))
20:56:42ricohello people. I need to upgrade the firmware of my samsung Q1 to enable UMS usb mode. I found a solution on internet using tcctool from rockbox site, but I have linux only on my computer. Do you know a way to upgrade my firmware on liinux ?
20:57:11 Quit Thundercloud (Read error: 104 (Connection reset by peer))
20:57:27bluebrothertcctool runs on linux ...
20:58:04ricoah great news :) do you know where I can find it ? I only find windows binary
20:58:33bluebrotheras the complete source for Rockbox and its tools ...
20:59:33ricook I'll try that, thx
20:59:33 Join merbanan [0] (n=banan@
21:01:25 Join Torne [0] (
21:02:33Tornehm. i seem to have done somethign to my ipod while experimenting with bootloaders
21:02:45Torneit doesn't boot up at all now without a hard reset, after being shut down from rockbox
21:03:10Tornejust does the thing where some weird icon flashes up for a moment with no backlight, in the apple rm
21:04:19Tornedoes anyone actually know what causes that to happen?
21:05:24 Quit miepchen^schlaf ()
21:05:51 Join tvelocity [0] (
21:06:04Tornehm, no, there it goes again working
21:06:09Tornefour times in a row seemed odd
21:06:15*Torne glares at it
21:12:41 Join miepchen^schlaf [0] (
21:13:29 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:14:18ricobluebrother: I find tcctool only here : but it is only about windows. Is there another repository ?
21:16:24bluebrotherrobin0800: no it's not. checked the Makefile?
21:16:38BigBambirico: The linux version is there...
21:17:34ricotcctool.c ? I only have to compile that and no need other driver like windows needs ?
21:17:34 Join karashata [0] (n=karashat@
21:17:52bluebrotherchecked the Makefile? You need libusb
21:17:59 Quit nibbler (Read error: 113 (No route to host))
21:18:23bluebrotherI suggest simply running "make" ...
21:18:25BigBambirico: Why not do "make" seeing as there is a makefile? :)
21:18:42 Join nibbler [0] (
21:18:43*bluebrother wonders why he asked if rico checked the Makefile
21:19:32ricohmm I checked but didn't really understand it ^^ I think I need libusb-dev in fact (I got some errors when I ran make)
21:20:05ricoah, it works ! thanks ;)
21:20:20bluebrotheryou need the development package of libusb obviously. But that is the cause for anything you're trying to build that needs the library
21:20:49pixelman1s: I guess microphone holes should be labelled then too?
21:21:03 Join tarbo [0] (n=me@unaffiliated/tarbo)
21:21:17 Quit ze ("building new system without even all the parts")
21:21:46ricoyes bluebrother, I was not sure which library I needed
21:25:30 Quit reacocard (Remote closed the connection)
21:27:50 Join funman [0] (n=fun@rockbox/developer/funman)
21:31:26 Join Aurix_Lexico [0] (
21:31:38funmanhello, i've been away for some times, could you sum up the progress on sansav2 for me ? (I noticed that basic c200v2 support was added, and still no FM radio for the clip -can't remember the FS number-, nor playback fixes)
21:34:53Zagorfunman: playback works fairly well with FS #9703 applied, as long as you don't get the sd crash
21:36:03 Join MethoS-- [0] (
21:38:15 Quit MethoS- (Read error: 145 (Connection timed out))
21:38:39 Quit kachna (Remote closed the connection)
21:41:11 Quit rico ("ChatZilla 0.9.84 [Iceweasel 3.0.5/2008122010]")
21:41:14 Join kachna [0] (
21:41:46funmanhas someone looked at SD code yet, or is it something i could start?
21:43:02 Join Jaykay [0] (
21:45:16Zagorfunman: I don't think anyone is looking at it
21:45:45 Join timc`` [0] (n=aoeu@
21:46:02 Quit timc (Connection timed out)
21:46:33 Join japc [0] (
21:46:41funmani'll try to have a look at it tomorrow
21:49:21 Quit fredddy (Remote closed the connection)
21:50:33 Join Thundercloud [0] (
21:52:20domonokyfunman: for fm, i think we only need the correct audio routing and as3514.c using line-in2 instead of line-in1, then it should work.
21:52:50funmanwhat's audio routing, if not selecting the correct line input ?
21:52:54 Quit karashata ("G'bye everyone!")
21:54:29domonokyi mean the switiching of the inputs. the code which should go into audio-as3525.c
21:55:00 Join B4gder2 [0] (
21:55:29funmani had thought about it, something like #ifdef CLIP || E200v2.. #define FMLINE LINE2 #elif defined(E200) #define FMLINE LINE1 #endif
21:57:20 Nick B4gder2 is now known as Bagder (
21:59:08 Quit LambdaCalculus37 (" ajax IRC Client")
22:01:47 Nick Bensawsome is now known as DOMO (n=Bensawso@unaffiliated/bensawsome)
22:01:56jhMikeSn1s: the capacity setting has no effect on charging decisions on the beast
22:04:43 Quit merbanan (Remote closed the connection)
22:05:19Jaykayis there any way to produce a diff from a single file?
22:05:37ZagorJaykay: compared to what?
22:05:44Jaykayto svn
22:05:52Zagorsvn diff?
22:07:12 Nick DOMO is now known as Bensawsome (n=Bensawso@unaffiliated/bensawsome)
22:07:12funmansvn diff ./my/file
22:08:34Jaykayworks, thanks!
22:15:22 Part Torne
22:17:21 Join Strife89 [0] (n=michael@
22:17:29 Join gartral1 [0] (n=Gartral@
22:24:10 Quit Strife89 ("I need to head off, I'll come back another time.")
22:25:48 Quit japc (Read error: 110 (Connection timed out))
22:26:40 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
22:26:56 Join Jaykay [0] (
22:28:12 Join akur [0] (
22:28:15 Part akur
22:29:48ZagorI want to change "modelname" for x5, m5 and m3 in tools/configure so they match the filenames used in firmware/export/config-*. can anyone think of anything that would break?
22:30:18Bagdernot as long as you don't touch the actual shortcut name you can select in the menu
22:30:42ZagorI won't
22:31:44*linuxstb thinks it's about time there was some model name/number cleanup...
22:32:14Zagorgigabeatf mismatches too. its' config file is called gigabeat.h
22:32:26linuxstbe.g. target_id and MODEL_NUMBER
22:33:06Zagormodel number?
22:34:10linuxstbI think it's being used as the initialisation for the checksums in binary headers. Not sure about other uses.
22:34:27linuxstbThe scramble -add option.
22:34:45linuxstbWhich also uses a four character model identifier
22:35:06jhMikeSwhy should the gigabeat F/X be "the generic gigabeat"?
22:35:33ZagorjhMikeS: I guess it was simply because it was the only gigabeat for a while
22:35:38ZagorI'll rename it
22:37:28linuxstbAh, is modelname used in the lang files?
22:38:34Zagorah, they are. but wrong!
22:39:22Zagorno, I'm confusing my changes with svn :-/
22:39:26jhMikeS# Gigabeat Fxx gigabeatf
22:39:26jhMikeS# Gigabeat Sxx gigabeats
22:43:05Jaykaywhat does the "user:" field in the lang files do?
22:43:26 Join m0f0x [0] (n=m0f0x@
22:43:28Jaykayrasher: pping
22:43:37Jaykayrasher: ping
22:44:33 Quit funman ("leaving")
22:46:21kugelseems like funman has more time again? \o/
22:47:17Jaykayyay. only the reason is not the best.
22:47:26jhMikeSJust reminding about c200/e200 charging (resynced into newer powermgmt framework):
22:48:35UnhelpfulZagor: so... the only thing that changes, in my API() cleanup, is that PREFIX is gone? :)
22:49:39bertrikjhMikeS, thanks for the resync
22:49:54bertriknot so much a resync, rather a rewrite I suppose
22:50:54***Saving seen data "./dancer.seen"
22:50:58 Join HellDragon [0] (
22:51:38jhMikeSI don't know. It's strictly on the powermgmt thread in the target framework sine it was done just for that sort of thing.
22:55:52ZagorUnhelpful: with the PREFIX gone, simple redefining of open works in sims. so that's all you have to do.
22:56:40ZagorI found that doom and rockboy actually already do that. Although they had to jump through hoops to bypass the PREFIX
22:57:08mooslinuxstb: hi, I will get a try to document beast installation for windows in the wiki
22:57:36Unhelpfulwell, they're ports of non-rockbox applications... it'd be a pain without *some* kind of fixup
22:58:19*Unhelpful sees that he doesn't even need to #undef POSIX file functions any more
22:58:33 Quit miepchen^schlaf ()
22:59:25linuxstbmoos: OK. Although I'm hoping I'll find time to streamline the install process a little - e.g. by combining mknkboot and sendfirm. Or we just ignore dual-booting for now...
22:59:29Unhelpfuli've got the wrapper stuff defined once in apps/plugins/lib/core_wrappers.h. i still need to add declarations for resize_init and bmp_init, and then this should be ready to go :)
23:00:11mooslinuxstb: nice, then maybe wait you do this first? /me no more use dualboot now :)
23:00:41UnhelpfulAPI will be completely gone, there will still be a few checks against PLUGIN, but only where code actually needs to differ... i don't think that's a whole lot worse than what's already there for various display types.
23:00:47mooslinuxstb: we don't have dualbooting either for the X5, and long time now it was suported...
23:02:01linuxstbmoos: Yes. Because we have charging and usb access on the beast, I don't consider dual-boot as vital.
23:02:41Unhelpfuli don't have time to bugfix if something breaks, right now, so i'll look for any comments when i get to work
23:02:41*Bagder agrees with linuxstb
23:02:43BigBambiIt is a bit like the F to me - pointless, as you can't use music from either RB or the OF on the other
23:02:54mooslinuxstb: and faster boot
23:03:07BigBambiand given radio too is in RB
23:03:14 Quit petur ("Zzzzz")
23:03:24linuxstbSo does the OF do anything Rockbox doesn't do? I guess USB host?
23:03:30 Join miepchen^schlaf [0] (
23:03:34BigBambiTV out I suppose
23:03:45 Join miepchen^schla [0] (
23:03:46ZagorUnhelpful: I think you should avoid calling any file "core". it's gets rather confusing any way you do it. better to have a clarifying comment inside the file.
23:03:50Unhelpfulusb host, TV out, video in newer codecs than mpeg2
23:04:08BigBambinothing vital
23:04:18BigBambi(IMO, of course!)
23:04:19UnhelpfulZagor: ok, i'll look into that when i get to work, as well :)
23:04:21linuxstbWhat codecs? I've never tried the beast's OF video playback...
23:04:44BigBambiExcept I could never produce one that worked properly
23:05:11linuxstbAt least you know where you are with mpeg1/2...
23:05:30BigBambiand it is easy to produce on all platforms
23:05:32moosiaudio X5 is a good example of bootloader rockbox only, and there was no to much complaints (even if we have a patch)
23:05:55linuxstbBut hasn't there always (or almost always) been an unofficial dual-boot bootloader?
23:06:00 Quit evilnick (" ajax IRC Client")
23:06:04linuxstbHence the lack of complaints.
23:06:15BigBambiand that exists too for the S
23:06:22BigBambiIn fact, it is much more official
23:06:32BigBambiif people want to do it
23:06:37linuxstbBut going back to the beast, I think it's fine to start with a single-boot "release" of docs/tools, and then aim to do dual-boot asap.
23:06:53mooslinuxstb yup a dualboot patch, but we don't suport it ;)
23:07:25mooslinuxstb: +1
23:07:59linuxstbIs someone willing to make a bootloader release for the beast? i.e. compile it with a version number (1.0?), tag in svn, then give to Bagder?
23:08:06*linuxstb doesn't have his beast with him
23:08:09jhMikeSlinuxstb: moos says you pinged me about sendfirm?
23:08:14 Join BrianRowe [0] (
23:08:31linuxstbjhMikeS: I don't think so. Although do you know anything about the dll the windows version uses?
23:08:39BigBambilinuxstb: I clearly can't do that, but I can test any
23:08:51moosyup this, I was wondering to share this to wiki
23:08:58jhMikeSIIRC it's a redistributable DLL
23:09:12linuxstbBut not open source? Meaning we can't call the windows sendfirm GPL?
23:09:40BigBambilinuxstb: I saw this today:
23:09:51jhMikeSor is it not what I'm thinking...have to look.
23:10:08 Quit Jaykay (Read error: 110 (Connection timed out))
23:10:21jhMikeSBTW, what prevents GPL code from using a DLL that isn't open source? I wouldn't think it would be possible to GPL anything using the OS then.
23:10:37 Quit {phoenix} (Remote closed the connection)
23:10:42 Quit BrianRowe (Client Quit)
23:10:43 Quit jhulst (Read error: 110 (Connection timed out))
23:10:56linuxstbBigBambi: Yes, mcuelenaere said that in the source code somewhere, or maybe a README.
23:10:57 Join BrianRowe [0] (
23:10:58 Join Hillshum [0] (n=chatzill@unaffiliated/hillshum)
23:11:14jhMikeSDid he build that DLL?
23:11:39BigBambiThat's what I was wondering
23:11:49linuxstbThe source to the DLL is in SVN, but to build it you need to link with a .lib from a MS SDK.
23:11:57BigBambiah, OK
23:12:02 Quit miepchen^schlaf (Connection reset by peer)
23:12:08moosindeed bad
23:12:10jhMikeSGoogling it only brings up mostly rockbox info
23:12:24jhMikeSah, ok
23:13:01Zagorwill anything bad happen if e200 and e200r have the same MODELINFO?
23:13:20Zagorerr, MODELNAME
23:13:30gevaertsjhMikeS: jhMikeS the GPL allows to link to non-GPL files that are distributed with the OS or compiler, but not if you distribute them with your binary
23:14:12linuxstbgevaerts: So you agree that the windows sendfirm can't be GPL'd?
23:14:28gevaertslinuxstb: that's my opinion, yes.
23:14:34jhMikeSgevaerts: then don't. what about the other side? can they be ditributes separately (as in not in the same .zip)?
23:15:23HillshumZagor: only bootloader install-wise iirc
23:15:36gevaertsjhMikeS: you're really getting into grey areas then.
23:16:23obowhat is being distributed with the dll? Doesn't it just call the WMDM dlls at runtime?
23:16:54jhMikeSgevearts: You bet. If they're gray areas, then someone can bother to spend their time arguing, possibly futilly.
23:17:05ZagorHillshum: do you mean that something will break regarding bootloader installation, or that the modelname is only used when building the bootloader?
23:17:27Hillshumthat the only difference is BL install
23:20:33BrianRoweHello everyone - a request from a new wiki user...
23:20:37 Quit bluebrother (Nick collision from services.)
23:20:42 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
23:20:54linuxstbobo: Have a look inside utils/MTP
23:21:27BrianRoweI'm looking to add my findings for a CF mod for the F series, so if someone could give me write access, I'd appreciate it
23:21:49linuxstbBrianRowe: One moment...
23:22:08BrianRowethanks... at your convenience
23:22:21BrianRoweThanks again
23:22:30linuxstbZagor: BTW, is there any way that user list could be split into one line per user? diffs are not helpful...
23:23:14 Quit bmbl ("Woah!")
23:23:14ZagorI don't think so. twiki variables have to be on one line. :-(
23:23:18 Join japc [0] (
23:24:15Zagorunless there is a way to do VAR = $(VAR) + string
23:28:23jhMikeSjust what is the problem here? I don't see this .lib in utils/MTP
23:28:37obolinuxstb: yup, MTP_DLL.dll... but why can't that be redistributed? It's compiled as MDd/MTd - so it only links against the WMDM libs
23:29:32bluebrotherZagor: is there no way to continue a line by terminating it with a \ or similar?
23:29:43Zagorbluebrother: not that I know, no
23:29:51bluebrothertoo bad
23:30:58 Join ajb`` [0] (
23:31:10bluebrotherI quite dislike that MTP_DLL.dll (btw, why dll two times?) requires WMP10 due to the MTP stuff −− that would make rbutil require that too :/
23:31:24gartral1if i change the SOURCES files in ./apps/bitmaps/native to make it use a different file for the boot image, it should work, right?
23:31:39 Join fml [0] (n=4fd3fd7e@gateway/web/cgi-irc/
23:32:07obobluebrother: the only other way is libmtp and libusb - but aren't there issues with libusb - having to register it for each device?
23:32:27bluebrotherbut there really seems to be no MTP implementation for windows except the WMP SDK :(
23:32:47bluebrotherobo: does libmtp work on windows at all?
23:32:55oboI did get libmtp to compile...
23:33:13BigBambiMaybe we could supply rbutil on a cut down live cd :)
23:33:13bluebrotherlibusb is a bit nasty on windows, but we would need that for e200r installation too
23:33:26obobut I couldn't get it to see any devices, but I'm not sure why - I think it was a libusb issue
23:33:33bluebrotherwe definitely could, but how big would such a CD get?
23:34:01BigBambiI don't know, I was pretty mch joking :)
23:34:06bluebrotherIIRC you need to install the libusb filter driver for each device you want to access.
23:34:18bluebrotherhaven't looked into that for ages though
23:34:20kugelno so much, with damn small linux, <100MB I think
23:34:37bluebrotherkugel: remember that you need X and Qt for Rockbox Utility
23:34:39BigBambiQuite large a download just to install rockbox
23:34:48kugelbluebrother: cli? ;)
23:34:51obobluebrother: that's as far as I got... I decided that end users weren't likely to jump through those hoops
23:34:56kugelwhat happened to that btw?
23:35:05jhMikeSIf you have a gigabeat S and you have been using it, chances are you have MTP installed already, correct?
23:35:19oboit's part of windows media player, so yes
23:35:23 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
23:35:26bluebrotherto what? The cli interface?
23:35:57bluebrotherit sits in the tracker, waiting for cleanup and resync.
23:35:59BigBambijhMikeS: Good point - you need WMP10 to get music on the S OF, so I guess so
23:36:08kugelwhat's bad about rbutil requiring software that's is installed anyway (the beast's OF requres it too I suppose)
23:36:17bluebrotherI'd like to get that in some time and bump to 1.1 or even 2.0
23:36:21fmlHello. In the RB manual, there is the macro "\reference" which produces something like "section X (page Y)" If the macro itself is placed into () then we get nested () which doesn't look good IMHO. I once made a macro that produces something like section X_Y, i.e. the page number is typeset as a subscript. Could it also be used in the RB manual?
23:36:54bluebrotherkugel: other players that don't have those requirements now getting such requirements too? I don't want WMP10 just to install Rockbox on an Ipod ...
23:37:23kugel"Requires WMP for Gigabeat S installation", no word about ipod
23:37:42*bluebrother gives up
23:37:48fmlAll references in that document were typeset like that. And I quite liked it. I didn't invent it but rather saw it in a book, liked it, and took it over for my document.
23:38:30bluebrotherfml: IMO that looks quite weird −− either leave out the page number or put it a normal-sized way. Having page numbers as subscripts looks ... well, strange.
23:38:42toffe82songbird is gpl, this perhaps can help ?
23:38:48gartral1if i change the SOURCES files in ./apps/bitmaps/native to make it use a different file for the boot image, it should work, right, or is there something else i should change
23:38:59obokugel: but if rbutil depends on WMP10 for the beast, it will need it installed to run.. at all
23:39:00toffe82but it needs wm11 installed :)
23:39:12bluebrotherkugel: rbutil 1.0.8 was linked against mingwm10.dll. The functions of that dll were never used, still it had that dependency.
23:39:25fmlbluebrother: you now think it does. But once you've seen it you'll ike it, I promise! :-)
23:39:40bluebrotherfml: got an example somewhere around?
23:40:12kugelhm, I thought it could load MTP stuff only if someone wants to use it with the beast, and not for other players
23:41:04gartral1i should also not that the sources file laks the standard rockbox header comment
23:41:05kugelbut what's better? not requiring wmp, or not supporting windows?
23:41:05fmlAnd it even makes the reading flow more smooth. Since normally you only want to see (read) the section number. But, if you're also interested in the page number you can have it as well.
23:41:05bluebrothernot if you link against a library.
23:41:35bluebrotherfml: use the html manual −− that doesn't show any page numbers ;-)
23:42:24 Quit kugel (Remote closed the connection)
23:42:26Zagorgartral1: I recall we do something special with the filename somewhere, but I can't remember what. try it and see what breaks :-)
23:42:30BigBambicould rbutil just call an external appropriately licenced windows sendfirm?
23:42:38gartral1make: *** [/home/gartral/Rockbox/trunk/build/apps/bitmaps/native/rockboxlogo.176x83x16.c] Error 1
23:42:38gartral1make: *** Waiting for unfinished jobs....
23:42:38gartral1rm /home/gartral/Rockbox/trunk/build/apps/bitmaps/native/rockboxlogo.176x83x16.c
23:42:38DBUGSent KICK gartral1 to server
23:42:39Kick(#rockbox gartral1 :No flooding!) by logbot!n=bjst@gateway/web/cgi-irc/
23:42:41toffe82ther is no way to run sendfirm with some cygwin library
23:42:42 Join kugel [0] (n=chatzill@unaffiliated/kugel)
23:43:06bluebrotherBigBambi: it could. But isn't that kinda weird? Plus, we still have the license issue
23:43:22BigBambiDoes that not escape the licence issue?
23:43:24bluebrotherunless rbutil downloads this sendfirm binary itself
23:43:32 Join gartral1 [0] (n=Gartral@
23:43:43gartral1oops, ment to alt tab and use pastebin, sorry
23:43:47bluebrotherand the license issue for sendfirm still remains
23:43:58BigBambiI know it'd be nice to have a GPL win sendfirm, but I thought the bigger problem was not being to include it within rbutil
23:44:08gartral1did that go through, or did logbot filter it out?
23:44:16BigBambisome of it did next time please
23:44:41gartral1i know, i tryed, and my fingers hit the wrong keys
23:44:51bluebrotheras a last resort we can simply make rbutil present the user a link to sendfirm to download and run himself
23:45:19fmlbluebrother: how can I paste an image
23:45:29BigBambiyeah, not too pretty, but a damn sight better than making rbutil dependent on wmp
23:46:17bluebrotherI remember there was a pastebin-like page for images. Just need to remember the URL ...
23:46:28Zagorfml: upload it to one of the gazillion photo hosting sites?
23:46:57pixelmaisn't there an "imagebin" also?
23:46:58 Quit ajb` (Read error: 110 (Connection timed out))
23:47:00Zagorgartral1: that just says make failed because something earlier produced an error. you need to include the earlier error.
23:47:20gartral1thats spat at me at make
23:47:28bluebrotherfound it:
23:47:31gartral1and tere is no other error
23:47:46gartral1theres no jobs
23:48:07Zagorgartral1: yes there is. make doesn't do anything by itself other than compare file dates. it pretty much cannot fail on its' own.
23:48:37gartral1this is with my edited file, if i revert the change, it works
23:49:06Zagortry make V=1 for full command lines
23:49:18fmlbluebrother: here is an example:
23:49:41Zagorfml: I agree with bluebrother. that looks very strange.
23:49:58Zagorit is definitely not clear that the 9 is a page number
23:50:15bluebrotherfml: it looks like a broken chemical formula or mathematical expression
23:50:39fmlZagor: in my document, there was a note about that in the beginning. We could place it where other icons are explained.
23:50:56bluebrotherI don't expect the average user reading such a note
23:51:02Zagorfml: having to explain the documentation is ... bad
23:51:34bluebrotheryou gotta explain what you gonna explain :P
23:51:56fmlZagor: but the icons are still explained :-) But if you don't like it I won't insist. I just saw the nested () and thought...
23:51:59BigBambifml: I'm not a fan of that I'm afraid
23:53:12 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
23:53:32Zagorgartral1: it looks like bmp2rb doesn't like your file. and it's being rude by only reporting an error code and not printing anything...
23:53:41 Quit fml ("CGI:IRC 0.5.9 (2006/06/06)")
23:53:43bluebrotherwe should rather consider dropping the page numbers completely. They were introduced to allow easier searching in a printout, but who prints out the manual anyway?
23:53:58bluebrotherplus, you can rather easily search for section number.
23:54:06gartral1ohh, umm, sould it be that it doesnt like starting with a 16 bit color image?
23:54:19bluebrotherI never missed page numbers in any kind of (technical) documentation referring to section numbers
23:54:46Zagorgartral1: it could. I don't remember the limitations of bmp2rb but that could indeed be one.
23:54:52 Join Seed [0] (
23:55:00BigBambibluebrother: me neither
23:55:02gartral1whats "safe"
23:56:07bluebrotherhmm, fml already gone.
23:57:14Zagorgartral1: well whatever format the original is, I wold consider safe
23:58:10*gartral1 is having extreme difficulties in determining what the bit depth of the originals are
23:58:42*gartral1 found a clue in registered PPI

Previous day | Next day