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 2009-11-03

00:00:21darkhamhi people, what's new on fuze's v1 front?
00:01:41JdGordon|a few scratches?
00:02:26Bob_Cthanks folks, he'll be here all week
00:03:47 Quit domonoky1 (Read error: 54 (Connection reset by peer))
00:04:29Stephen_darkham, seriously tho i think people nearly have recording working so that to me would be sorta new
00:05:35darkhami'm only a little bit (but ever... :) ) curious of what's new on rockbox's develop on the fuze...
00:06:36darkhamfrom the time i installed it in my fuze, i became a rockbox fanatic
00:07:12darkhami've almost all the original battery life , on rockbox too
00:07:37darkhama little bit less, but i play almost only flac files, from sd cards
00:08:15 Quit AEnima1577 ("Leaving.")
00:09:24 Quit Highlander ("Quitte")
00:10:28 Part froggyman
00:11:30 Quit mirak ("Ex-Chat")
00:14:18 Quit kugel (Read error: 113 (No route to host))
00:14:58 Join ender` [0] (
00:15:51 Quit ender (Read error: 60 (Operation timed out))
00:16:09darkhamare archos av3xx series planned for a new port?
00:18:05krazykitno ports are planned
00:18:27krazykitinterested owners who do the hard work of porting get things done. as it's done in people's free time, there can't really be any timelines, either
00:20:36krazykitthere may be information about the device on the wiki or the forums, if any effort has been started. the NewPort wiki page is a good place to start if you're interested in at least compiling information about it
00:22:01Bob_C lists as "preliminary", no indication of any current activity
00:22:24 Quit DerPapst ("Leaving.")
00:22:52 Join Thundercloud [0] (
00:22:53 Join phanboy4 [0] (
00:22:54 Quit ender` (" cd /pub && get beer")
00:24:41darkhami'm not a developer, i don't think i can help in anyway...
00:24:59darkhami don't know how
00:28:11linuxstbdarkham: You may want to look at archopen, if that project is still alive.
00:28:30 Join webguest10 [0] (
00:29:10 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
00:29:16webguest10hi, what is the displayresolution of the mini2440 touchscreen
00:29:32 Quit fildi (Remote closed the connection)
00:30:22 Quit bubsy (Client Quit)
00:31:24webguest10with rockbox or in general, as the friendlyarm site tells: Max: 1024x768 do I miss something? or is this when you attach another display?
00:31:43linuxstbIsn't there also a 7" 800x480 touchscreen available?
00:33:41Bob_CCurrently we support the 3.5 inch screen, other screeens are available. 1024x768 is probably the VGA adaptor
00:34:18webguest10which will not be supported i guess :D
00:34:51Bob_Cthere is no particular reason why not, it just needs someone to write some code :)
00:34:57webguest10ok thanks
00:35:44darkhamlinuxstb, yes, it seems it's alive
00:37:14webguest10how are buttons realized with the mini2440?
00:39:03Bob_Cthere are 6 buttons on the board, but they are not very useful, the touch screen is probably more useful
00:39:29Bob_Cif you have a hardware skills, you could make an external keypad.
00:40:48webguest10yes thats what I would want to know, where to attach and how to register in rockbox
00:41:23webguest10is this offtopic? i can head over to the lyre forum
00:41:41Bob_Cwhen I have time I will try to put up a schematic
00:43:27webguest10would be nice, thanks:)
00:44:31Bob_Cbasically there is a 34 way GPIO header on the mini2440, and push switches short GPIO to ground. The current s/w supports 8 buttons
00:45:42 Join axionix [0] (
00:46:14axionixhi, can someone tell me an affordable hard drive based player that is rockbox compatible that is still available without using ebay?
00:47:47evilnick_Baxionix: No hard-drive based players are still being produced that are supported with Rockbox
00:47:59evilnick_BHaving said that, you sometimes get lucky with refurbed players
00:48:12evilnick_BOr the odd ipod video that's unopened.
00:49:07 Quit evilnick_B ("Page closed")
00:49:17axionixok, yeah ive been waiting for a couple years now
00:49:22axionixseems they wont make one again
00:49:44 Quit webguest10 ("CGI:IRC")
00:50:37TheSevenaxionix: Flash is a far better suited storage media for portable devices
00:50:50axionixnot when you have 200gb of music :/
00:51:12TheSeventhere never were such big hdd-based players, right?
00:51:19axionixmy friend has a 120
00:51:29axionixactually its goffa from here
00:51:40TheSevenand in a few years, the average cheap flash-based player will have that much storage
00:52:09TheSevenyou know, SSDs are even attacking the PC market nowadays...
00:52:29axionixyeah i haave an SSD...not cheap. 32gb cost me $100 for a decently fast one
00:53:09Bob_CI am hoping to design a HDD player after mini2440
00:53:36TheSevenBob_C: what about using a 2TB 3.5" hdd?
00:53:50TheSeven(or probably 10TB by the time it's done)
00:54:18Bob_C3.5" is a bit big/heavy, but interface wise no different
00:54:20 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
00:54:27axionixso how big of a flash based plyer could i get for say $100?
00:54:32axionixrockbox of course
00:55:33JdGordon|8GB + 16GB microSD capable..
00:55:51axionixeek. all my shit is flac
00:56:29 Join AEnima1577 [0] (
00:56:36Bob_CTheSeven: my last player took 5 years to make, so you are probably right ;)
00:57:13Bob_Caxionix: what is a usable storage size for you? I want to use FLAC as well
00:57:44TheSevenif it ever actually worked, no matter after which magnitude of time, it's great!
00:57:58axionixi'd say at the very least 40gb
00:58:17Bob_Cthats about 100 cds?
00:58:54TheSevenaxionix: you should really go for something hdd-based then
00:59:04TheSevenalthough the big ones are still quite costly
00:59:33TheSevenyou could try e.g. a 240GB-modded ipod 5.5g :-)
01:00:21 Quit AEnima1577 (Client Quit)
01:01:07Bob_Ca 40GB 1.8 HD is about £45 so thats erm, $80?
01:01:07axionixyeah, i WANT an hdd player. i'd rather pay less for mechanical than wait for new tech to go down. my NAS cant even keep up with hdd space...waiting for bigger than 2tb atm
01:02:21Bob_CI think there is a 240GB 1.8 in PATA drive, bigger ones are SATA which are more difficult to interface
01:05:50 Join lifeless__ [0] (n=lifeless@
01:08:58 Quit MethoS- (Remote closed the connection)
01:09:00***Saving seen data "./dancer.seen"
01:09:26 Join phanboy_iv [0] (
01:09:34 Join T44 [0] (
01:10:23 Quit phanboy4 (Read error: 104 (Connection reset by peer))
01:10:27 Quit JdGordon| (Read error: 104 (Connection reset by peer))
01:10:34 Quit J-23 ("ZNC -")
01:10:37 Join J-23_ [0] (
01:10:49 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-7ef11674334d7f19)
01:10:59 Quit niekie (Client Quit)
01:11:04 Nick J-23_ is now known as Guest69461 (
01:11:20 Join niekie [0] (
01:11:46 Join AndyI [0] (n=pasha_in@
01:12:09*TheSeven glares at the build table
01:12:26TheSevenhow did my commit fix that red? :-P
01:13:08TheSevenor rather why was the build system doing things like
01:13:10TheSevenCreated Makefile
01:13:12TheSevenmake: *** No targets specified and no makefile found. Stop.
01:13:14TheSevenfor the bootloaders before?
01:14:41 Join AEnima1577 [0] (
01:17:46linuxstbTheSeven: Looks like a problem with one specific build server
01:18:19TheSevenoh, i see
01:18:46TheSevenlinuxstb: btw, what should we do about the 4g nano and classic?
01:18:56 Join barrywardell_ [0] (
01:19:00linuxstbWhat do you mean?
01:19:36TheSevenshould we already start a port, or should we wait until we find a proper (non-ibugger) way to boot?
01:20:07TheSeventhe 4g nano is fully supported by ibugger now
01:21:10linuxstbI would say start a port - for the 4g nano at least.
01:21:42 Quit AndyIL (Read error: 110 (Connection timed out))
01:21:51linuxstbI guess so.
01:21:53TheSeventhis is going to be confusing once we add the 8702, which seems to be very similar to the 8720
01:22:25TheSevenand there are a few things that could even be shared with s5l8700 despite diffent base addresses
01:22:32TheSevene.g. i2c, lcd, ...
01:22:36TheSevenprobably a lot more
01:23:01linuxstbMaybe rename 8700 to 87xx then.
01:23:21TheSevenbut implementing all 4 socs in a single tree will get an #ifdef hell...
01:23:36TheSevenmaybe somehow further subdivide it for the individual differing components?
01:23:51 Quit lifeless_ (Read error: 110 (Connection timed out))
01:24:01*TheSeven thinks that s5l8700.h needs some major rework anyways
01:26:17*TheSeven curses apple for preventing us from dumping the bootrom using some unknown method, and for using pki all over the place
01:27:08 Quit barrywardell (Read error: 110 (Connection timed out))
01:27:09 Nick barrywardell_ is now known as barrywardell (n=barrywar@rockbox/developer/barrywardell)
01:27:45 Quit reid04 (Connection timed out)
01:28:00 Quit seani ("Leaving")
01:28:22linuxstbTheSeven: It's almost as if they want to prevent reverse-engineering...
01:28:41 Quit Topy44 (Read error: 113 (No route to host))
01:28:53 Join reid04 [0] (
01:28:57 Join aaron_ [0] (
01:29:17TheSeventhe funny thing is that the bootrom was easily dumpable on the ipod touch 2g which uses the same soc (but a slightly newer revision), and is a much more important target for apple
01:29:56*TheSeven also considers renaming usb-s3c6400.[ch] to usb-synopsys.[ch], now that he knows who designed that controller...
01:30:08 Quit aaron_ (Remote closed the connection)
01:30:23 Join aaron424 [0] (
01:31:44 Join bubsy [0] (n=bubsy@
01:32:17TheSevenah right, is there some kind of USE_UNDERVOLTING define?
01:36:15 Quit aaron424 (Remote closed the connection)
01:39:51 Join lifeless_ [0] (n=lifeless@
01:40:21 Join aaron424 [0] (
01:41:31 Join aaron424_ [0] (
01:42:01 Quit aaron424 (Remote closed the connection)
01:42:48 Join Strife89 [0] (
01:43:19 Join Ingrid_Nirvana [0] (n=opera@
01:44:18 Part Ingrid_Nirvana
01:45:26 Join seani [0] (n=seani@
01:45:47 Quit barrywardell ()
01:46:02 Part axionix
01:49:24 Quit lifeless__ (Read error: 60 (Operation timed out))
01:50:46 Quit Thundercloud (Remote closed the connection)
01:56:52 Part toffe82
01:58:37 Quit seani ("Leaving")
01:59:07 Quit zu_ (Read error: 60 (Operation timed out))
01:59:48 Join zu [0] (
02:01:36 Quit GeekShadow ("The cake is a lie !")
02:02:20 Quit lifeless_ (Remote closed the connection)
02:02:27 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
02:02:38 Join lifeless_ [0] (n=lifeless@
02:11:30 Join stripwax [0] (
02:12:43 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
02:12:48 Join amiconn_ [0] (
02:12:49 Quit pixelma (Nick collision from services.)
02:12:50 Quit amiconn (Nick collision from services.)
02:13:01 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
02:13:08 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
02:13:12 Quit stripwax (Client Quit)
02:16:52 Join JdGordon| [0] (
02:18:56 Join polobricolo [0] (i=5c93487d@gateway/web/freenode/x-7e8b7968b6f79bf5)
02:19:49 Join kugel [0] (n=kugel@rockbox/developer/kugel)
02:19:52polobricolois there any function in the Rockbox API i could use to list the different files in a dir (with Lua)
02:20:34JdGordon|there is... I would assume they are avilable to lua
02:20:43JdGordon|if they arnt, im sure thye can be added
02:21:35 Quit lifeless_ (Remote closed the connection)
02:21:54 Join lifeless_ [0] (n=lifeless@
02:24:26polobricolook i found it in C, it is the dirent.h (POSIX) method
02:26:26JdGordon|no, there is a better one
02:26:33 Join daggett [0] (
02:26:45JdGordon|umm... ft_load() or umm..
02:30:23 Quit Stephen_ ("Leaving")
02:37:51 Quit polobricolo (Ping timeout: 180 seconds)
02:39:23 Join ShapeShifter499 [0] (
02:40:59 Join Strife1989 [0] (
02:41:36 Quit darkham (Remote closed the connection)
02:44:00 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
02:44:36 Quit lifeless_ (Remote closed the connection)
02:44:53 Join lifeless_ [0] (n=lifeless@
02:46:10 Quit Strife89 (Read error: 60 (Operation timed out))
02:59:32 Quit LambdaCalculus37 ("Leaving")
03:05:00 Nick Strife1989 is now known as Strife89 (
03:09:03***Saving seen data "./dancer.seen"
03:11:59 Join Strath [0] (
03:15:56 Quit panni_ (Read error: 113 (No route to host))
03:19:25 Join seani [0] (n=seani@
03:21:04 Join funman [0] (n=fun@rockbox/developer/funman)
03:21:18kugelhrm, how can I make the linker to do this: put function into section A, but address them as if they were in section B (the startup code would copy the code from A to B), code that calls functions from A should think they were in B
03:22:03funmankugel: look how uncached addressed are handled in the e200v1/c200v1 linker script
03:23:13kugelI see ".ncdata . + NOCACHE_BASE :", do you mean that?
03:23:16linuxstbkugel: Also look at the icode section - that sounds like what you want to do.
03:23:25 Join evilnick_ [0] (
03:23:52funmanyes but perhaps linuxstb advice is better
03:24:46kugel"} > IRAM AT> DRAM" ?
03:25:00kugelthanks for the tips
03:25:24funmanI think it means the code is logically in IRAM (so callers use IRAM), but physically in DRAM (in the binary)
03:28:02kugelthat seems to work (going by the .map)
03:30:18 Quit aaron424_ (Read error: 110 (Connection timed out))
03:31:21 Nick fxb is now known as fxb__ (n=felixbru@
03:36:05ShapeShifter499I keep getting errors trying to build the new rockbox utility
03:38:34 Join CaptainKwel [0] (
03:40:14ShapeShifter499anyone know what I could be doing wrong?
03:41:44 Quit evilnick (Read error: 113 (No route to host))
03:44:19krazykitShapeShifter499, it's hard to say without knowing what these errors are, really
03:46:25ShapeShifter499here is output of error
03:46:42 Quit daggett ("Ex-Chat")
03:46:48funmanProject ERROR: Qt >= 4.3 required!
03:48:01 Join Stevie8227 [0] (n=Stevie82@
03:49:19ShapeShifter499oops thanks for pointing that out
03:50:05 Quit Stevie8227 (Client Quit)
03:52:05 Quit seani (Read error: 110 (Connection timed out))
03:53:41ShapeShifter499wouldn't 4.5 work?
03:53:57funman4.5 >= 4.3 here
03:55:35 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
03:55:50 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
03:55:50ShapeShifter499well according to my ubuntu package manager I have 4.5.3
03:55:51 Join Topy [0] (
03:57:22kugelhm, the map seems correct. I think my copy asm is correct too, but it freezes at boot :/
03:58:15ShapeShifter499so I have the correct QT version??
03:58:18funmanShapeShifter499: do you have qt4-dev-tools installed?
03:58:56kugel and if anyone wants to have a look
03:59:17ShapeShifter499well I might be installing it now, I went to my package manager and clicked to install anything to do with qt
03:59:49kugelI basically want to move initialization-only code out of the main binary. so that it doesn't waste space. It would be copied to the codec buffer in crt0.S, later overwritten by codecs
04:00:18 Quit AEnima1577 ("Leaving.")
04:05:38funmankugel: _initstart is 0x1e80000 so it doesn't point to lcd_init_device in the binary
04:05:57funman_initstart should be before .init_text: in the .lds i believe
04:06:04kugeloh, indeed
04:07:03 Quit lifeless_ (Remote closed the connection)
04:07:21 Join lifeless_ [0] (n=lifeless@
04:08:39 Quit Rondom (Nick collision from services.)
04:08:49 Join Rondom [0] (
04:11:46kugelfunman: thanks, it works now
04:12:14kkurbjunkugel, do those pp chips have the ability to suspend/resume?
04:12:25kkurbjunthat's what the apple fw does doesn't it?
04:12:33kkurbjunI mean suspending that is
04:12:55kugelI haven't thought that. But I think *we* don't do it on any PP
04:13:33 Quit T44 (Read error: 110 (Connection timed out))
04:13:36funmanif we define a new attribute for functions only needed when booting, we can still remove it when those functions need to run more than once, should suspend/resume be implemented
04:14:05kkurbjunno, we don't currently, but I guess the only thought I had is that depending on the savings, the interference doing this might cause more hassle then it's worth if someone ever did decide to go for that feature.
04:14:11kugelI'm playing around, I'm not sure if we want that in SVN. there's also the problem that the binary is bigger now (it contains all data sections doo)
04:14:30kugelit's as big as the ram usage now
04:14:38funmankugel: NOLOAD
04:14:51kugelfunman: the init code must be loaded
04:14:55kkurbjunfunman, you need the data section defined though
04:15:00kkurbjunI mean loaded
04:15:33funmanyou need to put the .init_text section right after the last data section
04:16:14kugelmoving init code before data would not yield any saving. we could try to move the entire data at boot if the init code is right after normal code, but that may slow things down
04:16:16kkurbjunI think you would need to copy it in crt0.s and then make the no load sections re-use that space
04:16:39kkurbjunI mean, you would make the linker point the noload sections to before the init section
04:16:51kkurbjunor at the begining of it
04:17:49kugelhrm, NOLOAD data wouldn't need to be copied, right?
04:18:13kugelso directly after the last initialized data section (which is part of the binary anyway) could work
04:18:47funmankugel: see how .iram is declared, we reclaim the space after it has been copied
04:18:50kkurbjunyeah - you would need to do something similar to the iram sections
04:20:11kugelyea, it really seems to be the same actually
04:22:13kugelthanks for the tips, I'm going to sleep :)
04:22:41JdGordonany suggestions what the "is it a HWCODEC?" tag should be?
04:22:51JdGordonthe letter combo i mean
04:24:49 Quit TheSeven (Nick collision from services.)
04:25:07 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
04:25:18 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
04:25:24 Quit AaronM ("Emo Time In My Corner... //_-")
04:26:16 Quit kugel (Read error: 60 (Operation timed out))
04:27:27kkurbjunis it legal for vieprots to be defined at negative locations
04:27:34kkurbjunor locations larger than the screen
04:28:00kkurbjunJdGordon: what do you need a HWCODEC tag for?
04:28:13JdGordonits not
04:28:19JdGordon(to your question)
04:28:38JdGordoni need it so a recording skin can be shared with SWCODEC and HWCODEC targets
04:29:15 Join AEnima1577 [0] (
04:30:17kkurbjunso it would just append to the beginning/end of the skin name?
04:30:48JdGordonno, it would be a conditional inside the skin
04:30:55JdGordonthis may end up being a bad idea though :p
04:31:08JdGordon*may* :p
04:33:37kkurbjunJdGordon: is there a function that can be called to setup a viewport properly? I found the cause of that crash I think - on the remote for the mr500 the viewport y (and I think x) are setup with negative numbers
04:34:03JdGordonyou should always setup viewports with viewportmanager_get_defaults()
04:34:04kkurbjunthat causes what appears to be a buffer overflow
04:37:26 Join krazykit` [0] (
04:37:43AEnima1577what does a buffer overflow look like ?
04:37:59kkurbjunJdGordon: I didn't see that function - in viewport.c
04:38:22kkurbjunAEnima1577: in this case it is causing a data abort in an unrelated portion of code
04:38:22 Quit ShapeShifter499 ("ChatZilla 0.9.85 [Firefox 3.5.4/20091028153816]")
04:38:23JdGordonsorry.. viewport_set_defaults()
04:39:00kkurbjunit appears to be overwriting portions of the thread handling functions
04:39:17 Join krazykit1 [0] (
04:39:27AEnima1577kkurbjun, the only other time i have heard of a buffer overflow was when i was researching modding my xbox, creating a buffer overflow in that situation allowed a user to gain access to the bios
04:39:29kkurbjunstricktly it's not exactly a buffer overflow
04:40:39kkurbjunbut there are probably plenty of opportunities to do similar things with rockbox in malicious ways
04:40:45 Quit advcomp2019 ("Leaving")
04:41:12 Quit krazykit1 (Client Quit)
04:41:28 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
04:42:03 Quit krazykit (Nick collision from services.)
04:42:08 Nick krazykit` is now known as krazykit (
04:42:13kkurbjunAEnima1577: yes, that is common with alot of exploits involving running code on hardware developers did not otherwise have access to
04:43:14JdGordonthis hwcodec switch tag wont work
04:43:17JdGordonblast :p
04:44:56kkurbjunJdGordon: I'm looking at set_defaults now - it's not clear to me how this will do proper boundary checking though
04:45:16 Quit bubsy ("Party time, EXCELLENT! *wring wring wring!!* Party on, Wayne! Party on, Garth!")
04:45:33JdGordonthat will set the viewport you pass it to the "ui area" (whatever that may be) which will be correctly positioned, sized and coloured
04:45:35kkurbjunthe area that's the problem is in usb_screen.c
04:45:41JdGordonwhat do you want to do?
04:46:25JdGordonbecause the image is larger than the remotes display?
04:46:26kkurbjunwell, there is a fix_viewports function that is being called. On the mr500 the usb logo is larger than the screen
04:46:44JdGordonso create a new bitmap for it :)
04:47:13kkurbjunI could fix the logo, but I wanted to make sure that I'm not masking a underlying problem, or a section of code that should have better boundary checking
04:47:17kkurbjun: )
04:48:09 Join bubsy [0] (n=bubsy@
04:48:19kkurbjunI guess the only sections that the user has access to define their own viewports is through the skin though which checks the bounds there
04:49:01JdGordonthere is a check in usb_Screen so it uses full screen if the bitmap is larger than the user area
04:49:39kkurbjunahh.. hmm
04:50:08kkurbjuneverything works if I comment out the logo..
04:50:27kkurbjunand the revision that made this viewportified is the one that causes the data abort
04:51:12kkurbjununfortunately a seg fault does not happen on the sim
04:51:22JdGordonline 151 is where the fullscreen gets set
04:51:33JdGordonthis obviously assumes that the screen is big enough for the logo though
04:51:36JdGordonwhich obviously it isnt
04:52:06kkurbjunyeah, but the bitmap drawing functions I think check to make sure that the bitmap is located within the screen
04:52:13kkurbjunI'll double check that
04:53:42kkurbjunhmm, it does check, but then it adds the viewport offset afterwards :)
04:54:13 Quit bubsy ()
04:56:27kkurbjunJdGordon: what is set_defaults doing - it just returns the currently active viewport for the screen specified?
04:56:59JdGordonit is broken in svn... the discussion last week between me and kugel was to get it working correctly
04:57:27JdGordonwhat it should do is set the viewport you give it to the ui area set by a combination of the setting and statubsar skin info viewport thing
04:58:02kkurbjunI hmm
04:58:33kkurbjunthat's that mail thread?
04:58:54JdGordonyeah, and irc
04:59:11JdGordonit does that now, as long as you dont have the user setting *and* a sbs with %Vi set
04:59:14kkurbjunI have to admit I would have participated, but I have a hard time wrapping my head around this viewport stuff :-D
05:00:30kkurbjunso, what about the section that is starting on line 161?
05:00:41kkurbjunwhere it sets the title->y
05:01:02kkurbjunthat seems like it would be problematic for this setup
05:02:35kkurbjunat the moment I'm thinking that this could be somewhat fixed and prevented if the bitmap drawing function properly accounted for the viewport offset in the bounds checking
05:03:15kkurbjunallowing a negative viewport location for the title section
05:03:37JdGordonviewports must never have negative values
05:04:06kkurbjunare there other areas this would fail?
05:04:55kkurbjunone use that I could think of offhand is screen transitions where you could have one viewport sliding in and another out.. but that might not be the right way to do it...
05:05:37kkurbjunnot that it is done now or would be implemented I guess
05:06:20 Join bubsy [0] (n=bubsy@
05:06:45kkurbjunI was thinking about messing around with something like that on the MR500 though with the hardware accelerated screen positioning :-D
05:08:11kkurbjunhmm, it looks like alot of the bitmap functions have boundary checking but whenever the viewport stuff was added in it was not accounted for...
05:08:48kkurbjunI guess it was assuming that the values would be somewhat sane
05:08:49 Quit AndyI (Read error: 104 (Connection reset by peer))
05:08:50 Join AndyIL [0] (n=pasha_in@
05:09:05***Saving seen data "./dancer.seen"
05:09:36 Join AndyI [0] (n=pasha_in@
05:09:37 Quit AndyIL (Read error: 104 (Connection reset by peer))
05:10:40 Quit lifeless_ (Remote closed the connection)
05:10:58 Join lifeless_ [0] (n=lifeless@
05:22:27 Quit adiroiban (Read error: 110 (Connection timed out))
05:29:46 Quit Strife89 (Read error: 60 (Operation timed out))
05:38:06 Quit BHSPitLappy (Remote closed the connection)
05:38:48 Join Tomers [0] (
05:40:27kkurbjunJdGordon: I have a kind of fix in hand, but what about adding a boundary check to lcd_set_viewport?
05:41:05 Quit Tomers (Client Quit)
05:41:30kkurbjunthat could at least make sure the viewport has proper bounds - it wouldn't render the way the user might expect it to though
05:42:22 Quit CaptainKwel (Remote closed the connection)
05:49:49 Quit Horscht ("Verlassend")
05:55:33 Quit lifeless_ (Remote closed the connection)
05:55:51 Join lifeless_ [0] (n=lifeless@
05:59:55 Join lifeless__ [0] (n=lifeless@
06:05:15 Join shai [0] (
06:09:10 Quit kadoban (Read error: 54 (Connection reset by peer))
06:13:39kkurbjunJdGordon: when you get a chance could you let me know what you think of this as a fix to ensure that the viewport has a reasonable location and width?
06:15:28 Quit Strath ("Leaving")
06:16:54JdGordonkkurbjun: viewports shouldnt have negative values
06:17:00JdGordonits just asking for trouble
06:17:25 Join Guest23293 [0] (
06:17:25JdGordonim a bit sleepy apparently
06:17:41kkurbjunit fixes the viewport so that it doesn't have a negative value
06:17:45JdGordonwhy does the width need to change if x is <0?
06:18:01 Quit n17ikh (Nick collision from services.)
06:18:07 Quit lifeless_ (Read error: 113 (No route to host))
06:18:07 Nick Guest23293 is now known as n17ikh (
06:18:11kkurbjunI need to make sure that the x isn't larger than the screen size too
06:18:35JdGordonthose sort of bad values should be caught before it gets that far anywya
06:18:39kkurbjunI guess it doesn't if it is going to be drawn "wrong" anyway
06:19:13kkurbjunthe problem is that if they are not it will cause some really nasty side effects that could be difficult to debug
06:19:26kkurbjunI mean I was seeing data aborts in the threading code
06:19:56kkurbjunwithout a hint that it was related to the lcd screen I might have been trying to debug the USB driver
06:20:04kkurbjunI mean the remote screen
06:23:02kkurbjunthe usb screen has bad calculations for the logo and the title, I wouldn't be supprised if other screens turn up with bad calculations for really small screens
06:24:04kkurbjunthe proper fix would be to correct the screen, but there is no way that a remote logo and a title are going to properly fit on the mr500 remote screen
06:25:44kkurbjunI thought that checks could be added to each screen definition, but it would be a catch all if you put the check in the set_viewport function
06:26:47 Quit Llorean (Read error: 104 (Connection reset by peer))
06:27:07 Join Llorean [0] (
06:49:49 Join evilnick [0] (
06:50:14 Join uflops [0] (
06:56:16 Quit evilnick_ (Read error: 60 (Operation timed out))
07:04:45 Join beta2k_ [0] (
07:05:57kkurbjunJdGordon: I updated the patch: I think this would be good at the least for a check to ensure data corruption does not happen, but I don't know how often set_viewport is called - offhand I don't think these checks would be too much of a penalty to the system performance
07:06:08 Quit mikroflops (Read error: 110 (Connection timed out))
07:06:35 Join dkopelove [0] (
07:07:30 Part dkopelove
07:09:06***Saving seen data "./dancer.seen"
07:14:08 Join lifeless_ [0] (n=lifeless@
07:15:35 Join dkopelove [0] (
07:17:09dkopeloveHello, is anyone here a dev working on the Fuze v2 port?
07:18:24 Quit HBK (Read error: 60 (Operation timed out))
07:19:21funmandkopelove: Clipv2/Fuzev2 ports are a bit stalled at the moment
07:19:34funmando you want to help?
07:20:19 Quit beta2k (Read error: 110 (Connection timed out))
07:22:30dkopeloveappologies for breaking one of the irc best practices already
07:22:44funmandkopelove: which practice ?
07:23:01 Quit liar (Read error: 113 (No route to host))
07:23:59dkopelovethe practice of not asking questions that start with "Does/Is anyone ... " generally ambigious questions
07:24:22funmandkopelove: All Clipv2 code is on , I don't remember how far kugel had gone on the Fuzev2, I think he could control backlight
07:24:25dkopeloveanyway I'm new and I want to get involved, where can I start reading up
07:24:58funmanNext breakthrough would be to manage to make DMA transfers from the internal storage (we can already send commands to the SD but not transfer data)
07:25:09 Quit lifeless__ (Read error: 110 (Connection timed out))
07:26:04dkopelovewhat do we know about the hardware of the Fuze v2 (or any of these devices), do I assume correctly they are some arm uP
07:27:38funmanthere is an arm926-ejs
07:28:15funmanmyself i only have a Clipv2 but i assume the Fuzev2 use the same hardware
07:29:41funmanthe SoC is made by austriamicrosystems and is very very similar to AS3525 (the OF is developed with a AS3525 SDK, you can find strings of it in the binary)
07:29:51 Quit xavieran_ (Read error: 60 (Operation timed out))
07:29:56JdGordonkkurbjun: (sorry was watching tv)... set_viewport() is called *alot*... pretty much every time part of the screen is updated... I don't see whats wrong with pushing that error checking up to viewport_set_defaults() (which should always be used)... the problem you found is with bad code in usb screen... not something in the viewport stuff...
07:30:06funmanWe do not know which SoC it is exactly but I suppose it's an extended AS3525 (some registers added, some modified)
07:30:18JdGordonyes it sucked that it was a pain to find, but putting that check in the lcd driver is too far down the stack
07:31:25 Quit Shaid (Read error: 60 (Operation timed out))
07:33:24 Join kadoban [0] (
07:35:28 Quit bluebrother (Nick collision from services.)
07:35:31 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
07:38:39 Join AEnima15771 [0] (
07:40:51dkopelovefunman: thanks for you help, got my dev environment up, and building trunk. Now just need to figure out the status of the fuze v2 project: is kugel a good person to contact about fuze v2
07:42:06funmandkopelove: well i'll tell you the status on fuzev2: we can run code, we need to reverse engineer pretty much everything else from OF disassembly
07:42:12 Quit AEnima1577 (Read error: 60 (Operation timed out))
07:42:45funmani advice you to check the Clipv2 code (especially for backlight) and Fuzev1 code (lcd driver should be similar if not identical)
07:43:43dkopelovefunman: I'll look at the link you provided, thanks; one last Q: is there a seperate branch for v2 porting
07:44:03funmanno, all the code is either in SVN (for mkamsboot), either in the patch on FS #10047
07:45:21funmani'm pretty sure kugel had code to enable backlight though i can't find it on the FS #10047
07:45:30 Quit gevaerts (Nick collision from services.)
07:45:41 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
07:48:48funmandkopelove: ecrips made some work on fuzev2 (he found gpio for usb & left button)
07:48:58funmanhowever his git tree seems abandoned now
07:50:54 Join saratoga [0] (i=9803c6dd@gateway/web/freenode/x-22dd4d46a295e244)
07:52:41saratogafunman: i'm interested in eventually doing more AMS power testing in hopes of getting battery life improved like we did on the e200v1
07:53:11saratogai'm curious how hard it is to change the device clocks for DBOP, IRAM, IDE, etc
07:53:22funmansaratoga: just edit clock-target.h
07:53:35saratogai saw that but will the drivers be ok with different clocks?
07:53:56funmanyou can only tell by running them ^^
07:54:13saratogai want to eventually hook up the DMM and see what the watts/MHz looks like for the various devices
07:54:36funmanFlynDice managed to change ide clock i think, dbop could change without much problems as well
07:54:40saratogathen try to work out where all the power goes
07:55:13saratogahow hard would it be to change the main arm core to 31MHz? I think that means DBOP and IRAM change too?
07:55:34dkopelovefunman: I think I'll clone his tree unless you think that his work is too far behind the tip to be useful?
07:56:58saratogai don't think it matters, resyncing will be easy compared to reverse engineering
07:57:06funmansaratoga: i don't remember if fastbus is mandatory (pclk == fclk) or not in current clock-target.h
07:58:54funmandkopelove: I would clone rockbox tree, apply fs#10047 and manually apply
07:59:29funmanso if you find something we can add it to fs#10047 and it stays in sync with rockbox svn
08:00:27funmansaratoga: currently the cpu freq == pclk
08:01:24saratogaand most other devices are also run off pclk, correct?
08:03:05saratogadkopelove: cleaning up FS #10047 and getting it commited might also be a good idea
08:05:39saratogabasically theres all sorts of work to be done!
08:05:47 Quit Rob2223 (Read error: 60 (Operation timed out))
08:05:49dkopelovedo you all have the ability to commit or do you submit patches
08:07:45saratogadkopelove: funman and I can commit patches on the tracker
08:08:06saratogaand people who get patches on the tracker accepted tend to get SVN access pretty quickly
08:10:11 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:14:04funmansaratoga: yes their frequency is based off pclk, but clock-target.h helps keeping the same frequency if pclk is changed
08:14:47funmansaratoga: about fs#10047 what needs to be cleared is which SoC name to give to Clipv2/Fuzev2 (and Clip+?) : AS3525+ ?
08:15:03saratogafunman: thats the only hurdle?
08:15:18saratogaAS3525+ or AS3525v2 sounds fine to me
08:15:36saratogaits apparently some custom part so you can call it whatever you like
08:15:41funmani wanted to know how much code we could share with as3525, and without docs the answer is hopefully a lot but not all
08:15:51JdGordonfurther proof noone uses HWCODEC targets.... "filetype array full" after updating my AJBRv1
08:16:04saratogawell in the meantime you could still commit it ;)
08:16:14funmanwhat blocks me from cleaning it up is still no SD access
08:16:19funmaneheh yes
08:17:34 Join beta2k [0] (
08:17:58dkopelovefunman + saratoga: thanks I'm working on patching, have a good night
08:18:43 Quit saratoga ("Page closed")
08:21:40 Quit dkopelove ("leaving")
08:22:21 Join kugel [0] (n=kugel@rockbox/developer/kugel)
08:22:49 Quit kadoban (Remote closed the connection)
08:22:59funmankugel: hi, had you managed to enable backlight on fuzev2 ?
08:23:10funmansorry i can't remember nor find any code ..
08:24:30*JdGordon thinks he is almost ready to commit rec statusbar changes...
08:27:14 Join kadoban [0] (
08:33:00 Quit beta2k_ (Read error: 110 (Connection timed out))
08:37:31kugelfunman: yea, but quite weak backight, couldn't manage to set it higher. the code was copied from clipv2
08:38:52funmansaratoga: the patch on fs#10047 also contains lot of ugly debugging code for SD so i'm still a bit reluctant to commit it
08:39:59kugelJdGordon: haha @ filetype array error
08:40:23 Join biv [0] (
08:40:35JdGordonits not sbs though... could be .lua
08:42:44CIA-8New commit by funman (r23490): Sansa AMS: fix system_exception_wait() ...
08:43:45 Join ender` [0] (
08:44:35 Quit amiconn (Remote closed the connection)
08:44:36 Quit pixelma (Read error: 54 (Connection reset by peer))
08:46:47bivI have a problem with my favorite game. Games aren't mission critical but should I ask here?
08:47:32 Join flydutch [0] (
08:47:44funmanbiv: yes
08:48:23bivok funman, thanks and thanks for all your work (esp on the AMS stuff) here goes
08:48:34bivI have several Sansa e250 (both v1) have been running rockbox very well for months. Recently, with the current build (NOT the 3.4 release), brickmania doesn't allow play with wheel. Turning the wheel doesn't move the 'paddle' ... any ideas?
08:49:11bivwhen I change from current build back to 3.4 it works again, but then when I go to current build it ceases again. Just don't want this introduced into a general release
08:50:52bivI hadn't changed builds for a while but the past few days 'current build' it's done that. Not sure which build it started with
08:50:59JdGordonanyone want to give this a gander before I check it in?
08:51:14funmanthere have been a lot of changes to brickmania for touchscreen targets, the change might have been in there
08:51:22bivOh ok.
08:51:35kugelkkurbjun: viewport.c doesn't give negative values, and the lcd driver should do boundary checking
08:51:51bivmakes sense.. I've never submitted a bug report before, should I do so over this?
08:52:20 Join pixelma [50] (n=pixelma@rockbox/staff/pixelma)
08:52:26 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
08:52:55kugelif one lcd driver doesn't, it's a bug there
08:54:30funmanbiv: if kkurbjun or kugel can spot the bug quickly perhaps you don't need to ;)
08:55:07 Quit pixelma (Remote closed the connection)
08:55:07 Quit amiconn (Remote closed the connection)
08:55:23bivfunman ok thanks :> I'm not even registered so that would be preferable for all involved . :>
08:55:37funmanhum button_status() doesn't send scrollwheel status
08:56:14kugelwhy me?
08:56:21funmanbiv: I think i found the problem
08:56:23 Join pixelma [50] (n=pixelma@rockbox/staff/pixelma)
08:56:24 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
08:56:42funmankugel: i just thought you'd know more about buttons than me
08:56:42 Join pixelma_ [0] (i=58d97f15@rockbox/staff/pixelma)
08:56:43bivoh great funman!
08:56:51kugeloh that's possible :)
08:57:20kugelit shouldn't be calling button_status() anyway
08:58:06 Quit kadoban (Remote closed the connection)
08:58:12CIA-8New commit by jdgordon (r23491): Initial recording screen skin support (usable in .sbs skins only for now) ...
08:58:32 Join kadoban [0] (
08:58:33funmankugel: perhaps that's needed for touchscreens ?
08:58:37 Quit beta2k (Read error: 104 (Connection reset by peer))
08:58:48 Join beta2k [0] (
09:00:28funmankugel: can I commit or do you want to check if button_status() use needs to go ?
09:01:20kugelJdGordon: stereo/mono could be done with %St?
09:01:36JdGordonprobably... but so what?
09:01:45JdGordonindividual tags is much nicer
09:02:45kugelfunman: that probably works for now
09:02:57kugelJdGordon: they're waste
09:03:03 Join petur [50] (n=petur@rockbox/developer/petur)
09:03:12JdGordonwe waste plenty of tags like that
09:03:15CIA-8New commit by funman (r23492): brickmania: fix scrollwheel use
09:03:36funmanbiv: next current build will work, thanks for the report!
09:03:40kugelJdGordon: because they were before %St, that's all
09:03:53bivfunman wow thanks, that was quick.
09:04:07kugel"we already do it" is never a good excuse
09:04:09bivand thanks for all the work overall and especially on the AMS versions
09:04:36bivto all the coders.
09:04:52JdGordonthen we should remove all the shortcuts...
09:05:41kugeland break themes? hardly a good idea for 3-4 tags. but we shouldn't be adding more to it
09:09:07***Saving seen data "./dancer.seen"
09:11:16bivFunman, installed the new build and it indeed fixed it. Thanks! (also thanks to all who keep adding games.. I'm on jury duty and they pass the time).
09:11:24bivthanks again... 'night
09:11:47 Join di7ra [0] (
09:13:05 Quit biv ()
09:13:51 Part di7ra
09:14:30CIA-8New commit by funman (r23493): Sansa AMS PCM : cleanup ...
09:16:48 Join einhirn [0] (
09:17:45 Join Stevie8227 [0] (n=Stevie82@
09:22:59 Quit kugel (Read error: 113 (No route to host))
09:23:02 Join Thundercloud [0] (
09:24:29 Join BigNoob [0] (
09:25:25 Quit funman ("free(random());")
09:26:31BigNoobI have a question, about people who use rockbox regulary
09:26:33 Join maruk [0] (
09:28:47BigNoobAre there themes that change the menu layout, not just the background in rockbox?
09:30:00linuxstbIf you mean the order of items in the menus, then no.
09:30:20BigNoobbut the layout for example
09:30:36BigNooblike the original firmware has the circle spinning type thing
09:30:45BigNoobbut this just has a boring list
09:30:54BigNoobis it possible to change that
09:33:03BigNoobCan rockbox run those types of themes ^^^^^
09:34:02linuxstbLook at for the kind of themes Rockbox can run.
09:35:02BigNoobOk, so I guess they can't change the menu layout, into something rather than a list
09:35:04 Join swilde [0] (
09:37:29 Quit pixelma_ (".")
09:38:45GodEatercorrect, they cannot
09:41:07 Quit BigNoob ("CGI:IRC (EOF)")
09:44:54 Quit Thundercloud (Remote closed the connection)
09:47:38 Quit Stevie8227 (Read error: 110 (Connection timed out))
09:52:28pixelmano-one uses hwcodec is wrong. I still like using my Ondio from time to time, just don't update that often. It currently runs a 23xxx build but before .sbs I think and I didn't see a filetype array issue - there were also two "bug reports" in the forums with problems introduced by recent changes on the RecorderV1 and even the Player
10:05:36 Quit Llorean (Read error: 54 (Connection reset by peer))
10:11:56 Quit reid04 (Read error: 110 (Connection timed out))
10:12:02 Join reid04 [0] (
10:21:23 Join kugel [0] (n=kugel@rockbox/developer/kugel)
10:23:30 Join FOAD_ [0] (
10:24:02 Join seani [0] (n=seani@
10:34:50 Quit FOAD (Read error: 110 (Connection timed out))
10:34:51 Nick FOAD_ is now known as FOAD (
10:36:28 Join funman [0] (n=fun@rockbox/developer/funman)
11:00:46 Join Stevie8227 [0] (n=Stevie82@
11:02:03 Join mikroflops [0] (
11:03:36 Quit phanboy_iv (Read error: 60 (Operation timed out))
11:09:08***Saving seen data "./dancer.seen"
11:13:59 Quit uflops (Read error: 110 (Connection timed out))
11:28:59 Part Stevie8227
11:30:20 Quit kugel (Read error: 110 (Connection timed out))
11:30:47 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
11:33:53 Join DerPapst [0] (
11:39:04 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
11:39:19seaniCould I have a nudge in the right direction for changing the code that gets executed when the PLAY / PAUSE buttton is pressed? I'm assuming there's a chance to add something in there before the functions in playback.c get called?
11:44:20 Nick Guest69461 is now known as J-23 (
11:53:03 Join Bob_C_ [0] (
12:01:20 Quit beta2k (Read error: 60 (Operation timed out))
12:01:31 Join beta2k [0] (
12:05:45 Join MethoS- [0] (n=clemens@
12:16:41seaniAh, looks like wps.c - wanders off to fiddle
12:16:44 Quit seani ("Leaving")
12:24:07 Quit MethoS- (Remote closed the connection)
12:32:19 Join Omlet [0] (
12:32:23 Join Omlet05 [0] (
12:37:50 Join kugel [0] (n=kugel@rockbox/developer/kugel)
12:42:22kugelweeh, -5.5K by moving init functions
12:43:35funmannot bad! which functions did you move ?
12:45:32kugela lot, but not all yet
12:47:21funmankugel: main()
12:48:01funmanhum I wonder why there is code after the call to app_main() in main() ..
12:48:16 Join adiroiban [0] (
12:50:09kugelnot sure, root_menu() cannot return
12:50:19 Quit Sajber^ (Read error: 54 (Connection reset by peer))
12:50:57funmanLED_REAL is only defined for archos, perhaps it's a left over from good old times
12:51:07funmanamiconn: do you have an idea ?
12:56:08kugelI think I have almost all now. -6.7K on my e200
12:56:55funmanbtw main / app_main / root_menu could be added the attribute noreturn
12:57:25funmanand perhaps naked? I don't know if regs are saved on the stack for noreturn functions
12:58:43 Join seani [0] (n=seani@
12:58:59amiconnLED_REAL isn't a leftover; it is used
12:59:50funmanamiconn: in main() ?
13:01:36amiconnIn various places
13:01:41 Quit Omlet05 (Client Quit)
13:01:41 Quit antil33t (Read error: 104 (Connection reset by peer))
13:01:41 Quit Omlet (Client Quit)
13:01:47 Join antil33t [0] (n=Mudkips@
13:01:47 Quit DerPapst ("Leaving.")
13:01:56amiconnThe loop in main() is just an indicator, should app_main() ever return (erroneously)
13:01:56funmanI wonder about this specific place
13:02:18kugela comment would be a better indicator
13:02:19amiconnIt will blink the LED in such a case
13:02:32amiconnA comment won't help the user
13:02:58funmanroot_menu() can't return, so app_main() neither
13:06:01funman__attribute__((noreturn)) effectively saves registers on the stack at function entry
13:06:56amiconnI know it shouldn't happen normally
13:07:25funmanI can't see a way it could happen erroneously
13:07:57funmanIf it's due to random memory corruption, the function might jump to anywhere, not specifically the caller (app_main)
13:09:08funmanthe only way I can think it would return is if the source code is modified
13:09:12Tornefunman: er, last i checked declaring a functoin as noreturn didn't actually make it optimise away the prolog/epilog unless it could statically see that it was impossible for the function to return (and if it can see that it can optimise it anyway)
13:09:13***Saving seen data "./dancer.seen"
13:09:19amiconnIt can happen if a developer goofs somewhere
13:09:35Tornefunman: i may be wrong but last time i fiddled, noreturn basically only affects whether you get various warnings or not :)
13:09:57amiconnThis is old code, it could probably be removed
13:10:37Tornee.g. to let you get away with writing int foo() { for(;;); }
13:10:37funmanTorne: then here attribute naked makes more sense? (to remove unneeded prolog/epilog)
13:11:55Tornemanual says it does do that optimisation
13:12:11funmanit = noreturn? that's what I understand from the gcc manual too
13:12:32Torneyes. but i'm pretty sure i've seen it fail to do it :)
13:12:38Torneof course these things may both be true
13:12:53Tornenaked means something entirely different and shouldn't ever be on a C function
13:13:18Tornenaked breaks local variables, amongst other things (it removes any handling of the stack pointer, so there's no memory alloced for locals)
13:13:31 Join Biont [0] (
13:13:41funmanoh I thought it would only remove registers saving/loading
13:14:02Torneno, it kills the prolog/epilog entirely
13:14:09Torneyou can only use asm() inside a naked function
13:14:12Torneno C code at all
13:14:41Torneit's for doing asm functoins which need nonstandard prolog/epilog code without having to write a seperate .s
13:15:43kugelis -6.7K considerable for SVN, or is this too dirty?
13:16:15Torneanyway, poking around on the internet suggests that gcc is probably just not very *good* at following through on the implications of noreturn, so whether you get particular optimisations or not may be somewhat random :)
13:16:29Torneits primary purpose is to kill warnings, not enable optimisation
13:20:33BiontHello everyone. I had an idea regarding touchscreen WPS and would like to know if there would be any troble implementing it before suggesting it on the forums. It's about defining custom tags/actions
13:20:56CIA-8New commit by funman (r23494): root_menu() can't return, tell gcc about it ...
13:21:14kugelthat will make rockbox so much faster :P
13:21:15funmankugel: I don't think it's dirty, perhaps you should bring the discussion on the dev list
13:21:41BiontIt should be something like the "repmode"-command which cycles through different values
13:22:59kugelfunman: I foresee many reds
13:23:16funmanwhy ?
13:23:25kugelno main() for sims
13:23:59funmanit's still called app_main() in the sim
13:24:00kugelunless that main() was never called from sims and the real main is elsewhere
13:25:34kugelalright, ignore me then. sim's main() is in uisdl.c
13:25:54funmansince it runs on a 'real' OS main() is reserved
13:27:36kugelevery C program needs a main()
13:28:05funmanyes that's what I mean, on the simulator main() will initialize SDL and stuff, and then call rockbox's main() (well, app_main() )
13:28:27 Nick fxb__ is now known as fxb (n=felixbru@
13:32:00 Join teru [0] (
13:33:19kugel-6.7k is less than I hoped
13:33:59kugelit'll be even less for non-ARM (those have usually more compact code)
13:40:58 Quit kugel (Read error: 60 (Operation timed out))
13:50:36 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:57:17Biontwould custom tags/actions for the (touchscreen) WPS be feasible? They could work like the repmode command and cycle through 2 or 4 values
13:58:27 Join Strife89 [0] (n=michael@
13:59:33BiontThat way, you could easily create a bunch of different popup menus to display settings ....or make buttons that change the appearance of the wps and many other things
14:01:16BiontI'll suggest it on the forums as well, but I wanted to ask if it is possible at all (or possible without too much hassle) first
14:02:26 Join Strife1989 [0] (n=michael@
14:02:56*gevaerts is a bit confused. "repmode"?
14:03:18Biontrepmode - cycle the repeat mode
14:03:43kugeldeleting themes doesn't work anymore?
14:04:17BiontIt's just to illustrate how I'd imagine it to press it and the value goes from 1 to 2 and so on
14:04:36kugelor does it take a while until its removed?
14:05:18gevaertsah, ok.
14:05:30*gevaerts doesn't know much about the GUI code, so he can't help
14:08:24seani336Mb of downloads to build the manual. Crikey. There isn't a question in there, just saying.
14:10:17BiontI don't know anything about coding at all. I'd love to do it myself if I could (because then, I could also make a theme that shows off all the cool stuff that would be possible with custom actions)
14:10:36Biont(which would surely help getting it committed)
14:15:13 Nick fxb is now known as fxb__ (n=felixbru@
14:17:46 Quit funman ("free(random());")
14:18:49preglowZagor: got a mail saying i'm bouncing mails again...
14:19:19*Zagor sighs
14:19:29 Quit Strife89 (Read error: 110 (Connection timed out))
14:35:23 Join Jaykay [0] (
14:38:05Jaykayi'm wondering about the ascending and descending values in the file size delta tabel, e.g. the zvm 30gb. is this due to different build clients, and can someone explain it to me?
14:50:54CIA-8New commit by BobC (r23495): Improvements to mini2440 audio; change CPU freq to 406MHz
14:51:22 Join froggyman [0] (
14:56:39kugelBagder, Zagor: The font page doesn't show Bob_C's real name for me
14:57:05 Join mcuelenaere [0] (
14:59:28Bob_C_Yes, I wondered that.
14:59:55gevaertskugel: I suspect something fishy in COMMITTERS. His login is the only that starts with an uppercase letter. I don't know how that is handled, but I do know that that file is supposed to be ordered, and I suspect that systems relying on that use strict ascii sorting
15:00:23*gevaerts does think that waiting for Bagder or Zagor to comment is best :)
15:00:27Zagorgevaerts: almost. it doesn't rely on sorting, but on the case being the same as the svn login
15:00:28kugelstrcmp vs strcasecmp failure? :P
15:00:34Bob_C_I noticed the COMMITTERS file is not in strict order anyway, but uppercase may be problem
15:00:47ZagorI fixed that now
15:01:21CIA-8New commit by zagor (r23496): Fixed case dependency
15:02:36gevaertsBob_C_: as far as I can see it is in strict order
15:04:26 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
15:04:34Bob_C_ok, I did a sort/diff, I thought some out of order were. I guess it might depend on locale
15:09:17***Saving seen data "./dancer.seen"
15:09:54*mcuelenaere agrees with kugel (recording tags mail)
15:10:26*kugel hands mcuelenaere a "Reply to email" window :)
15:10:46*mcuelenaere considers an "I agree" mail pretty pointless :)
15:10:56kugelI wonder why I'm always the last who gets his email from the ML
15:11:30kugelmcuelenaere: I think we still had 3 Swedish committers only if "+1" mails were useless
15:12:54kugelan "I agree" just as contributes well to a discussion as anything else (excluding troll...)
15:12:55 Quit AEnima15771 (Read error: 110 (Connection timed out))
15:13:36 Join evilnick_B [0] (i=0c140464@rockbox/staff/evilnick)
15:15:54 Join AEnima1577 [0] (
15:16:54FlynDicesaratoga:(for the logs)re: AMS frequencies, The problem with changing the core frequency to 31MHz is the restriction that PCLK <= FCLK so we need to lower PCLK also. If we do that then the IRAM and Memory are clocked lower also and it actually make battery performance worse.
15:17:08 Quit Strife1989 (Read error: 113 (No route to host))
15:18:19 Quit kugel (Read error: 60 (Operation timed out))
15:18:33FlynDiceI'll try to add some more presets to FS #10308 to make it easier to try different cominations if you're pulling out the test equipment....
15:18:50FlynDiceer combinations even
15:20:46mcuelenaerekugel (logs): sure, but I thought agreeing in here could result in an discussion while agreeing with the mail wouldn't have much result..
15:24:01 Nick fxb__ is now known as fxb (n=felixbru@
15:27:40 Join DerPapst [0] (
15:35:47 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:35:52 Quit kugel (Read error: 104 (Connection reset by peer))
15:42:56 Quit Jaykay (Read error: 60 (Operation timed out))
15:45:36 Join kugel [0] (i=kugel@rockbox/developer/kugel)
15:59:34 Part LinusN
16:05:40 Join Llorean [0] (
16:07:49 Join explore_ [0] (
16:07:59 Quit explore_ (Client Quit)
16:08:08kkurbjunkugel: so what is your take on the viewport issue.. currently none of the lcd drivers do proper bounds checking when a viewport is used
16:08:23kkurbjunthey all assume that the viewport is defined to be within the screen
16:09:05kkurbjunbut the more I think about it the more it seems like either a check when setting the viewport or clipping when drawing the viewport is appropriate especially with lua
16:09:11 Quit teru ("Quit")
16:09:57kugelkkurbjun: where's the exact failure? lcd_update_rect should check, bitmap_*[_part] too
16:10:44kkurbjunlook at lcd_bitmap_part in lcd-16bit.c
16:11:04kkurbjunit clips the bitmap with respect to x/y, but then it adds the viewport offsets in
16:11:51kkurbjunso if the viewport is negative or wider than the screen then it will draw into memory that it shouldn't be
16:12:47kkurbjunit seems to me that the logic for clipping is just broken with viewports - someone added that clipping logic in for a reason and then it was broken when viewports were added
16:13:34kkurbjunI mean, we could make the requirement that bitmaps have to be defined within the screen too, and just remove all of the clipping logic :-D
16:13:58 Join teru [0] (
16:14:20kugelah I see
16:14:55kkurbjunbut this also means that all of the lcd drivers need to be fixed with additional logic
16:14:57kugelwell, IMO the drivers shouldn't be doing all these checks anyway, but the callers instead, but I may be the only one thinking that
16:16:03kkurbjunkugel, the concern I have is that it adds a ton of checks all over the code, and when you look at things like plugins it could cause all kinds of problems if we let people accidentally (or maliciously) overwrite areas of memory
16:16:53kkurbjunnot with lua people don't even need to compile their own plugins you could do all kinds of things to the player just with a script
16:18:20kugelyea but the checks are done repeatedly even if the viewports are properly set up (which is true for (assuming) 99% of the core)
16:20:14kkurbjunwps's might not have the bitmaps setup properly without any clippnig though, what if the viewport is properly defined and the bitmap they are using is wider than the viewport
16:20:51kkurbjunthat is currently taken care of, but why neglect the viewport definitions if we're already doing clipping
16:22:09 Join toffe82 [0] (n=chatzill@
16:22:38kugelthe wps checks
16:22:54kugelI'm not saying the checks shouldn't be done at all, but transfered to the caller
16:23:05kkurbjunwhere does the wps check?
16:23:44kugelI don't think the wps checks if the images are wider than the vp currently
16:23:53kkurbjunso the wps should do all the clipping?
16:24:13 Join HBK [0] (
16:24:27kugelI'd be in favor of that yes
16:24:42mcuelenaerecan't we just expose "checked" calls to plugins and "unchecked" (aka performance-related) calls to core-only?
16:24:43kugelbut that's just me, I'm not really proposing the change the whole system
16:25:33kkurbjunI think clipping at the driver adds a small impact that is worth it for simplicity all over the code base
16:25:51kkurbjunplus sanity checking against user created input
16:25:56mcuelenaereI disagree, drivers should be as fast as possible
16:26:04*gevaerts really thinks that viewport clipping should be done by the "system", not the caller
16:26:15mcuelenaerealso isn't there some kind of unwritten rule Rockbox shouldn't have much defensive code?
16:29:16kkurbjuncurrently I'm planning on implementing clipping at the driver with a macro that can be called in each instance - if someone wants to run fast and dirty they can disable it, but I think if you pull clipping all together it will cause more complexity and problems than it's worth
16:30:21Bob_C_Fast is good, but not crashing is better imho
16:30:56kugelkkurbjun: remember we have a new "no code in macros" rule (I don't like it)
16:31:10kkurbjunplus these functions have loops that are much larger than the checks that are present
16:31:26Bob_C_s/crashing/going wrong generally/
16:31:41kkurbjununless you are talking about really small bitmaps or small drawing functions it's practically negligible
16:31:58 Quit Bob_C_ ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
16:32:03kkurbjunkugel: thanks for the heads up, I'll keep that in mind
16:39:18terukugel: you seem to forget to do vol_changed = false; in r23436. when i change volume in wps, it keeps showing info that volume is changed.
16:41:33kugelteru: oops, can you correct that?
16:43:23terukugel: can't do it now. i'm going to sleep..
16:45:21 Quit teru ("bye")
16:51:22DerPapstwth... the battery of my gigabeast just balloned. i hope i didn't damage anything while ripping that thing out. o.O
16:55:46 Quit AEnima1577 ("Leaving.")
16:58:21 Join Strife89 [0] (n=michael@
16:59:53 Quit bubsy ()
17:06:14 Join jgarvey [0] (
17:06:15 Quit niekie (Read error: 104 (Connection reset by peer))
17:06:15 Quit Zagor ("Don't panic")
17:06:15 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
17:06:15 Join evilnick_ [0] (
17:06:16 Join toffe82_ [0] (n=chatzill@
17:06:17 Join Xerion_ [0] (
17:06:20 Quit Topy (Connection reset by peer)
17:06:29 Join Topy [0] (
17:06:31 Join Blue_Dude [0] (n=chatzill@rockbox/developer/Blue-Dude)
17:07:14Blue_DudeI was fooling around with logf yesterday. I noticed that in a number of files, LOGF_ENABLE is defined. Shouldn't all these be commented out unless the dev specifically enables them?
17:07:35 Join saratoga [0] (i=9803c6dd@gateway/web/freenode/x-68a0c043251f94f4)
17:08:39DerPapstYAY! it still works :D
17:09:13saratogaFlynDice: is the restriction that PCLK <= FCLK explained in the manual or experimentally determined?
17:09:17 Quit Xerion (Read error: 60 (Operation timed out))
17:09:18 Nick Xerion_ is now known as Xerion (
17:09:20***Saving seen data "./dancer.seen"
17:12:10Blue_DudeAny objections to systematically commenting out all LOGF_ENABLE's?
17:12:54 Join bubsy [0] (n=bubsy@
17:13:05 Quit toffe82 (Read error: 60 (Operation timed out))
17:17:53 Quit bubsy ()
17:21:32 Join stoffel [0] (n=quassel@
17:22:22kugelBlue_Dude: just don't use evil // :)
17:22:42Blue_Dudekugel: I wouldn't think if it.
17:22:47*Blue_Dude is offended!
17:23:18 Quit jfc^3 (Read error: 110 (Connection timed out))
17:23:36 Join jfc^3 [0] (
17:24:59 Join Omlet [0] (
17:25:02 Join Omlet05 [0] (
17:25:06CIA-8New commit by blue_dude (r23497): Comment out LOGF_ENABLE defines everywhere, replace evil comments
17:25:14 Quit Horscht (Read error: 113 (No route to host))
17:25:20 Quit evilnick (Read error: 110 (Connection timed out))
17:25:53 Join stoffel_ [0] (n=quassel@
17:26:09 Quit stoffel (Read error: 104 (Connection reset by peer))
17:26:35saratogaBob-C: would you update TargetStatus with the mini2440 info?
17:26:57Bob_Csaratoga: sure, good idea
17:30:09 Join funman [0] (n=fun@rockbox/developer/funman)
17:30:44 Join AEnima1577 [0] (
17:31:17 Join niekie [0] (
17:32:35FlynDicesaratoga: 7.3.14 CGU - Clock generation unit, p100 in r1.13 of the datasheet, read ARM922T and AMBA bus clock it explains fastbus, asynchronous, & synchronous.
17:33:17Blue_DudeWTF? How did I end up with a red for that?
17:35:01 Join link4u [0] (
17:35:43link4ui cant boot my rockbox from the sd card on my Iaudio cowon D2
17:35:53link4uany 1 know solution?
17:36:15link4ui can boot from NAND, but not from the SD.
17:36:33link4uHelp pleease ...
17:38:28link4uDoes any one have the D2?
17:38:37link4upleease heeelp
17:39:06gevaertslink4u: people aren't watching this channel all the time, and not everyone knows about everything...
17:41:43funmanBlue_Dude: it's a "normal" build client failure, not your fault
17:41:49Blue_DudeI can't find the problem. Why did r23497 break?
17:42:13funmanit didn't
17:42:18Blue_DudeThank goodness. I thought maybe I'd forgotten to close a comment somewhere.
17:42:31funmanyou can check the specific errors by clicking on the red for each model
17:42:55Blue_DudeI did, but not closing a comment can sometimes cause strange build errors.
17:43:24Blue_DudeLeaving out expected includes, etc.
17:43:36funmanyes but here it doesn't happen when parsing C code
17:43:52 Quit stoffel_ ("No Ping reply in 90 seconds.")
17:44:17FlynDicesaratoga: re:ams clock restrictions, Even though it says it is not allowed we were running ansync bus with PCLK=62MHz & FCLK=31MHz, before we knew better, and it ran but caused problems when we tried to enable the mmu.
17:44:24 Join stoffel [0] (n=quassel@
17:44:45 Join bubsy [0] (n=bubsy@
17:48:25 Join Jaykay [0] (
17:53:29saratogaFlynDice: I was looking at that but it seemed to me that in async mode you could run fclk<pclk
17:54:30saratogaoh nevermind
17:54:31saratogai see now
17:54:33FlynDicesaratoga: "Within asynchronous mode, the ARM922T frequency must be higher than the AHB bus frequency,"
17:54:49link4uDoes anyone have the cowon D2?
17:55:37 Join Rob2222 [0] (
17:56:05FlynDicesaratoga: we run PCLK = FCLK by using fastbus
17:56:27 Quit Jaykay ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
17:59:32link4uDoes anyone know what is wrong when "ata mount error" or something shows up? I have a cowon d2 with files on SD card.
18:00:49FlynDicesaratoga: In other arm targets there is usually a divider allowing you to run AHB bus at HCLK and APB bus at PCLK=HCLK/2 but we don't have that with as3525, at least by what the datasheet says.
18:02:35 Join stoffel_ [0] (n=quassel@
18:02:35 Quit stoffel (Read error: 54 (Connection reset by peer))
18:05:41 Quit mcuelenaere (Remote closed the connection)
18:07:37 Join liar [0] (n=liar@
18:11:22 Quit Strife89 ("Lunch.")
18:12:20 Quit link4u ("CGI:IRC")
18:13:02 Join helpme_d2 [0] (
18:14:11 Part froggyman
18:14:38helpme_d2Hi. I have problems booting rockbox on D2 "SD" mode. It comes some sort of "ata mount Error" Pleease help a n00b like me :P
18:15:42 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:16:13helpme_d2Do i have to insert SD card while i boot, or can it be in before?
18:16:35Bob_CIs your card formatted as FAT32
18:18:12evilnick_Bhelpme_d2 / link4u: Have you read ?
18:24:46helpme_d2i can try format it again?
18:25:15helpme_d2but it can be inserted before turnon right?
18:25:53helpme_d2i have read those
18:26:00helpme_d2no sulution there
18:26:37evilnick_Bhelpme_d2: From my reading of that page, you should insert the SD card before powering it on. But I don't own that player.
18:27:04helpme_d2ok, i have tried both, but the same error
18:27:37helpme_d2exFat or fat?
18:28:03helpme_d2i have vista (yes, it sucks)
18:28:24 Quit gevaerts (Nick collision from services.)
18:28:36 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
18:29:01evilnick_BFAT32, and what is the exact error message that you get?
18:29:08 Join Casainho [0] (n=chatzill@
18:29:45Casainhohello everyone :-)
18:30:04evilnick_BCasainho: Congratulations!! (also to Bob_C)
18:30:42Casainhosome of you may already know, but others not, but Lyre Mini2440 project plays audio now :-) −− I shot a small video:
18:31:03Casainhoevilnick_B: thanks, but we need more ;-)
18:31:22Bob_Chelpme_d2: no, exFAT is not compatible
18:31:26Casainhoevilnick_B: I mean, more developers and users :-)
18:32:17helpme_d2wait a minute and i check exactly wich error
18:32:33helpme_d2ata mount something
18:32:46 Join bertrik [0] (
18:34:35 Quit funman ("free(random());")
18:35:12helpme_d2currently i do NAND boot, wich works fine
18:35:23 Quit petur ("*plop*")
18:35:26gevaertsdoes the SD card work then?
18:35:49 Quit maruk ("Leaving.")
18:35:55helpme_d2i dont think so, the file browser cant go higher than the nand root.
18:36:08helpme_d216 mb very slow card
18:37:09 Nick YPSY is now known as Ypsy (n=ypsy@
18:37:48gevaertsis there a <microSD> entry in the file browser?
18:38:05helpme_d2in the nand root?
18:38:20gevaertsin the root
18:38:46helpme_d2ATA mount file not found
18:39:08helpme_d2and the version ofcoarse
18:40:51helpme_d2DAMN. i hate vista, now i cant rename it to ".rockbox".
18:41:45 Quit DerPapst ("Leaving.")
18:42:16 Join dkopelove [0] (
18:44:06helpme_d2where is the "microSD" entry?
18:44:29gevaertsit should be in the root, i.e. when you go to Files you should see it
18:44:37 Quit Horschti ("Verlassend")
18:44:46helpme_d2nope, i dont
18:44:55helpme_d2i can try another SDHC card?
18:45:03gevaertsWhich version are you running? You can see it in "System->Rockbox Info"
18:45:04 Quit kugel ("exit(0);")
18:45:21helpme_d2that works
18:45:27helpme_d2now i see microsd
18:45:54helpme_d216 Mb cards dont work then?
18:47:04gevaertshm, is it formatted correctly? Rockbox only supports FAT16 and FAT32, but a 16MB card could be FAT12 I think
18:47:50helpme_d2as i said earlier, i have vista, and have 3 opt. when formatting : exFAT, FAT and ntcs
18:48:58domonokyhelpme_d2: vista probably chooses some silly FAT format for you. try a better format program :-)
18:48:58evilnick_Bhelpme_d2: Maybe try formatting it with so that you know for sure which FS it is.
18:49:26gevaertsFAT is definitely the right choice there, but I think windows just selects a FAT variant based on size, and for really small cards this might not work with rockbox. If that's indeed the problem, a third party formatting tool might help
18:49:45*gevaerts seconds evilnick_B's suggestion
18:49:47helpme_d2ok, thanks guys
18:52:25helpme_d2the sdhc card works fine, it boots
18:54:16helpme_d2swissknife thingey doesnt work though
18:54:32saratogaFlynDice: what do we clock DRAM at?
18:54:45saratogampmc_clk that is
18:54:49helpme_d2but see ya
18:55:07 Quit helpme_d2 ("CGI:IRC")
18:55:08*Bagder points out that the COMMITTERS file says that the user name should be in lower case...
18:56:00FlynDiceWe clock DRAM at PCLK freq because any attempt to use a divider for PCLK= mpmc_clk results in a white screen. Nothing I have found avoids this so far...
18:56:04bertrikCan we use real names in the commit messages?
18:56:30FlynDiceor for PCLK=mpmc_clk/2 I meant
18:56:32bertrikI'm seeing some commits attributed to 'myself' lately and I think that's a bad idea
18:57:14saratogaFlynDice: do you think its a problem with DBOP or more general?
18:58:05saratogagiven the poor performance of the IRAM on AMS, running fclk at 31Mhz and dram at 62 MHz might save a good bit of power while still performing acceptably, assuming we ever figured out how to use it
18:58:51FlynDiceI don't think that's a DBOP problem, we got a white screen failure when voltage scaling was reimplemented and it had to do with voltages a hair low getting stuck in the adc voltage check loop.
18:58:54 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
18:59:05saratogaok so its just a general failure to service the lcd
18:59:12CIA-8New commit by BobC (r23498): Change username to lower case
18:59:24FlynDiceI guess so
18:59:42saratogathe bad IRAM performance is itself odd given that the spec sheet says it has single cycle latency and 128 bit access bus(!), which should make it many times faster then DRAM
19:00:19*Bob_C hangs head in shame at first commit fail
19:00:43linuxstb_bertrik: You mean people saying "myself" instead of using their own name?
19:00:51 Join Cor33 [0] (
19:00:55bertriklinuxstb, yes
19:01:14linuxstb_Why is that bad?
19:01:42bertrikbecause I don't always know who did the commit
19:01:48rasherSurely any work is assumed to be by the committer if nothing else is stated
19:01:55rasherso "myself" is a step up
19:01:58bertrikWhy don't people just use their real name?
19:02:05saratogaits grammatically ackward
19:02:13saratogausually one does not refer to themselves in the third person
19:02:14linuxstb_Why don't you know who did the commit?
19:02:16rasherbertrik: Most commits don't have *any* names on it
19:02:34rasherIt'd be awkward to start putting your own name on all commits by default
19:02:40*domonoky thinks there is no name needed for commiting your own work.
19:02:52rasherBesides, if you're looking at the commit message, the committer's name should be *right there*
19:02:52*linuxstb_ agrees with rasher and saratoga
19:03:02bertrikWhy don't people just use their real name?
19:03:02domonokyonly for commiting patches from others, you should use the recommended commit format
19:03:18rasherbertrik: because it's awkward
19:03:29rasherbertrik: And you have your name on it by virtue of being the committer
19:03:37linuxstb_bertrik: That's why commits have a name attached automatically
19:04:07 Join Link4u [0] (
19:04:10rasherbertrik: When exactly is this an issue?
19:04:11bertrikI don't see how it's awkward, I put my name under an e-mail for example, instead of 'myself'
19:04:17amiconnImo saying 'myself' or even one's full name is superfluous in commit messages
19:04:34amiconnsvn always shows the committer as well, so it would be redundant
19:04:50Link4uhi again. Does rockboy work on iaudiio d2?
19:04:51domonokythe only reason we have the recommended commit message format, is to track people who dont already have commit access.
19:04:54 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:05:00bertrikyeah, either put the real name there, or just omit it
19:05:03linuxstb_amiconn: I would use it when part of the commit is mine, and part is from FS, for example.
19:05:12domonokyLink4u: yes, but it is in a early state.
19:05:16Bagderdo we really want uncommented LOGF_ENABLE lines?
19:05:22Bagderhm, commented I mean
19:05:29amiconnlinuxstb_: That's a special case
19:05:38rasherbertrik: surely it doesn't do any harm
19:05:54BagderI would rather see them removed than within comments all over
19:06:22bertrikrasher, ok, but why have any FS-commit message format at all then?
19:06:41rasherbertrik: huh?
19:06:42 Quit Cor33 ("CGI:IRC (EOF)")
19:06:43bertrik'myself' becomes a special case
19:07:04linuxstb_Ah, you're only talking about FS commits?
19:08:10rasherI still don't see what's wrong with writing "Myself". The committer's name is right next to it..
19:08:20linuxstb_Bagder: one advantage is that it makes it clearer how logf works to a newcomer. But I agree its messy...
19:09:21***Saving seen data "./dancer.seen"
19:09:35 Quit Biont ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
19:09:38 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:10:48FlynDice^^ +1 or I'd still be trying....
19:10:56bertrikWell I'll be using my real name
19:12:14 Quit linuxstb_ ()
19:14:24 Join phanboy4 [0] (
19:15:36 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:15:41 Quit Link4u ("CGI:IRC")
19:22:23 Join Corey [0] (
19:24:24 Join phanboy_iv [0] (
19:27:39FlynDicesaratoga: re AMS bad IRAM performance, I see it says there's a 128 bit wide data bus for the TRAM but that still gets to the 922 core via AHB at 32 bits wide and HCLK freq correct?
19:34:45*Llorean just got an email from a writer at Forbes regarding Rockbox.
19:37:03LloreanApparently Lee Gomes (does technology articles for their website) is doing something on Digital Audio and may be including some about Rockbox.
19:43:05 Quit phanboy4 (Read error: 110 (Connection timed out))
19:44:33 Quit Corey ("CGI:IRC (EOF)")
19:48:37 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-f7e7780859ef60f7)
19:48:46 Join GeekShado_ [0] (
19:49:01 Join Stephen_ [0] (
19:50:19 Nick bluebroth3r is now known as bluebrother (n=dom@rockbox/developer/bluebrother)
19:55:13 Join GeekShad__ [0] (
19:55:53 Join stoffel [0] (n=quassel@
19:55:56 Quit explore ("leaving")
19:56:12 Quit stoffel_ (Read error: 104 (Connection reset by peer))
19:58:13 Quit amiconn (Remote closed the connection)
19:58:13 Quit pixelma (Remote closed the connection)
19:59:36 Join Thundercloud [0] (
20:03:20 Join pixelma [0] (i=quassel@rockbox/staff/pixelma)
20:03:21 Join amiconn [0] (i=quassel@rockbox/developer/amiconn)
20:05:30 Quit GeekShadow (Read error: 110 (Connection timed out))
20:07:34Stephen_anyone noticed this ?
20:07:44Stephen_I just tested and it is indeed the ondio build
20:07:49 Join pyro_maniac [0] (
20:08:00pyro_maniacbetrik: ping
20:08:02 Quit dkopelove ("leaving")
20:09:00pyro_maniacbertrik: ping (now without typo in name :) )
20:10:27 Join panni_ [0] (
20:10:46 Join domonoky1 [0] (
20:11:42 Quit GeekShado_ (Connection timed out)
20:13:58domonoky1uh, what happend with the current builds ? somehow the contents of the zips are swapped. for example the ondio fm zip contains a ipod-nano build ?
20:14:19Stephen_yeah sansa has ondio in it
20:14:27FlynDiceStephen_: Yes the e200v2 download appears to be the ondio, the fuze is correct
20:14:55domonoky1the contains a ipod video build... is a ipodcolor.. looks pretty messed up :-/
20:15:44gevaertsthe builds are correct. The problems start after uploading
20:16:02gevaertsthe build log for e200v2 does indeed build for e200v2
20:16:47FlynDiceondio is ipod nano...
20:16:54 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.5.4/20091016092926]")
20:16:58*gevaerts sees that Zagor has run away
20:17:22Stephen_a new game sort of like deal or no deal for rockbox builds ?
20:18:04linuxstbThat's an impressive bug...
20:18:09FlynDicenano is ipod video...
20:18:48domonoky1FlynDice: you dont need to repeat what i already said :-)
20:19:32FlynDicewhoops... just following the trail that started with my beloved e200v2 ;-)
20:19:32*linuxstb wonders about making a commit to see if it will fix itself next round...
20:21:05 Join casainho_ [0] (n=chatzill@
20:21:23*FlynDice loves ambitious volunteers!
20:21:49 Nick casainho_ is now known as Caainho (n=chatzill@
20:24:10 Join darkham [0] (n=darkham@
20:25:09*amiconn is kinda surprised that the __attribute__((noreturn)) change didn't cause warnings
20:25:09darkhami never thinked that doom would have run properly in my fuze...
20:25:17darkhamit's unbelieveble
20:25:51amiconnLast time I tried __attribute__((noreturn)), all that happened was gcc spitting "Warning: noreturn function does return"
20:25:54CIA-8New commit by dave (r23499): Cosmetic commit (removing empty line from end of file) to test build system
20:26:09 Join perfectdrug [0] (
20:26:30 Join mcuelenaere [0] (
20:26:35darkhamwill the mantainer never implement the fuze's scroll?
20:26:57domonoky1darkham: which maintainer ? :-)
20:26:57linuxstbThere's no single maintainer.
20:27:05 Quit domonoky (Read error: 110 (Connection timed out))
20:27:07mcuelenaereCasainho: I hope you know of absolute touch mode?
20:27:10domonoky1darkham: maybe you could volunteer ?
20:27:10Lloreandarkham: It's entirely volunteer. Will you implement it, perhaps?
20:27:27darkhamdomonoky1, linuxstb Llorean , why not?
20:27:47pixelmaand what's the Fuze's scroll, what's so special about it?
20:27:49linuxstbdarkham: Because that's not how Rockbox works. Everyone is a volunteer, doing what they can, when they want to do it.
20:27:51perfectdrugany X5 user around to verify a bug report I'm going to file for someone else? or maybe its better to post the bugreport first, isn't it?
20:28:24darkhampixelma, the fuze's round bottun
20:28:58gevaertsAre you saying that the scrollwheel doesn't work for you?
20:29:24pixelmaperfectdrug: if it doesn't have to do with colour, I could test on my M5 (the greyscale version of the X5)
20:29:46saratogaactually i was going to disable doom on the fuze so people stop reporting that it crashes due to lack of memory
20:30:37domonoky1saratoga: just add a splash a doom startup: "may crash, please dont report" :-)
20:31:36linuxstbsaratoga: That may just cause more questions, as people will still have doom.rock installed (and it will eventually give a plugin version error...)
20:32:06linuxstb(unless of course you write a "null" doom plugin to replace it...)
20:32:09saratogadon't they have to have it listed in plugins somewhere for it to show up in the menu?
20:32:15saratogaor is that only viewers?
20:32:33linuxstbJust viewers AFAIK
20:32:41LloreanHow much memory does the Fuze have?
20:33:13domonoky1looks like the buildsystem fixed itself..
20:33:16LloreanDoes Doom 1 crash then or just Doom II-based WADs?
20:33:20linuxstbDoes it actually crash? i.e. it doesn't check if its run out of RAM?
20:33:22domonoky1pixelma: bits ofcourse
20:33:52Lloreanlinuxstb: The PC version won't start if not enough free RAM is detected, so I guess that's the only measure they felt necessary?
20:34:00mc2739linuxstb: it locks at the end of the first level
20:34:26gevaertsHow hard would it be to make the doom malloc check for out of memory and exit gracefully?
20:34:36gevaerts(assuming that that's what happens)
20:34:42amiconn8MB RAM should be sufficient for running doom
20:34:45 Quit flydutch ("/* empty */")
20:34:51saratogalinuxstb: i crash pretty quick in the second level of doom 1
20:34:56saratogathough it depends what I do
20:35:04*linuxstb downloads a 2nd gen Nano build, and it's a 2nd gen Nano build...
20:35:12saratogayou can watch in the sim as it depletes the malloc buffer and then dies
20:35:48linuxstbThat does sound like a valid bug though...
20:35:59saratogathe bug is that our doom port is inefficent
20:36:07perfectdrugpixelma: no it hasn't to do with colour. So let me describe: the X5 has as far as I understand no stopbutton, so long play is used for this. whenever he presses play too long he gets a terrible earbleeding audio distortion, when he resumes playback via the mainmenu. He also uses most of the time the remote, at which the play button serves 3 functionalities play/pause, stopp and shutdown. But he says the problem occurs even without the remote attached
20:36:15saratogaeither due to using too much memory, or not having an efficient malloc, or perhaps both
20:36:36linuxstbPlus the bug of not failing gracefully.
20:37:45pixelmaperfectdrug: with all audio formats (MP3 or something else) and does he use voice or something?
20:37:47 Quit Thundercloud (Remote closed the connection)
20:38:48pixelmaperfectdrug: and could he check if he has keyclicks enabled?
20:39:26pixelmathat caused an aweful beeeeeep on my c200 for a while
20:41:12perfectdrugpixelma: not sure about audio formats but he says it also occurs with a clean install. but I'm not sure with this neither. I let him check thanks. do you know the combo to clear settings at startup?
20:41:40gevaertsperfectdrug: please backup the settings first. We want a chance to find what's going on...
20:41:50pixelmano, I don't
20:42:17gevaertsIIRC there's also a reset settings menu item somewhere
20:42:32 Join Jaykay [0] (
20:43:39Jaykaysecond try: i'm wondering about the ascending and descending values in the file size delta tabel, e.g. the last commit. is this due to different build clients, and can someone explain it to me?
20:43:42*pixelma has to build an M5 build first to have full support of the radio
20:44:20 Quit phanboy_iv (Read error: 104 (Connection reset by peer))
20:44:40 Join Biont [0] (
20:45:06CaainhoI have a question to ask: will Mini2440 port be able to be listed on first Rockbox page near to Stable/unstable ports?
20:45:54pixelmaperfectdrug: if he backs up the config.cfg from the computer he could also remove it from the place in .rockbox, then settings will also be reset to default at the same time
20:46:16pixelmaI mean if he saves it to another place
20:47:04 Quit Casainho ("ChatZilla 0.9.85 [Firefox 3.5.3/20091020122601]")
20:50:28 Nick Caainho is now known as Casainho (n=chatzill@
20:50:43 Join dos1 [0] (n=dos@unaffiliated/dos1)
20:51:30 Join funman [0] (n=fun@rockbox/developer/funman)
20:52:33 Join kugel [0] (n=kugel@rockbox/developer/kugel)
20:53:51linuxstbJaykay: I'm not sure anyone understands exactly why, but yes, I think it's just different build servers.
20:56:41 Join Tomers [0] (
21:03:15Bionthas anyone ever found out why the database refuses to build properly in some cases? I already erased all the tags that aren't neccessary on my D2 and it still won't build
21:03:58Tomershas anyone EVER compiled rbutil in Windows?
21:04:17*Tomers pulling my hair out trying to compile rbutil with Qt Creator
21:04:46pixelmaI think one of the two Domoniks (or even both) did
21:04:57BiontIf bad tags are the problem, perhaps someone knows what exactly is wrong with them
21:05:30pixelmasorry, Dominiks
21:05:30bluebrotherTomers: I frequently compile rbutil on windows. Why?
21:06:14Tomersbluebrother: I followed RockboxUtilityDevelopment#How_To_Compile
21:06:19pixelmaBiont: what files (type), how are they tagged, maybe which target?
21:06:23Tomersand didn't make it
21:06:32linuxstbBiont: No-one ever seems to file a useful bug report about that. You can try adding "database.ignore" files into some of your directories, to see if it any specific file or sets of files causing the problem.
21:06:42TomersI have latest Qt SDK, I've installed Cygwin... What else is needed?
21:07:00TomersI've opened the project in Qt Creator, and get lot's of reds
21:07:03bluebrotheryou don't need cygwin. You need MinGW and msys
21:07:24TomersDoes Qt SDK installs msys? I think it also has MinGW, right?
21:07:41 Quit Utchybann (Read error: 113 (No route to host))
21:07:53Biontpixelma: mp3 and flac, I don't know about how they're tagged, cowon d2
21:08:03linuxstbTomers: "MINGW You need a recent version (current is 5.1.4), the package offered for installation by Qt is not sufficient." (from the wiki page)
21:08:05bluebrothernot sure about that. I never used Qt SDK myself (it wasn't around when I started using it, and for building statically you need to rebuild Qt yourself anyway
21:08:42bluebrotherbut I guess Qt SDK installs an appropriate version of MinGW. I doubt it installs msys though
21:08:49Biontlinuxtb: how does that work? I just copy that file into directories and they'll then be ignored?
21:08:56bluebrotherfurthermore, I'd suggest compiling from the command line, at least until everything works.
21:09:03 Quit antil33t (Read error: 110 (Connection timed out))
21:09:04pixelmaby the way, compiling an M5 build (with radio mod enabled though I doubt it's important), I get following message right after "generating dependencies": m68k-elf-gcc: /home/rockbox/apps/plugins/boomshine.lua: linker input file unused be cause linking not done
21:09:07Tomersbluebrother: So you pull Qt's tree and recompile it all?
21:09:12linuxstbBiont: Yes.
21:09:17pixelmaon cygwin that is
21:09:20mcuelenaerepixelma: that's a known issue
21:09:23***Saving seen data "./dancer.seen"
21:09:37bluebrotheralso, you need to put the sources in some location without spaces in the path. Putting it somewhere below "My Documents" is usually a bad idea, at least up to XP
21:10:05pixelmamcuelenaere: what does it mean?
21:10:14Biontlinuxstb: Cool, I didn't know that
21:10:23bluebrotherwell, if you only want to build it but aren't interested in creating statically built binaries you can use the provided version. I still prefer getting Qt Creator and Qt itself separately.
21:10:31mcuelenaerepixelma: when generating dependencies, the preprocessor gets run for each source file
21:10:36mcuelenaereand it doesn't skip .lua files
21:11:07linuxstbmcuelenaere: So the build system isn't showing that as a warning/error ?
21:11:16mcuelenaerelinuxstb: nope
21:11:39*bluebrother powers up the VM to check if current svn rbutil builds on windows
21:11:55 Join Bob_C_ [0] (
21:12:30bertrikpyro_maniac, hi
21:14:58pyro_maniacbertrik: did you have any contact to meizu people?
21:15:14bertrikpyro_maniac, no
21:15:29bertrikexactly what do you mean by meizu people anyway :)
21:15:50 Join hillshum [0] (
21:16:16pyro_maniaci don't know exactly but people who may help you.
21:17:29pyro_maniacor do you think there is no chance in getting help of them?
21:18:57bertrikI don't expect any help, but I haven't tried either
21:20:15pyro_maniacmy ceo visited a producing partner of meizu. i don't know how NDAed wie are but maybe i can give you some contact information if you want.
21:20:51pixelmaperfectdrug: I couldn't reproduce the issue on my M5 yet, tried with voice enabled and disabled, keyclicks on and off, MP3 and Ogg Vorbis files
21:21:06pixelmaI don't have a remote for it though
21:22:43 Join Thundercloud [0] (
21:23:08bluebrotherTomers: current svn builds fine on my windows box. How exactly does your problem look like?
21:23:10BiontOkay I put database.ignore files in every folder that has no music in it. what is there to avoid concerning tags?
21:24:04linuxstbBiont: I don't know. You should try putting database.ignore files in some of your music folders, and seeing if that helps. i.e. narrow down which files cause problems.
21:24:05 Quit Casainho (Remote closed the connection)
21:24:08Tomersbluebrother: I probably didn't read the web site deep enough, and missed the remark regarding the wrong version
21:24:31TomersI thought I could compile it using the Qt suite, and I couldn't
21:24:32 Quit perfectdrug ("CGI:IRC (EOF)")
21:25:03TomersI'm going to use MSYS, MinGW and compile Qt from source, then compile Rbutil. Is that the right path to go?
21:25:29 Join perfectdrug [0] (
21:25:47bertrikpyro_maniac, a lot of information is available already, like a datasheet of the main SoC and an M6SP schematic. The hard part is the FTL in my opinion and I guess we'll never get official specs for that.
21:26:03kugelwow, the sound quality of the mini2440 is quite *horrible*
21:26:30bluebrotherTomers: well, the version issue with w32api is probably not an issue anymore with current Qt releases. Haven't tried since a while though.
21:26:39Biontlinuxstb: Okay....doesn't sound like too much fun, I'll try that tomorrow
21:26:45Biontthank you
21:27:17kugelinteresting, the watermark caluclation is so totally wrong on it
21:27:40Tomersbluebrother: Thanks for wasting your time for this :-)
21:28:13pyro_maniacbertrik: sounds right to me. but if i should ask my boss just tell me.
21:28:19linuxstbkugel: A limitation of the hardware, or bugs?
21:30:35kugellinuxstb: I assume the hardware, it wasn't any better on the preinstalled OS
21:31:13linuxstbSo, it's not portable, and has horrible sound quality. Great DAP potential...
21:31:31kugelI don't think anyone bought it for using it as a dap
21:31:51kugelexcept casahino maybe
21:31:57linuxstbThen why is there a Rockbox port?
21:32:25JdGordon|kugel: how horrible?
21:32:37kugelJdGordon|: pretty
21:32:48kugelinstall a current build and see yourself
21:33:04*JdGordon| hasnt even installed a bootloader yet :(
21:33:53kugellinuxstb: why not? :) it's requiring any new apps/ code
21:34:24bertrikpyro_maniac, thanks for the offer, but no need to ask your boss I think
21:35:07CIA-8New commit by torne (r23500): FS #9728: Battery current measuring on the Video iPod ...
21:35:11 Quit Biont (Remote closed the connection)
21:35:52bertrikkugel, what's so horrible about the sound quality?
21:36:22bertrikis it the codec, or maybe the board layout causing interference or something like that?
21:37:18kugelit sounds like a 50s radio without the crackling noises
21:37:26Torneprobably layout
21:37:52kugellike a bag of shit, basically :P
21:38:07TorneMany devboards have exceptionally poor sound quality
21:38:18Torneit's too much of a pain to route those lines better :)
21:39:39kugelTorne: is there any plans for making the battery runtime estimation based on the power it uses?
21:39:59kugelthe nano2g could do that too
21:40:01Tornethat would be a rather large departure from how we work on every other target.
21:40:07 Join froggyman [0] (
21:40:10Tornei mean it's possible
21:40:15Tornebut i'm not particularly bothered by it :)
21:40:19bertrikamiconn had some ideas to make runtime estimation more dynamic IIRC
21:40:30Tornei committed that because it's useful to test other power related changes
21:41:00 Join guest001 [0] (
21:41:52Torne(and to reproduce FS #10719)
21:42:04kugelthe mini2440 is probably going to be very useful for debugging the eabi breakage
21:42:17TorneThat is a good idea :)
21:42:22kugelserial logf/printf is nice!
21:42:28Torne..serial logf?
21:42:34Tornefull ICD surely
21:43:04Tornein-circuit debug
21:43:10Tornemuch more useful than printf.. ;)
21:43:18*kugel hates those font where you can't distinguish I and l
21:44:04kugelTorne: that too of course. I think casahino managed to make ocd with eclipse and stuff already
21:47:59 Part guest001
21:52:33 Join AaronM [0] (
21:54:55TorneUnfortunate news: I can't see anything in the beast's FAT code to indicate that it's in any way not standard
21:55:12Torneit's a rather trivial implentation..
21:57:14CIA-8New commit by blue_dude (r23501): Makes sure codec API builds correctly if ROCKBOX_HAS_LOGF is defined
21:59:01 Quit jgarvey ("Leaving")
22:04:31seaniCould someone have a look over my second patch for FS #10754 ? I think this removes the intrusion into playback.c, and adds a note to the manual. Thanks!
22:07:08 Quit darkham (Read error: 110 (Connection timed out))
22:07:48 Join darkham [0] (
22:09:57 Quit darkham (Remote closed the connection)
22:11:57 Join n1s [0] (n=n1s@rockbox/developer/n1s)
22:12:43 Quit perfectdrug ("CGI:IRC")
22:13:16 Quit Jaykay (Remote closed the connection)
22:15:23 Quit stoffel (Remote closed the connection)
22:16:40kugelsprintf is a standard file right?
22:16:45kugelsprintf.h I mean
22:17:04*kugel wonders if it should contain a prototype of a non-standard function
22:18:50kugelhmm, maybe it's not a standard file
22:19:00funmanman sprintf says it's declared in stdio.h (never seen a sprintf.h)
22:20:04CIA-8New commit by kugel (r23502): Fix a #elif without condion (is an error with newer gcc/cpp versions).
22:20:11CIA-8New commit by kugel (r23503): Rename vfnprintf to vuprintf (u stands for user for the user callback and data passed to it) to avoid confusion with file and buffer size related ...
22:20:29 Quit shai ("Leaving")
22:23:30 Join faemir [0] (n=faemir@
22:23:53 Join antil33t [0] (n=Mudkips@
22:25:55*domonoky1 tests the audio on the mini2440, and thinks it doesnt sound too bad. But SD (or uart?) accesses cause very bad interfernces..
22:27:18 Join LambdaCalculus37 [0] (n=LambdaCa@rockbox/staff/LambdaCalculus37)
22:28:04 Quit antil33t (Read error: 54 (Connection reset by peer))
22:28:05CIA-8New commit by kugel (r23504): Simplify uart_printf() a bit by using vuprintf(), that also makes removing a static buffer possible.
22:28:45kugeldomonoky1: yours is better than mine then?
22:29:17domonoky1kugel: maybe i just have lower standards ? :-)
22:30:23kugelpossibly :)
22:30:25 Join MethoS- [0] (n=clemens@
22:30:39kugelalthough I wouldn't call myself an audiophile
22:30:58domonoky1kugel: is your lcd mounted ontop of the board ?
22:31:24kugelI actually want to access the buttons there ;)
22:33:40*kugel assumes splashes could be simplified with vuprintf too
22:35:35 Join antil33t [0] (n=Mudkips@
22:36:03saratogafuze current consumption: 550mah/16 hours = 34 mA
22:36:41saratogathe spec sheet says 0.25mW per MHz for the arm9 core, so at 62MHz and 4v: 62*.25/4 = ~4 ma
22:36:56saratogaso the rest of the chip is using 30 mA or the specs are way off
22:37:09saratogai wonder where all that power goes
22:38:11kugeldid you make a bench?
22:38:23funmansaratoga: did you try without playing or with lowest volume ?
22:38:49funmanor even with HP muted (there's a bit in as3514 regs)
22:39:42*kugel us reminded that he wanted to have a look at clocking dbop down when backlight is off
22:40:25saratogafunman: thats just estimated from battery benches and the numbers in the spec sheet :)
22:40:45 Quit Tomers (Read error: 113 (No route to host))
22:40:48saratogai'll try and find time to go and actually measure things as soon as i can
22:40:56kugelwould be nice if it could be turned off altogether, although that makes waking it up on button presses hard (especially with bertrik's patch) I assume
22:41:01funmanoh ok i thought you had measured with a dmm
22:41:24funmankugel: btw what's the status on this patch?
22:41:55funmani thought it was working on fuze
22:42:24kugelthe latest adds some display shifting and sporadic blue pixels because it access dbop while the lcd is working
22:43:02saratogathe spec sheet says 22 mA is typical without DRAM but with audio playback
22:43:25kugeldidn't the OF have an isr for erroneous dbop transfers?
22:43:32GodEateris charging working on the beast now ?
22:43:36 Join jgarvey [0] (
22:43:46kugelmaybe that's used to detect button presses when the dbop is off/in a low power state
22:43:47funmani was understanding that the OF only uses DRAM for USB, and everything else fits in the 384kB
22:44:24funmanoh no, some SD code is located in DRAM
22:44:51 Quit funman ("free(random());")
22:45:03kugelsaratoga: 30 isn't far off then, considering that we use dram (I assume the OF does so too)
22:45:38kugelalso, 30 would be an avarage. it's probably a bit lower when the battery is full
22:47:52*kugel is changing languages to stop audio playback on his mini2440 :/
22:48:12n1sGodEater: it has worked for a long time
22:48:31n1ssingle boot would be kind of stupid otherwise :)
22:49:02saratogathe voltages are regulated so the power consumption should be nearly constant
22:54:21 Join r0b- [0] (
22:54:30r0b-how complete is rockbox for the Sansa Fuze?
22:54:47saratogacheck the front page's link
22:55:04saratogai wonder what dropping the pclk to 31MHz does to power consumption
22:55:53bluebrotherdomonoky1: is there a reason libmkamsboot.a doesn't contain md5.o and dualboot.o? This seems to cause link issues for me −− strangely only in one setup, which should be identical to the one that's working.
22:57:27domonoky1bluebrother: at the moment i cant think of any reason.. maybe problems with universal libs on Mac ? But i am not sure.
22:57:34 Quit jgarvey ("Leaving")
22:57:48bluebrotherI'm quite surprised it does work without those objects at all.
22:57:50 Quit bertrik ("Leaving")
22:58:28*domonoky1 checks the Makefile..
22:59:27 Quit gevaerts (Nick collision from services.)
22:59:37 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
22:59:49domonoky1it got lost in the last revision. thats the reason..
23:00:05bluebrotherhuh? On my linux box it seems to have those objects in the .a, but not on the newer windows setup
23:00:24bluebrotherhmm, I've svn up'ed all checkouts first.
23:00:26*bluebrother rechecks
23:00:35domonoky1tomers broke it :-)
23:01:20*bluebrother now has the issue on linux too. Seems an old .a was lying around
23:02:25bluebrotherdomonoky1: you're planning to fix? I've already changed it in my local tree
23:02:40domonoky1then commit it. :-)
23:05:52bluebrotherhmm, the way dependencies are handled in that Makefile is broken IMO.
23:06:30bluebrotherdepending on objects and relying on them to depend on other sources is really confusing.
23:07:20 Quit lifeless_ (Remote closed the connection)
23:07:38 Join lifeless_ [0] (n=lifeless@
23:08:50*hillshum notices LambdaCalculus37's bday on his GCal
23:09:23*LambdaCalculus37 will continue with hillshum in -community :)
23:09:27***Saving seen data "./dancer.seen"
23:10:51bluebrotherhmm. Why was there no notification by CIA?
23:11:10domonoky1bluebrother: you should also add it to the depencies ?
23:11:32 Quit Omlet (Connection timed out)
23:11:43bluebrotherdomonoky1: I'll recheck the whole Makefile. Those way of handling dependencies is quite broken.
23:11:59 Quit Omlet05 (Connection timed out)
23:12:19domonoky1oki, just make sure to not break the trickery for mac :-)
23:12:50kugelCIA-8: ???
23:12:55 Join TopyMobile [0] (
23:13:37bluebrotherdomonoky1: the current "way" is that libmkamsboot.a depends on libmkamsboot.o, which in turn depends on dualboot.o and md5.o. ..oO
23:13:40*bluebrother screams
23:14:33bluebrotherwhile this does work it's completely confusing and non-KISS.
23:14:39 Quit n1s ("Lämnar")
23:14:43bluebrotherwhich means: bad. IMO :)
23:15:56 Quit pyro_maniac ("Leaving.")
23:16:20 Join DerPapst [0] (
23:16:33*kugel doesn't find that confusing
23:17:17amiconnsaratoga: There are more chips consuming power, not just the SoC
23:20:30*domonoky1 wants automatic building of rbutil on all 3 OS, to find such breakages earlier :-)
23:22:20bluebrotherwell, if you have to look dependencies up because it depends on it depends on it depends then that's nothing thats a good way if this kind of dependency isn't needed at all
23:22:50bluebrotherplus, depending on objects breaks parallel building
23:22:53 Quit dos1 (Remote closed the connection)
23:23:07bluebrotherat least if those objects are created by the same build
23:23:38amiconnDepending on objects is usually not avoidable for libraries and final binaries
23:23:41domonoky1doesnt the depency just specify a target ?
23:24:17domonoky1and this Makefile isnt easy, with all this mac things.. but feel free to improve.
23:24:23bluebrotheryes, but that means that the other target has to be built first.
23:24:55gevaertsbluebrother: what's the problem with that? Make is designed for that sort of thing
23:24:57domonoky1sure, but there is no way around this, when you link things together.
23:24:58bluebrotherbut building the source _itself_ does not require that object (which is the target) at all
23:25:55hillshumdomonoky1: Linux and OS X could be added to the build system without too much difficulty...
23:26:34domonoky1bluebrother: ah you mean the mkamsboot.o rule ?
23:26:39bluebrothergevaerts: sure it can handle it. But it's simply overly complex in this case
23:26:46bluebrotherdomonoky1: yes.
23:27:12bluebrotherhillshum: I've had a setup building rbutil on linux and windows a while ago. Though that was based on buildbot
23:29:07 Quit DerPapst (Read error: 54 (Connection reset by peer))
23:29:29 Join DerPapst [0] (
23:29:41domonoky1bluebrother: yes, that could be improved.. but at the moment its nicely central, otherwise you have to take extra care to get all depencys right for all cases this Makefile is called.
23:34:18 Quit bmbl ("Bye!")
23:35:11 Quit bubsy ()
23:43:57 Quit domonoky1 (Read error: 104 (Connection reset by peer))
23:47:42 Join MG_Man [0] (
23:49:17 Join bubsy [0] (n=bubsy@
23:50:47 Quit Stephen_ ("Leaving")
23:51:25 Quit mcuelenaere ("Gnight")
23:53:23MG_ManWell. if any of you remember me, I finally got my iriver
23:53:42MG_ManDoes it matter if we put on the 1.28 or 1.29 firmware before we put on rockbox?
23:54:00MG_ManI know the USA version is unsupported, but does the choice of 1.28 or 1.29 make any difference?
23:55:35MG_ManWhat are the changes between them anyway?
23:56:01MG_Manoh, and 1.30 EU
23:57:42evilnick_B1.30 is the most recent, but if you're putting Rockbox on it, then you won't be too worried about the OF anyway
23:58:03 Join fyrestorm [0] (

Previous day | Next day