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 2008-12-10

00:00:00rasherBetter stay in hiding then. Commit'n'run is very much frowned upon
00:01:48Unhelpfuli'll be around, although not continuously at my computer.
00:01:48 Join DerDome [0] (
00:02:15gevaertsUnhelpful: are you aware of ?
00:05:06Unhelpfuli was aware of the existence of a build table. anyway, things are about to get busy again at home, i'll commit when i'll be around a bit more, in case there's any fallout.
00:05:51gevaertsUnhelpful: if you check back about five to ten minutes after committing, you're fine
00:06:45Unhelpfuli can probably manage that.
00:07:37gevaertsAt least don't wait a week before checking
00:09:09 Join Hillshum [0] (
00:10:38BigBambiUnhelpful: \o/
00:11:00BigBambiAnother of the most often requested features goes in :)
00:11:22rasherNow for jpeg-in-core!
00:12:00BigBambiyeah - Unhelpful, get working :P
00:12:03*rasher still thinks a sort of image-codec thing that used the plugin-buffer might work
00:12:20 Quit alexbobp ("leaving")
00:12:28 Join alexbobp [0] (n=alex@
00:12:50rasherBut might mean more trouble than simply fusing it into the core, I guess
00:12:57Lloreanrasher: Using the plugin buffer would be a problem, since the images should be decoded and converted on buffer probably, and that can happen while you're in a plugin.
00:13:07 Join funman [0] (n=fun@rockbox/developer/funman)
00:13:18rasherLlorean: If you're in a plugin, you don't need AA anyway!
00:13:24Lloreanrasher: Until you leave it.
00:13:25funmanHostile: did you just say you have a 4GB Clip?
00:13:37*gevaerts found why using the flash buffering patch on a gigabeat breaks USB
00:13:47 Quit itcheg (" ajax IRC Client")
00:13:53Lloreanrasher: Maybe image codecs get their own memory on the assumption that maybe we'll support other image formats in the future (GIF for wpses maybe?)
00:13:54rasherLlorean: Sure, that might just be the cost though.
00:14:25rasherNot getting jpeg AA if you were in a plugin during buffering, that is.
00:14:50rasherI don't know if this would be worth it, but it *is* an alternative to putting it all in the core
00:15:16LloreanBut it'd lead to similar brokenness as "asking them to manually stop+resume if they change WPSes"
00:15:34rasherYes indeed
00:15:34funmanI just received my JTAG cable, I have a 20% off if someone wants to buy one on the ebay shop where I got it.
00:15:38*gevaerts gets overexcited and throws the file browser out of the core ;)
00:15:42LloreanI think if we do jpeg in the core, we should at least be firm about it and make it work. All or nothing.
00:17:17 Join _lifeless [0] (n=lifeless@
00:17:40kugelUnhelpful: Good work! Really
00:17:49funmanUnhelpful: welcome :)
00:17:57Unhelpfulgrah, yellows :P
00:19:00LloreanUnhelpful: What are the restrictions on image resizing?
00:19:23gevaertsUnhelpful: is the #if LCD_PIXELFORMAT set around line 455 in resize.c supposed to have all possible cases? If so, maybe an #else\n#error unsupported pixel format\n #endif may be a good idea
00:20:04funmankugel: did you have time to work on the fuze buttons yet (study the OF isr)?
00:20:22 Quit robin0800 (Remote closed the connection)
00:20:34amiconnUnhelpful: (1) Why is there an increase for the mono targets (though it's a moderate one), and more importantly, why does it differ between the archos recorders and the Ondios?
00:20:35UnhelpfulLlorean: up-to, or down-from, no smaller than 2x2
00:20:47LloreanNo upper bound?
00:20:54LloreanFor source image, I mean.
00:20:57UnhelpfulLlorean: it'll take me a minute ;)
00:21:03rasherYeah, I was wondering how the ondios got away cheaper than recorders as well
00:21:07kugelfunman: not really. I've had a look but my studies is keeping me busy these days
00:21:09Unhelpfullarger than is reasonable on these CPUs
00:21:13 Join Erlend^SE [0] (
00:21:40funmankugel: i'll try to take a look then perhaps i can find some interesting stuff
00:21:40LloreanUnhelpful: Okay, that's fine.'
00:22:03LloreanUnhelpful: Just wanted to add "Album Art Resizing" to the MajorChanges page (I think the associated work probably doesn't need to be mentioned since it's not changelog, it's more user-focused)
00:22:39Unhelpfulupscaling is target-size-limited, should be able to go to targets roughly 2^24 each dimension?
00:22:40bertrikgevaerts, how did it break USB?
00:22:53*bertrik finally makes sense of the radio in the clip
00:23:17Lloreanamiconn: It's 128B difference in size, I would assume this is significant?
00:23:18gevaertsbertrik: it doesn't handle SYS_USB_CONNECTED messages
00:23:22Unhelpfuldownscaling is source-size-limited, should go to roughly the 2^24 each dimension as well
00:24:01amiconnLlorean: The recorders and Ondios all have the exact same lcd, so I fail to see why this happens if everything is ifdefed properly
00:24:08Unhelpfulamiconn: the bmp loader is rewritten. i'm still looking for more ways to get it down without forking the loader outright. as to the difference between targets, i'm clueless.
00:25:02 Quit BigBambi ("Please insert girder")
00:25:16 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
00:25:18 Join gromit` [0] (
00:27:20_Auron_yay got my sansa
00:27:33 Part akur
00:28:40 Quit mofux (Read error: 110 (Connection timed out))
00:28:49Unhelpfuli can do archos and ondios builds later and go through the preprocessed source - bmp.c should be the same on both, there's a mistake if it's not
00:31:03_Auron_it's a v2
00:31:15rasherIs this note enough, or does anything else need changing on AlbumArt: ?
00:31:28ameyer_Auron_: which model?
00:31:45Hostile_Auron_, sansa v2 support has been progressing very quickly, just hold on and before you know it it will be supported
00:31:47funman_Auron_: read the "target status" wiki page linked from the front page
00:31:47_Auron_sansa e260, got it from for $30 onblack friday
00:32:14_Auron_it's a cool device anyways, I can wait :P
00:32:39ameyerif the e200 series is anything like the clip, it's theoretically possible to make a rockbox build that's almost usable
00:33:00kugelrasher: the aligning stuff is obsolete too now (iiuc). Also, I'd change the "no resize part" into "Resize[..]Rockbox does resize now"
00:33:08Hostilecan't wait to get rockbox on my clip =)
00:33:27funmanHostile: I read that you have some C skills
00:33:35Lloreanrasher, kugel: Alignment still matters I think
00:33:37rasherkugel: The aligning is still used I think, since the loaded cover's aspect might not match the target size
00:33:44LloreanSince it maintains aspect ratio... yeah.
00:33:50funmanwe still don't know how to access internal storage for devices with >2GB
00:33:52rasherAnd yeah, I've found some other stuff
00:34:22 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:34:45funmansince we have 1GB / 2GB / 4GB Clips perhaps we can examinate how they differ and find how to access the whole data
00:34:50kugelgoing by the diffs, there's much read (i.e. removed) in the aligning sections
00:35:08kugelbut yea, there's still some, so sorry if that was wrong
00:35:31_Auron_what's keeping the > 1 OFW from being rockboxed?
00:35:38_Auron_just software security
00:35:43***Saving seen data "./dancer.seen"
00:35:48ameyercompletely different hardware
00:36:16gevaertsNot completely. There are rumours that the plastic casing is the same
00:36:45ameyerwent from a PP SoC (ARMv4?)to an AMS SoC (ARMv9?), mainly
00:37:06kugelameyer: it's still armv4 afaik
00:37:11funmanAMS is ARMv4 , arm9tdmi
00:37:44Hostilefunman, remember I said my C skills extended to I made a blackjack CMD line program in windows
00:37:47Hostilehardly skills
00:38:03 Quit herrwaldo ("Konversation terminated!")
00:38:06 Quit bertrik ("Leaving")
00:38:23 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
00:38:28HostileI'm afraid the only way I could contribute was by testing things on my 4GB clip for you guys =\
00:38:37funmanoh ok, but still you have some
00:38:43kugelso, what's happen to the resizing in the pluginlib?
00:39:14funmanHostile: is it a clip v1 or v2 ? (first digit of firmware version)
00:39:18Nico_Pwow, resize on load :)
00:39:19Hostileits a v2
00:39:23Nico_Pcongrats to Unhelpful
00:39:27Unhelpfulamiconn: the yellows are all targets with a grey/mono remote + grey mainscreen - probably because the initialization and use of dest vs rdest are both branched on remote, the compiler maybe imagines it might change, somehow?
00:39:51funmanHostile: be careful! we have no code running on clipv2 yet, we first need to find at least one button to enable dual boot
00:40:17ameyer4 GB can be clipv1
00:40:34ameyerthen again, 1 GB and 2 GB can also be clipv2, apparently
00:40:36Hostileah, I never even thought about it, I knew there were 2 hardware revisions. It figures you only have code running on v1
00:40:46funmanameyer: i was only aware for 4GB and 8 GB
00:40:46amiconnUnhelpful: It's in fact all targets with vertically packed 2bpp main. Dunno which of those conditions causes it, as I didn't study the code yet
00:40:50Hostilewhats the difference between the revisions?
00:41:06funmanHostile: no idea
00:41:18ameyerflash access has to be different
00:41:30Unhelpfuli can hide it easily enough, but i want to make sure that it *should* be hidden first.
00:41:31ameyerthere may be more RAM on the V2
00:41:42ameyerbeyond that, *shrug*
00:41:52Hostilehmm, yeah. I wish I had a way of helping you guys
00:42:14funmanHostile: perhaps you can help figure (by opening it and taking pics for example) ;) sorry I have to go now
00:42:19 Quit funman ("leaving")
00:42:23Unhelpfuldefinitely safe on H100
00:43:15Nico_PUnhelpful: how come there is added code in bmp.h?
00:44:37Unhelpfulget_totalsize and friends are now used in resize.c. there are a few ways i could change that, probably, if it's a problem.
00:45:13Hostilehaha. yeah I'm definitely not tearing apart my clip. I actually use it as an mp3 player. It's so small, way too easy to kill
00:45:37gevaertsYou use it as an mp3 player?
00:45:43*gevaerts can't imagine such a thing
00:45:50Unhelpfulbmp loader has already decided on total size, when it picked resize dimensions... that could be passed to the scaler, instead of having the scaler calc it again. small binsize savings, too, probably.
00:46:52Hostilegevaerts, right, well saying as I didn't get it as just a toy. People tear apart engines on their classic cars and rebuild them, but they wouldn't do such a thing on their main vehicle they drive to work everyday. Same kinda thing
00:47:11 Join HBK [0] (
00:48:17_Auron_what clockspeed is the cpu in the v2?
00:48:58kugel_Auron_: Have you read the wiki pages? They contain much information you ask for
00:50:39_Auron_ah okay, nm
00:52:34Unhelpfulamiconn: those are all targets that branch inside scale_nearest on whether the bitmap is for the remote or not. there's a branch inside the outermost loop to set up the destination pointer for the current row, which is of a different type on remote than on main display, and one in the innermost loop to write pixels the correct way based on the value of remote
00:52:45 Join LambdaCalculus37 [0] (i=41585845@rockbox/staff/LambdaCalculus37)
00:52:55Unhelpfuli'd think gcc would catch that one, but it misses it even if i const remote :/
00:53:18LambdaCalculus37Unhelpful: Welcome to the crowd. :)
00:53:42kugelrasher: re sorting tags: I haven't much time lately so I just put up a sync. And yea, I'd also be interested in using the sorting tags only for "inner sortings", but it looked like a mess to implement
00:53:55Nico_PUnhelpful: it might be a good idea to move the code into a c file. it depends on what the others will say, really.
00:54:09kugelI might just have not looked hard enough though
00:55:24 Quit tyfoo2 ("Carpe diem")
00:55:39 Quit bluebrother ("leaving")
00:55:41kugelI think I'll try to work on it. It'd be a proper solution to solve the "The" sorting, and also language-independent
00:56:21amiconnThat fix looks, umm, nasty....
00:57:23_Auron_there does not appear to be any note of the processor speed on the e200 pages, for either v1/v2. only place I've ever seen where the clockspeed was listed was on wikipedia
00:57:33_Auron_nor on the links on those pages either
00:57:38_Auron_but I was just wondering if someone knew
00:57:55 Join itcheg [0] (i=62db4767@gateway/web/ajax/
00:58:03Unhelpfulamiconn: hence my comment in the other channel.
00:58:48Unhelpfuli really can't imagine any other reason for gcc to behave this way, other than that pattern being a way to tell it you know better than it does. it doesn't seem to change the code at all.
00:59:42ameyer_Auron_: v1 is dual-core 80 MHz ARM7TDMI, I think v2 is single core 248 MHz ARM9TDMI
00:59:54ameyersomebody please correct me if I'm wrong
01:00:09rasherLooks like we have fluctuating builds now
01:00:25UnhelpfulDeviceChart says 100MHz, which i'd assume is for v1.
01:00:45ameyerrockbox only uses 80MHz, for some reason
01:00:58 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
01:00:59 Quit Nico_P (Read error: 104 (Connection reset by peer))
01:01:01 Nick pixelma2 is now known as pixelma (n=pixelma2@rockbox/staff/pixelma)
01:01:06ameyerbut PP says 100MHz, I think
01:01:11_Auron_probably battery life and stability concerns
01:01:20_Auron_unless it's just always underclocked or something
01:01:27LloreanSome PP chips are 80mhz, some are 100mhz.
01:01:42LloreanRight now we just use them all at 80 when boosted.'
01:01:56 Quit CaptainKwel (" ajax IRC Client")
01:02:01 Quit neddy ("Leaving.")
01:02:24ameyernot sure what the v2s can be clocked at, but my clip does 248 MHz boosted
01:02:31gevaertsActually, did anyone ever test if the shorter boost times from using 100MHz would give longer runtimes?
01:02:47amiconnameyer: Only PP5022 is known to be specced 100MHz, and the only targets having that are the Mini G2 and the Sansa c200 v1.
01:02:50 Quit DerDome ("Leaving.")
01:03:37ameyeris ARM9TDMI faster clock-for-clock than ARM7TDMI?
01:04:27gevaertsamiconn: according to the product briefs, PP5024 as well
01:04:47amiconnSansa e200 (PP5024) is most probably identical, but not confirmed afaik. PP5020 (ipod 4th Gen, Mini G1, iriver H10) is specced 80MHz, PP5021 (ipod Nano, Video) is a PP5022 software wise, but its clock specs aren't know
01:05:19amiconnThe assumption is that it is a PP5022 that isn't guaranteed to be reliable at 100MHz
01:05:55amiconnipod 1st..3rd gen are PP5002, which is specced 90MHz. So these are the only ipods which are known to be underclocked (a bit)
01:06:28ameyera google search finds which suggests PP5024 can do 100 MHz
01:06:45gevaertsAlso linked from the wiki
01:07:13_Auron_I don't suppose there's any custom firmware for zunes
01:07:40amiconnIt may also very well be the case that even PP5022/PP5024 can't be run reliably at 100MHz because of heat issues in those small cases
01:08:14gevaertsGood point
01:08:21ameyerit only matters for .APE playback, anyway
01:08:39ameyerother than potentially a little bit of battery life
01:08:44gevaertsAlso midi I guess
01:08:53gevaertsAnd heavy DSP use
01:08:55amiconnAnd video on the ipod video
01:12:05ameyervorbis only uses 1 core on dual-core targets, right?
01:13:26amiconnMost codecs use only 1 core
01:14:20amiconnThe only exceptions so far are mpa and spc, and as a non-codec, mpegplayer
01:14:45pixelmamidi playback got a lot better on PP recently, I was impressed when trying again last week after some time
01:14:47amiconnFurthermore the greylib uses the 2nd core
01:15:00ameyerwould it even be feasable to use dual-core on vorbis?
01:16:07amiconnIt might help a whole number of codecs, but it's non-trivial work
01:16:21*amiconn will try to mape the ape codec use dual core
01:19:01 Part toffe82
01:26:25_Auron_jeeze, just added some music to my microSDHC, and it's taking a long time to refresh the database
01:28:45 Quit massiveH ("Leaving")
01:30:53 Quit itcheg (" ajax IRC Client")
01:31:31 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:34:34 Quit JdGordon (Remote closed the connection)
01:36:55 Quit jhulst (Read error: 113 (No route to host))
01:37:05 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
01:41:08 Join CaptainKewl [0] (
01:42:33 Quit LambdaCalculus37 (" ajax IRC Client")
01:42:59 Join RoC_MasterMind [0] (
01:43:54 Quit itcheg (" ajax IRC Client")
01:43:56RoC_MasterMindI have used rockbox on the toshiba gigabeat f40 and ipod 4g, and was surprised by the additional 20GB of space and color screen on the f40 (good value for money!). I can't find any f40's on ebay, nor X series, and was wondering if there are other popular models people get
01:46:44 Quit lasser (Read error: 110 (Connection timed out))
01:51:26 Quit XavierGr ()
01:51:56 Join fdinel [0] (
01:54:43 Quit ender` (" Never trust a computer you can't throw out a window. -- Steve Wozniak")
01:54:43 Quit faemir (Connection timed out)
01:59:13 Join itcheg [0] (i=62db4767@gateway/web/ajax/
02:02:56saratogadual core vorbis is on my todo list, but first more single core optimizations would make sense
02:03:24saratogait may not even be neccesary, I think <30 MHz decode on PP would be fairly easy with only one core, so unless we drop the clock speed it shouldn't matter
02:08:50 Join neddy [0] (n=john@nat/sun/x-4b545bed0fc3442d)
02:09:02rasherI wonder what the point of comments in WPS files eating newlines is
02:09:44rasherIt produces some rather bizarre behaviour if you place a comment after WPS code
02:09:50 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
02:09:51pixelmahumm, I used it to break a very long line
02:10:13pixelmanot that it's needed
02:10:26 Quit saratoga ("CGI:IRC (EOF)")
02:10:33rasher%it# This is the title
02:10:37rasher%ia# This is the artist
02:10:42rasherWill produce "titleartist"
02:11:13pixelmaI usually put the comments above and don't comment... I know
02:11:39pixelmadon't comment "obvious" things I meant to say
02:11:40rasherQuite unexpected if you ask me. I shall have to ask Nico_P - he's responsible for this behaviour (r13140)
02:12:05JdGordonthe only sensible option is to only eat the \n if the whole line is a comment.... but that might not be possible
02:12:21 Quit perrikwp (" ajax IRC Client")
02:12:26pixelmain case that's changed, there's one SVN wps that's using ot as I said to break a long line
02:12:53rasherImagine my surprise when my carefl
02:12:57JdGordonmaybe we should add line breaker tag then?
02:13:00rashercareful commenting completely broke my wps
02:13:40JdGordonactually, a line breaker tag should be really easy to add
02:13:52rasherPlease let it be \
02:13:59JdGordonof course
02:14:20rasherAnd remember to add a %\ tag to display a literal \
02:14:29rasherIf you bother, that is
02:14:32pixelmathen it needs another special treatment again in case someone w...
02:14:37rasherDon't know if it's worth it
02:15:20pixelmamaybe really ask about the reasoning first in case there is something we couldn't think of
02:15:35JdGordonfixing the comment tag to only eat the \n for new lines is proabbly just as easy
02:17:01rasherTwo different issues, really
02:17:20rasher1) Make #'s behaviour less surprising 2) Add a line-breaker char
02:18:35 Join perrikwp [0] (i=98214de9@gateway/web/ajax/
02:19:09 Quit pentahoax (Remote closed the connection)
02:25:34LloreanWouldn't it be bad for comments *not* to eat newlines?
02:25:48LloreanEvery comment line would then make the WPS one line further down the screen if there were comments on their own lines.
02:26:04LloreanUnless there's a way to "close" comments.
02:27:51kugelrasher: I think I found a way
02:28:16kugel(re sort tags). I'll have a closer look tomorrow
02:29:29 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111319]")
02:29:30rasherLlorean: Full-line comments, yes. But if you put a comment after some wps code, you expect the newline to have effect
02:30:28rasherOr the next line of wps code will be put after that one, with no linebreak, which may or may not break things horribly
02:32:31Unhelpfulis something funny in the binsize logging? i note that my one-liner in the 2-bit scaler changed binsize on targets that don't compile it...
02:33:30rasherIt looks like you've introduced some code that compiles to different sizes on different build servers
02:35:44***Saving seen data "./dancer.seen"
02:37:24JdGordonUnhelpful: if its a really small diff it can be ignored (probably different comps geting the job, or some randomness in gcc..)
02:39:16 Join MethoS-- [0] (
02:40:48 Join __lifeless [0] (n=lifeless@
02:41:04 Quit _lifeless (Remote closed the connection)
02:43:59 Join massiveH [0] (
02:49:45 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
02:51:02 Quit MethoS- (Read error: 113 (No route to host))
02:57:31 Quit rasher (Read error: 54 (Connection reset by peer))
03:00:24 Join rasher [50] (n=rasher@rockbox/developer/rasher)
03:06:29 Nick fxb is now known as fxb__ (
03:08:29 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
03:10:26 Quit perrikwp (" ajax IRC Client")
03:21:01 Quit itcheg (" ajax IRC Client")
03:21:43 Join itcheg [0] (i=62db4767@gateway/web/ajax/
03:30:11 Quit amiconn (Nick collision from services.)
03:30:12 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
03:32:10 Join leachim6 [0] (
03:32:36leachim6I have all of my album art resized and ready to go on my pc in my music library
03:32:39leachim6the right size and format and everything
03:32:46leachim6but how do I transfer it to my ipod easily
03:32:53leachim6I want to put it all in my .rockbox dir so it's portable
03:33:05leachim6is there a script for that ?
03:35:23krazykit`if there is, it'd be on the album art wiki page. if not, write your own
03:38:05leachim6...I dunno how
03:38:41rasherThese kinds of scripts tend to be tied well into people's folder structure, and people usually stop working on it once it works for them, so I don't think there are any general purpose scripts for it
03:39:15rasherIt's not terribly on topic here, really, #rockbox-community is probably better suited
03:39:34leachim6what is this ... the dev channel ?
03:40:35ameyerAnd resizing should no longer be strictly necessary. Conversion to .bmp is still necessary, though.
03:42:31leachim6when will jpeg be supported ?
03:42:34leachim6is that even on the list
03:43:12ameyerwhen someone gets around to implementing it
03:43:30leachim6which could be next week or next year...
03:45:47ameyerCould be (although probably not next week because the freeze for 3.1 is coming soon. Could also be "never"). According to the wiki "JPEG album art is currently not supported. Supporting JPEG requires adding a JPEG decoder to the Rockbox core. This is wanted, but won't be done until someone actually decides to. The decoder already exists as a plugin. PNG support seems more unlikely as there currently is no PNG decoder in Rock
03:48:01 Quit pixelma (Connection timed out)
03:50:12 Join EspeonEefi [0] (i=eefi@STRATTON-SIX-FOURTY-SIX.MIT.EDU)
03:56:48soapcould the support for cover.<horiz size>x<verti size>.bmp be removed now to save a little bin space?
03:57:28rashersoap: I think people should be able to pre-scale if they so chose, and have a "full size" as cover.bmp
03:58:22*ameyer wonders what sort of battery/cpu hit scaling causes
03:58:38ameyerI'd think it wouldn't exactly be computationally cheap
03:58:58rasherAlso, it'd ruin weirdo wpses like mine ( - which might be the only one of its kind, I'll admit)
04:02:02soapdoes the scaler ignore aspect ratio?
04:02:19soapI mean, would the scaler butcher your WPSs album art?
04:02:26krazykit`soap, i believe Unhelpful said it preserves aspect ratio
04:02:28leachim6will rockbox handle spaces in filenames in .rockbox/albumart ?
04:02:43leachim6like /rockbox/albumart/Band-name of album.bmp
04:03:04rashersoap: It preserves aspect ratio, so a square album art becomes square, even when my %C tag specifies a rectangular image.
04:03:42rashersoap: But if <horiz size>x<verti size> support was removed, I couldn't mix the two, and my covers would only work in either this WPS or "all the others"
04:04:36soapahh - yes, I didn't consider the "two bitmaps for two situations" case.
04:06:08soapthough could the search for <horiz size>x<vert size> be replaced with a search for <WPS specified prefix>cover.bmp w/o adding a bin hit? It would solve your problem and open up new options.
04:07:40Unhelpfulameyer: the greyscale one is cheap. the color one is definitely not.
04:07:43Lloreansoap: So have "wide" somewhere in the WPS tag and it searches for wide-(normal AA string).bmp?
04:07:51rasherI imagine it'd reduce the binsize to replace one with the other, and yeah, it might be interesting
04:08:36soap"wide" or "dark" or "XXX" or what have you.
04:09:00rasherI've been toying with the idea of creating a script+wps combo that added ipod-style reflected album-art
04:09:18rasherCould be done without this feature, through a specific dimension of course
04:09:26soapI don't think I've seen such pretty pictures.
04:10:00LloreanI like the idea of dropping size suffixes entirely to be replaced with a custom prefix string.
04:10:20 Join AndyIL [0] (i=AndyI@
04:10:24LloreanEspecially since you can then use it conditionally again: If a prefix version exists use it, else shift the display around and use the default.
04:10:31LloreanSize checks don't make much sense any more, since scaling is in.
04:12:22soapI understand the corner-cases, but I do think reclaiming the binsize lost to size checks would be the "ethical" thing to do.
04:12:57rasherI'm not sure there's any corner-case left, if you replace it by a prefix thing
04:13:00soapMake any "feature shift" stand on its own vs. slipping in through a back-door.
04:13:17soap(I'm arguing against my proposal now)
04:13:19LloreanI think the prefix thing preserves the existing featureset nitcely
04:13:42LloreanAnd expands it a bit
04:14:30 Quit CaptainKewl (Remote closed the connection)
04:14:32 Join blkhawk- [0] (
04:14:51 Join CaptainKewl [0] (
04:15:17soapTo summarize my counter argument: Would it not be better to trim the now-fat of size-specific art and IF WPS art prefix checking is a good idea let it go in on its own merits through the front door instead of slipping it in through the back door by replacing (unneeded) code with (arguably) a new feature?
04:15:47 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
04:16:11Lloreansoap: Well, I think dropping size checking would be a lost feature right now (conditional display based on the availability of different sized arts)
04:16:13 Quit AndyIL ()
04:16:18LloreanI'm not sure if scaling already nuked that though
04:16:39LloreanSo your feature would in fact be a more abstracted way of doing the same thing. It would be "replacing one feature with a better version of the same feature"
04:16:55Unhelpfuli'd have to check, that part was all akio
04:16:56 Join AndyIL [0] (i=AndyI@
04:17:22soapI agree the concept is a more abstracted way of doing something which already existed - but rasher's cool WPS be damned, it is a corner case.
04:17:54soapThe intent of the feature was to allow users to switch between WPSs w/o having to load new album art on their player.
04:17:57ameyerUnhelpful: I see. Not that it matters, as my main music player is grayscale and I don't use album art anyway
04:18:16Lloreansoap: Possibly, but it's a "feature" we would be removing. I'm not really comfortable with removing things you could do previously, just in general.
04:18:51 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
04:19:00soapThat case has been solved by Unhelpful's militantly cool feature - I think an argument can be made that a "hack" which was not the original intent of now-bloat code is hardly justification for keeping the bloat.
04:19:19rasherI think it should be done in two steps (that way we also get to see the binsize impact of both), but I do think both steps should be done (by Someone)
04:19:47LloreanI think that if the binsize can be reduced *without* losing any existing functionality, it's a win.
04:20:00Unhelpfulameyer: actually, it might be worth you doing a battery_bench run before and after my checkin, since greyscale bitmaps are always "scaled"... but the hit is probably not measurable without AA, since everything else loads once
04:20:14LloreanIf the binsize would be increased more than a tiny, tiny amount by swapping size checks for prefixes, then some discussion should certainly happen.
04:20:25soapMind you - this is devil's advocacy from someone who thinks next-album's art is a good idea ;)
04:20:42Unhelpfulsoap: i'm on board for next-album. ;)
04:21:22soapI am aware of no DAP on the market with such a feature.
04:21:35Unhelpfulhowever, people seem to expect me to do jpeg next, and i had kind of thought i should *before* the commit
04:22:35 Quit AndyI (Read error: 110 (Connection timed out))
04:23:15LloreanNext-cover display should be a pretty minor addition of a tag, anyway, right?
04:24:06*ameyer wonders what the scaler does when the bitmap is the same size as the output
04:24:21ameyerlike, if the scaler is told to resize a 100x100 bitmap to 100x100
04:25:12*ameyer dives into the code to try to figure it out
04:25:23Unhelpfulameyer: the greyscale one? it just outputs each pixel as it would without scaling.
04:25:30soapLlorean, what if the next track's art is not yet in the buffer?
04:25:50Unhelpfulthe color has a special case for direct-output, since those scalers are full of maths
04:26:05soapYou would have to give audio data the boot in deference to album art if such a tag were to be implemented.
04:26:36Lloreansoap: If the next track's not in the buffer, we don't display it's other ID3 info either.
04:26:49LloreanI don't see any harm in not showing it until the next round of rebuffering starts.
04:27:11Unhelpfulindeed - if you queue next in the middle of a large playlist that's already buffered, you'll probably have no next-song data
04:28:23soapLlorean, I know I only own "big memory" targets, but I also tend to listen to long tracks. You're telling me that if I'm listening to track 3, and it itself is not fully in the buffer (because it is at the end of the buffer) the WPS won't show track 4's info?
04:29:57ameyerWouldn't the buffer load the metadata and then fill the rest of the buffer with compressed audio and
04:30:10ameyererm, add "PCM" to the end of that
04:30:31 Quit blkhawk (Read error: 110 (Connection timed out))
04:30:53Lloreansoap: At least it did this in the past. It *may* be possible that this got changed sometime after MoB, but that used to be the standard behavioru.
04:31:30 Nick blkhawk- is now known as blkhawk (
04:35:45***Saving seen data "./dancer.seen"
04:45:36 Quit MethoS-- (Remote closed the connection)
04:46:46 Quit ChadFM (" HydraIRC -> <- Now with extra fish!")
04:49:03 Join pabs_ [0] (
04:51:40 Quit neddy (Remote closed the connection)
04:52:30JdGordondo bmps get scaled on load or on display now?
04:53:06LloreanOn load.
04:53:46JdGordonso next track AA might not be so good... unless it can be resized again... 200x200 this track with 200x200 next track wont work
04:54:19LloreanIt'll work just fine as long as you're not trying to display them on the same screen. :-P
04:59:06 Join neddy [0] (n=john@nat/sun/x-45d7f1c219cad5d1)
05:01:02 Quit pabs (Read error: 110 (Connection timed out))
05:07:12 Quit itcheg (" ajax IRC Client")
05:10:41 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:14:23 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
05:15:32 Join sbhsu_ [0] (
05:16:36 Quit Seedy (Read error: 60 (Operation timed out))
05:23:33UnhelpfulLlorean: just looked, the size stuff is still there
05:27:55 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
05:31:15 Quit sbhsu (Read error: 113 (No route to host))
05:35:00Unhelpfulthere's a lot of stuff from bmp.c that got moved into bmp.h and declared inline so that resize.c could also use it. i think it needs to either go back, or into a new file, and be non-static
05:35:24Unhelpfulget_totalsize will be needed by any other scaler frontend, and brightness might be useful as well
05:35:51Unhelpfulalthough brightness should probably retain an inline version
05:36:56Unhelpfuldither_mat and vi_pat should probably just be replaced by directly indexing the arrays they return values from
05:37:58Unhelpfuland both of those may belong in a general "internal bitmap handling" file, that should probably be a different one from the "windows bitmap loading" one
05:38:04 Quit Horscht ("We don't make mistakes, we just have happy little accidents")
05:39:03 Quit sbhsu_ (Read error: 110 (Connection timed out))
05:39:57 Join tofor [0] (
05:50:34 Quit massiveH ("Leaving")
05:58:02 Join hobbs [0] (n=nandrew@p3m/member/hobbs)
06:00:02hobbsMy player died, and I doubt I'll be able to fix it; what are opinions on the "best" compatible player that can be fairly easily acquired?
06:02:57advcomp2019hobbs, what player did you have?
06:02:59Unhelpful it looks like just stripping out the size-based search saves 128B on coldfire, 192B on arm. a more thorough removal would get rid of the size_string parameter entirely, but that would touch a lot of stuff - plugin api, pictureflow plugin, probably sliding_puzzle...
06:03:25advcomp2019hobbs, plus what is wrong
06:03:43hobbsadvcomp2019: Gigabeat S30 (yeah, I know it's not fully supported, but it worked very well other than having to boot to stock to charge)
06:04:40hobbsit doesn't do anything −− no LCD or hard drive activity when I press the power button or plug in the charger.
06:05:33hobbsI'll open it up and have a look if I can find a screwdriver that will work, but nothing is simple with gadgets nowadays so it's probably just dead.
06:06:38Unhelpfulwell, it's not any of the fixable S problems i know :/
06:08:07hobbsIf I have to buy, I'm considering an F40, since it's a pretty similar device and it's a fully supported target. I see a good number of them on eBay.
06:08:47advcomp2019hobbs, o ok.. i think the easiest is the sansa or one of the other gigabeats from what i am seeing
06:13:26hobbsout of those I think I'd definitely take the gigabeat. Installation procedure looks easy too, simpler than the S :)
06:14:11 Quit AndyIL ()
06:19:05 Join AndyI [0] (i=AndyI@
06:19:34 Quit courtc (Read error: 113 (No route to host))
06:22:27hobbswell thanks guys. hasta.
06:23:02 Part hobbs ("Leaving")
06:27:03 Quit RoC_MasterMind ("Leaving")
06:27:15 Quit __lifeless (Remote closed the connection)
06:27:34 Join __lifeless [0] (n=lifeless@
06:28:59 Quit J-23 (Remote closed the connection)
06:29:23 Join J-23 [0] (
06:31:41 Quit tofor ("Leaving")
06:35:48***Saving seen data "./dancer.seen"
06:43:19 Quit CaptainKewl (Read error: 110 (Connection timed out))
06:43:44 Quit jfc (Read error: 54 (Connection reset by peer))
06:44:07 Join jfc [0] (
06:45:07 Join Darksair [0] (n=user@
07:08:21 Quit ameyer (Remote closed the connection)
07:31:30 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
07:54:33 Quit GodEater (" ajax IRC Client")
07:54:40 Join GodEater [0] (i=c2cbc962@gateway/web/ajax/
08:00:13 Join n1s [0] (n=nils@rockbox/developer/n1s)
08:03:10 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
08:09:17 Quit BHSPitMonkey (Read error: 148 (No route to host))
08:19:51*n1s spots 2 space indentation in bmp.c
08:21:47Unhelpfuli think i'm looking at the same. switch statement?
08:24:26 Join kugel [0] (n=chatzill@unaffiliated/kugel)
08:24:30n1soh, and congrats on committing your patc :)
08:24:34Unhelpfulpreferred is... cases 4 in from switch, blocks four in from case?
08:25:16Unhelpfulif it's worth a commit to clean it up, i can, otherwise i can wait until i have something else in there, comments or such :)
08:25:27amiconnFor cases I've seen several variants in rockbox
08:26:08n1syes, we are a bit inconsistent but any indent should be 4 spaces at least
08:26:18amiconn(1) cases by 4, statements by another 4. (2) cases at the same level as switch(), statements by 4. (3) cases by 2, statements by 4
08:26:47*n1s prefers (1)
08:26:50Unhelpfulif your #3 is statements by 4 from the switch, that's what's there
08:27:00*amiconn prefers (3)
08:27:54amiconnYes, that's what I mean
08:29:46amiconnImho (1) is too much indentation, especially if there's nesting, and (2) is somewhat confusing
08:30:35amiconnI know there is that general indent-by-4 rule
08:33:14Unhelpfulheh, got 16 bytes back on ondio
08:35:52***Saving seen data "./dancer.seen"
08:36:06Unhelpfulstill not near what the bmp.c refactor cost, though. i'm trying to think of ways to put the conversions done inside read_part_line back inside read_bmp_fd, but i can't think of any that don't involve forking the bmp reader, putting the code in a macro, or *cringe* including the code from another file
08:36:48 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:37:08 Quit AndyI ()
08:37:41 Quit BigBambi (Read error: 113 (No route to host))
08:39:55 Join Rob2222 [0] (
08:42:20 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:43:15 Join AndyI [0] (i=AndyI@
08:51:31 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
08:51:50amiconnUnhelpful: Did you find out why there's that binsize difference between recorders and Ondios?
08:52:40 Join pixelma [0] (n=pixelma2@rockbox/staff/pixelma)
08:52:55Unhelpfulstill no idea. and it seems to be consistent across the next couple of commits, unlike the other fluctuations
08:53:01 Nick Bagderr is now known as B4gder (n=daniel@gateway/web/cgi-irc/
08:54:15Unhelpfulbut i'm testing tweaks on ondio now, so let me do a recorder as well, and see
08:56:32 Join ender` [0] (
08:56:52 Join funman [0] (
09:02:12amiconnUnhelpful: I'm kinda worried because this size difference shows that some code changed in an inconsistent, unexpected way. Either there's a wrong #ifdef, or there's a weird alignment, but I can't spot either...
09:02:39Unhelpfulamiconn: i can't either. in particular, the compiled bmp.o for each is the *same* :/
09:03:36amiconnHmm, maybe there was such a problem *before* your commit?
09:03:44 Join bertrik [0] (
09:04:32 Quit gevaerts (Nick collision from services.)
09:04:37Unhelpfulit would still need to have been exposed *by* my commit, if it was somewhere else... unless you mean that the binsize diff was in the old bmp.c?
09:04:44 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
09:05:11pixelmamaybe the old bmp.c was smaller
09:05:21pixelmaon Recorder
09:05:48amiconnUnhelpful: I mean the old one, and that your commit fixed a hidden bug.
09:05:54Unhelpfulchecking right now
09:06:01 Quit Hostile ()
09:06:57 Join petur [50] (n=petur@rockbox/developer/petur)
09:07:24Unhelpfulit's possible that something elsewhere changed, but the changes in the rest of the tree don't depend much on LCD specifics, other than color vs not
09:07:52 Quit Rob2223 (Read error: 110 (Connection timed out))
09:08:24B4gderUnhelpful: congrats on your commits and fine work on the scaling code!
09:10:53Unhelpfulamiconn: nope, old bmp.o builds the same on both targets, as well :/
09:11:45UnhelpfulB4gder: thanks. now we just need to figure out why the delta was different on ondio than recorder/player
09:12:29B4gderI noticed, curious indeed
09:14:07ZagorI'm also curious about the unitialized warning. can anyone figure out why it was output?
09:14:33Zagorah, I got it
09:14:37pixelmaUnhelpful: didn't see something in the diff but asking anyway - is there something in the code that only checks for presence of a remote at all? I am asking because that's a difference between Ondio and Recorders, the latter can use a non-lcd remote. Just an idea...
09:15:57Zagorto nitpick, I think a pointer should be initialized to NULL rather than itself :)
09:16:05Unhelpfulnot a bad idea... i think everything i touched was sensitive only to remote LCDs
09:17:03UnhelpfulZagor: i can't speak for the pointer, but when i tried that trick with rgbval/rgbinc in scale_h_linear, gcc generated more code to initialize them
09:17:45 Quit EspeonEefi ("さよなら")
09:17:53Unhelpfulx=x seems to be code for either "i know it will be initialized" or "i really want stack garbage in my variable, please"
09:18:18ZagorI guess it's not any worse than (void)x;
09:18:36Zagorwhich we use in lots of places
09:18:57Unhelpfulit certainly looks nastier... and has less obvious reasons
09:19:34B4gderI'd say setting it to zero/NULL is nicer
09:19:46Unhelpfuleven if somebody sees it and know i'm hiding an unitialized data warning, they have to read the code to find out if the warning was legit or not.
09:20:10B4gderthe comment next to the code will explain surely? ;-)
09:20:12funmana comment could explicit the reason for this nasty looking initialization
09:20:49Unhelpfulwhat, comments? how did i let any of *those* get in?
09:21:23 Join Rob2223 [0] (
09:21:49 Quit Rob2222 (Read error: 104 (Connection reset by peer))
09:22:36*B4gder has done 32% comments in his rockbox C code!
09:23:04 Quit funman ("leaving")
09:23:06B4gderaccording to ohloh
09:23:18 Join funman [0] (n=fun@rockbox/developer/funman)
09:24:32B4gder amiconn is at 14%, LinusN is at 17%, Zagor at 16%
09:25:07*Zagor suspects Bagder once edited a pile of file headers :)
09:25:11Zagoradded even
09:25:15B4gderit certainly looks like that
09:26:20B4gderbut I'm above 20% on projects like curl
09:26:47Unhelpfulpixelma: that was a great idea, but if that's it, it seems not to be anything obvious that i can find by looking through the patch
09:27:36Unhelpfulbmp.c is already known to compile identically, resize.c isn't compiled for mono, and the rest has no references to REMOTE defines that aren't LCD-related :/
09:28:08 Quit bertrik (Remote closed the connection)
09:28:09pixelmaas I said, didn't see anything either
09:29:49Unhelpfulit's hiding. :/ perhaps comparing symbol size deltas would be useful
09:31:34 Quit agaffney (Read error: 110 (Connection timed out))
09:34:33*LinusN is holding two oscilloscope probes in his hands and pressing F5 using a pen in his mouth
09:34:50B4gderwhat about your nose? ;-)
09:35:02B4gderah you type here with that!
09:35:03LinusNnot long enough
09:35:31LinusNhmm, maybe i can use a pen on my nose for typing...
09:35:34LinusNin my nose
09:38:06 Quit perrikwp (" ajax IRC Client")
09:38:11 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
09:51:52Unhelpfulfound it, sort of, with python
09:52:33Unhelpfulthe ondio build drops a symbol, ___ashrsi3, size 130B
09:53:33Unhelpfulbut it stays in the recorder build
09:54:07 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
09:54:14funman"start in screen: db" config on Clip makes it reboot indefinitely, I believe because db isn't built ?
10:03:13funmanI do not know settings/database/tagcache : is this right ?
10:04:00JdGordonno, that wont work
10:04:17funmanwhat is 'db' ?
10:04:26JdGordonorder is important there, db is the database
10:05:15funmanand with respect to tagcache?
10:05:37funmanso it should be conditional, but keeping the original order?
10:06:16JdGordonyes... but there is also another conditional missing so ill fix that
10:06:53funmanthanks :)
10:06:58 Join EspeonEefi [0] (i=eefi@SAFFRONCITY.MIT.EDU)
10:07:04JdGordonit wont fix any problems though....
10:07:11JdGordonunless the clip doesnt have db
10:07:30funmanit doesn't, i don't know if RAM size permits it
10:07:45JdGordonbut the option is in the menu?
10:08:07kugelfunman: isn't there some ram free with the flash buffering patch?
10:08:28funmanJdGordon: yep it is
10:08:47funmankugel: sure, (~1MB) but perhaps it's better used as audio buffer?
10:09:08 Quit Darksair (Connection timed out)
10:09:22kugelthe audio buffer isn't that important on flash based daps imho
10:09:43funmanthat's true ;)
10:10:01kugeland the database doesn't take the whole ram.
10:10:04funmananyway i have no idea what these features require
10:10:25 Quit tarbo (Read error: 110 (Connection timed out))
10:10:47kugelhm you could just try to build a whatever target without HAVE_TAGCACHE and compare rockbox-info.txt to a build with HAVE_TAGCACHE
10:10:51 Quit kachna (Read error: 113 (No route to host))
10:11:45*funman bugs B4gder about the fuze firmware on his webpage not being compressed
10:12:20kugelfunman: you might download the .15 version from the sansa forums
10:12:41B4gderwhat you can compress files?!
10:12:46funmanthe version is not important
10:13:37Unhelpfulamiconn: i have your explanation. ___ashrsi3 is a libgcc function referenced by the old bmp.o, but not the new, when built on sh. ondio has no other core object files that reference it, but on recorderv2, it's used in firmware/drivers/ata.o
10:13:43JdGordoncan anyone think of a cleaner way to do this than ?
10:14:09kugelfunman: not too important no. But as far as my disassembly of the lcd driver went I noticed the .15 version seemed to have dropped the lcd_delay altogether (which I noticed to be not needed too)
10:14:29B4gderlzma is a fine compression...
10:14:30kugelit might just contain bug fixes worth to know about
10:16:55B4gderfunman: now there are compressed versions too
10:17:17 Quit jhulst (Read error: 113 (No route to host))
10:17:39B4gderkugel: do you have the link to the .15?
10:18:16funmangoogle://sansa+forum+fuze ;)
10:18:22kugelpick any, they're all the same
10:18:47kugelthe filename decides which features are active and which not
10:19:26B4gderfunny how that zip is much smaller...
10:20:14 Quit agaffney (Read error: 110 (Connection timed out))
10:20:31funmanhm.. my tool recognizes only 3 library blocks in fuze OF (drmndt_pers, otg_functio and trusted_fla)
10:21:25kugelthat tool which you gave me a while back? It used to find all (at least way more than 3) library blocks
10:22:53funmanyes, but i took an older version i had posted on the forum, perhaps i had enhanced it after that
10:23:27funmananyway i don't need these blocks now, just thought that was strange because i remembered having tested it on the fuze (perhaps only version 11)
10:25:10funmanwhat's the primecell number for the gpio controller already? (didn't think to take a copy of the as3525 datasheet with me..)
10:25:45kugelwhat do you mean by primecell number?
10:26:19funmanthe GPIO controller is a documented ARM PrimeCell module, i'm asking the reference number so I can download its datasheet on
10:26:45funmanthat's it, thanks!
10:29:34funmantagcache on the clip : +41kB binsize, +63kB RAM usage
10:30:02JdGordonarg... red rec build.. wtf?
10:30:10kugelfunman: worth it, imho
10:30:12JdGordonsoap's server again
10:30:32funmani'll just give it a try to see what is it ^^
10:30:50JdGordonkugel: get anywhere with the v2 buttons?
10:31:04kugelJdGordon: Sadly no, did you?
10:31:10funmanJdGordon: i'm trying to take a look at fuze code with the rockbox' arm disasm
10:31:14JdGordonnup, been doing other stuff
10:31:17kugellooks like funman is going to look at
10:31:50kugelfunman: objdump is slightly better than that isn't it?
10:32:25funmanobjdump doesn't show the values loaded from memory - arm_disasm does
10:32:37kugeloh, I didn't know that
10:32:50funmanarm_disasm doesn't support thumb however, but it looks like the isr and the functions it calls are ARM
10:34:20kugelfunman: are you more interested in the isrs or in the buttons? As you've noticed the majority of the buttons are without isr
10:34:33funmangreat.. data abort when accessing the first argument of utf8decode()
10:34:49funmankugel: i'm looking at the interrupt now
10:35:25funmanwhen looking for FM code i have understood a bit more the diagnosis mode however, does this mode exist on Fuze or e200v2 ?
10:35:54***Saving seen data "./dancer.seen"
10:37:19kugelfunman: yes it does. name the OF fuzet.bin and you get it. And there's one test which checks for all buttons
10:37:52funmanwhich is why I mentioned it ;)
10:38:35 Join einhirn [0] (
10:39:00 Quit JdGordon (Remote closed the connection)
10:39:35 Join lasser [0] (
10:40:26funmankugel: how is this mode called ? grep 'Diagn' gives no result
10:40:32 Join JdGordon [0] (
10:40:47funmanperhaps it's using UTF-16 encoding
10:41:24funmanOk I found test strings in otg_functio (like on the Clip)
10:49:42funmangpioa isr is "void INT_GPIOA(void) { return; }"
11:00:44 Join kachna [0] (
11:06:19kugelfunman: sounds useful!
11:06:36 Join idak [0] (
11:07:43 Join mofux [0] (
11:07:51amiconnUnhelpful: That explains a lot, thanks for checking :)
11:08:51Unhelpfulno problem. i've still no clue what that libgcc symbol does, probably implements some operation that's not provide by the hardware, or supports some gcc peculiarity, or such
11:11:27amiconnUnhelpful: Fyi, __ashrsi3 is the arithmetic-shift-right-by-n subroutine. SH1 has no single-instruction shift-by-n
11:12:00Unhelpful...i removed a shift-right-by-n from the bmp loader? there are several in my version...
11:12:21idakUnhelpful: congratulations and great thanks on your works!
11:12:29amiconn*arithmetic*. Maybe you're using those on unsigned values only?
11:12:40Unhelpfulamiconn: perhaps!
11:13:15amiconnThere's __lshrsi3 for that. Btw, on SH1 it's better to use unsigneds for that kind of stuff
11:13:39Unhelpfulidak: thanks for your work as well, it's a good chunk of what i committed - all of the album art rework, and much of the loader.
11:16:56 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
11:18:59 Quit kugel (Remote closed the connection)
11:34:30 Join hamhock- [0] (
11:38:02 Join kugel [0] (n=chatzill@unaffiliated/kugel)
11:38:29funmanhey kugel I start to understand how to read from gpiob (there is a lot of functions so it's just the beginning)
11:39:21kugelfunman: cool
11:39:27funmana change on B0 significates 'stuff happened' (because the isr only sets bit 0 of GPIOB_IC (at 0x41c)
11:39:39B4gderI'm disabling soap's build server for now
11:40:19funmanand I'm seeing: B1 = 1, and then B5 or B0 (depending of some value in memory) 1
11:41:12kugelI saw that too I think, but I couldn't interpret it
11:47:02funmanhm what else did you see?
11:49:15kugelsimilar stuff for gpioa
11:49:43funmancalled from the gpiob isr?
11:50:00kugeli haven't looked at the isrs
11:50:34funmanoh ok
11:50:35kugelalso, somehow my annotations are gone (wtf?), so I can't exactly tell what I saw
11:52:56 Join BdN3504 [0] (n=55b2074c@gateway/web/cgi-irc/
11:58:20BdN3504Hey, can someone help me install the utf-8 support for latex so i can compile the manual using the vmware build envrioment? I don't quite understand what i have to do here: i did the apt-get install latex-ucs command in the vm, but i don't understand the installation instructions on this page:
11:59:52funmanBdN3504: apt-get install latex-ucs wasn't enough ? apt-get should not make you need to do manual installs
12:00:42 Quit kugel (Read error: 131 (Connection reset by peer))
12:02:24BdN3504when i try to "make manual" i get the following error:
12:02:43BdN3504! LaTeX Error: File `utf8.def' not found
12:02:44 Join moos [0] (
12:03:31 Join kugel [0] (n=chatzill@unaffiliated/kugel)
12:04:47pixelmaBdN: read your post in the forums and I don't think a seperate list of settings make sense because it can get huge and then there are two (one is in the appendix). I'd suggest adding a column for the ability to set as quickscreen item to the latter and reference that in the quickscreen chapter. Also, there is a patch in the tracker for the quickscreen manual entry which I meant to look at...
12:05:38pixelmathe table in the appendix is supposed to list all available options (things you could see/set in cfg files)
12:05:41 Quit linuxstb (Read error: 110 (Connection timed out))
12:06:46BdN3504ok, fine. i think that's pretty reasonable. we just add a column like "available to be set as quickscreen" to the list in the appendix
12:06:56 Join courtc [0] (n=court@unaffiliated/courtc)
12:07:35BdN3504do you have a link to that manual entry you mentioned?
12:09:11 Join Jabone_ [0] (
12:10:27 Join flx [0] (
12:11:13pixelmaFS #9524
12:11:42 Quit flux (Read error: 104 (Connection reset by peer))
12:14:19BdN3504thank you, i will have a look at it, too. but i'm afraid, i'll only be able to help with the manual, if i am able to learn latex, which won't happen in a shorter period of time than say about a month or two... so long
12:14:22 Quit BdN3504 ("CGI:IRC")
12:17:43pixelmahe left...
12:19:49 Quit Jabone (Read error: 110 (Connection timed out))
12:22:57 Part B4gder
12:26:18funmanhum I think what bertrik took for FM i2c on the fuze is actually button code, and FM i2c is SDA = A7 and SCL = A6 (just like the e200v2, once again)
12:30:18funmankugel: could you try FS #9611 (FM) and adapt it to build on Fuze, using the pins I mentioned (editing fmradio-i2c-as3525.c) ? (I know you can resume playback by editing the config, not sure how you can resume to FM however)
12:33:04 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
12:35:49 Join MethoS- [0] (
12:35:57***Saving seen data "./dancer.seen"
12:39:35kugelfunman: I could resume if I started playback
12:41:12funmanresume settings is stored somewhere, so that can be edited manually
12:42:32funmanbut playback resume only resume files, not FM :/
12:43:20funmanperhaps it's possible to hardcode a start menu (like debug-> FM radio
12:43:42funmanor put a panic in si4700.c where the tuner is detected
12:43:45 Join MethoS-- [0] (
12:43:46pixelmaset start screen to radio be means of editing config.cfg?
12:44:16kugelI'm just gonna try grabbing a .playlist_control from my e200 to test sound, and I could use start_screen setting to get into fm
12:44:45funmansounds good (thanks pixelma i hadn't seen the radio option)
12:48:08 Join Aurix_Lexico [0] (n=comrade@
12:50:56 Quit MethoS- (Read error: 145 (Connection timed out))
12:54:24 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
12:56:44 Join funman_ [0] (
12:58:29 Quit hamhock- ()
13:04:03 Quit funman (Read error: 60 (Operation timed out))
13:10:25 Quit perrikwp (" ajax IRC Client")
13:12:39 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
13:17:01kugelfunman_: what about the .address?
13:18:31 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
13:20:09 Quit JdGordon (Read error: 131 (Connection reset by peer))
13:21:15 Quit funman_ (Read error: 110 (Connection timed out))
13:22:15 Nick fxb__ is now known as fxb (
13:26:04 Join funman [0] (n=fun@rockbox/developer/funman)
13:28:41funmankugel: what is .address ?
13:29:16funmanoh, 0x20 ? don't change it it's specific to the FM chip (general across all rockbox targets)
13:30:11 Quit bmbl ("Woah!")
13:30:20idakUnhelpful: I found my bug in SVN. apps/recorder/bmp.h: L.174"(LCD_REMOTE_DEPTH == 2) && (LCD_REMOTE_PIXELFORMAT == VERTICAL_INTERLEAVED)" in get_totalsize() is unnecessary for iriver H300 (1bpp remote).
13:30:26kugelfunman: ^
13:30:42funmankugel: use the Clip address (Clip/Fuze/e200v2 have the same si4700 FM chip, unlike m200v4)
13:30:56kugelok, I didn't know that
13:31:25kugelfunman: can you paste a playlist_control? The one of my e200 doesn't seem to work well due to other partition layout
13:33:25funmanP:2:/MUSIC/Pink Floyd - Dark Side Of The Moon:
13:33:50kugelP:2: didn't work for me
13:34:00funmannote i have a partition table, no super flopyp
13:34:37funmanadvcomp2019: do you have an unpartitioned clip? can you paste the playlist_control?
13:35:22advcomp2019funman, you mean without rockbox?
13:35:49funmannope, i mean original disk layout (1 FAT partition without a partition table)
13:36:21advcomp2019nope then
13:36:42 Join fredddy [0] (
13:37:12funmankugel: shouldn't you have sound if you start in FM ?
13:39:45 Quit fredddy (Read error: 113 (No route to host))
13:39:46 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
13:40:10 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
13:40:31 Join fredddy [0] (
13:42:49funmanhum the manual displayed page numbers do not correspond to pdf page numbers :/
13:43:15funmanoh yes they do >< *shame*
13:43:48 Quit fredddy (Read error: 113 (No route to host))
13:44:15 Join fredddy [0] (
13:44:17kugelfunman: I wanted to test general sound first
13:44:30funmani believe pitchscreen & quickscreen should be enabled on the clip as well
13:44:55funmankugel: do you mean general sound played by FM, or specific audio files playing? ;)
13:44:59kugelfunman: sure, I've recently optimized pitchscreen for small screens
13:45:33 Join freddy_ [0] (
13:45:41kugelfunman: sound as in headphone output. And given that audio files is known to work, it's the best way to make sure general sound is working
13:48:04funmanthe '2' in P:2: means PLAYLIST_CONTROL_FILE_VERSION , nothing about partitions. Did you check you were using a correct path ?
13:48:44 Quit Aurix_Lexico ("Leaving.")
13:51:12kugelfunman: yep
13:52:30 Join TheSphinX^ [0] (
13:53:38funmanso what happens instead of resuming playback?
13:53:47kugelhm, "Nothing to resume"
13:54:01funmanisn't the file/position stored somewhere else? (nvram.bin?)
13:54:14kugelit seems playlist_control only contains the path to the music played last, not the file and/or resume point
14:00:38 Part idak
14:00:47kugelfunman: well, I copied nvram.bin, but it only went to the wps (the very default one, as in parsing cabbiev2.wps failed) but didn't play
14:02:23funmanI've no clue where the file/position are stored, just supposed nvram.bin out of the blue
14:03:29kugelfunman: nvram.bin worked, it resumed the file
14:03:47kugelbut without actually playing it
14:06:43 Quit Erlend^SE ("Client closed by user.")
14:08:22 Join tyfoo [0] (
14:10:34 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
14:10:54 Join massiveH [0] (
14:13:39 Quit kugel (Remote closed the connection)
14:14:05 Join JdGordon [0] (
14:21:36 Join kugel [0] (n=chatzill@unaffiliated/kugel)
14:21:57 Quit JdGordon (Read error: 104 (Connection reset by peer))
14:23:49 Join itcheg [0] (i=62db4767@gateway/web/ajax/
14:24:39 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
14:29:54 Quit Zagor ("Client exiting")
14:30:25 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
14:31:39 Quit tvelocity (Read error: 60 (Operation timed out))
14:34:40funmankugel: could you check if you can go to the radio screen ? (to see if the radio is detected, sound output can wait until we have buttons)
14:35:58***Saving seen data "./dancer.seen"
14:37:24 Quit MethoS-- (Remote closed the connection)
14:41:00 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
14:41:36kugelfunman: oki
14:43:38funmanI've found some code which will read 8 bits from the GPIO using a quite complex protocol
14:43:50kugelfunman: I bet I get in the radio screen
14:43:56funmanif 8 bits = 8 buttons, we still have no idea how the wheel works
14:44:06kugelI also get in it using the start screen on my e200 (which has no radio)
14:44:24funmankugel: hm, but shouldn't it write 'no radio detected' or something?
14:44:29kugelfunman: 8 is better than 1 :)
14:45:07funmanwell i've no clue what it reads exactly
14:45:45kugelfunman: It apparently doesn't. Also, I can't get in the radio screen by browsing to it (it's not in the main menu). But radio screen is compiled, since the e200 series normally have it
14:46:23 Quit itcheg (" ajax IRC Client")
14:47:49funmancan you see the FM in main menu of e200v2 ?
14:48:27kugelfunman: it doesn't find a station. I mapped the power button to start scanning, but it doesn't stop scanning
14:49:04LambdaCalculus37Is FM scanning implemented on the AMS Sansas?
14:49:21kugelno wonder then!
14:49:27funmanevery frequency is 'tuned'
14:49:32funmankugel: what about sound?
14:49:41kugelno sound
14:49:58funmanhm can you try to revert as3514 to use line1 and not line2 ?
14:50:57 Join robin0800 [0] (
14:55:51kugelfunman: also revert the move of the HAVE_RECODING #endif?
14:57:50kugelfunman: You didn't change the else branch in audiohw_set_monitor(), it's still at line in 1 afaics
14:59:15kugelor maybe it was bertrik
15:00:18funmankugel: i meant drivers/audio/as3514.c not the audio-xx file
15:00:29funmanthis last one is just a hack atm
15:00:57kugelfunman: Yes, I also mean that file
15:05:10 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
15:05:34funmani don't know if frequencies display mean radio
15:06:18funmanyou can put a panic in firmware/drivers/tuner/si4700.c in si4700_init() to see if the chip ID is 0x1241, now we'll be fixed; )
15:14:01 Join Schmogel [0] (
15:15:34 Quit agaffney (Read error: 110 (Connection timed out))
15:16:12 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
15:20:06soapI'm shutting down to memtest. Pissing me off. Spend premium money for premium parts and grumble grumble.
15:20:47 Quit soap ()
15:24:35 Quit lasser ("ChatZilla 0.9.84 [Iceweasel 3.0.4/2008112309]")
15:25:17funmankugel: if you confirm that the FM chip is detected, please update the hardwaremappings wiki and define sda/scl the same than e200v2 (and remove the old sda/scl suppositions and the comment)
15:26:00funmani'll try to find some time this evening (or at worse tomorrow) to get some code for the buttons.
15:26:15funmandid you already set an interrupt on GPIO B0 changes?
15:28:21funmanthat'd be useful to see if it's called on button press and/or wheel scroll
15:29:11 Quit funman ("leaving")
15:32:18 Join _lifeless [0] (n=lifeless@
15:36:45 Quit kachna (Read error: 110 (Connection timed out))
15:44:42 Quit __lifeless (Read error: 110 (Connection timed out))
15:57:34 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
15:57:34 Quit Rob2223 ()
16:03:35 Join maffe [0] (
16:05:01 Quit tyfoo ("Carpe diem")
16:09:28maffeI just had an idea: using a supported player (like the X5) as a PC display (not for things like 3D games, but the display would be sufficient to show information like number of unread mails or progress of downloads)
16:10:39 Join tvelocity [0] (
16:11:38krazykit`maffe, many of the targets don't have a usb stack that allows for that. only the software usb targets COULD do that (portalplayer among others)
16:13:30 Quit moos ("Rockbox rules the DAP world")
16:14:27maffeWould it work file-based? So, for example, a program on the computer writes a file on the player and rockbox reads and displays it?
16:16:05krazykit`no. when connected via usb (storage), rockbox itself will not do anything with the files. doing so could introduce issues with reading and writing files by the host operating system
16:16:56krazykit`(further explanation for this could be found in a discussion about allowing rockbox to play music while connected and mounted by a computer. i think it was in irc)
16:19:02n1sand it has been asked many times :)
16:20:22 Quit robin0800 (Read error: 54 (Connection reset by peer))
16:20:50krazykit`maffe, plus you'd have to write the software to display on the device too
16:21:11n1swith hardware mods anything is possible though :)
16:22:21maffekrazykit`: i thought that would be a minor problem
16:23:06 Join MethoS- [0] (
16:23:23maffeas there already are text and image viewers
16:24:31n1smaffe: well, displaying may be easy but communicating with the computer is probably harder
16:24:52maffeyes, that's why I would use files
16:27:03*n1s wonders what the point of the exercise would be
16:27:40maffethen I could look at my X5 to see if there are new mails and don't have to wait for the CRT to warm up
16:28:29maffeor I could see when a download finished while playing a fullscreen game
16:29:03n1ssolution: buy another regular display...
16:29:15maffebut they are expensive
16:30:17maffeand there's no fun :)
16:31:19 Join lasser [0] (
16:34:39maffeI like hardware hacks, I read about using a very cheap digital photoframe for the keychain as a display
16:36:00***Saving seen data "./dancer.seen"
16:41:16 Join toffe82 [0] (
16:57:53 Join jgarvey [0] (
17:00:29 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111318]")
17:03:21maffeI can't find the discussion krazykit` mentioned
17:04:35 Join aarcane [0] (
17:07:57 Join tarbo [0] (n=me@unaffiliated/tarbo)
17:07:59 Quit aarcane (Client Quit)
17:09:27maffeanother regular display would need much space and energy
17:15:25 Join ShiftPlusOne [0] (
17:16:28ShiftPlusOneHello, on my ipod there's a connection next to the audio output which I suspect is for the 'remote' control bit that some headphones have.... was wondering how it works if I wanted to interface it with my computer... and does it work with rockbox?
17:17:05 Join funman [0] (n=fun@rockbox/developer/funman)
17:17:06 Part tmwsiy2012
17:17:22ShiftPlusOneit's an ipod mini 2g, btw.
17:17:23 Join robin0800 [0] (
17:18:49LambdaCalculus37The remote doesn't work, as far as I know. Then again, I don't have a remote so I can't say for sure.
17:19:45maffeLet's say I have a dedicated directory on the player. The host can write files in that directory. When there are 2 files, Rockbox knows the older one is not being written to anymore and reads it, then deletes it. The host does not write further files if there already are 2. Could something like that work?
17:20:24ShiftPlusOneI tried shorting some of the pins, nothing happens... don't know if that's how it works though
17:20:52ShiftPlusOnegrounding doesn't work either.
17:21:07n1sShiftPlusOne: i'm pretty sure rockbox does not have support for remotes on the ipods
17:21:46n1spatches are welcome ;)
17:22:30maffekrazykit`, can you comment about this approach?
17:24:08 Quit denes_ (Read error: 110 (Connection timed out))
17:24:12 Join denes_ [0] (
17:25:29ShiftPlusOnewhat does the I/O port screen show and what does GPIO and all that other stuff mean?
17:25:58maffeGPIO = general purpose input/output
17:27:37n1sShiftPlusOne: GPIO's are pins on the cpu/soc connected to various things, like buttons, other ic's connectors
17:27:40ShiftPlusOnewell... grounding some pins seems to make C flicker from FA to F2.... not that I can do anything with that
17:28:22ShiftPlusOnewether headphones are plugged in seems to be A... would the remote part be on that screen?
17:29:21 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
17:30:20n1smaybe this is interesting, no idea if that's true for minis though
17:31:03ShiftPlusOneAnother pin changes C to FE.... does that mean anything?
17:31:42n1syes, that some bits changed...
17:31:59ShiftPlusOneis ipl up properly now (with all the packages and the installer?)
17:32:13ShiftPlusOnen1s, in the context of trying to make the remote work...
17:32:25LambdaCalculus37ShiftPlusOne: It is.
17:32:32n1susually one bit corresponds to one pin, not all are connected to something though
17:33:52n1sbut IIUC this remote port is actually the same as the dock connector so maybe that half-finished patch for accessory support is interesting too
17:34:45n1ssee FS #8624
17:35:27amiconnmaffe: Storage access is exclusive. The dap firmware has no access to the files while it's connected to the host.
17:35:39maffeoh :(
17:36:38 Quit plus_M (Remote closed the connection)
17:38:29 Part B4gder
17:38:32ShiftPlusOnehmm.. well ipl doesn't support the serial port for the mini and the patch doesn't support the mini either.
17:39:43LambdaCalculus37ShiftPlusOne: Patches are welcome, as n1s said before. :)
17:39:56LambdaCalculus37If you'd like to work on remote support, go for it.
17:40:23ShiftPlusOneand the protocol is more complex than I had assumed =/
17:40:42ShiftPlusOneLambdaCalculus37, I have a theory... if I could do it, somebody would already have done it.
17:42:29ShiftPlusOnebut since it has come up... which part of the source should I look at if I was to use that 'C' GPIO thing to play/pause when triggered?
17:42:46ShiftPlusOne(no low level programming experience)
17:43:20funmankugel / dany_21_a : <- here is how the OF reads from GPIOB (at least how I read it ..)
17:48:17 Quit blithe (Read error: 110 (Connection timed out))
17:49:33 Join blithe [0] (
17:49:48 Join plus_M [0] (
17:52:16 Quit massiveH ("Leaving")
17:52:45 Quit kharo (Read error: 110 (Connection timed out))
17:54:14 Join kharo [0] (
17:56:00 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:58:36 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
17:59:28ShiftPlusOneerror running =/
18:00:56ShiftPlusOneah... no free space.
18:03:35ShiftPlusOnefreed up space... same problem
18:05:39n1sShiftPlusOne: which os?
18:06:45ShiftPlusOnewindows, andLinux (coLinux) seems to be because I was missing libc6-dev
18:10:24ShiftPlusOneyeah that was it... compiling away now.
18:10:41maffe there is the hack I mentioned above
18:12:00 Quit mofux (Remote closed the connection)
18:12:13gevaertsmaffe: all very well, but with a USB-ATA bridge this is not possible
18:13:34 Part LinusN
18:14:10 Quit linuxstb (Read error: 60 (Operation timed out))
18:15:22maffeyeah it seems I have to live with that
18:15:53 Quit petur ("work->home")
18:16:46 Join herrwaldo [0] (
18:23:15 Join karashata [0] (n=karashat@
18:25:40LambdaCalculus37Yay, FM radio on my Clip! :)
18:25:49LambdaCalculus37It sounds great!
18:26:59 Quit freddy_ ("Verlassend")
18:27:41maffeFM radio cannot sound great ;(
18:29:02LambdaCalculus37maffe: "It sounds great" in the sense that it works well and sounds like you would expect it to.
18:29:27maffeyeah sorry
18:35:48 Quit Bensawsome (Read error: 104 (Connection reset by peer))
18:35:52 Join Bensawsof [0] (
18:36:04***Saving seen data "./dancer.seen"
18:36:54 Quit funman ("leaving")
18:38:07 Join webguest94 [0] (n=4157b72d@gateway/web/cgi-irc/
18:40:58 Quit webguest94 (Client Quit)
18:41:37LambdaCalculus37Hmmm... MP3 playback on the Clip is still a little quirky, but Vorbis playback is very smooth now.
18:43:46 Nick Bensawsof is now known as Bensawsome (
18:44:20 Join bertrik [0] (
18:44:37 Join nk [0] (n=4157b72d@gateway/web/cgi-irc/
18:44:39 Quit ShiftPlusOne ("Leaving")
18:49:21 Quit nk (Client Quit)
18:55:25 Join nk [0] (n=4157b72d@gateway/web/cgi-irc/
18:55:31 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:59:54 Join faemir [0] (
19:00:14 Join miepchen^schlaf [0] (
19:04:10 Join soap [0] (
19:05:17 Quit soap ("Lost terminal")
19:06:17 Quit neddy ("Leaving.")
19:08:45 Join massiveH [0] (
19:13:30 Join Thundercloud [0] (
19:15:05 Quit nk ("CGI:IRC (EOF)")
19:21:40rasherShouldn't we freeze about now?
19:21:49LloreanI was about to say the same thing
19:26:14rasherI think the only objection last time was amiconn, because he thought it was too soon, but I think consensus on the ml was that we did in fact want a 3-month schedule, so I think it's time
19:26:21 Quit Thundercloud (Remote closed the connection)
19:28:08fredddy I just pasted the newest patch for fuze keymaps they now compile all
19:28:20 Join GodEater_ [0] (n=ge@rockbox/staff/GodEater)
19:29:44Topic"Please read before speaking: | Please direct offtopic/social chat to #rockbox-community | We are now in freeze for 3.1" by ChanServ (ChanServ@services.)
19:30:06rasherWell that was easy
19:31:44rasherI assume we need a swede to create the 3.1 branch?
19:32:03rasherOr is that not for another week?
19:32:05LloreanI believe so, but I thought it was freeze for a week, then branch for a week?
19:32:17 Join Thundercloud [0] (
19:33:11rasherAre you going to send a fixed-up copy of your 3.0 freeze message?
19:33:42LloreanNah. I've just sent a short email saying, more or less, "Alright guys. Frozen for 3.1"
19:34:05LloreanI don't think much needs to be said this time around.
19:35:01bluebrothera branch is just a copy in svn, just like a tag, so everyone with commit access should be able to do this
19:35:06 Quit robin0800 (Read error: 60 (Operation timed out))
19:35:57LambdaCalculus37Llorean: Should an announcement be posted to the forums as well?
19:35:59 Join Zagor [242] (
19:37:28LloreanLambdaCalculus37: Ideally I'd like it to be *not* a big deal, honestly.
19:37:46LloreanI'd like to, as soon as possible, just have it be "every 3 months, there's a new version number"
19:38:00LloreanBut maybe we should still, at least this time. I dunno.
19:38:03 Join robin0800_ [0] (
19:38:31bluebrotherannouncing it on the dev ml would be in place IMO. users ml isn't necessary ... IMO again ;-)
19:38:44LloreanI've sent a shot off to -dev already.
19:38:56LambdaCalculus37I think that should suffice then for now.
19:39:17 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
19:39:25LloreanAn announcement to the user ML when the release is actually *out* in a short "3.1 is now Released" type form makes sense, just as a sort of routine thing.
19:40:03 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
19:40:23rasherI still think there should be release notes. Basically take MajorChanges and flesh it out a bit.
19:40:44rasherZagor: Is there a way to get a list of bugs closed since last release?
19:40:48LloreanRelease notes would be nice.
19:41:27rasherHere we go :)
19:41:45 Quit miepchen^schlaf ()
19:41:51bluebrotherrelease notes would be good, especially to include them in rbutil at some later point
19:42:18bluebrotherthough I need to find the time implementing that. As with retrieval of the most current release info from the website
19:42:50 Join miepchen^schlaf [0] (
19:43:21 Join bertrik_ [0] (
19:44:54pixelmabluebrother: what do you think of a phrasing in our manual such as "...will happily play..."? Fine by me but since it sounds a bit informal so wanted to ask first in case someone things it looks too unprofessional
19:44:59 Quit bertrik (Read error: 110 (Connection timed out))
19:45:00 Quit nplus (Read error: 110 (Connection timed out))
19:47:04pixelmawill 3.1 be released on Christmas then?
19:47:18rasherMore or less
19:48:07pixelmaand we have a codename... ;)
19:48:38 Join nplus [0] (n=nplus@141.25.Globcom.Net)
19:49:43LambdaCalculus37pixelma: "Not for Workgroups"? ;)
19:51:16 Quit Thundercloud (Read error: 60 (Operation timed out))
19:54:50bluebrotherchristmas present? ;)
19:56:26bluebrotherpixelma: sounds ok to me −− we aren't making a professional product (and usually manuals of professional electronics can be quite informal) so I don't see a problem here
19:56:58bluebrotherplus, being a bit informal might make it look more "human" thus more appealing to users
19:57:05 Part domonoky
19:57:05 Quit miepchen^schlaf ()
19:57:15*bertrik_ is curious about the codename
19:57:29 Quit robin0800_ (Connection timed out)
19:58:45 Join Tangent1 [0] (
19:59:05 Join miepchen^schlaf [0] (
19:59:43Tangent1is the fifth generation Ipod a better platform for rockbox than the 5.5 or do they have similar issues with video playback?
20:00:29LloreanFor video they're basically identical.
20:00:39 Join kachna [0] (
20:00:51Tangent1does either have a performance advantage in any other field?
20:01:42LloreanThe only difference between the two is the disk. I'm not sure which one has an advantage there, but it shouldn't affect everyday use.
20:04:02*amiconn wonders whether we should provide new archos flash images to go with 3.1, again
20:04:40amiconnThere's one noticeable improvement - the asm optimised ucl decompressor, shortening boot time for flashed rockbox
20:05:01*amiconn is already running that on all his archoses
20:06:14LambdaCalculus37amiconn: I think we should provide them. Sounds like a good idea.
20:06:40LambdaCalculus37amiconn: Speaking of the Archoses, you have a JBRv1, correct?
20:06:59 Join Rob2222 [0] (
20:07:29 Join gregorovius [0] (
20:08:02amiconn(plus others)
20:08:19gregoroviusHi... I'm wondering, is there any music player or manager to sync music to my rockbox'd ipod? something that handles album art and such?
20:08:31Zagorrasher: yes we can. if not in the web interface, then by raw sql requests
20:08:38BigBambigregorovius: You can use whatever you like
20:08:56BigBambigregorovius: Just Sorry, I didn't read the album art bit
20:09:38BigBambigregorovius: I don't know if there is anything that will do both synching and set up the album art for you automatically in the way Rockbox wants
20:09:54BigBambiBut then I just drag and drop so I've never looked
20:10:10gregoroviusI can copy my music via rsync, but then how do I manage album art?
20:10:27BigBambiDepends if you have album art arranged
20:10:53BigBambiI have /artist/alub/tracks with a cover.bmp in each album directory, so I just copy that too
20:11:26gregoroviusthat's all that's needed?
20:11:38BigBambiyes, see
20:12:51gregoroviusokay, thanks
20:13:06gregoroviusin any case, if anyone knows of software that does what I was looking for, I'd like to hear it
20:13:14gregoroviuselse I'll just manage with rsync and some bash vudu
20:13:28Tangent1anybody here ever buy refurbished music playing hardware?
20:13:50BigBambigregorovius: Once you have album art set up, not much voodoo is required - just copy music and art
20:13:51Lloreangregorovius: Well, nobody's apparently wanted it enough to add support to one of the myriad open-source music sync tools.
20:13:59BigBambiTangent1: yes, many people
20:14:14Tangent1not anybody in the world, anybody here, that can tell me how they liked it
20:14:14 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
20:14:16gregoroviusLlorean: there's one guy doing a rockbox plugin for banshee, but it's not finished yet
20:14:29BigBambiTangent1: yes, many people *here*
20:14:39gregoroviusi'd like to see one for amarok, in my case
20:14:47BigBambiTangent1: And it, like anything, depends what you buy from where
20:15:03BigBambiI have never had a problem personally
20:15:29*rasher noticed what looked like an off-by-one error in the playlist, when trying to move a file
20:16:52LambdaCalculus37amiconn: Does your JBRv1 give you a reasonable amount of buffer space (i.e. > 439KB)? I can't roll a build without the database functionality due to a missing sh-elf-ld in my /usr/local/sh-elf/bin folder. For some odd reason, the binary was never placed into that folder.
20:17:11LambdaCalculus37I also want to disable the database to see if it'll give me back some buffer space.
20:17:59amiconnRockbox Info says I have 1.24MB of buffer
20:18:12amiconnThat's a build from 2008-12-02
20:18:33amiconnNothing disabled, just sane settings
20:18:48amiconnYes, but that doesn't help anything regarding free ram
20:18:59amiconn(due to no rombox)
20:20:11Tangent1hmm, have one item in my shopping cart, worth 280 dollars, but my subtotal is 560 dollars
20:20:14*Tangent1 is confused
20:20:22BigBambiTangent1: Please stay on topic
20:21:04LambdaCalculus37amiconn: I lowered the "Max Entries in File Browser" setting to 5000; now Rockbox Info is reporting 888KB free.
20:21:16*LambdaCalculus37 wonders why it was so high in the first place
20:21:25*amiconn has that at 400 ...
20:21:29Tangent1okay, sorry. not used to channels where that's an issue
20:21:44BigBambiTangent1: The guidelines are linked in the topic
20:22:06LambdaCalculus37Tangent1: This topic is for technical talk, and it's logged. Off-topic talk tends to clutter up the log.
20:22:28*LambdaCalculus37 sets his to 400
20:22:34*rasher has created ReleaseNotes31
20:22:35amiconnActually it's 500 right now, but I could lower it further
20:22:42LambdaCalculus37amiconn: 1.27MB free.
20:22:49BigBambiLambdaCalculus37: Why not set it to just more than the amount you need?
20:23:12LambdaCalculus37BigBambi: I must've set it by mistake and forgot about it.
20:23:15 Join wpyh [0] (n=william@
20:24:33*amiconn lowers it to 300, as 'stats' says the largest dir has 222 entries
20:25:36pixelmarasher: re. your off by one error when moving a file in the playlist - did you use shuffle or "insert shuffled" tracks or folders in the playlist?
20:25:51rasherpixelma: Pretty sure I did
20:25:52*pixelma bets
20:26:36pixelmaI see that too sometimes and there is a flyspray entry for that, let me dig it up
20:26:42 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
20:27:49 Join ebil|work [0] (n=andy@
20:27:50*LambdaCalculus37 also lowers to 300; 'stats' says the largest dir has 217 entries on his
20:28:01rasherpixelma: Yup, that'd be the one
20:28:11pixelmarasher: it can also be more than one off
20:28:15ebil|workIs it possible to have rockbox interface with my pioneer HU in my car? I've heard talk about it, but I can't find any info on the site...
20:28:16LambdaCalculus37amiconn: How much buffer now?
20:28:49*LambdaCalculus37 now has 1.28MB of free buffer on his JBRv1
20:29:45 Part Tangent1
20:29:49LambdaCalculus37Wonder where that 0.02MB difference is from...
20:31:22pixelmaused font?
20:31:50LambdaCalculus37pixelma: 08-Rockfont.
20:32:06BigBambiOther settings?
20:32:22amiconnLambdaCalculus37: What's your max playlist size setting? Mine is 4000
20:33:06LambdaCalculus37amiconn: 1000 here.
20:36:09***Saving seen data "./dancer.seen"
20:36:19 Join Thundercloud [0] (
20:37:35amiconnThe font in use doesn't matter as the font buffer size is fixed
20:37:36gevaertsrasher: I vote to ignore build system bugs and manual bugs
20:38:54LambdaCalculus37amiconn: Out of simple curiosity, if I were to build Rockbox for the JBRv1 without database support, how much extra buffer (if any) should that give me?
20:39:08amiconnI have no idea
20:40:03amiconnYou will at least gain the binsize difference. If you get it small enough to run rombox, you'll reclaim (approximately) the whole binsize for additional buffer
20:40:27amiconn(binsize means uncompressed here)
20:41:22 Join soap [0] (n=user@rockbox/staff/soap)
20:42:45 Join Seed [0] (
20:42:55soapBagder, you can go ahead and unban my buildserver if you want. It is not online yet, but when it comes back it will be fully operational - one way or another.
20:42:59*LambdaCalculus37 also remembered that he still wanted to get an 8MB RAM chip for his Archos
20:47:11 Join {phoenix} [0] (
20:47:13 Quit parafin ("So long and thanks for all the fish")
20:47:20 Join parafin [0] (i=parafin@
20:47:48 Quit parafin (Client Quit)
20:48:03 Join parafin [0] (
20:49:36Zagor103 bugs closed since 3.0
20:49:57rasherZagor: Can you generate a list that includes the close message?
20:50:08*scorche|sh reminds LambdaCalculus37 to get extras if he ever finds them
20:50:16 Quit TheSphinX^ ("XChat@Linux")
20:50:22rasherI bet a bunch of those are rejects or duplicates
20:50:30Zagorrasher: I suppose you've found how to get the list from the web interface? not with the close message though.
20:50:50rasherZagor: Nope, I've just been going through the messages linked from the svn log
20:51:24Zagorwhoa, that was longer than I thought...
20:51:32gevaertsI think it would be good if Mr Someone went through the list to check for rejects, post-3.0 bugs and other unwantedness
20:51:36*gevaerts volunteers
20:52:00Nico_Pgevaerts: for what purpose?
20:52:03rashergevaerts: you're aware that I checked some already
20:52:10gevaertsNico_P: release notes
20:52:17Nico_Pah, I see
20:52:24ZagorI think I can make a "taskid: closemsg" list from mysql
20:52:40Zagorif that would be more useful?
20:53:02rasherCertainly - and if you're outputting a html, make it a link to the task
20:53:33rasherWell, taskid tasksubject: closemsg perhaps
20:53:58Nico_P$ svn log -r 18614:HEAD | grep FS | wc -l
20:54:22rasherNico_P: Those aren't all bugfixes those. Plenty of patches among them
20:54:24ZagorI'll see what I can do. my sql skills are somewhat limited...
20:54:27Nico_Prasher: I know
20:54:57rasherZagor: Maybe this might even be interesting to create as a permanent page
20:54:58 Join petur [50] (n=petur@rockbox/developer/petur)
20:55:43Zagorrasher: hmm, that would mean mixing my poor sql skills with my limited php knowledge. what could possibly go wrong? :)
20:55:55gevaertsrasher: If you've checked some already, maybe it's best if you add all not-known-to-be-irrelevant ones to the wiki page, and then I'll go through the remaining ones?
20:56:15Zagorof course I could do it in perl
20:56:29rashergevaerts: I've already added them (from the SVN log)
20:56:30 Join neddy [0] (n=john@nat/sun/x-da80a02fa1048d7b)
20:57:28gevaertsrasher: I know, but that list misses some. Do you still have the full list from the svn log? Then I can remove your remove-list from the FS log, and see what remains
20:58:38rashergevaerts: I don't have anything but what I copied to the wiki page
21:00:22 Quit gregorovius (Read error: 60 (Operation timed out))
21:00:41gevaertsrasher: ok. I'll see what I can do then
21:01:46 Join gregorovius [0] (
21:02:34*gevaerts wants to tell FS to show him all 103 of them on one page...
21:03:38rasherI believe there's a preference you can set, but it might not go higher than 100
21:04:03Nico_Pit goes up to 250, and yes it's a pref
21:06:48 Quit massiveH ("Leaving")
21:07:24*Nico_P likes the ASCII art logo in the old release notes
21:07:47rasherIt's not too late to add it to 3.1 :)
21:07:58Unhelpfulok... i think get_totalsize and friends should go in a new bmpsize.c, and be moved out the header. recalc_dimension probably belongs there too.
21:07:59Zagorwas it missing from 3.0?
21:08:00gevaertsrasher: so the list on the wiki is the svn log with all irrelevant stuff filtered out?
21:08:22rasherIn fact..
21:08:41rashergevaerts: Pretty much. I haven't checked for post-3.0 bugs though
21:09:36Unhelpfulthese are all things that will be needed for calculating sizes of in-memory bitmaps, and none of them are specific to bitmap files. any other file loader frontend for the scaler will need them.
21:09:40gevaertsrasher: my plan is to do that :)
21:09:53rashergevaerts: not going to stop you
21:10:15 Quit Thundercloud (Remote closed the connection)
21:11:39Zagoradding links now
21:12:39 Quit gregorovius ()
21:12:53Nico_PZagor: closure reason too maybe?
21:13:14ZagorNico_P: that's what closure_comment is
21:13:23Zagoroh, you mean the field. right
21:13:30Nico_Pyes, "fixed" and such
21:13:51 Quit miepchen^schlaf ()
21:14:01 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:14:02Nico_Pwe don't have any new officially supported targets, do we?
21:14:15BigBambinot yet
21:14:24*BigBambi sees what Nico_P means
21:14:33BigBambiand the answer is still no :)
21:15:25Nico_Pdo the release notes really need to have the full credits?
21:16:02 Quit karashata ("G'bye everyone!")
21:16:23Nico_Pthe list has gotten seriously long though
21:16:29 Join miepchen^schlaf [0] (
21:16:53ZagorNico_P: it doesn't cost anything to include it though
21:17:03Zagorand I think it's a nice gesture
21:17:38rasherIt's not in the way either, down at the bottom
21:17:56Nico_Phmm so are we dropping the natural sorting in this release?
21:18:28 Quit fredddy (Remote closed the connection)
21:20:03Unhelpfulit would be nice to support the musicrbrainz sort order tags, for that
21:20:34saratogashould i delete bugs fixed in targets 3.1 isn't released for?
21:20:39saratogafrom the wiki that is
21:20:54LloreanUnhelpful: There's an in-progress patch for sort-order tags, but I believe it currently requires you to manually use them as filters in the database.
21:21:10LloreanI think the preference would be "If an artist sort-order tag is present, use it, otherwise use the normal artist tag" transparently
21:21:21 Join neptunepink [0] (
21:21:33rasherLlorean: Also, it displays the sort-order tag, rather than the artist tag. I think kugel had an idea and wanted to work on it
21:22:03rashernatural sorting seems more important/useful in the filebrowser to me
21:22:32maffeI have another idea :) Instead of USB, one could use the line-in to receive a modulated signal :D
21:22:32 Join karashata [0] (n=karashat@
21:23:12_Auron_it's completely possible :P
21:23:18Unhelpfulneeds work, then. i think that perhaps the best response to akio's report might be to replace the #ifdefs there with if (remote && sizeof(fb_data) != sizeof(fb_remote_data), since there are too many complicated ways for this to go wrong, and that's what the real condition is.
21:23:27_Auron_it's what TI graphing calculators do
21:24:20Unhelpfulrasher: and natural sorting would still be useful as an approximation of correct sorting in tag db when there are no sort order tags, i'd think
21:24:39rasherUnhelpful: I don't think the two are terribly related, really
21:25:22*gevaerts sees a comeback of the natural sorting order discussion and remembers his plans for when that happens
21:25:34gevaertsAll sorting orders are equally natural or unnatural!
21:25:47LloreanUnhelpful: A naming dispute.
21:26:19maffe_Auron_: you mean data transmission over line-in?
21:26:46Zagorresolution_reason added
21:29:14gevaertssaratoga: was FS #9410 present in 3.0?
21:29:25Unhelpfulthoughts on this?
21:29:35Zagorlinks added
21:29:57saratogagevaerts: no its related to dual core playback
21:30:12gevaertsok. Not for the release notes then
21:30:15 Join Aurix_Lexico [0] (n=comrade@
21:30:26Unhelpfulthe extra condition is what the complicated #if is really trying to express. unless i'm missing some way to test size of a typedefed data type in macro?
21:30:58gevaertsI'm not sure what to do with FS #9360
21:31:48rashergevaerts: Just include it, I'll say
21:32:20gevaertsOK. I'm going to add some bugs right now. Should I remove the build system and manual ones while I'm at it?
21:32:44rasherGo for it
21:33:09rasherOr make sections, perhaps
21:34:13gevaertsI think that at least build issues are irrelevant for releases, unless maybe those that are visible when building voices
21:34:43ZagorUnhelpful: it seems to me those two conditionals are rather different?
21:35:21 Quit soap (Read error: 110 (Connection timed out))
21:35:22UnhelpfulZagor: they are. the second is just to guard the test, since fb_remote_data is undefined on targets without an LCD remote
21:35:53BigBambiZagor: I like the title of - "SQL result - phpMyAdmin" :)
21:35:54 Join soap [0] (
21:36:07ZagorBigBambi: pretty huh :)
21:36:11BigBambiyep :)
21:37:07Unhelpfulthe if() condition represents what the original #if was trying to do - use sizeof(fb_remote_data) if the bitmap is for a remote and if the remote data size is different from the main display data size
21:37:33 Quit karashata ("G'bye everyone!")
21:37:38gevaertsrasher: I think the xpdf search issue can probably stay, but I very much doubt if people who use releases care about the other build and manual issues
21:38:31Lloreangevaerts: Maybe a simple note about "misc build system changes" or something?
21:38:41gevaertsLlorean: ok
21:38:43LloreanA way to note that they happened, but aren't really relevant to the final output?
21:38:52 Join funman [0] (n=fun@rockbox/developer/funman)
21:39:03ZagorUnhelpful: in that case I say use the code rather than the macro.
21:39:34 Quit neptunepink (Read error: 104 (Connection reset by peer))
21:40:06bluebrotherhmm. Maybe it would make sense to include build system / manual / etc bugs in the list but sort them out?
21:40:52 Join massiveH [0] (
21:42:05rasherI was thinking have the closed bugs split in "core / plugins / manual / misc" sections, like the Known Bugs section
21:42:33maffeI think I should write a DTMF decoder for Rockbox
21:43:31LambdaCalculus37funman: I've been trying out the Clip for a little while already, and I've noticed that while MP3 playback tends to skip and stutter a lot, Vorbis playback is quite smooth.
21:45:20 Quit perrikwp (" ajax IRC Client")
21:45:29 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
21:45:44funmanLambdaCalculus37: that's explainable if vorbis requests higher packets to the buffering thread; i understood that the flash buffering would only refill when a request couldn't be serviced without reading from the flash
21:46:32 Join liguhy [0] (n=46b3aebf@gateway/web/cgi-irc/
21:47:34funmanit would be nice to find if/where buffering.c bugs on us
21:47:46liguhyanyone know if it is possible to rotate an e270's screen output 180 degrees?
21:47:59liguhyhave a damaged lcd where you can't see the top 1/3
21:48:28BigBambiYou want the "Upside Down" mode
21:48:48liguhydeal is, I can't see most of the menu's cause they're usually all in the top 1/3 :(
21:49:02BigBambithe manual has the menu layout I think
21:49:10*BigBambi checks that statement
21:49:42maffemaybe you should activate voice output
21:50:04liguhystill have to find it in the menus. lol
21:50:09 Quit soap ("Leaving")
21:50:13BigBambihmm, it doesn't
21:50:16LloreanVoice output defaults to enabled.
21:50:30LloreanYou just don't hear it normally because it doesn't do anything until you copy a voice file over.
21:51:03gevaertsOK. All closed bugs still listed on ReleaseNotes31 were present in 3.0
21:51:29liguhyso I should hear my menus by default? don't remember that
21:51:39BigBambi<Llorean> You just don't hear it normally because it doesn't do anything until you copy a voice file over.
21:51:54LambdaCalculus37funman: I'll take a look at it later on.
21:52:15 Quit herrwaldo (Read error: 104 (Connection reset by peer))
21:52:31liguhyok, so I assume that's something I do manually through the computer and reboot the sansa? where's the voice file?
21:53:01BigBambiI'm pretty sure the manual tells you that
21:53:17 Quit wpyh (Read error: 110 (Connection timed out))
21:53:18leachim6is there any lyrics support in RB ?
21:53:31BigBambiThere is a patch on the tracker can I make that work ?
21:53:38BigBambiI don't know how up to date it is about ratings
21:53:50leachim6how do song ratings work
21:53:54BigBambiBy patching the source code with it, then building Rockbox
21:54:03Lloreanleachim6: Please, read the manual for what features Rockbox already has.
21:54:10leachim6of course...
21:54:11liguhyI'll check it out and look for the upside down mode and/or voice commands
21:54:20funmanLambdaCalculus37: I had tried some time ago to add sanity checks in various functions
21:54:21 Quit perrikwp (" ajax IRC Client")
21:54:43funmanbut didn't figure anything expect that the pcm buffer looked corrupted
21:55:00bertrik_funman, I think you had some doubt about the GPIO pins I assigned to i2c on the fuze
21:55:58funmanbertrik_: yes, I saw the pins you mentioned in GPIOB isr (probably related to buttons/wheel), and in wav_codec (where I had found FM code for m200v2/fuze) block I found the same pins than e200v2
21:56:17 Join herrwaldo [0] (
21:56:28funmankugel tried to use these pins and could access the FM menu and see frequencies change, but no sound to confirm that FM worked
21:56:50funmana panicf() in si4700_init() would tell for sure ;)
21:57:41bertrik_the latest clip radio patch also adds some more debugging info to the fm debugging screen, that may help too
21:57:58 Join ibseco [0] (
21:58:59funmanwe can't use buttons (yet) on the fuze so all the menu selection has to be done via config files
21:59:48 Quit LambdaCalculus37 (" ajax IRC Client")
22:00:09bertrik_funman, I'm pretty sure that the protocol done on those pins was i2c, and not button code, I can look it up for you if you want
22:00:18funmanI have (with a lot of difficulty) soldered 8 wires to the (JTAG?) connector of my bricked Clip, I'll look for JTAG wiggler mapping and try to establish a connection
22:00:36funmanbertrik_: I have posted an analysis earlier, let me find the link
22:01:01funmani thought (a lot) also about i2c but it seems to miss ack bytes
22:02:30bertrik_funman, if you look closely at the PCB picture from the ABI disassembly and compare the traces with the ball layout, you can trace most JTAG pins
22:03:03funmani was actually going to trust that the e200v2 mapping was the same ^^
22:03:41bertrik_funman, I think it is
22:04:17liguhygot the voice stuff working just by updating, guess it wasn't in my older version. found what I needed in the manual. thanks guys!
22:04:47funmanI think my cable is this one:
22:05:11liguhyguess I don't have to ebay it for 15 bucks, learn how to solder and try doing a new lcd, or just trashing it. rockbox rocks my socks off :)
22:08:25leachim6I can't find song ratings in the manual
22:08:25 Quit liguhy ("CGI:IRC (EOF)")
22:08:27leachim6can someone link me ?
22:10:06funman search "rating"
22:10:06bertrik_funman, I think it does look very much like i2c, for example a start condition on line 18, on line 41 it reads back the ACK after writing a byte, but I would expect B1=0 on line 61 as an ACK
22:11:04 Join MethoS-- [0] (
22:11:06 Join slckr [0] (n=44204a41@gateway/web/cgi-irc/
22:11:08funmanbertrik_: i probably messed up when reading
22:11:14slckrhey everyone
22:11:41funmanI just think : there is no address (or rather i didn't find any)
22:11:55slckrso i just got this player called the Slacker Portable and im wondering if anyone thought id be perfect for rockbox
22:12:27slckrits got a 4 inch crisp screen and a touch pad, im thinking of startign up a website with its innards splayed out
22:12:36bertrik_funman, BTW, I think it is common to NOT ack the last byte read, so when reading only 1 byte, it is normal to NAK it
22:12:39funmanslckr: you could put pictures on the wiki
22:12:53saratoga slckr: feel free to port rockbox to it
22:12:55Lloreanslckr: Rockbox happens when owners of the players decide they're interested enough to do the work.
22:13:10LloreanThe best way to get Rockbox on a player is to find people who already have the player, and get them interested in working on it too.
22:13:11jhMikeSbertrik_: you are correct, sir
22:13:29slckrok. ive used rockbox since the beginning on my archos brick, then i used it on the ipod, then my gigabeat. I can't wait until i start to open this thing
22:13:48slckrfrickin wifi built in. imagine applications being downloaded through rockbox or something
22:13:58funmanbertrik_: it's right, the code ack the x-1 bytes, and nak the last one (but always only read one)
22:14:38Lloreanslckr: Rockbox isn't really designed to be a multifunction OS. If you're thinking like that, you may prefer working on a distro of Linux for it.
22:14:38bertrik_hmm, the i2c address is 0x61, so it's definitely not an si4700 radio that it communciates with
22:14:48funmanoh of course ..
22:15:03funmani guess the generic i2c driver would fit perfectly then
22:15:09gevaertsrasher: can you check if the closed bugs lists on ReleaseNotes31 are ok now?
22:15:16funmanwe just need kugel now ^^
22:15:25 Join MethoS_ [0] (
22:15:51rashergevaerts: I'd add a header like in Known bugs, so it gets an entry in the TOC
22:16:11funmanby the way I see the address sent most significative bit first, and the read is made less significative bit first, does it sound normal?
22:16:18rasherAnd remove the note at the bottom about introduced bugs, if we've weeded those out
22:16:26slckrthanks for the linux tip, i just googled and it runs linux
22:16:42 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
22:16:54gevaertsrasher: once saratoga is done I'll do those
22:16:54funmanslckr: you could run linux, and then build a simulator with libsdl ;)
22:16:55slckrdoesn't this make it easier rockbox to port?
22:17:13slckri wish i had an idea how to use commands in linux :p
22:17:18gevaertsIt may make it easier to find documentation
22:17:51slckrthese things are going for $60 seems very frickin cheap for wifi player
22:18:53saratogaslckr: another good thing to remember is not to annoy developers by telling them how great a port would be
22:19:54 Nick bertrik_ is now known as bertrik (
22:20:17funmansaratoga: did you find some JTAG software which runs on Linux?
22:20:36BigBambigevaerts: Perhaps a note on the known bugs bit to say that not all of these affect all targets (or omething imilar)
22:20:49gevaertsBigBambi: good idea
22:20:56 Join Nibbler [0] (
22:21:02slckrhaha, great. im just excited about the posibilities, with any player with rockbox :)
22:21:23funmanbertrik: thanks
22:21:29slckrthanks you all, ill be back here in a few days. ill have pictures of the innards :p
22:21:45saratogafunman: karl recommended OpenOCD based on his work on the gigabeat
22:22:15bertrikit works with a lot of JTAG "cables". I have a simple parallel port "wiggler" and it worked with that too.
22:22:18saratogai'm sure he'd be willing to help you with it
22:23:02gevaertsrasher: done
22:23:10gevaertsBigBambi: can you check my wording?
22:24:04BigBambiI'd still mention that they don't all apply to all players
22:24:17bertrikit looks like the buttons on the fuze go straight to the ribbon connector without any i2c chip in between, so I don't know what the i2c is for yet
22:24:37BigBambiPerhaps "Please have a look at the individual bug reports if you think that some may apply to you, as often a bug can depend on very specific circumstances and many only exist on certain DAPs"
22:24:46BigBambior whatever is the preferred term for DAP :)
22:25:10rasherI try to avoid using DAP
22:25:20BigBambiyes, that is why I mentioned it
22:25:23funmanbertrik: the wheel as well?
22:25:34 Quit MethoS- (No route to host)
22:25:40gevaerts"players", as used elsewhere on the page
22:25:42BigBambibut some don't like Player as Archos helpfully named one it
22:26:06gevaertsBigBambi: done
22:27:00BigBambiThe page looks good
22:27:19bertrikfunman, I think so, I assume that the wheel is sensed by the two small 4-pin black rectangular thingies
22:27:23rasherWhat does "Configurable mono recording mode" mean?
22:27:47gevaertsThat's about some things petur did I think
22:27:58Lloreanrasher: "Left" "Right" "Left+Right"
22:27:58funmanI see 20 pin JTAG headers, but I'm not sure if I must read from top to bottom or the reverse, and which side is the left or right. Does the rule for IDE wires apply here ? (red wire is/next to power)
22:28:09*BigBambi makes the most vital edit of all - "these page is" to "this page is" :)
22:28:11rasherLlorean: Ah
22:28:41 Quit jhulst (Remote closed the connection)
22:28:46LloreanI hoped you'd understand that, because I think if I tried to explain it with words I'd have mangled it horribly.
22:28:51LloreanWith more words, rather.
22:29:02 Quit EspeonEefi ("さよなら")
22:29:05funmanbertrik: reading from the wheel will be enough to higher the moral of our troops !
22:29:08gevaertsOf course we need to remember to review these lists in a few weeks when we actually release
22:29:17saratogafunman: get a breadboard and then wire the jtag into that so you can flip the order around
22:29:27 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:29:35 Quit MethoS-- (Read error: 113 (No route to host))
22:29:36saratogathough the order is probably the same as atomicpunk found on his m200
22:29:41funmansaratoga: i have no breadboard and not enough wires
22:30:05saratogathey're a few dollars in the US, might be worth getting one
22:30:07funmanbut well i'll use the JTAG connector as a breadboard and put the wires directly into it without solder/connector
22:30:13rasherLlorean: Think I got it.
22:30:22funmanUnfortunately I don't know where to get one, and i critically miss time
22:30:28saratogaalso, i managed to solder my fuze finally using a microscope and a very fine soldering tip
22:30:42bertrikmaybe the i2c goes to a kind of expander chip to read all the buttons on the fuze
22:30:42saratogabut have not yet had the courage to take it off the microscope mount and into my work bench
22:31:03saratogai2c is really odd, they only need 2 pins for the wheel and they have plenty for that
22:31:08bertrikI can't find any other chips on the fuze photo's that could do that
22:31:34funmansaratoga: perhaps all the pins go to a converter and this one is linked to the SoC with i2c ?
22:31:55saratogajust seems like an odd thing to do
22:32:17funmani've seen odder: some manufacturers use multicolor screens on monochrome daps :/
22:32:18saratogathe wheel on the v1s was just directly attached to GPIO
22:32:35saratogaand its essentially the same type of sensor
22:32:49bertrikthe bus address could be a clue to the type of device, unfortunately there's only 127 device addresses possible, so probably a lot of devices share device addresses
22:33:44funmanis the address 0x60 or 0x30 << 1 ?
22:35:32 Quit saratoga ("CGI:IRC (EOF)")
22:35:56 Quit neddy (Connection timed out)
22:36:12***Saving seen data "./dancer.seen"
22:36:39 Join obo [0] (n=obo@rockbox/developer/obo)
22:38:43 Quit slckr ("CGI:IRC (EOF)")
22:38:58 Nick amiconn is now known as lazyconn (n=jens@rockbox/developer/amiconn)
22:39:09 Nick lazyconn is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:40:49bertrikfunman, I think it is (address << 1) +W/R bit, but this is usually very confusing in device datasheets
22:43:11bertrikFig 10 in the Philips I2C specification (UM10204) shows that the address and the W/R bit are distinct
22:44:01 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
22:44:51 Quit Nibbler (Read error: 60 (Operation timed out))
22:48:08 Quit massiveH ("Leaving")
22:48:40 Quit miepchen^schlaf ()
22:51:40rasherAre we aiming for December 24th?
22:52:26 Quit {phoenix} (Remote closed the connection)
22:52:30Bagdernot if I'm gonna build the binaries...
22:52:58Bagderthen we're going for 25th or 26th
22:54:05LloreanI'll be happy as long as there's a RC in a week.
22:54:33LloreanIf there's a little delay in building after the second week, it's not a big deal.
22:54:34bluebrotherwe need an updated Rockbox Utility as well, and I won't be able doing them over the holidays
22:54:37BagderI'll make a round of RC builds immediately after the branch I think
22:55:01bluebrotherunless doing them before. Shouldn't be a problem as there isn't much activity around that at all
22:55:26Bagderright, I doubt rbutil should need much change for 3.1
22:55:35gevaertsbluebrother: I'd suggest making the full install install the release, not the current build
22:55:48LloreanCouldn't an updated RButil be basically done now, and we just upload a binary for it when 3.1 is actually out?
22:56:12*rasher should have another look at compiling a static 64-bit build
22:57:40 Quit XavierGr (Nick collision from services.)
22:57:51 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
22:59:04 Quit petur (Read error: 110 (Connection timed out))
22:59:51 Quit _lifeless (Remote closed the connection)
23:00:04 Join _lifeless [0] (n=lifeless@
23:00:06 Quit Zagor ("Clint excited")
23:00:18funmanno response on the JTAG connection, and the '0 bytes recovery' USB mode shows loads of errors now
23:00:59funman+ one of the wires unsoldered so i won't try anymore for today
23:01:37Nico_Pis there a way to get a wiki index as a tree?
23:02:01bluebrotherLlorean: it could.
23:02:02 Quit itcheg (" ajax IRC Client")
23:02:19rasherNico_P: I doubt it'd be useful. Most pages don't have a parent set, or a completely random page as parent
23:02:35 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
23:03:06Nico_Prasher: that's the point actually. I'd like to be able to easily spot which pages could get a relevant parent
23:03:23bertrikfunman, it would be interesting to see what we read back from i2c address 0x61 and if buttons or the wheel influence the reading. Indeed I think we can re-use generic_i2c for that.
23:03:59*funman desperately looks around for kugel
23:04:49rasherNico_P: Ah. No idea anyway. Btw, why does the # character in a wps eat the following newline? Shouldn't that only happen if the comment is the full line?
23:05:14Nico_Prasher: IIRC it does
23:05:32Nico_PI think it does what you say
23:05:44funmanbertrik: do you know how we can use the seek feature of SI4700 to look for channels ? All frequencies are 'tuned' so we can't detect proper channels that way
23:06:00rasherNico_P: Nope - try this wps: %ia#artist\n%it#title
23:06:26Nico_Partist and title show on the same line?
23:06:43Nico_PBagder: here?
23:06:59BagderI am
23:06:59bertrikfunman, no, I think the seek function of the si4700 does not really match with the radio code in rockbox
23:07:01rasherNico_P: Yup
23:07:16funmanbertrik: i was afraid of that :/
23:07:25Nico_PBagder: any chance we could get TWiki's Tree and/or TreeBrowser plugin installed?
23:07:49 Quit funman ("leaving")
23:07:52Bagderquite probably, let's talk with Zagor about that when he gets around next time
23:10:28Nico_PI think we would gain readability by using TreePlugin
23:11:41bertrikfunman, maybe the AFC rail can be used to determine if the currently tuned frequency corresponds to a valid channel with a radio station on it
23:12:01 Join massiveH [0] (
23:13:19LloreanFor people who've held the clip: Does it seem unreasonable to use the volume +/- buttons to move up and down in the list?
23:13:33LloreanDoing that would make the keymap _much_ simpler, and it's not uncomfortable at all for me at least with either hand.
23:14:31LambdaCalculus37Llorean:Works for me.
23:14:44bertrikI find it much easier to use the buttons around the select button to move up and down
23:15:01 Join miepchen^schlaf [0] (
23:15:24gevaertsI'm a bit dubious about this. How will volume setting work?
23:15:32Lloreangevaerts: Same as always.
23:15:32bertrikthe headphone plug gets a bit in the way for comfortable navigation
23:15:44Lloreangevaerts: It's only applicable to the list, where on most targets you can't adjust the volume anyway.
23:17:38LloreanAnd it would, I believe, allow for the clearing up of every core keymap issue on the clip all in one go.
23:18:09 Quit jgarvey ("Leaving")
23:18:15bertrikwhat core keymap issues?
23:18:16 Quit LambdaCalculus37 ("Ka-chunka")
23:18:16*amiconn would think that using the volume buttons for up/down would be rather unintuitive
23:18:20Lloreanbertrik: If I whip up a patch later, would you mind trying using it for a few days, see if it's one of those "uncomfortable now, because I'm not used to it things" or if it's a "always uncomfortable" thing?
23:18:36amiconnWhat would you put on the up/down buttons?
23:18:41Lloreanamiconn: There aren't one.
23:18:51LloreanIt's like the c200, except the volume button is vertical rather than horizontal
23:18:52gevaertsLlorean: I would still strongly prefer to keep volume buttons for volume everywhere, but I haven't tried the alternative (on any player)
23:19:10amiconnIiuc the wheel-like thingy is actually a 4-directional button?
23:19:27bertrikamiconn, exactly
23:19:34Lloreanamiconn: In the same sense the iPod wheel is a 4-directional button. They're labelled and everything.
23:19:52*amiconn doesn't remember having seen any labels on them
23:20:04LloreanThey light up, you don't often see them in powered-down ones.
23:20:28bertrikLlorean, I don't think I can be convinced easily about this, but I won't mind trying it
23:20:55amiconnWell, on the M3 we have to use Vol-Up/Down for list navigation...
23:21:01Llorean"Up" is "Play/Pause" and down is "Submenu" like other Sansas.
23:21:28*amiconn doesn't understand why they did that weird assignment
23:21:53*gevaerts clarifies that "strongly prefer" doesn't mean "strongly object to the alternative"
23:22:10LloreanThe volume button doesn't seem to do anything in the list at the moment, actually.
23:23:02gevaertshm, the volume buttons seem to work properly in lists in my gigabeat, but the status bar doesn't update properly
23:23:23Lloreangevaerts: Just referring to the clip here
23:23:28LloreanIt definitely works in the lists on the Gigabeat F
23:23:38bertrikI would like to have the volume buttons control volume in every menu if possible
23:23:53gevaertsLlorean: does the status bar update properly on your F?
23:24:14Lloreanbertrik: I see "Volume" as being secondary to having buttons for resuming and stopping playback, and other features we've always had a button for.
23:24:45Lloreangevaerts: It updates slowly. So not entirely properly.
23:25:04gevaertsLlorean: good point. I hadn't realised that this move would free up the play->WPS combination
23:25:32 Quit ebil|work (Read error: 60 (Operation timed out))
23:26:02Lloreangevaerts: That's more or less my goal. The other alternative is to keep the "pad" as directional in the list, and assign the missing functions to the volume buttons. Less intuitive mapping, but possibly easier to use once you know it.
23:26:14bertrikI like to volume to always work no matter what menu I'm in
23:26:24Lloreanbertrik: You have already said this...
23:27:26LloreanBut, generally speaking, very few Rockbox targets have universal volume control, and in the case of the F it's because it had those buttons completely spare anyway.
23:27:35gevaertsLlorean: that alternative would mean breaking _every_ label instead of some of them
23:27:41LloreanIn the case of the c200, there's still some discussion over what to do with them.
23:27:50Lloreangevaerts: That's why I proposed using them for list movement instead.
23:28:00 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
23:28:10*gevaerts slightly leans over to Llorean's proposal
23:28:27Llorean"Up" and "Down" at least, in a sense, has an intuitive relation to how the buttons are positioned out still.
23:29:42*bertrik still thinks it's a very awkward idea
23:29:43LloreanAnyway, at the very lease we can try it out.
23:31:21leachim6where do you guys get all of your themes from ?
23:31:37*domonoky thinks the m200v4 has a similar key lay-out, but i would not want to use the volume_keys for up/down, better reuse the vol only where needed, and keep the volume on other screens..
23:31:38leachim6when I used RB before....there was a site that had hundreds of them
23:31:44leachim6but I can't seem to find it again
23:31:57domonokyleachim6: take a look at the wiki under WpsGallery
23:32:14Lloreandomonoky: The m200 has a different physical layout though. They aren't up/down there. It should probably reflect the c200's final mapping, whatever that is, only without a Record button.
23:33:51leachim6thanks, that's not the site I used before, but it has tons of themes anyway
23:33:54leachim6what themes do you guys use
23:33:57leachim6is that off-topic ?
23:34:33domonokyLlorean: did you find time to rockbox your m200v4 ? :-)
23:36:13Lloreandomonoky: Ah, that reminds me. How do you dual boot an m200?
23:36:19_Auron_will there be an announcement when sansa v2s are supported?
23:36:33gevaerts_Auron_: of course
23:36:36domonokyLlorean: press left, while booting..
23:36:42Llorean_Auron_: It'll be in MajorChanges
23:36:50Lloreandomonoky: I thought the m200 was one of the ones where keys weren't supported yet?
23:36:55*Llorean can't apparently keep track of these things
23:37:22bertrikmaybe we need a more detailed port progress table for ams sansas :P
23:37:22domonokynope keys on m200v4 work since a long time.
23:37:37 Quit kachna (Remote closed the connection)
23:37:39domonokyi even have fm-radio running locally in my tree :-)
23:37:46LloreanThen I'll Rockbox my m200 tonight, when I get home with a battery for it.
23:37:50_Auron_how difficult would it be to start working on a plugin for rockbox? (ie a game)
23:38:15Llorean_Auron_: There are people whose first C experience ever has been in working on a plugin for Rockbox.
23:38:29_Auron_well that sounds rather pleasant actually
23:38:36domonoky_Auron_: depends on your coding skills, but plugins are easy.. take look at some simple plugins to see how they work..
23:39:16_Auron_most I ever did in C/C++ was a basic software 3d engine using SDL to access the framebuffer and create a windows app
23:39:54_Auron_which used text-output milkshape 3d models for 3d models
23:39:58gevaerts_Auron_: you probably can manage a plugin then
23:40:09 Join kachna [0] (
23:41:39_Auron_will probably look into it once sansa v2s are supported
23:41:59 Quit domonoky (Read error: 104 (Connection reset by peer))
23:47:02 Join Hillshum [0] (
23:48:54 Join itcheg [0] (i=62db4767@gateway/web/ajax/
23:50:01alexbobpI have a philosophical question. If the clip breaks off of my sansa clip, is it still a sansa clip?
23:50:40BigBambialexbobp: Please stay on topic
23:50:52 Join EspeonEefi [0] (i=eefi@STRATTON-TWO-EIGHTY.MIT.EDU)
23:50:52 Join Thundercloud [0] (
23:50:53alexbobpokay, sorry >_>
23:52:18 Quit EspeonEefi (Client Quit)
23:53:15*Llorean definitely sees a few keymap improvements for the clip already.
23:53:23LloreanThat i feel will be non-controversial.
23:53:46LloreanSpecifically, the quickscreen is a bit hard to leave right now.
23:54:35bertrikoh I think QS support for Clip was added only today or yesterday, so very possible that it doesn't work very intuitively yet
23:55:53LloreanWell, it allows both "up" and "down" to control the bottom one, which is unnecessary and makes it hard to leave the screen (though select could also be used for leaving it instead of Up)
23:58:51 Quit bertrik ("zZz")

Previous day | Next day