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

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

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

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

#rockbox log for 2009-05-29

00:02:56CIA-38New commit by 03mcuelenaere (r21123): Lua: make sure action.h gets C preprocessed before processing it
00:03:10 Join mt [0] (n=MTee@rockbox/developer/mt)
00:04:39 Quit domonoky (Read error: 104 (Connection reset by peer))
00:05:17mcuelenaeregit-push SVN post-commit hook still isn't fixed?
00:06:16 Quit petur ("Zzzzz")
00:06:53 Join JdGordon| [0] (i=836b0070@rockbox/developer/JdGordon)
00:07:30JdGordon|mcuelenaere: hey, i saw your q last night.... i wont have a chance to loo at it till sunday, so can you put it on FS?
00:07:38JdGordon|assuming its still happening?
00:08:26 Quit bertrik ("Leaving")
00:08:31mcuelenaerehowever this could also be caused by the fact that the Onda is still a very new port..
00:11:37 Quit amiconn (Nick collision from services.)
00:11:40 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
00:11:40 Quit pixelma (Nick collision from services.)
00:11:41 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
00:11:58 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
00:11:59 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
00:14:10 Quit trisiak (Read error: 113 (No route to host))
00:16:25mcuelenaereJdGordon: FS #10259
00:17:43JdGordon|ok, i love the svn rev.... ish :)
00:21:00 Quit matsl_ (Read error: 110 (Connection timed out))
00:24:39 Quit Lss (Read error: 110 (Connection timed out))
00:25:10mcuelenaerewill look into that later ;)
00:30:43 Join itcheg [0] (i=62db4c46@gateway/web/ajax/
00:31:03 Join HellDragon [0] (
00:32:45 Quit JdGordon| (" ajax IRC Client")
00:42:01 Quit n1s ("Lmnar")
00:42:32 Quit fyre^OS (Read error: 54 (Connection reset by peer))
00:42:44 Quit itcheg (" ajax IRC Client")
00:42:47 Join fyrestorm [0] (
00:44:28 Quit HellDragon (Client Quit)
00:47:05 Quit Lynx_ (" HydraIRC -> <- Like it? Visit #hydrairc on EFNet")
00:50:56 Join AndyIL [0] (i=AndyI@
00:53:12 Join perfectdrug [0] (n=5b0ed723@gateway/web/cgi-irc/
00:55:42perfectdrugpixelma: I just put the finished svg image of the sansa clip to the patchtracker, feedback would be nice so that I may begin with the fuze maybe
00:57:27perfectdrugFS #10260
00:58:20 Join Thundercloud [0] (
00:58:43funmanperfectdrug: very nice
00:59:26funmanperfectdrug: perhaps you could draw something on the yellow part of the screen ? (where the revision number usually is)
00:59:45pixelmaperfectdrug: what happened on the right side of the svg (watching in my browser only currently though)
01:00:36perfectdrugpixelma: i thought there is a button :D isn't it?
01:01:02funmanperfectdrug: there is, but the right border looks black
01:01:14pixelmathat's not what I mean... what funman said
01:01:33mtanything inside apps/codecs could be included like this : #include <codecs/foo/bar.h> right ?
01:01:49perfectdrugfunman: yeah i know something yellowish would be nice but the revision is usually left out in the manual
01:03:40perfectdrugah i see...i don't know, have to check it in inkscape, maybe it's just the browser
01:04:03pixelmathe black area on the right seems to only be a problem in the browser... which is not a good sign though
01:05:08 Join HellDragon [0] (
01:05:19perfectdrugyeah tested it too, hmm
01:05:26 Quit mcuelenaere () sometimes fails at radio
01:08:14ThundercloudWhy the hell is Flo Rida - Low playing when I asked for the Dance tag?
01:08:27krazykitThundercloud, why are you asking in an on-topic channel about rockbox?
01:08:45ThundercloudThat is because my scroll button went off
01:08:52ThundercloudI went for amarok, went one too far
01:08:54Thundercloudmy bad
01:09:43 Quit AndyI (Read error: 113 (No route to host))
01:09:45pixelmaperfectdrug: did you use the logo from the other target svgs?
01:10:26pixelmalooks like it but with thicker lines
01:11:01 Join golephish [0] (n=4b8b8f93@gateway/web/cgi-irc/
01:11:26perfectdrugi fixed the previous svg there was something mixed with the layers
01:12:31perfectdrugpixelma: yes i took it from the ondio, but changed it a little because it was strange arranged, have i messed with the line thickness?
01:13:28pixelmaheh, did Sansa really exchange the TM in the writing for an ® ?
01:13:37pixelmaor Sandisk...
01:15:08 Quit perrikwp (" ajax IRC Client")
01:15:31perfectdruggoogle images seems to show it this way
01:15:33Unhelpfulmcuelenaere: that's... very strange. i really can't imagine why buflib_buffer_out would hang. it compacts the buffer, then it shifts the contents up in memory, and returns a pointer. is there anything weird i should know about alignment, or such, on this target?
01:16:31 Quit golephish ("CGI:IRC (Ping timeout)")
01:17:21Unhelpfulkugel: which yield?
01:17:22pixelmaperfectdrug: the "strange arranged" had a reason - for some reason epstopdf doesn't like holes in filled objects and got me unpredictable results - so I made the "holes" own objects and put them on top
01:18:24 Quit Mathiasdm ("Yuuw!")
01:18:41linuxstbmt: Yes, that looks right. How are things going?
01:18:43perfectdrugmy sansa v1 has the TM also but the v2 has the ® ;)
01:19:01perfectdrugpixelma: oh thats a shame because i fixed that :O
01:19:24mtlinuxstb : Still rewriting the metadata parser. I didn't like how it looked the first time.
01:20:18perfectdrugi unfortunately haven't tried epstopdf i just lokked at the eps file
01:20:21funmanperfectdrug: my Clipv1 has no TM and no ® , only the Sansa label
01:20:57Unhelpfulpixelma: as i recall it there are a few different ways for ps/pdf to handle self-intersecting outlines, so a mode needs to be set properly to create outlines with holes.
01:21:02mtcollege-wise , I'm finally done with my first (and ugliest) subject !
01:21:06 Join Mathiasdm [0] (
01:21:22perfectdrughm so maybe leave it out at all?
01:21:47 Quit Mathiasdm (Client Quit)
01:22:48mtlinuxstb : is it OK to memcpy in the metadata parser ?
01:23:02pixelmaUnhelpful: there was no problem when the object wasn't filled, so only outline. My workaround was good enough for me but if you know how to set the modes...
01:23:11funmanperfectdrug: I have put the keymap on FS #10260
01:24:25perfectdrugfunman: thanks i look into it
01:25:07Unhelpfulpixelma: i learned PS for a project a while ago. i don't remember everything, but there were a few different fill modes, something along the lines of "everything inside the outline", "odd number of intersections" and a handedness-based one. if you're talking about converted from SVG, i'd say it's probably the converter's job to get these things right. :/
01:26:18 Quit kachna (Remote closed the connection)
01:26:50 Quit Thundercloud (Remote closed the connection)
01:27:07pixelmaso it's all Inkscapes fault.. ;) Anyways, I'm fine with the workaround and it would be easy to use the old way again here
01:27:47perfectdrugok so i revert it back :D
01:28:43 Join kachna [0] (
01:29:14 Quit thegeek (Read error: 104 (Connection reset by peer))
01:29:33 Join thegeek [0] (
01:30:14pixelmaperfectdrug: otherwise it looks really nice, only one minor nitpick still: there is a small gap in the lewer left "corner".
01:30:53 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
01:31:35Unhelpfulpixelma: the last time i used inkscape, i couldn't get it to save a SVG background such that KDE could render it properly. i was never able to figure out who was "wrong"
01:32:43 Quit kachna (Remote closed the connection)
01:32:50pixelmaI wonder if Inkscape provides the possibility to "snap" nodes (?) of one object to nodes of a different one, haven't figured that one out yet
01:33:19perfectdrugpixelma: i try to clean everythin up tomorrow, Unhelpful: gnome seems to have problems with rendering with my file too, the buttons are too light
01:33:27linuxstbmt: Yes, that's fine.
01:36:13 Join kachna [0] (
01:37:17***Saving seen data "./dancer.seen"
01:39:20pixelmaperfectdrug: thanks for your effort :) If you add the button labels it would be nice if you can put them (and related lines etc.) on an own layer
01:40:13 Join kugel [0] (n=kugel@rockbox/developer/kugel)
01:41:29mtlinuxstb : If I want to compile and link librm to rm.c metadata parser, should I add a rule in codecs.make ?
01:42:42perfectdrugbtw: i just learned inkscape this evening, the other times i just couldn't figure out how to do gradient and such :D thanks to your excellent templates pixelma it was much easier, i learned a lot today, awesome work from you doing all the players that are already there:)
01:43:25linuxstbmt: The other metadata parsers are standalone - they don't link with the codec libs at all.
01:44:02perfectdrugah and why don't we use the svg templates for the site? rather than photos?
01:44:46mtlinuxstb: hmm .. I wanted all the parsing to be done in librm.
01:44:49kugelJdGordon: there was talk about clearing the area around the ui viewport, what are your thoughts?
01:45:18kugelI think that's unneeded if the UI is done properly, and it also means keeping 4 additional vp structs in ram
01:45:29pixelmaperfectdrug: I also only learned Inkscape doing these player svgs for the manual :)
01:46:17JdGordonkugel: it doesnt, and it doesnt....
01:46:23pixelmabut that's getting off-topic and would be better suited in the other channel - glaring at the topic
01:46:31linuxstbmt: But doesn't librm (for the codec) need to do different things to the metadata parser?
01:46:32JdGordoni mean... yes, the clearing is needed, the rest of your comments was wrong
01:46:43kugelJdGordon: also, did you try to use the REFRESH_STATUSBAR flag of gui_wps_refresh, and assign a statusbar flag to *1* wps ?
01:46:55kugelJdGordon: can you please elaborate?
01:47:38mtlinuxstb: librm mainly has 2 functions : 1. get_packet - for the codec. 2- parse_header - which I want to use in the metadata parser as is.
01:47:54 Quit robin0800 (Remote closed the connection)
01:48:15JdGordonkugel: generally speaking, there are 4 viewports around the ui which can be easily figured out so no need to keep them in memory... and the issue is if the wps changes track (so it wants to do a full screen clear) while the ui is sitting in its loop waiting for a key press... it doesnt do a full redraw there unless it has to
01:48:33linuxstbmt: Then just move parse_header to the metadata/rm.c file?
01:48:57kugelJdGordon: recalculating the 4 vps seems more wasteful to me
01:49:19pixelmaspeaking of statusbar and updates... in the set background and foreground screens: wasn't there a time the whole screen changed colour life (or at least I seem to remember that there was a small sample rectangle)?
01:49:28Unhelpfulkugel: were you referring to error_wait?
01:49:53mtlinuxstb: I'll copy it, because I have to leave it there for the test program. (was just looking for a 'nicer-looking' solution)
01:49:57JdGordonpixelma: the colour chooser screen has the colour rectangle
01:50:03kugelUnhelpful: yes
01:50:15pixelmaJdGordon: not on my c200
01:50:53kugelJdGordon: ok, that seems easy. I haven't thought of screen_clear_area
01:51:35JdGordonpixelma: screen too small maybe? it needs to be fixed up?
01:51:36pixelmaalso - if I actually set a different colour and you are returned to the "Colours" menu, sometimes the statusbar vanishes or keeps the old colours until leaving that screen again
01:51:52pixelmaJdGordon: I'm quite sure it once was there
01:51:55JdGordonwhack that on fs...
01:52:10 Part wincent ("Kopete 0.12.7 :")
01:53:50pixelmavanish is a bit imprecise - the space is still kept so I think it "just" uses same fore- and background colours
01:56:01 Quit perfectdrug ("CGI:IRC (EOF)")
01:56:16 Join trisiak [0] (
01:57:03pixelmaand "until leaving the screen" is a bit wrong too, seems to fix itself with the next statusbar update
01:57:28JdGordonif you want it fixed... put it on FS
01:57:28mtinstead of printf - logf or debugf ?
01:58:16Unhelpfulpixelma: so, archos devices have neither ram DB nor dircache...
01:58:43Unhelpfuli'm timing a cache build on my beast with both features turned off. it's taking quite a while. :/
01:59:08pixelmaJdGordon: yeah, will do tomorrow if I don't forget (I'm off to sleep shortly)
01:59:29 Quit funman ("leaving")
02:00:07Unhelpful153s without scaling any cover images :/
02:00:13pixelmaUnhelpful: it's true... I don't mind much with my Ondio though, HD based players are worse (disk spin and much more storage possible)
02:01:04Unhelpfulpixelma: right, that's why amiconn had such rotten times with no cover images :/
02:02:22kugelUnhelpful: what are you benching? with resizing and Load to RAM toggled respecitvely?
02:05:56 Part toffe82
02:06:17CIA-38New commit by 03FlynDice (r21124): AMSSansa: Change comment to describe FCLK as input more accurately
02:06:19 Quit perrikwp (" ajax IRC Client")
02:06:51 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
02:07:09JdGordondid someone do a nautghy commit and break ipodpatcher's compile?
02:07:21JdGordonfat32format.c:59: error: expected ‘)’ before ‘x’
02:08:40pixelmaUnhelpful: could pictureflow itself take care of that? Although I don't know how much space the database could need on a large HD target with a possibly high amount of entries...
02:09:23Unhelpfulkugel: i'm benchmarking the time to search for covers. i've disabled load-to-ram and dircache, and commented out the portion that loads the images - all it does is collect album names, and search for cover images.
02:09:53 Join webguest51 [0] (n=47ebd77b@gateway/web/cgi-irc/
02:09:59Unhelpfulpixelma: searching for covers during DB build and adding a cover file "tag" would probably be the easiest way to speed this up.
02:10:42Unhelpfulalthough, i can't really see how we might justify doing that on archos...
02:10:48 Quit webguest51 (Client Quit)
02:12:00pixelmaI don't know if it's really worth it, especially since pictureflow has no real use currently other than showing off ;)
02:13:57Unhelpfulpixelma: that, too. on targets with album art, though, using the ramcache to get an album art "tag" for a file would be much faster than searching, too...
02:16:13 Quit gregzx (Read error: 110 (Connection timed out))
02:16:24Unhelpfulthe issue i'm trying to deal with is that cached album art and actual album art can get out of sync with each other. this happens because pictureflow uses a simple indexed list of albums, and generates the cache filenames from the index. if you add or remove albums, of course, the index of nearly every album can easily change.
02:16:57JdGordon(04:46:47 PM) kugel: JdGordon: also, did you try to use the REFRESH_STATUSBAR flag of gui_wps_refresh, and assign a statusbar flag to *1* wps ? <- no
02:18:18Unhelpfulso, i'm thinking about storing the AA path in the cache files, *always* doing the AA search at startup, and generating new cache entries whenever a cachefile doesn't have the "right" cover path. but, the cover search alone took 6-7min for amiconn :/
02:20:03pixelmarecreating the cache (including resizing) takes quite some time already on my c200 (I usually don't enable dircache or load database to RAM though)
02:20:44pixelmafor 10GB of lossy music (2GB + 8GB microSD)
02:27:13 Join evilnick_home [0] (
02:31:59 Join webguest44 [0] (n=4012a942@gateway/web/cgi-irc/
02:32:13webguest44$ svn co svn:// rockbox
02:32:16 Quit webguest44 (Client Quit)
02:33:45 Quit mt (Read error: 104 (Connection reset by peer))
02:36:35 Quit agaffney (Read error: 104 (Connection reset by peer))
02:39:14 Join miepchen^schla [0] (
02:41:38 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
02:44:12 Quit Llorean ("Leaving.")
02:45:01 Join Llorean [0] (
02:47:18 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
02:50:46 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
02:51:03 Quit moos ("Rockbox rules the DAP world")
02:59:03 Quit chandoo ("Leaving")
03:03:25 Quit lymeca (Read error: 54 (Connection reset by peer))
03:03:40 Join lymeca [0] (
03:04:24 Join DaCapn [0] (
03:36:07Unhelpfulpixelma: ideally, filenames would also be based on a hash of the cover path, or something else that won't change unless a cover changes. as long as search time can be made short, i think that cache "check" time can, as well.
03:37:01 Quit obo (Read error: 110 (Connection timed out))
03:37:21***Saving seen data "./dancer.seen"
03:41:21Unhelpfulit may have to be something that is only available on targets without dircache, though... i'd imagine that the file_exists and therefore album cover search is not very fast without dircache.
03:43:20 Quit kugel (Read error: 110 (Connection timed out))
04:07:30 Quit froggyman ("CGI:IRC (Ping timeout)")
04:38:57 Quit miepchen^schla (Read error: 101 (Network is unreachable))
04:50:36 Join kamlurker [0] (
04:51:29 Join James79 [0] (n=4573e0ea@gateway/web/cgi-irc/
04:52:08James79Hey all - I don't know if this is a Rockbox problem or just an Ipod HD problem - but it is saying my iPod is full (10 megs free) - even though I just deleted close to 500 megs of material.
04:52:33James79An error correction scan found nussing.
04:53:00James79Any thoughts/ideas would be fantastic.
05:01:45James79Hmm - everyone sleeping?
05:01:54Unhelpfulwhat, precisely, is claiming that it is full? a computer? the device itself?
05:02:20James79Both the Rockbox and the computer are saying there are only 10megs free.
05:02:58James79After seeing that, I went and deleted an additional 500megs of material. It is STILL claiming only 10 megs is free.
05:04:01James79Is there possibly some sort of DUMPFILE that is just growing and growing or something?
05:09:25 Join shadoxx [0] (
05:10:21James79It doesn't matter how much I delete - it still says I only have 10megs free...
05:10:43Unhelpfuli don't think that rockbox creates any such thing. i would run checkdisk with *all* of the checks turned on. if that doesn't work, i'm not really sure what to try next.
05:11:06shadoxxJames79: have you tried moving it to another PC?
05:11:30shadoxxOr rebooting to a different OS if you have it
05:11:52James79Not yet - that's next step - as I only have one the one.. but can borrow my roommates if need be.
05:11:57James79Running a full chkdsk now.
05:12:51shadoxxAlso, just wanted to direct anyone that's working on the YH-82X/92X ports.
05:13:33shadoxxGeneral inquiry. Been bugging me for a day now, and I still can't figure it out. I was really excited to try and help test the new port. But I can't even get any firmware to flash on the player.
05:22:00 Nick JdGordon is now known as JdGordon|afk (n=jonno@rockbox/developer/JdGordon)
05:23:13James79I think my ipod just had harddrive click - death... which would explain the problem
05:24:16shadoxxWhat gen?
05:24:33shadoxxJust drop it onto a carpeted surface a few times
05:25:10shadoxxTaking care not to crack the screen of course.
05:25:20James79yep - apparently it's also at mega low battery *my guess is all the futzing with scans and error checks and whatnot* - so I'm going to let it sit and charge up before I try anything else...
05:26:11James79Might not be clickdeath - that might have been just low battery CYCLE DIE CYCLE DIE... so, got it to the low battery charge screen.. gonna let it sit for a few hours before I go back to it. Ugh.
05:26:17James79It might be time to move to a new Mp3 player.
05:27:53shadoxxGet a YH-820 and help with the new port. :) Soon as I get mine working I'm seeing if I can throw a 32GB CFII card into it.
05:27:54James79Appreciate the help all
05:28:19James79Heheh, all I could do is bugfind - I'm not much of a programmer type.
05:28:32James79But damn if I'm not good at finding interesting ways to crash things =)
05:30:55James79Ok - I'm going to go - thanks again
05:30:55 Quit James79 ("CGI:IRC")
05:36:19*Unhelpful thinks that he can get the PF cache build to be *quite* a bit faster, but only by accessing the tagcache files directly. should be pretty safe as long as it checks the file versions, i'd think?
05:37:24***Saving seen data "./dancer.seen"
06:07:15 Quit intrados (Read error: 104 (Connection reset by peer))
06:08:21 Join intrados [0] (
06:14:12 Join fyre^OS [0] (
06:16:46 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
06:21:45 Quit kamlurker (Read error: 110 (Connection timed out))
06:22:58 Join jfc^3 [0] (
06:25:39 Quit jfc (
06:25:39 Quit Ridayah (
06:25:39 Quit Chex_ (
06:25:39 Quit tmzt (
06:25:39 Quit J-23 (
06:28:07NJoinRidayah [0] (
06:28:30NJointmzt [0] (
06:28:33 Join Chex [0] (
06:30:57 Quit fyrestorm (Read error: 110 (Connection timed out))
06:43:41cool_walking_Gigabeat S doesn't build with HAVE_MULTIVOLUME -
06:44:11Unhelpfulcool_walking_: does LBA48 work in rockbox? and what about over USB?
06:44:37cool_walking_Yes, I can get at the whole HDD from Rockbox and USB.
06:44:56Unhelpfuli'm not sure why you'd want HAVE_MULTIVOLUME, then?
06:46:52cool_walking_The OF creates the storage partition at ~127GB. I have enlarged it, but the OF loader (before Rockbox runs) doesn't like that, and forces me to do sendfirm again to get it working again. So now I am attempting to leave the OF-created storage partition alone and create a second one, hoping the OF loader will be okay with that.
06:48:07 Quit Horscht ("Verlassend")
06:48:18 Quit AndyIL (Read error: 110 (Connection timed out))
06:48:31Unhelpfulah. i didn't realize the OF had a problem there. we need to get rockboot in flash.
07:16:20 Nick fyre^OS is now known as fyrestorm (
07:24:23 Quit intrados (Read error: 104 (Connection reset by peer))
07:25:30 Join intrados [0] (
07:30:44 Nick flx_ is now known as flux (
07:31:03 Quit jmillikin (Read error: 110 (Connection timed out))
07:34:19 Quit r0b- (Read error: 110 (Connection timed out))
07:34:23 Join r0b- [0] (
07:37:27***Saving seen data "./dancer.seen"
07:40:49 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:44:14 Join bertrik [0] (
08:10:21 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:14:30 Join ender` [0] (
08:24:39 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:27:11 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:28:55 Join AndyI [0] (i=AndyI@
08:29:18 Join J-23 [0] (
08:32:24 Join Rob2222 [0] (
08:32:48 Join kugel [0] (n=kugel@rockbox/developer/kugel)
08:37:33 Quit safetydan ("Leaving.")
08:37:49 Join matsl_ [0] (
08:43:41CIA-38New commit by 03FlynDice (r21125): AMSSansa: clock-target.h and debug-as3525 now use AS3525_FCLK_PREDIV correctly. Default frequency scheme remains 248/62/62.
08:44:03 Quit bertrik (Read error: 113 (No route to host))
08:45:07 Quit perrikwp (" ajax IRC Client")
08:45:25 Quit intrados (Read error: 104 (Connection reset by peer))
08:46:34 Quit advcomp2019 (Read error: 113 (No route to host))
08:50:44 Quit Rob2223 (Read error: 110 (Connection timed out))
08:57:04 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
08:57:50 Join flydutch [0] (
09:01:53 Quit tvelocity (Remote closed the connection)
09:04:42 Quit r0b- ("( :: NoNameScript 4.22 :: )")
09:14:47 Join Thundercloud [0] (
09:18:36 Quit BHSPitLappy (Remote closed the connection)
09:20:45 Join FrankTM_ [0] (
09:32:00 Quit shadoxx (Remote closed the connection)
09:36:52 Quit FrankTM (Read error: 110 (Connection timed out))
09:37:28***Saving seen data "./dancer.seen"
09:40:54 Join Mathiasdm [0] (
09:41:20 Quit tessarakt ("Client exiting")
09:41:28 Quit kugel (Read error: 113 (No route to host))
09:54:29 Quit martian67 (Read error: 104 (Connection reset by peer))
09:57:08 Join robin0800 [0] (
10:02:31 Quit kachna (Read error: 113 (No route to host))
10:04:32 Quit Thundercloud (Remote closed the connection)
10:04:40 Join kugel [0] (n=kugel@rockbox/developer/kugel)
10:14:42 Quit flydutch (Read error: 110 (Connection timed out))
10:32:52 Quit kugel (Read error: 113 (No route to host))
10:34:40 Join kugel [0] (n=kugel@rockbox/developer/kugel)
10:44:14 Quit cool_walking_ (" ajax IRC Client")
10:54:28 Join joa_ [0] (
10:57:25 Join pyro_maniac [0] (
11:01:17 Part pyro_maniac ("Leaving.")
11:01:28 Join pyro_maniac [0] (
11:01:57Unhelpfulfor the logs: for things that may need to look at the index data, and then fetch several tag values for the same index, i'm adding some new bits to tagcache. struct index_entry is now declared in tagcache.h and named tagcache_index_entry, retrieve is now public and named tagcache_retrieve_index, and a new function tagcache_get_index fills in a tagcache_index_entry given an index ID and a struct tagcache_search.
11:06:04kugelUnhelpful: one more try: any opinion to FS #10258?
11:07:16kugelI realized it didn't come through when I asked it for the first time due to shaky internet here :/
11:10:02Unhelpfuli had *no* idea this would be this controversial, or even a little controversial.
11:10:18pondlifeIt's not *that* controversial ;)
11:10:24pondlifeJust a little
11:11:13pondlifeI'd go for a (bool album) parameter to the AA search, that changes the prioirity from track-first to album-first.
11:12:11 Quit joa (Read error: 110 (Connection timed out))
11:12:41Unhelpfulpondlife: one more paramater will mean using stack arguments on some targets. i think the same method for passing the info can stay, and it can just change whether the track-specific search is first or last, instead of skipping it.
11:13:05pondlifeAh, ok
11:13:11Unhelpfultrying every track if we don't find AA on the first one will probably never happen, though ;)
11:13:19pondlifeNo :)
11:13:41Unhelpfulthe idea of it taking 20x as long to start PF on a device *with no covers* frightens me ;)
11:14:05pondlifeWould it make sense to only check track art if the number of tracks = 1 ?
11:14:22pondlifei.e. Covers the cuesheet case..
11:16:04Unhelpfulbut "test track art last instead of first" will cover that, too, and probably be easier than checking the track count.
11:27:37pondlifeUnhelpful: Have you also seen ?
11:27:44pondlifeI hadn't until today.
11:28:30Unhelpfulpondlife: can you try this?
11:29:49Unhelpfuli've seen it before... i'm not sure that all of the pieces are the right way to do things. AA-in-DB is not a great idea because of size tagging
11:30:41Unhelpfuli think that playing/queuing a track or album from PF should happen *in* PF not, via a return value :)
11:31:08pondlifeBut the idea of being able to use PF for non-album selections is very appealing to me
11:31:16pondlifeI have too many albums!
11:31:45pondlifeAt least use of A-Z pre-selection would help lots..
11:33:24pondlifeUltimately, I'd like PF to be a third (database-based) browser (for targets with enough memory).
11:34:06 Join cool_walking_ [0] (
11:35:26pondlifeJust building a sim with your patch, what should I look out for?
11:36:06Unhelpfulyour track covers should work in PF again ;)
11:37:07pondlifeI don't use them ;)
11:37:19pondlifeI only use per-album art
11:37:30***Saving seen data "./dancer.seen"
11:37:49pondlifeMy input was mainly that PF and WPS AA should be as consistent as possible.
11:38:12pondlifeWith a view to PF being less album-based in the future.
11:39:58Unhelpfuli don't think that it will be supporting arbitrary selections soon. for one thing, every change of selection would mean redoing the covers cache. :/
11:40:19pondlifehehe, that's not practical ;)
11:41:09 Nick fxb__ is now known as fxb (
11:44:35Unhelpfulindeed it's not. after i work out the kinks in the new cover search optimization i'm working on, my next step was going to be coming up with some more-or-less invariant key to use for cache filenames in place of the index in the album list, so that *may* become feasible.
11:45:07pondlifeSHA-1 time (or some simpler hash)?
11:47:23pixelmapondlife: isn't the point of pf to show of a lot of albums (re. your A-Z pre-selection)? ;)
11:47:32Unhelpfulwell, actually, as i'm transitioning to using tag_filename to find every file in the master DB index (though only the first one found per album is used to search for cover art), i was thinking that the index in that table might work. as i understand it, we don't normally delete entries there, we just mark them as dead. i'm not sure whether inserts will always go at the end, though... otherwise, yes, some hash of the album name, with a
11:47:33pixelmashowing off too
11:47:33Unhelpfulfew bits for collision handling.
11:48:21pondlifepixelma: It's just a bit too slow to get to letter M albums when you've got lots.
11:48:48Unhelpfulpondlife: how about zoom-to-letter rather than show-only-letter, then?
11:49:12pondlifeYes, that would be useful indeed
11:49:48pixelmayeah, thought so now too. Also: it's not slow, it's just giving you time to enjoy all the art... ;)
11:50:16Unhelpfulit wouldn't even be hard, you could just pass a seek for the album table as an argument, since it has those for all albums ;)
11:51:03pixelmaand there already is a menu
11:51:16pondlifeCertainly having a <Covers> option in the tagnavi for album selections would be s nice thing...
11:52:46kugelpondlife: pf and wps weren't consistent before either
11:53:02pondlifeI never said they were, just that they ought to be...
12:00:53 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
12:02:42Unhelpfulthe trouble with tag_filename seems to be that if ramsearch is enabled, tagcache_search doesn't open any FDs.
12:04:07 Join VytenisS [0] (
12:04:50Unhelpfulfor some reason, get_next is *not* properly hitting the ramcache case, so that when the needed index file is found not to be open, it calls open_files, which doesn't read in the header or seek past it. then ecread tries to read in an entry from the current position, which is invalid because it's not positioned at an entry.
12:12:44kugelpondlife: from db to pf is easy
12:13:05kugelbut there's no sense in doing that until we can control playback
12:13:51 Join wincent [0] (
12:14:26pondlifekugel: For some reason I assumed that had already been done :/
12:14:51pondlifeMust have misread one of the commit messages a monh or so back
12:15:10Unhelpfulpondlife: it now functions during playback. that's pretty much it.
12:15:29pondlifeSo what's stopping it using the plugin playback control stuff?
12:15:54Unhelpfuli thought that stuff was just start/stop/pause and the like?
12:16:02pondlifeAh, ok. I guess new playlist/insert playlist are needed too.
12:16:26 Join l403 [0] (n=l@
12:16:39kugeli meant actually starting playback by selecting a cover
12:17:01pondlifeYes, or a track
12:17:07Unhelpfulkugel: hrm, but selecting a cover takes you to a tracklist.
12:17:27 Join gregzx [0] (
12:17:32Unhelpfuli think a "play" item in the context menu would be ok... and a play button could be added on targets that can spare one.
12:17:39kugelwe could start the album when selecting a cover, but still show the tracklist (allowing to pick single tracks too)
12:17:39 Quit Mathiasdm ("Yuuw!")
12:17:48pondlifeOR even a way to take you from the cover to the db browser for the tracklist - then use the existing context menu?
12:18:11pondlifeSeems to be less work than implementing a brand new track context menu
12:18:16Unhelpfulpondlife: but PF has its own fancified tracklist view.
12:18:25kugelI think that's more complicated than letting pf start the playback
12:18:32pondlifeYup, but I'd forego that for the intgration.
12:18:50pondlifeAs long as entry/exit into the picturelist could be achieved nicely..
12:18:58pixelmaI don't think selecting a cover should start playing the album - in every other browser it does not do that to. Selecting a track from the track list or a context menu on the cover
12:19:08pondlifeI agree.
12:19:13pixelmawould equal other browsers' behaviour
12:19:47Mikachuif you have lots of extra buttons, one could start playback and one could view the tracks?
12:19:51pondlifeThe cover = the album (so a context menu coiuld still play/insert it). But normally we just list the tracks.
12:19:54Unhelpfulsurely we could do that the same way that the DB browser does, though it might need a few new API exports...
12:20:16pondlifeThat's what I was thinking.
12:20:43pondlifeIf PF works nicely during playback, then we probably have much of the code in the DB browser
12:21:01kugelI imagine we just do another tagcache search when selecting an album to retrieve the tracks
12:21:32kugeland upon selecting the track, start playback in the same way as the db browser does
12:22:12pondlifeThe tricky bit would be exiting from the DB track list back into the plugin's picturelist...
12:22:23Unhelpfulpondlife: we already have a context menu mapping for track and album list. it does nothing right now. ;)
12:22:39Unhelpfulalso, there's a select mapping in the tracklist, but it does nothing.
12:22:41kugelwhy go to the db tracklist?
12:23:09pondlifeI don't think you need to reinvent the track list again. The db one is good, with a useful context menu...
12:23:59kugelI mean, why should we use the db browser track list, and not the pf one?
12:24:08pondlifeA seperate copy to do all of the playlist manipulation??
12:24:28Unhelpfulpondlife: we already *have* a reinvented tracklist, though... why not continue using it, but use the core DB browser code to handle the context menu and select action.
12:24:44kugelwe surely can just export the db browser context menu to pf
12:24:49pixelmaI would like if the pf tracklist wouldn't do this stupid scrolling, it really looks weird (and depending on display can blur the text a lot) if it only has to scroll a few pixels
12:25:10pondlifeThe "pretty" tracklist could certrainly be used for the DB and file browaer, but that's a seperate UI issue.
12:25:23pondlifeI don't presonally like it much though
12:25:32pondlifeMuch prefer the normal one
12:25:36pixelmame too
12:25:42kugelI'm talking about the case when an album was selected with pf
12:26:07pondlifeYes, rhen you go into a track list and want to do an Insert Shuffled, say?
12:26:07pixelmafor pf it's ok, but not for file browser/database view
12:26:36kugelpondlife: Sounds doable?
12:26:42pondlifeI don't see the need to have multiple versions of a text browser, that's all.
12:27:15pondlifePF brings the picture browser, in my mind.
12:27:44pondlifeDoesn't PF depend on the DB anyway? Or did I get that wrong?
12:28:16kugelyes it does
12:28:31 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
12:28:52pondlifeOK, so it's just a picture version of the album list we need, not even a seperate browser
12:29:30kugelwhat for?
12:29:48pondlifeTo integrate PF
12:30:04pondlifePromote it to being a proper browser, not a demo plugin
12:30:36kugelI thought we just do Database->Album->By cover
12:30:45kugelthat opens pf directly
12:30:45Unhelpfulpixelma: i've not actually forgotten that that is an annoyance in PF ;)
12:31:07pondlifeBut to select a track, use the DB tradk listing and then we get the context menus etc. all for free.
12:31:34pondlifeAnd return seamlessly to the album cover, naturally! ;)
12:31:35kugelwe should be able to export the context menu without rewriting
12:31:59CIA-38New commit by 03unhelpful (r21126): Change the search_albumart_files "ignore track art" option to check track art last.
12:32:33pondlifeWhy have dfferent track listing code? Just for the "pretty" thing?
12:33:17kugelof course
12:33:22kugelthat's the whole reason for pf
12:33:37Unhelpfulindeed, doesn't PF exist to be shiny?
12:33:39pondlifeYes, but the track listing is IMHO less pretty than the normal one.
12:33:56kugelI disagree
12:34:07pondlifeWhy not then have the shiny list everywhere (optionally). Would you not prefer that?
12:34:37pondlifeI like the shaded cursor bar more than the PF blue one, for example.
12:34:38kugelI would just prefer to not exit pf directly just after selecting an album
12:34:45pondlifeBut this is just opinion
12:34:55*Unhelpful only thinks that the wonky title scrolling is less pretty in PF
12:35:05Unhelpfuland that's a problem in the album list level as well
12:35:31 Join mt [0] (n=MTee@rockbox/developer/mt)
12:36:05pondlifeI don't see a great reason for a seperate list browser, that's all. But, hey, I don't have time to contribute much, so it's not up to me ;)
12:36:17pixelmapondlife: the blue bar is shaded too I think but less visible... I also dislike the gradient within the list too (but it's not that visible on the c200's display at least)
12:37:00Unhelpfulpondlife: i don't see any reason to chop it out when it's already there, though. this started as a clone of a desktop app, is there a compelling reason to change that part of it?
12:37:30pondlifeJust that I'd love to see PF feel part of Rockbox, not so plugin-y.
12:38:01pondlifeI could see myself using it as a browser, with a little work.
12:38:39kugelIn my view, pf could completely replace the db browser once you hit the album node
12:39:16kugelstart playback in pf, go directly to the wps, return to pf again when select is hit
12:40:23pondlifeI don't see a really good reaon that it should display in a differrent style, and not play well with configurable display settings etc..
12:40:25Unhelpfulkugel: PF would need to be a TSR to allow popping in and out of it, and TSR plugins are not allowed to use the plugin buffer.
12:40:26pixelmaonly in an "all albums" node, you could have your "album" part divided how you want with tagnavi (e.g. A-Z preselection)
12:41:34Unhelpfuli would much rather see a special item in tagnavi for PF. i generally want to see many album titles on the screen at once, and to have the random and all options at the top.
12:42:09 Quit martian67 (Remote closed the connection)
12:42:14pondlifeUnhelpful: Yes, I agree. Just have a <Covers> option where PF can cope with it.
12:42:17kugelUnhelpful: no, not really
12:42:37kugel"TSR plugins are not allowed to use the plugin buffer." and that is wrong, IIUC
12:43:15kugelother plugins cannot be started as long as a TSR runs since they block the plugin buffer
12:43:18pixelmait's not, TSR plugins should allow other plugins to run so as I understand it should not use it themselves
12:43:18Unhelpfuli was told that a month or so ago, when vis plugins were being discussed.
12:43:25pondlifeIs there a reason to unload a plugin until another plugin is loaded?
12:43:42kugelnot really
12:44:02kugelunless the plugin in question boosts which might not be desirable
12:44:09pondlifeCould PF perhaps also work with other background colours? Or is the black essential for the effect?
12:44:15kugelbut pf can be exited just fine and restarted by the wps
12:44:17Unhelpfulpixelma: other plugins can't load anyway, since the part of the buffer where the TSR plugin loaded is still occupied. i was told, however, that plugin_get_buffer must not be used by TSR plugins.
12:44:39Unhelpfulpondlife: there is no blending done on the reflections, that would be a bit of a problem.
12:44:46 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
12:45:02pixelmaI thought you should be able to play some games while running a battery_bench
12:45:13pixelmafor example
12:45:32pondlifeUnhelpful: Yup, thought so. Pity.
12:46:12Unhelpfulthe reflections *could* be blended, instead of merely faded, but it would cost a bit
12:47:00kugelpf doesn't need to be tsr, and I don't think it would be useful if it was a tsr
12:48:05Unhelpfulpixelma: but battery_bench is still loaded in the plugin buffer. the stack for its background thread is there, its code is there, etc.
12:49:11Unhelpfulif you try to run another plugin, the tsr hook that battery_bench set up will run, and tell you that it's running, and ask if you want to quit it so that you can run the other plugin.
12:49:32pixelmahmm, ok
12:52:04pondlifeI'd think it could be unloaded/loaded as necessary, but that the plugin handler could be lazy and just reuse the in-memory image if it's the same plugin that's already loaded. This requires that all plugins explicitly zero their variables on init, of course...
12:52:24pondlifeSo it's like unload/reload, but disk spin-up might not happen.
12:53:01 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
12:53:03Unhelpfulpondlife: or that they detected being pre-initialized, and jump to a special entry point :)
12:53:37pondlifebtw your patch above - seems to have stopped the WPS picking up folder.jpg
12:53:54Unhelpfulhuh? i don't see how... :/
12:54:35pondlifeHmm, I'll unpatch and rebuild to check
12:55:06Unhelpfuloh, the one i linked to? i tested in a sim, found some bugs, and committed a fixed version :)
12:57:33Unhelpfulthe one i linked to will actually fail all cover art searches, and only find track art. :)
12:57:42pondlifeI'll svn up
12:58:26CIA-38New commit by 03unhelpful (r21127): Don't recheck for track art when checking track art first.
13:06:43 Quit ender` (Read error: 110 (Connection timed out))
13:09:23 Join sandhya [0] (n=sandhya@
13:11:23sandhyaHi, I installed rockbox on my 1st gen nano a few days ago... and I don't know if I unplugged it incorrectly or not, but I'm getting mountpoint readonly error messages when I try accessing it now via rockbox
13:12:18sandhyaany ideas how I make it writeable again?
13:12:51sandhyaI currently cannot see any text on the rockbox menu system on the nano and was looking to install (again) the fonts package
13:13:56pixelmadid you try to reset the Ipod and get it into disk mode? And for the second problem, you can reset settings at startup, should be described in the manual
13:14:46pixelmaI don't think it's a font package issue, guess it's more likely same fore- and background colour
13:14:49gevaertsyou probably have a corrupted filesystem
13:15:32*pixelma leaves help to people who actually own an Ipod ;)
13:16:17Horschtsound like he's using linux.
13:16:38HorschtI can't remember the command right, but i think fsck is what he'd need to run
13:18:27tmztsandhya: you mean linux is remounting it readonly?
13:19:11tmztwhat do you mean, accessing it now via rockbox?
13:21:01sandhyatmzt: yes, it appears to be remounting it readonly... I should also have probably mentioned that I'm using Kubuntu Jaunty
13:21:17sandhyatmzt: accessing it via the rockbox installer rather
13:21:23tmztdmesg |grep remounting |tail -n1
13:21:28tmztdmesg |grep remounting |tail -1
13:22:09Bagdersandhya: remounting read-only is a clear indication the fs driver found problems with it
13:22:12 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:22:56sandhyatmzt: that returns nothing
13:23:21tmztinteresting, check mount
13:23:35sandhyaBagder: any fixes?
13:24:16sandhyatmzt: /dev/sdc2 on /media/KUMAR'S IPO type vfat (rw,nosuid,nodev,uhelper=hal,uid=1000,utf8,shortname=mixed)
13:24:27Bagdersandhya: yes, fsck your partition
13:24:44Mikachuthe kernel remounting it won't change what mount says
13:24:47Bagderrw is read+write
13:24:51tmztoh, thanks
13:24:56tmztdidn't know that
13:25:14tmztcat /proc/mounts should though
13:25:20 Quit perrikwp (" ajax IRC Client")
13:25:25tmztmakes sense since mount uses /etc/mtab by default
13:25:26 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
13:25:29sandhyals output: drwxr-xr-x 7 sandhya root 4096 1970-01-01 05:30 KUMAR'S IPO
13:25:41tmztthat won't help
13:26:05tmztumount /media/KUMAR's\ IPO
13:26:11Mikachuwow, keyboard.c is a mess
13:26:14tmztfsck.vfat /dev/sdc2
13:26:22Mikachuhangul joining intermingled with input handling
13:26:41Mikachuyou'll need -r or it won't actually fix anything
13:27:06sandhya"FATs differ but appear to be intact. Use which FAT ?"
13:27:17Mikachuselect one or the other
13:27:22sandhyafirst or seconf FAT?
13:27:31Mikachuyour guess is as good as ours
13:27:50 Quit kugel (Read error: 110 (Connection timed out))
13:28:04sandhyaI get a reclaimed clusters message.. and a correct? prompt
13:28:16cool_walking_just say yes to everything
13:28:57tmztMikachu: thanks again
13:28:58Mikachuif you gave -r, it will ask if it should write at the end, otherwise it will quit without doing anything
13:29:14Mikachutmzt: for what? :)
13:29:47sandhyaI reran the fsck and selected the other FAT and got a similar message with a list of (corrupt?) files as well.. four fnt files and two log files
13:29:48tmztMikachu: reminding me, I didn't know it was different from fsck.ext2 in that way
13:30:22Mikachuthe tricky thing with fsck.vfat is even if you run with -r and fix everything, running it again can find more errors
13:30:34Mikachuand even if it doesn't, i'm not sure if you can be sure there's nothing wrong :)
13:31:09sandhyaI reran it twice with -r and no errors returned now
13:31:30sandhyaumount and remount now?
13:31:48cool_walking_it shouldn't have been mounted when you did the fsck...
13:31:58Mikachuit's okay if it was ro mounted
13:32:00Mikachubut yeah :)
13:32:20sandhyaoh :o
13:35:17sandhyareinstalled the fonts package and now the menus are visible :)
13:36:32 Join __lifeless [0] (n=lifeless@
13:36:45 Quit _lifeless (Remote closed the connection)
13:37:33***Saving seen data "./dancer.seen"
13:52:38 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:53:20 Quit robin0800 (Read error: 104 (Connection reset by peer))
13:59:18 Quit pyro_maniac (Remote closed the connection)
14:05:29 Join MarcGuay_ [0] (
14:05:31 Join einhirn [0] (
14:06:05 Join kachna [0] (
14:25:14amiconnkugel: There is a reason why TSR plugins must not use the plugin buffer. Other plugins aren't the issue, the core is
14:25:35amiconnThe core uses the plugin buffer for a few things (e.g. playlist viewer)
14:26:50amiconnUnhelpful: DB indices will change if the database is rebuilt (and I always rebuild the full db when disk content changed, as I don't trust the 'update function)
14:29:06linuxstbamiconn: Any opinion on a solution to ? It looks like we need more ARM defines to deal with differences within an architecture version.
14:29:51sandhyaI've been playing with rockbox for the last hour or so... and it's simply amazing :)
14:30:29sandhyathanks for everybody's help in getting me up and running! and thanks to all the devs too :) Highly nifty stuff!
14:30:40 Quit sandhya (Remote closed the connection)
14:38:42amiconnlinuxstb: I think 'configure' specifies the wrong -mcpu option for it
14:39:35 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:40:03linuxstbamiconn: I thought the problem was that the arm946e-s was ARMv5TExP, which doesn't have ldrd and strd - so gcc is correct to error out on those instructions.
14:42:19 Join flydutch [0] (
14:43:11Tornelinuxstb: that's correct, yes
14:43:55Torneif it's a rev 0
14:44:02Torneit was fixed in rev 1
14:44:08Tornebut SoC vendors rarely bother to update :)
14:44:28linuxstbTorne: Hmm... That's even worse... :(
14:44:38Torne <- is a reference
14:44:45 Join robin0800 [0] (
14:45:00Tornebut it's also confirmed by my local arm guru
14:46:16Tornemy experience of ARM SoCs is that the vendors are in such a hurry to get to the market they are always using r0 of the core and they never want to change it later because it'll break all the bugs their customers are depending on ;)
14:46:36Tornewe have to turn off loads of undocumetned weird bits in crazy control registers to make half the ARM11 cores here work right
14:46:53Tornebecause TI are happy with r0 :)
14:48:41Torneso yah, you'll probably need a new set of defines for v5TExP :)
14:50:43 Quit kugel (Read error: 110 (Connection timed out))
14:50:58linuxstbI guess one solution would be to add a #define ARMv5TExP and change the #if around the use of strd and ldrd to be "#if ARM_ARCH >= 5 && !defined(ARMv5TExP)"
14:51:16Tornesymbian has a good 200+ lines of header for setting defines for ARM processor features
14:51:28n1s#define CPU_ARMv5TE_WITH_CRAZY_BUGS
14:51:34Tornewe generalyl don't do stuff conditional on arch because of things like this, and go with per-feature macros
14:51:35 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:51:47Tornen1s: yeah that happens
14:52:01Torne#ifndef __CPU_ARM1136_ERRATUM_353494_FIXED :)
14:52:09LambdaCalculus37MarcGuay: Ping
14:52:29linuxstbI think currently most of our ARM code (apart from CPU/SoC specific inits) is just ARMv4, although there's a little ARMv5+ - such as that troublesome code in libdemac
14:52:38linuxstbSo I guess this will get worse in the future.
14:52:46Torneyah, you have a smaller range of CPUs for now
14:52:55Tornei guess rockbox has never had to run on XScale for example
14:53:10Tornewhich is Weird And Different in loads of fun ways and is halfway between v4 adn v5
14:53:21Torneand even then has the manzano variant which is different again
14:53:32linuxstbNo, but it does run on one ARMv6 chip - iMX31
14:54:00Tornewe're up to, hm
14:54:02Tornefourteen processors
14:54:09Torneplus defines for various CPU errata
14:54:17Tornenwo that's fun to write code for :)
14:54:39linuxstbTorne: BTW, who are you? ;)
14:54:50Tornei'm a kernel developer for symbian
14:54:54Tornewell technically now nokia as they bought us
14:54:58Tornebut i still work on the symbian kernel
14:57:51MarcGuayLambdaCalculus37: You rang?
14:58:39Torneeventualyl nokia will actually release our code
14:58:49Tornethen you could crib all our hacky ifdefs! :)
14:59:06LambdaCalculus37MarcGuay: Got my m200 with me today.
14:59:08kugelamiconn: what does battery_bench use?
14:59:26MarcGuayLambdaCalculus37: You said it's v3/SD, right?
14:59:38LambdaCalculus37MarcGuay: Yeppers.
14:59:57MarcGuayLambdaCalculus37: I think the goal with those is to study what's been done with the Sansas and port it.
15:00:12MarcGuayCurrent code doesn't get past the NAND ID check.
15:03:50 Join pyro_maniac [0] (
15:04:54LambdaCalculus37MarcGuay: On neither the c100 nor m200.
15:05:41linuxstbTorne: Any interest in porting Rockbox to run as a Symbian app? I'ld use it...
15:05:53Tornei've never written a symbian app
15:06:01n1sI'd try it at least
15:06:06linuxstbNow's your chance! ;)
15:06:11Tornewell, not one that runs on {ei,av,ui}kon
15:06:26Tornei only ever boot the textshell on our devboards, and write command line unit tests :)
15:06:45Torneactually programmign for the higher levels of symbian is so painful i gave up
15:06:52Tornei am not proud of our end user APIs :)
15:07:00Tornefortunately i'm also not *responsible* for them so hey
15:07:07Tornei can just throw my hands up and say "at least our kernel is nice"
15:07:20linuxstbThat was my experience - I tried Symbian programming in the very early days, and just gave up, very confused...
15:07:22Torneand generally blame S60 for all the slow crappiness of actual phones
15:07:33Tornemumble grumble 80kb stacks mumble
15:07:48amiconnkugel: The plugin itself is of course loaded into the plugin buffer, and that is no problem. However, a TSR plugin must not use plugin_get_buffer(), it can only use static allocations
15:08:10kugelah, I see
15:08:14amiconnThe core functions which use the plugin buffer use plugin_get_buffer(), so it is safe this way
15:08:16Tornelinuxstb: it was less horrible in the EPOC32 r5 era, Psion 5 and so on
15:08:19LambdaCalculus37Torne: While I feel your pain, come over to #rockbox-community if you want to further tell us any war stories about programming Symbian. ;)
15:08:27Torneok, sorry
15:08:32Tornei'll stop being offtopic :)
15:08:51*LambdaCalculus37 shows Torne the way to #rockbox-community
15:09:06amiconnlinuxstb: We might need ARM capability bits
15:09:40linuxstbamiconn: That could be nice. Although any idea what to call them?
15:10:05amiconnARM_FEATURE_LDRD, ARM_FEATURE....
15:11:02amiconnOr s/FEATURE/CAP/g
15:12:11linuxstbamiconn: From the ARM ARM - "Some early implementations of the E variant omitted the LDRD, STRD, MCRR, MRCC and PLD instructions." - are you thinking of one bit for each instruction?
15:12:36Torneshouldn't need one for each, but you might want a seperate bit for PLD anyway
15:12:45Tornebecause even on chips where PLD is implemented it doesn't necessarily do anything
15:13:41Torneas far as the ARMv5TExP thing goes though, it should either have all of those or none
15:23:01 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
15:23:51 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
15:25:16 Quit joa_ ("( :: NoNameScript 4.22 :: )")
15:25:54 Join evilnick [0] (i=0c140464@gateway/web/ajax/
15:29:19 Quit MarcGuay (Read error: 60 (Operation timed out))
15:29:39 Join MarcGuay [0] (
15:36:16CIA-38New commit by 03pixelma (r21128): Make 'Backlight mod' an Advanced build option for the Ondios (largely inspired by the RTC mod related changes in r12555). Remove the defines in the ...
15:36:43Horschtdoes somone have an idea about FS #10257?
15:37:35***Saving seen data "./dancer.seen"
15:39:55 Quit perrikwp (" ajax IRC Client")
15:40:17 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
15:40:26 Quit kachna (Read error: 60 (Operation timed out))
15:44:47linuxstbHorscht: So it's fine in the jpeg viewer, but corrupted when used as album-art? Is it being resized, or being displayed 1:1 ?
15:45:38Tornecan I get write access to the wiki (TorneWuff)?
15:45:51linuxstbOne moment...
15:46:18pixelmareminds me of a similar problem the scaler had with bmp on greyscale once... let's see if I can find the fix
15:46:43linuxstbTorne: That should be done
15:47:04Torneyup, thx.
15:53:31 Nick fxb is now known as fxb__ (
15:54:38 Quit froggyman ("CGI:IRC")
16:02:00 Join AndyIL [0] (n=pasha_in@
16:05:34 Quit AndyI (Read error: 60 (Operation timed out))
16:09:03 Join kugel_ [0] (
16:09:11 Quit kugel (Nick collision from services.)
16:09:17 Nick kugel_ is now known as kugel (
16:09:49kugelUnhelpful: any idea how we can get a struct tree_context from the album selection?
16:18:13 Part LinusN
16:23:41 Quit domonoky (Read error: 104 (Connection reset by peer))
16:26:05MarcGuay_"undefined reference to 'get_context_mapping'" even though I've created keymap-c100.c in apps/keymaps?
16:28:00MarcGuay_Ping JdGordon|afk
16:29:26cool_walking_MarcGuay, I think he said he won't be back till Sat night.
16:33:09MarcGuay_Forgot SOURCES - thanks anyways.
16:49:21 Join Guest81638 [0] (n=quassel@
16:49:59 Part Guest81638 (" - Chat comfortably. Anywhere.")
16:55:06 Join petur [50] (n=petur@rockbox/developer/petur)
16:58:36*kugel has a very hackish way to start pf in the db browser AND start playback within pf
16:59:25evilnickkugel: Ooooh, that's exciting!
17:01:52 Quit pyro_maniac ("Leaving.")
17:02:00 Join JimBastard [0] (i=42f65302@gateway/web/ajax/
17:02:06 Quit matsl_ (Read error: 110 (Connection timed out))
17:02:07JimBastardSup party people
17:02:14kugeljust need to figure out how to play the actually selected album
17:02:23JimBastardI met one of your people last night at the NYC hacker news meetup
17:02:29JimBastardcool project i must say
17:02:35evilnickJimBastard: Was that Robert?
17:02:37pondlifekugel: heh, play a random album :)
17:02:49JimBastardi have his name in my netbook, i forgot already
17:02:54JimBastardhe was kinda intense, but cool
17:03:51LambdaCalculus37JimBastard: That would be me. Join #rockbox-community and we'll chat more.
17:04:06JimBastardholler back
17:04:10 Join funman [0] (n=fun@rockbox/developer/funman)
17:07:25funmanLambdaCalculus37: could you test mkamsboot OSX binaries?
17:07:48LambdaCalculus37funman: Only briefly, but they work.
17:07:59LambdaCalculus37Oh wait.. you mean now?
17:08:12funmani mean with the test script you downloaded from me
17:08:38LambdaCalculus37funman: I was able to build successfully with the test script.
17:08:49funmanjust to be sure they generate the same files than linux and win32 version (did you check the md5sums?)
17:10:20 Quit cool_walking_ (Remote closed the connection)
17:10:43martian67I’d just like to interject for a moment. What you’re referring to as Linux, is in fact, GNU/Linux, or as I’ve recently taken to calling it, GNU plus Linux. Linux is not an operating system unto itself, but rather another free component of a fully functioning GNU system made useful by the GNU corelibs, shell utilities and vital system components comprising a full OS as defined by POSIX.
17:10:43martian67Many computer users run a modified version of the GNU system every day, without realizing it. Through a peculiar turn of events, the version of GNU which is widely used today is often called “Linux”, and many of its users are not aware that it is basically the GNU system, developed by the GNU Project.
17:10:43martian67There really is a Linux, and these people are using it, but it is just a part of the system they use. Linux is the kernel: the program in the system that allocates the machine’s resources to the other programs that you run. The kernel is an essential part of an operating system, but useless by itself; it can only function in the context of a complete operating system. Linux is normally used in combination with the GNU operating syst
17:10:48martian67em: the whole system is basically GNU with Linux added, or GNU/Linux. All the so-called “Linux” distributions are really distributions of GNU/Linux.
17:10:59linuxstbmartian67: ?
17:11:38*evilnick throws martian67 into either #rockbox-community or back to 4chan
17:11:54*martian67 flashes a toothy grin
17:12:02LambdaCalculus37funman: No, I haven't. Want me to check the md5sum of just the binary?
17:12:04martian67sorry, i was just waiting for somone to say linux :D
17:12:17*LambdaCalculus37 flashes a big shiny boot at martian67 :D
17:12:31funmanLambdaCalculus37: should generate md5sums of 5 binaries produced with mkamsboot, just compare them with the md5sums already in the 'sums' file
17:12:39LambdaCalculus37funman: Okay.
17:17:47 Join sansa [0] (n=405afa2f@gateway/web/cgi-irc/
17:17:50Horschtlinuxstb, sorry was afk
17:17:54Horschtyes, correct
17:18:13Horschtthe jpeg works fine in the rockbox jpeg viewer, but not when used as album art
17:18:21sansadoes anyone know how the sansa clip rockbox project is going
17:18:26Horschtnot being resized
17:18:34Horschtbeing displayed 1:1
17:18:54funmansansa: have a look at
17:19:18sansalooking now
17:20:46sansaoh looks difficult i am a complete nood and cant compile
17:21:42 Quit sansa (Client Quit)
17:21:50Horschtalso, linuxstb, it does not seem to realy be related to all overly dark albumarts. I have other covers almost only black (Black album by Metallica, Spinal Tap) which, as mentioned, are almost fully black, but they show up correctly. It must be either a coincidence that those 2 albumarts that I attached are very dark, or something special about them
17:22:51linuxstbHorscht: OK. I can't help, I was just probing you for more info...
17:23:08linuxstbHopefully Unhelpful will see it.
17:29:44MarcGuay_linuxstb: Could you propose a temporary solution to the problem described in
17:30:13Torneifdef it for that platform, as a temporary hack?
17:30:29Torneor do you mean "what instructions to use instead"? :)
17:30:42linuxstbThere is already an alternative (for ARMv4)
17:31:08 Quit petur ("connection reset by beer")
17:31:12Torneoh, yes, of course. so just hack up the ifdef, until someone decides on how to deal with this nicely? :)
17:32:15MarcGuay_Should I disable this codec, then? I'm unsure where to do that's what's being suggested.
17:32:21linuxstbSo yes, changing the #if to something like "#if ARM_ARCH >= 5 && CONFIG_CPU != CPU_TCC770" should work. (but isn't nice to commit...)
17:33:08linuxstbMarcGuay_: Look in apps/codecs/demac/libdemac/predictor-arm.S and change the "#if ARM_ARCH >= 5" lines to what I suggested.
17:33:27MarcGuay_linuxstb: Thanks.
17:33:32*MarcGuay_ stops grepping
17:34:37linuxstbMarcGuay_: Make that "TCC770", not "CPU_TCC770"...
17:34:56*MarcGuay_ stops making
17:34:59MarcGuay_:) thanks
17:35:32linuxstbMarcGuay_: Or even better, use "... && !defined(CPU_TCC77X)"
17:35:40*linuxstb thinks he's finished now...
17:37:36***Saving seen data "./dancer.seen"
17:40:19*LambdaCalculus37 wonders if that'll help the m200 as well
17:41:59 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:43:33 Join ender` [0] (
17:45:52linuxstbLambdaCalculus37: Yes, you'll get the same problem there.
17:46:38linuxstbI guess that fix isn't too bad, but it seems nasty to commit it instead of solving it properly...
17:47:44 Join jmillikin [0] (
17:52:28 Join {phoenix} [0] (
17:53:20LambdaCalculus37linuxstb: For now, it'll make do as a patch, but not for final commit.
17:55:59MarcGuaylinuxstb: Still getting a bad checksum error. Checksum: 63313030, Sum: 2c60271 mean anything to you?
17:57:49MarcGuay_modelname is a garbled character, as well...
17:59:27 Join mcuelenaere [0] (
18:02:55 Join Lss [0] (
18:03:03linuxstbmodelname is read from the start of the rockbox binary, so that would point to NAND trouble.
18:04:46 Join stoffel [0] (
18:15:48 Part pondlife
18:19:19 Quit kugel (Read error: 110 (Connection timed out))
18:19:36 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:20:14 Join _lifeless [0] (n=lifeless@
18:20:49 Join Corto [0] (n=d4f6430a@gateway/web/cgi-irc/
18:24:33CortoHi! I just built bootloader & rockbox and installed them on my Sansa Fuze 8GB for the first time...
18:24:41Cortoit dual boots all ok, screen saver (black screen on sleep) works and reacts keys, but all I can see is the Rockbox logo and version info..
18:25:10Cortohave I missed something? (I have never used Rockbox before)
18:25:59linuxstbWhat does the version info say?
18:26:09 Quit SirFunk__ (Read error: 110 (Connection timed out))
18:26:19 Quit {phoenix} (Remote closed the connection)
18:26:43CortoVer. r21128-090529
18:27:37linuxstbYou should only see that screen for a second or so, then the Rockbox main menu should appear. So it sounds like something is wrong (which is to be expected on the Fuze at the moment I think).
18:28:21 Join SirFunk__ [0] (
18:28:23Cortoah, ok...
18:28:47MarcGuaylinuxstb: I had the wrong OFFSETs in the config. Rockbox boots on the c100.
18:29:06MarcGuayVery neat.
18:29:35MarcGuayOh keymapping, how I dread you.
18:30:05linuxstbDoes it have similar buttons to any other target?
18:30:24MarcGuayPretty similar to the c200. Which doesn't help much!
18:31:33Cortolinuxstb: ok, thanks! I guess I'll keep checking the forum and status page on wiki now and then and try again later
18:33:17 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
18:33:38 Join cmike90 [0] (n=3f419a1a@gateway/web/cgi-irc/
18:35:06cmike90whats with the mp4 files causing my sansa to go slow
18:36:46 Quit cmike90 (Client Quit)
18:37:17 Quit __lifeless (Read error: 101 (Network is unreachable))
18:37:21 Join cmike90 [0] (n=3f419a1a@gateway/web/cgi-irc/
18:37:45 Quit funman ("leaving")
18:37:47 Join __lifeless [0] (n=lifeless@
18:38:52 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:39:05 Quit cmike90 (Client Quit)
18:41:57 Quit _lifeless (Read error: 110 (Connection timed out))
18:42:28 Quit kugel (Read error: 110 (Connection timed out))
18:49:02 Join SirFunk_ [0] (
18:55:00 Quit l403 (Remote closed the connection)
19:01:20 Quit parafin (Remote closed the connection)
19:01:48MarcGuayHow does audio work on the D2/Iaudio7 if "Due to this an "Error Accessing Playlist Control" will occur when starting playback"?
19:01:58MarcGuayJust PCM beeps?
19:02:15 Quit SirFunk__ (Read error: 110 (Connection timed out))
19:02:16 Quit flydutch ("/* empty */")
19:06:42MarcGuayGentleman mails says otherwise...
19:06:54 Join {phoenix} [0] (
19:15:13 Join stripwax [0] (
19:15:26 Join funman [0] (n=fun@rockbox/developer/funman)
19:17:00 Quit SirFunk_ (Read error: 60 (Operation timed out))
19:21:32 Join SirFunk_ [0] (
19:23:22*mcuelenaere thinks Lua really needs bitlib..
19:28:26gevaertsMarcGuay: it just doesn't keep state permanently
19:30:39 Join PaulJam [0] (
19:31:56 Quit perrikwp (" ajax IRC Client")
19:32:08 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
19:34:22 Join miepchen^schla [0] (
19:37:40***Saving seen data "./dancer.seen"
19:37:44 Join tvelocity [0] (
19:46:58 Join alexbobp_ [0] (
19:47:25 Join bertrik [0] (
19:51:01 Quit alexbobp (Nick collision from services.)
19:51:08 Nick alexbobp_ is now known as alexbobp (
19:55:34 Quit jfc^3 (Read error: 104 (Connection reset by peer))
19:56:43MarcGuaygevaerts: I don't understand. Could you say it again another way?
19:59:18MarcGuaygevaerts: Do you mean it will play the song after the error?
20:00:08 Join mirak [0] (
20:05:07gevaertsthings like resuming after stop won't work
20:05:20 Join jfc [0] (
20:08:47 Quit alexbobp (Read error: 60 (Operation timed out))
20:08:59 Join alexbobp [0] (
20:10:31 Quit bubsy (Read error: 60 (Operation timed out))
20:19:14MarcGuay_#error "re-run configure this just so wrong" - is this a joke or bad english? top of config-h100.h
20:20:18 Join toffe82 [0] (n=chatzill@
20:20:54gevaertsBagder should be able to tell us
20:22:30BagderI have no sense of humor so it must've been a mistake then!
20:22:43funmanit was a safety when h120 was renamed it seems
20:22:54Bagdereven though I think those three top lines should be cut out
20:23:19gevaertsmost people probably have re-run configure since then
20:24:01 Quit mcuelenaere (Read error: 60 (Operation timed out))
20:24:29 Join xnyhps [0] (n=xnyhps@2001:470:1f14:da:219:e3ff:fed7:c57c)
20:24:41Bagderprobably, yes
20:27:52CIA-38New commit by 03gevaerts (r21129): Remove an old safety making sure that people re-ran configure after the h120/h100 split. People now have had nearly four years to do so, which is ...
20:28:08 Join moos [0] (i=mustapha@rockbox/staff/moos)
20:33:23LambdaCalculus37MarcGuay: You're able to read the internal NAND on the c100 now?
20:33:33*LambdaCalculus37 is referring to the TargetStatus wiki page
20:35:35MarcGuay_LambdaCalculus37: It loads Rockbox, so yessir, she reads.
20:36:14LambdaCalculus37MarcGuay_: I can't wait for the "Ladies and Gentlemen..." mail. ;)
20:36:19MarcGuay_Seeing if i can get radio rocking now...
20:36:55LambdaCalculus37MarcGuay_: Is it using the slightly kludgey method we talked about earlier?
20:37:51MarcGuay_LambdaCalculus37: Iffing out the TCC770 more that codec? Yes.
20:39:34LambdaCalculus37MarcGuay_: We should look into a cleaner, more "commit-ready" solution for that.
20:40:27MarcGuay_LambdaCalculus37: That's so far out of my reach I can't even see it. Somebody who's comfy with compilers will have to take care of it.
20:41:11*Torne suggests just defining on/off feature macros for subsets of the ARM instruction set, instead of relying on comparisons with the architecture version :)
20:41:30Tornehave a big header file somewhere that turns the right things on and off depending what cpu it is
20:41:52Tornetry and structure it so the first letters of each macro spell out dirty words insulting ARM, if you can
20:43:09 Join _lifeless [0] (n=lifeless@
20:43:12 Join mcuelenaere [0] (
20:47:47 Quit __lifeless (Read error: 60 (Operation timed out))
20:55:22 Quit BHSPitLappy (Read error: 60 (Operation timed out))
20:56:48 Quit Corto ("CGI:IRC")
21:03:09 Quit AndyIL (Remote closed the connection)
21:03:24 Join AndyI [0] (n=pasha_in@
21:03:24 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
21:03:26 Join SirFunk__ [0] (
21:03:39 Quit SirFunk_ (Read error: 60 (Operation timed out))
21:03:47 Join mcuelenaere [0] (
21:05:46 Quit stripwax ("")
21:11:02 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
21:11:46*MarcGuay_ tries to imagine why the radio code is in the "recorder" directory...
21:12:26funmanhistorical reasons?
21:15:02 Quit xnyhps ("Leaving.")
21:15:07 Join xnyhps [0] (n=xnyhps@2001:470:1f14:da:219:e3ff:fed7:c57c)
21:15:35 Quit xnyhps (Client Quit)
21:16:09 Join xnyhps [0] (n=xnyhps@2001:470:1f14:da:219:e3ff:fed7:c57c)
21:16:55 Quit xnyhps (Client Quit)
21:17:24 Quit perrikwp (" ajax IRC Client")
21:17:47 Quit saratoga (" ajax IRC Client")
21:18:51 Join saratoga [0] (i=9803c6dd@rockbox/developer/saratoga)
21:23:24 Quit AlexP (Read error: 110 (Connection timed out))
21:28:26 Quit SirFunk__ (Read error: 110 (Connection timed out))
21:36:20 Join SirFunk__ [0] (
21:37:43***Saving seen data "./dancer.seen"
21:41:05 Join obo [0] (
21:46:28 Quit moos (Read error: 110 (Connection timed out))
21:49:25 Quit funman ("leaving")
21:52:21 Join kugel [0] (n=kugel@rockbox/developer/kugel)
21:54:18 Quit mirak (Remote closed the connection)
21:55:54 Join froggyman [0] (n=47ba0b80@gateway/web/cgi-irc/
21:56:15 Quit SirFunk__ (Read error: 110 (Connection timed out))
21:57:00 Join SirFunk__ [0] (
21:59:10 Quit LambdaCalculus37 (" ajax IRC Client")
22:00:02 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
22:05:09 Quit ender` (" We know the speed of light... so, what's the speed of darkness?")
22:06:56 Join merbanan [0] (
22:08:41 Join __lifeless [0] (n=lifeless@
22:09:07 Join mirak [0] (
22:14:10 Join ender` [0] (
22:15:01 Join perfectdrug [0] (n=5b0ed874@gateway/web/cgi-irc/
22:15:59perfectdrughi i just created FS #10262 svg image for the sansa fuze :)
22:19:17 Quit froggyman ("CGI:IRC")
22:19:39CIA-38New commit by 03FlynDice (r21130): AMSSansa: Add AS3525_DRAM_FREQ as a configurable frequency. Attempts to use PCLK != DRAM still fail but a method is now in place. Default scheme ...
22:20:02 Join froggyman [0] (n=47ba0b80@gateway/web/cgi-irc/
22:20:34pixelmaperfectdrug: I'll look into committing them. About the side button - I usually drew them even if they wouldn't be visible from the top in reality. That makes it easier to label them
22:22:28perfectdrugpixelma: ok so let me add the keymaps for this file, too. i also changed the clip-front.svg a little as you may have seen
22:22:43pixelmayes, saw it
22:23:50perfectdrugi made the fuze more like your files
22:25:32 Quit _lifeless (Read error: 113 (No route to host))
22:27:30 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
22:29:20kugelyou made a fuze one \o/
22:29:59amiconnHmm. 'make reconf' doesn't seem to work for advanced builds (at least it doesn't for voice builds)
22:30:51 Quit n1s ("Lmnar")
22:32:44 Quit stoffel (Read error: 113 (No route to host))
22:36:18 Quit bmbl ("Woah!")
22:44:12Unhelpfulamiconn: you'd say it's worth hashing the album name, then?
22:44:59 Join moos [0] (i=mustapha@rockbox/staff/moos)
22:45:22kugelI'm not far from having pf in the db browser
22:46:15Bagderamiconn: right, reconf is not quite complete (yet)
22:47:00UnhelpfulHorscht: thanks for the heads-up on FS #10257
22:47:50 Join bubsy [0] (i=Bubsy@
22:47:52 Quit bubsy (Remote closed the connection)
22:48:33 Join bubsy [0] (i=Bubsy@
22:48:35perfectdrugpixelma: i added the keymaps and the button, i have no LaTex installed so I can't check the pdf output of both files unfortunately
22:55:55pixelmaI'll do that (hopefully it works)
22:56:13 Join petur [0] (n=peter@rockbox/developer/petur)
22:57:58 Quit froggyman ("CGI:IRC")
23:01:32 Quit merbanan (Read error: 110 (Connection timed out))
23:04:39Unhelpfulwouldn't tagcache_is_X_tag be faster if we used a bitfield for the membership and just tested, for example, (1 << tag) & NUMERIC_TAGS ?
23:05:10Unhelpfulright now they iterate over a list of tag values that belong to the group they test for, and return true if they find the passed tag in the group
23:05:46 Quit FlynDice (Remote closed the connection)
23:05:46Unhelpfulit would also be smaller, as a table of ints would be replaced with one int.
23:07:13pixelmaperfectdrug: in the clip-svg there was just a small text object to the left (reading "ddd") which made the first conversion result a bit weird, after deleting it everything went fine
23:08:20 Join FlynDice [0] (
23:08:50perfectdrugpixelma: oops :D thanks for spotting
23:10:29 Quit BHSPitLappy (Read error: 110 (Connection timed out))
23:11:31 Join spiral1 [0] (
23:11:51*pixelma also has such a small png now which can be used in those target tables
23:12:03moosUnhelpful: bug Slasheri about...
23:12:55Unhelpfulmoos: it's one of the few parts of tagcache i think i understand well enough to go poking at it. the issue of tagcache_search failing with ramcache on and type tag_filename is a bit more complicated.
23:13:08 Quit perfectdrug ("CGI:IRC")
23:13:18 Join funman [0] (n=fun@rockbox/developer/funman)
23:13:28 Part spiral1
23:13:47 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
23:14:12kugelUnhelpful: I'm currently having problems with the sorting of the search results again :(
23:14:28moosUnhelpful: pitty that the full tagcache beast is that big, not a lot of people tried to poke it
23:14:32Unhelpfulkugel: what kind of problems?
23:14:39 Join perfectdrug [0] (
23:14:54kugelfilenames for a album don't come in sorted by track number
23:15:12pixelmaI recently stumbled across a weird database initialise bug, think it still exists but will try with a more recent build
23:15:45Unhelpfulkugel: i noticed that, too. i had to store a sort key for each and then post-sort them, see create_track_index in pictureflow for how that worked.
23:15:52funmanFlynDice: i'm running some tests on my fuze with caches
23:15:59 Quit JimBastard (" ajax IRC Client")
23:16:12*moos likes those thought to have a pf browser, sweaty :)
23:16:37perfectdrugpixelma: so on which player should I start working next? :D do you have any in progress?
23:16:39kugelUnhelpful: any idea if how I can add the file name to struct track_info and fill it in create_track_index?
23:17:43Unhelpfulkugel: i'm a bit confused, what exactly do you want the filename for?
23:17:52kugelto add them to the playlist
23:18:09pixelmaperfectdrug: no I don't really (well I have a started beast rework lying around here for a while...) are there any other AMS Sansas that aren't covered yet?
23:18:23kugelI can create a playlist and start playback within pf mostly fine, but the playlist is sorted weird
23:18:33kugel(and not in the same way as the track list, that is)
23:18:44Unhelpfulkugel: it saves the seek offset in the track table, surely you could use that to look up the filename via tagcache?
23:19:38perfectdrugpixelma: m200 as it seems
23:20:02kugelI tried adding a tag_title filter with that seek, without success, but maybe I just did it wrong
23:21:15 Quit evilnick (" ajax IRC Client")
23:21:33FlynDicefunman: do you have them working?
23:21:37pixelmadatabase initialise bug: "initialize now" while playing music -> database will update in the background, return to WPS while it's running -> if it's ready try to stop playback with still something to play -> you get the "Please reboot to enable" splash + music resumes and you're back into the WPS
23:21:38Unhelpfulkugel: i think maybe you should hang on to what you have so far. i'm adding some new tagcache methods, and i think that one or two more would make it easy to fetch an indexid via the seek into one of the tables that has them, and then use the indexid to grab the index entry, or any other tag entry.
23:22:26funmanFlynDice: well no, but i'm making comparisons. I thought I had good performance and no crashes with data cache disabled, but running it several times shows crashes
23:22:49pixelmaperfectdrug: you could try that one then, or maybe those GoGears?
23:22:53funmanWhat I noticed is that there is no more wps corruption if I make the aligned_buffer aligned on 32 bytes (size of a cache line)
23:22:57 Join parafin [0] (
23:23:19FlynDicefunman: I'm headed off to work shortly and ironically that's where I can spend the most time on rockbox!
23:23:51kugelCan I be your co-worker?
23:24:22funmanFlynDice: ok, it's a bit late there and i'll work tomorrow and sunday, so I'll just post what I have on FS #10048
23:24:24Unhelpfulsomething like an int tagcache_find_seek_index(struct tagcache_search *tcs, int tag, long seek)
23:24:35FlynDicekugel: not unless they give me an airplane with wifi...;)
23:25:12Unhelpfulor it could work like the existing tagcache_find_index and store the index in tcs and return a bool for success/failure
23:26:39FlynDicefunman: OK I 'll take a look when I get in tonight, I think I've run clock-target.h into the ground for now!
23:27:42funmanI tried to lower the storage clock but with divider set to 0 or 1, rockbox fails to boot (didn't try higher dividers)
23:28:44FlynDicefunman: I tried too with same results and finally gave up. I dis get fclk prediv working nicey thoufg
23:28:56FlynDiceer though
23:29:21 Quit HellDragon (Client Quit)
23:29:22 Join kachna [0] (
23:29:26amiconnThere seems to be a bug wrt voice file generation for hwcodec here... need to investigae :\
23:30:21 Join stoffel [0] (
23:33:32*pixelma wonders how up-to-date CREDITS-MANUAL is
23:33:44LambdaCalculus37toffe82: Do you think it would be a good idea to split the GoGear HDD6330 off in the build system into its own target? Since it's mostly all of the same code as the HDD1630, this would make builds easier as well.
23:34:22gevaertsIf they need different options, they should be different builds
23:36:16LambdaCalculus37gevaerts: Yeah, the HDD6330 needs a 220x176 LCD and has an RTC.
23:36:33funmanSansa Fuze, MMU operational, DRAM & IRAM cached, Data Cache disabled : mp3 215% realtime, 115.10MHz needed for decoding. Stable (so far)
23:36:34*LambdaCalculus37 makes that "has a 220x176 LCD"
23:36:43LambdaCalculus37funman: Nice work! ;)
23:37:06funmanwell it would be nice if I understood what is going on
23:37:47***Saving seen data "./dancer.seen"
23:38:10funmanmpegplayer stucks on 'loading' if I don't use a cacheline-aligned buffer for sd transfers, although the data cache is disabled.
23:39:07gevaertsLambdaCalculus37: seems pretty clear to me then
23:39:14FlynDicewell it would be nice if I understood what is going on <−−−− hey that's my line You're the one that's supposed to know ;)
23:39:47 Quit ender` (Read error: 104 (Connection reset by peer))
23:40:31LambdaCalculus37gevaerts: Guess I'll create a new folder for the GoGear HDD6330 in the SVN trunk and try to add it into the build system.
23:40:48LambdaCalculus37I think it'll also make it easier to add int any HDD6330-specific code, too.
23:41:17 Join bitf [0] (
23:42:33CIA-38New commit by 03pixelma (r21131): Commit FS #10260 by Marko Pahlke: Clip SVG for the manual (used to add a pdf and update the png as well). Add him to the credits.
23:43:08bitfSansa e280 can't delete files, files gone computer but not on play and can't figure out how to delete on player
23:43:18 Quit mirak (Remote closed the connection)
23:43:20bitfany help?
23:43:22funmanand then with this setup, sometimes I can't boot (white lcd screen)
23:43:44perfectdrugpixelma: thanks! :)
23:44:02bitfSorry, files gone on computer but not on player
23:45:56funmanbitf: which files?
23:45:58 Join HellDragon [0] (
23:46:14funmantry running scandisk
23:47:05bitfWhat is scandisk?
23:47:18bitfthe windows utility?
23:47:41 Quit stoffel ("leaving")
23:48:14FlynDicefunman: about the white screen.. it's a pretty common failure mode, do you have a feeling whether it's a common cuase or any number of different things?
23:48:18funmanor any utility to verify the filesystem
23:48:38funmanFlynDice: no idea.. I don't have white screens with a SVN build
23:49:20bitfI'm using linux but I'll see what I can do
23:49:32funmantry fsck.vfat -a /dev/sdX
23:49:56pixelmabitf: are those files in the "music" folder?
23:50:40pixelmaand does your computer's file browser (or whatever you use) show hidden files/folders?
23:50:43Mikachubitf: i would use -r rather than -a
23:50:55FlynDicefunman: Wow, I 've been getting the "white screen of death" all the time for mmu & pclk !=DRAM
23:51:34 Quit kugel (Remote closed the connection)
23:51:49bitfpixelma: yes Mikachu: What do you mean?
23:52:09Mikachufor the fsck comand
23:52:12funmanFlynDice: usb serial support would definitely help a lot!
23:52:27 Quit perfectdrug ("Leaving.")
23:52:28FlynDiceyes I think so :(
23:52:48funmantest_disk passes on my setup
23:53:30funmani don't know if it's a bit slower or a bit faster than SVN code though.
23:53:38 Join perfectdrug [0] (
23:53:43funmanonce it booted it doesn't seem to crash
23:54:31funmanhum but playback stops early
23:54:41funmanpcmbuf is empty
23:55:42bitfNothings working, is there a way to delete files on the player
23:55:59toffe82LambdaCalculus37: there is 2 different hdd 63, one with radio the other without
23:56:06funmanbitf: format it
23:56:09gevaertsbitf: tell your file manager to show hidden files
23:56:51bitfI'm in over my head
23:58:26 Quit {phoenix} (Remote closed the connection)
23:58:53 Join maddler [0] (

Previous day | Next day