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 2005-08-30

00:00:08 Join stripwax_ [0] (
00:00:21linuxstbBut that uses #define, and checking my Linux system includes, they use typedef.
00:01:41 Join Paul_The_Nerd [0] (
00:08:23stripwax_hey - so, richersounds in the uk are selling iriver ihp 120s for £149.95 Website states they have *5* in stock ..
00:09:12stripwax_(also *8* h140s)
00:09:22stripwax_(for £100 more!!)
00:09:50Wettcan order from rest of europe ?
00:10:06stripwax_Wett not sure
00:11:00stripwax_Hrmm.. "prices valid in store" ...
00:11:40stripwax_I'm sure they'll accept phone calls and will deliver
00:14:23 Join tvelocity [0] (
00:20:24 Quit muesli- (Read error: 113 (No route to host))
00:21:17 Quit Sucka ("a bird in the bush is worth two in your house")
00:31:56 Join rasher [0] (
00:33:26rasherI created a patch for this bug: "Should prevent saving 'filename.' " (attached), but I'm not sure it's correct. It's pretty short though.
00:35:12rasherAnyone have time to look at it?
00:35:13 Quit Maxime` (Read error: 104 (Connection reset by peer))
00:36:31 Join Maxime [0] (
00:36:39 Join webguest39 [0] (
00:40:11rasheramiconn: Is it necessary to sync the other language files?
00:40:23 Nick Maxime is now known as Maxime` (
00:40:44amiconnrasher: It's not strictly necessary, but recommended
00:40:58amiconn(especially for the ones used for voice file building)
00:41:35 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.6/20050716]")
00:41:41 Quit DangerousDan (Read error: 104 (Connection reset by peer))
00:42:08***Saving seen data "./dancer.seen"
00:42:32 Part stripwax_
00:42:33rasherIt saves space - anything else? (not that that's not worth it)
00:43:29 Quit preglow (
00:43:29 Quit Slasher (
00:43:29 Quit JoeBorn (
00:43:29 Quit ansivirus (
00:43:29 Quit pabs (
00:43:29 Quit crashd (
00:43:29 Quit Hadaka (
00:43:29 Quit HCl (
00:43:29 Quit Wett (
00:43:29 Quit AdNauseam (
00:43:29 Quit hicks (
00:43:29 Quit Asku (
00:43:29 Quit linuxstb (
00:43:29 Quit Bagder (
00:43:29 Quit TCK (
00:43:29 Quit igor47 (
00:43:29 Quit courtc (
00:43:29 Quit crwl (
00:43:29 Quit ze (
00:43:29 Quit solex (
00:43:29 Quit Nibbler (
00:43:29 Quit cYmen (
00:43:29 Quit lostlogic (
00:43:29 Quit merbanan (
00:43:29 Quit christel (
00:43:29 Quit Febs (
00:43:29 Quit thegeek (
00:43:29 Quit CoCoLUS (
00:43:29 Quit dwihno (
00:43:29 Quit mbr (
00:43:29 Quit odd (
00:43:29 Quit silencer (
00:43:29 Quit Strath (
00:43:29 Quit crash (
00:43:29 Quit Ismo (
00:43:29 Quit godzirra (
00:45:18amiconnrasher: (1) It saves little bit of space in the .lng - not significant unless it would grow too big otherwise
00:46:00amiconn(2) Deprecating the voice entries saves space in the voice file - very desirable on archos, and also good on iriver
00:46:10amiconn(3) It reduces confusion
00:47:16rasherYeah. I was mostly wondering if it was strictly needed, or just a very good idea.
00:47:49NJoinSlasher [0] (
00:47:56rasherAnyway, do you have time to look at my small patch? You seem to be the one most into the fat driver if I understand right
00:47:56NJoinpreglow [0] (
00:47:58NJoinpabs [0] (
00:47:58NJoinHadaka [0] (
00:47:58NJoincrashd [0] (
00:47:58NJoinansivirus [0] (
00:47:59NJoinJoeBorn [0] (
00:47:59NJoinHCl [0] (i=hcl@2001:610:1908:8000:290:27ff:feca:8029)
00:48:38preglowi'll sync norwegian for fun and giggles
00:49:02 Quit HCl (
00:49:02 Quit JoeBorn (
00:49:49preglowthree deprecated entries, that's that?
00:49:49NJoinHCl [0] (i=hcl@2001:610:1908:8000:290:27ff:feca:8029)
00:49:55 Quit HCl (
00:50:25NJoinHCl [0] (i=hcl@2001:610:1908:8000:290:27ff:feca:8029)
00:51:09 Quit HCl (
00:51:37Moosrasher: what's your patch do?
00:51:37NJoinHCl [0] (i=hcl@2001:610:1908:8000:290:27ff:feca:8029)
00:51:59rasherIt adds a little error-checking to the add_dir_entry
00:52:08rasherTo avoid creating filenames ending in a period
00:52:21rasherwhich is invalid according to the fat specs
00:52:36Moosah ok :)
00:52:56NJoinNibbler [0] (
00:52:56NJoinsolex [0] (
00:52:56NJoinmerbanan [0] (
00:52:56NJoinchristel [0] (i=christel@freenode/staff/christel)
00:52:56NJoincYmen [0] (
00:52:56NJoincrwl [0] (
00:52:56NJoinlostlogic [0] (
00:52:56NJoinze [0] (
00:53:01 Quit crwl (
00:53:01 Quit merbanan (
00:53:01 Quit lostlogic (
00:53:01 Quit Nibbler (
00:53:01 Quit cYmen (
00:53:01 Quit solex (
00:53:01 Quit christel (
00:53:01 Quit ze (
00:53:01 Quit HCl (
00:53:01 Quit pabs (
00:53:01 Quit crashd (
00:53:01 Quit Hadaka (
00:53:01 Quit ansivirus (
00:53:01 Quit Slasher (
00:53:01 Quit preglow (
00:53:17amiconnwhat the **** is freenode doing today???
00:53:30rasherBlowing up the internet, it seems.
00:53:30Moossh**t :)
00:55:12NJoinHCl [0] (i=hcl@2001:610:1908:8000:290:27ff:feca:8029)
00:55:12NJoinze [0] (
00:55:12NJoinlostlogic [0] (
00:55:12NJoincrwl [0] (
00:55:12NJoincYmen [0] (
00:55:12NJoinchristel [0] (i=christel@freenode/staff/christel)
00:55:12NJoinmerbanan [0] (
00:55:12NJoinsolex [0] (
00:55:12NJoinNibbler [0] (
00:55:12NJoinFebs [0] (
00:55:12NJointhegeek [0] (
00:55:12NJoinStrath [0] (
00:55:12NJoinsilencer [0] (
00:55:12NJoinodd [0] (
00:55:12NJoinmbr [0] (
00:55:12NJoincrash [0] (
00:55:12NJoindwihno [0] (n=dw@
00:55:12NJoinCoCoLUS [0] (
00:55:12NJoinIsmo [0] (
00:55:12NJoingodzirra [0] (
00:55:12NJoinWett [0] (
00:55:12NJoinAdNauseam [0] (
00:55:12NJoinTCK [0] (i=TCK@
00:55:12NJoinBagder [0] (n=daniel@
00:55:12NJoinlinuxstb [0] (i=dave@
00:55:12NJoinAsku [0] (n=aksu@
00:55:12NJoinhicks [0] (
00:55:12NJoincourtc [0] (n=court@
00:55:12NJoinigor47 [0] (
00:55:22NJoinSlasher [0] (
00:55:27NJoinpabs [0] (
00:55:27NJoinHadaka [0] (
00:55:27NJoincrashd [0] (
00:55:27NJoinansivirus [0] (
00:55:31NJoinpreglow [0] (
00:56:55preglowwhat the hell is up with freenode today
00:57:12Maxime`we are beta-causalties ^^
00:57:19linuxstbI don't know, but it doesn't seem to be affecting me...
00:57:29Moosme too :)
00:57:42 Quit ze (
00:57:42 Quit crwl (
00:57:42 Quit merbanan (
00:57:42 Quit cYmen (
00:57:42 Quit solex (
00:57:42 Quit Nibbler (
00:57:42 Quit lostlogic (
00:57:42 Quit christel (
00:57:42 Quit CoCoLUS (
00:57:42 Quit dwihno (
00:57:42 Quit mbr (
00:57:42 Quit odd (
00:57:42 Quit silencer (
00:57:42 Quit Strath (
00:57:42 Quit Febs (
00:57:42 Quit crash (
00:57:42 Quit godzirra (
00:57:42 Quit Ismo (
00:57:42 Quit thegeek (
00:57:42 Join godzirra_ [0] (
00:57:45NJoinsilencer [0] (
00:57:45NJoinIsmo [0] (
00:58:09NJoinFebs [0] (
00:58:12NJoinodd [0] (
00:58:12NJoincrash [0] (
00:58:12NJoinNibbler [0] (
00:58:12NJoinsolex [0] (
00:58:12NJoinmerbanan [0] (
00:58:12NJoinchristel [0] (i=christel@freenode/staff/christel)
00:58:12NJoincYmen [0] (
00:58:12NJoincrwl [0] (
00:58:12NJoinlostlogic [0] (
00:58:12NJoinze [0] (
00:58:16NJoinmbr [0] (
00:58:21NJoinCoCoLUS [0] (
00:58:21NJoinStrath [0] (
00:58:22NJoindwihno [0] (n=dw@
00:58:44NJointhegeek [0] (
01:01:05rasherI just synced the remaining 100% languages. Let's see if I didn't mess that up
01:01:36Maxime`what does "synced" means by the way?
01:02:12rasherBring into the same shape as something else (in this case, english.lang)
01:02:37MoosMaxime: ~mettre à jour
01:02:49Maxime`yes, but
01:02:57Maxime`there's "updated" and "synced"?
01:03:10Moosit's just words :)
01:03:37rasherJust synonyms
01:10:38rasheramiconn: Anyway, do you have time to look at my small patch? You seem to be the one most into the fat driver if I understand right (repeating myself because I don't trust Freenode)
01:12:22 Quit tvelocity ("Leaving")
01:12:26amiconnI'll try to, but there are more things that need fixing/ polishing...
01:12:39NJoinJoeBorn [0] (
01:13:01amiconnAPart from that, the fat driver isn't my work, although I did quite some things to it
01:13:19rasherSure, just wanted to make sure someone knew about it.
01:13:53rasherYeah, but then, I mostly know of what happened during the last 6 months
01:13:56Moosit's Zagor territory no?
01:14:27rasherI thin kit is, judging by the viewcvs page
01:14:51rasherHe seems busy though
01:15:04Moosyes :(
01:16:00linuxstbAnyone know anything about AAC?
01:16:20rasherNot the slightest.
01:16:37linuxstbNor me - not sure why I've spent the last couple of hours playing with libfaad....
01:17:17rasherBecause Rockbox would end up being a real iPod alternative by all means if it plays aac?
01:17:26rasherWhich would be sortof cool.
01:17:39linuxstbI know that iTunes and the iPod use AAC files in an "m4a" (Quicktime?) container. Are there any other common containers?
01:18:23 Quit crashd ("leaving")
01:18:39rasherI don't think so
01:18:48dpassen1m4a is just a renamed mp4 container
01:18:51 Quit dwihno (
01:18:51 Quit mbr (
01:18:51 Quit crwl (
01:18:51 Quit crash (
01:18:51 Quit merbanan (
01:18:51 Quit lostlogic (
01:18:51 Quit Nibbler (
01:18:51 Quit cYmen (
01:18:51 Quit solex (
01:18:51 Quit christel (
01:18:51 Quit ze (
01:18:51 Quit odd (
01:18:51 Join odd_ [0] (
01:18:52NJoinze [0] (
01:18:52NJoincrwl [0] (
01:18:53NJoinsolex [0] (
01:18:53rasheroh right, mp4
01:18:55NJoinchristel [0] (i=christel@freenode/staff/christel)
01:19:00dpassen1AAC comes either in the container or as raw data .aac
01:19:00linuxstbI've got another slight problem in that both ALAC (Apple Lossless) and AAC both use .m4a as an extension. But Rockbox needs to cope with that anyway (e.g. for Ogg Vorbis/Speex/FLAC).
01:19:02NJoinmerbanan [0] (
01:19:02NJoinlostlogic [0] (
01:19:07 Join crash_ [0] (
01:19:13 Join Nibbler [0] (n=sven@
01:19:14***Server message 505: 'logbot :Unregistered users cannot currently send private messages due to spambot problems. please register! ( )'
01:19:16linuxstbIs .aac commonly used?
01:19:20NJoinmbr [0] (
01:19:38dpassen1shouldn't be, anymore
01:19:51dpassen1it was used before there was any sort of attempt at a standard
01:19:57linuxstbSo just supporting Apple's format should be enough, at least to start with?
01:20:14dpassen1Apples and Nero's, if there are any differences
01:20:28linuxstbI was just about to ask about Nero. Any idea what they use?
01:20:31Moosrasher: for the french translations, "décibel" is the right term for dB you could remove your comment :)
01:20:54rasherNot my coment
01:21:09Moosa ok, but is it the good term
01:23:26rasheramiconn: does playcount work on archos after your addition?
01:23:52NJoincYmen [0] (
01:24:23NJoindwihno [0] (n=dw@
01:24:37amiconnrasher: I tested it with a search filter, that worked
01:24:38rasher(changelog quesions coming up)
01:24:55amiconnI didn't try the wps tag, but I think it will work too
01:25:21rasherwhat about the multiple partition-thing? How does that work for non-Ondio?
01:25:29rasherDoes it work at all?
01:25:52amiconnIt's not activated for other targets than ondio
01:26:01amiconnIt would work if activated
01:26:07amiconn(config option)
01:26:34amiconnDoesn't make much sense without hotplug, except maybe for special applications
01:26:51amiconn(aka special geek power users)
01:27:29rasherYeah, not sure why you'd do it
01:27:48 Quit ender` (No route to host)
01:28:13linuxstbSlasher: If you read this in the logs, do you think we could move the call to get_metadata() to before the code that loads the codec? That way, get_metadata() (with some modifications that I would be happy to do) could be used to identify the codec needed to decode the file.
01:28:20amiconnFAT16 support would also work for hd based models, but doesn't make much sense either
01:29:29amiconnIn fact I tested some aspects of both FAT16 and multivolume on my player back then
01:30:17rasherOh right, my stupid "stats" plugin - how would that work on an Ondio?
01:30:25rasherIt opens "/" and then recurses from there
01:30:26 Join solex_ [0] (
01:30:35 Quit silencer (Read error: 110 (Connection timed out))
01:31:23amiconnrasher: It works, and it works faast
01:31:36amiconnHowever, it has an oddness (maybe on other units too)
01:31:43rasherDoesn't surprise me
01:31:58rasherWhat is it?
01:31:58amiconnIt only displays the final stats *after* the splash that it finished
01:32:09rasherOh yeah
01:32:34amiconnThe stats plugin doesn't make that much sense on Ondio, as it always gives the stats for the whole file system
01:32:50rasherThe whole plugin doesn't make sense if you ask me
01:32:54amiconn(meaning it adds internal flash and MMC if such is plugged)
01:34:42rasherThere's something about selecting battery type for Ondio - what's the point of that?
01:35:47amiconnBattery runtime calculation
01:36:10amiconnThe Ondio is designed to run from AAA primary cells, so there is no charging circuit
01:36:27amiconnIf you put in NiMH AAA cells the discharge curve is different
01:36:43amiconnThe option adapts the runtime calculation to the discharge curve
01:38:38rasher"Low battery level handling has been improved" - what was wrong?
01:38:43rasher(I do ask a lot of questions..)
01:38:54rasherLast one though.
01:39:06zeimproved doesn't mean something was wrong.. just that its gotten better than however it was
01:39:06rasherWell, it's something that changed on Ondio
01:39:11zefixed means something was wrong
01:39:14rasherWell, what was "not-good"
01:39:26zeno matter good something is, it can almost always be improved
01:39:57rasherYes, well, low battery handling improved probably means some corner-case got elliminated or similar
01:40:14zeyeah i dunno
01:41:20rasherActually, the commit message was "Proper 'battery level dangerous' handling.", so something probably was wrong
01:42:08rasherHrm, that doens't seem to be strictly an Ondio thing
01:42:17zeah well then
01:42:22zethat does imply that something was improper
01:42:25amiconnrasher: Yes.
01:42:26rasherOr, it was.
01:42:36rasherThe changes just looked more general
01:42:50rasherWas it expecting player/recorder values?
01:43:01rasherthat's the change
01:43:10amiconnplus this:
01:44:18amiconnThere were 2 things on Ondio: (1) The 'dangerous' threshold was too low (2) The threshold has to depend on the battery type
01:45:33amiconnThis threshold is mainly used in 2 places (1) It makes the battery icon blink, that's only cosmetic. (2) The flash plugins use it to check when it's better not to flash
01:45:52amiconnThe latter is especially important on a device without charging...
01:45:53rasherHow does hot-swapping MMCs work?
01:46:12amiconnJust plug in a card, and it is automatically mounted
01:46:18 Quit solex (Read error: 113 (No route to host))
01:46:25amiconn...under /<MMC1>
01:46:28rasherNice, how about removal?
01:46:44amiconnYes, that dir uses characters illegal in an ordinary filesystem...
01:46:55amiconnJust rip out the card
01:47:09amiconnIt gets unmounted and all filehandles get invalidated
01:47:28amiconnOf course better don't do this if there is a file open for writing...
01:47:35rasherI guess that's almost an advantage (using illegal characters)
01:47:47rashersince it ensures that it can't exist
01:47:55amiconnThe playback engine is robust enough to handle the vanishing files
01:47:58preglowgotta go, guys, later
01:48:09 Quit preglow ("s")
01:48:11amiconnIt just skips them as if they weren't present from the beginning
01:48:48rasherThis reminds me of a bug I saw.
01:48:48linuxstbI wonder if the new "iriver" playback engine is as tolerant.
01:49:13rasherDeleting a playing file would make rockbox play from the next file when the buffer ran dry
01:49:15amiconnThe iriver playback engine handling of absent files is bad
01:49:35amiconnIt removes them from the playlist, while it should only skip them and leave the playlist alone
01:49:48 Quit Moos (" HydraIRC -> <- The professional IRC Client")
01:50:10rasherLooks like I'm all questions now. Does the Sim sound output work in cygwin?
01:50:50linuxstbThe sim has sound? Is this for both Archos and iriver? (sorry, I'm adding to rasher's questions...)
01:51:06rasherX11 on Linux. Iriver only I think
01:51:08rasherusing the codecs
01:51:27rasherNot sure if cygwin emulates /dev/dsp
01:51:56amiconnIt works on cygwin with a little adjustment
01:52:19amiconnIt also worked in win32 when I tried it
01:52:50amiconncygwin emulates a lot, also /dev/dsp
01:53:01rasherNo wonder it's such a beast
02:01:06rasherah well, got a segfault when trying
02:03:45rashermpa works though
02:03:58rasher(on linux)
02:07:12 Quit hicks (Remote closed the connection)
02:08:25rasherAnyway, I'm off.
02:08:26 Quit rasher ("CGI:IRC")
02:39:29 Quit Wett ()
02:42:11***Saving seen data "./dancer.seen"
02:52:12 Quit cYmen ("zZz")
03:29:34 Join pike [0] (
03:40:47 Quit pike (Read error: 104 (Connection reset by peer))
03:45:50 Join webguest23 [0] (
03:46:56webguest23are there any instructions for compiling on linux?
03:48:23webguest23anyone alive
03:56:41webguest23hello mcfly
04:02:23 Quit courtc (Read error: 110 (Connection timed out))
04:03:25 Join courtc [0] (
04:12:00 Quit webguest23 ("CGI:IRC (Ping timeout)")
04:12:05zecompiling what? last i did anything involving rockbox, compiling on linux was what instructions were for
04:42:12***Saving seen data "./dancer.seen"
05:45:37 Join pike [0] (
05:53:32 Quit JoeBorn ("Leaving")
06:23:34Slasherlinuxstb: Hmm, that should be possible. I can take a look later today
06:30:25 Join LinusN [0] (
06:41:31 Join paugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
06:42:15***Saving seen data "./dancer.seen"
06:48:05 Quit dpassen1 ()
07:27:20 Join StrathAFK [0] (
07:31:14 Quit Strath (Read error: 110 (Connection timed out))
07:33:11 Join B4gder [0] (
07:41:36 Join ender` [0] (
08:10:17 Join tvelocity [0] (
08:19:43 Quit tvelocity ("Leaving")
08:24:37 Join webguest20 [0] (
08:42:16***Saving seen data "./dancer.seen"
08:48:29 Quit webguest20 ("CGI:IRC")
09:02:45 Join einhirn [0] (
09:33:44LinusNi intend to remove the "line selector" option, objections?
09:33:50B4gdergo go go
09:34:24B4gderI intend to remove the database objections from th ReleaseTodo
09:34:35LinusNthe what?
09:35:15B4gderthe problems mentioned about the database status
09:35:43B4gderthey're not big enough to warrant any action before release
09:36:08LinusNi have just fixed one of the onplay issues
09:36:28B4gderin db browsing mode?
09:36:44B4gderhow did you solve it?
09:37:35B4gderbtw, the code size problem is no longer really a problem for the release, is it?
09:38:10LinusNthe onplay handling in tree.c wasn't db aware, and there was no way in dbtree to get the filename for a given track
09:38:29 Quit linuxstb (Read error: 110 (Connection timed out))
09:39:13 Join linuxstb [0] (
09:44:46linuxstbSlasher: Thanks. When you look at it, I think you should move the call to probe_file_format() into get_metadata() - so get_metadata() is responsible for both detecting the filetype (and setting track.id3.codectype) and reading the metadata. Unless you can think of a reason not to.
10:02:57 Quit linuxstb ("Leaving")
10:06:02 Nick paugh is now known as AliasCoffee (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
10:07:23solex_LinusN: why do you want to remove the line selector? I use it all the time.
10:08:35Slasherhmm, what was that line selector option?
10:08:43solex_or will "bar" be the default?
10:09:17LinusNbar will be the only option
10:09:22Slasherif it is the bar as line selector instead of an arrow, please do not remove that!
10:09:45Slasheror make the bar default..
10:10:19Slasherit's pain to navigate with the arrow only because iriver lcd is quite slow and with fast scrolling the arrow is invisible
10:10:49B4gderthat reminds me, the credits list is truly unreadable on iriver
10:12:27LinusNSlasher: i don't understand what you mean, first you tell me not to remove it, and then you tell me that it is a pain to navigate with the arrow
10:12:54SlasherLinusN: if you remove it and make the bar the only option (default), then it's completely ok
10:13:17LinusN10.09.17 # <LinusN> bar will be the only option
10:13:26Slasherah, yes good :)
10:13:35Slasherjust go and remove it then :)
10:13:42AliasCoffee(just kidding)
10:13:49CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
10:13:49*B4gder slaps AliasCoffee ;-)
10:15:14solex_I don't understand the commit message from 2005-08-30 05:40:14
10:15:22solex_does it concern ordinary users?
10:15:32 Nick solex_ is now known as solex (
10:19:56LinusNsolex: it means that you can't use ROMBox until 2.5 is released
10:20:25LinusNi.e when the BootBox is released
10:20:31LinusNArchos stuff
10:20:38solexah, thanks.
10:21:15*B4gder does the secret oh-my-driver-works-with-select() dance
10:22:05LinusNi know the feeling
10:22:27B4gderits a great feeling when HW and SW starts to actually work together
10:22:40B4gderand the concepts seem to work
10:23:26 Join DangerousDan [0] (
10:24:52B4gderour new code is roughly 1000 times faster than the previous product ;-)
10:30:27 Join cYmen [0] (
10:34:14 Join koniu_ [0] (
10:35:27 Join kurzhaarrocker [0] (
10:35:57kurzhaarrockerThank you for your great job with the peak meter, amiconn
10:37:17 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
10:39:43 Quit koniu_ (Read error: 104 (Connection reset by peer))
10:42:18***Saving seen data "./dancer.seen"
10:44:46 Join ashridah [0] (
11:02:50amiconnkurzhaarrocker: thx, I hope my changes make at least some sense
11:02:57amiconnI intend to do some more before release
11:04:36kurzhaarrockercleanup or performance twiddling?
11:07:56amiconnI intent to have 3 performance levels, actively selectable by the code using the peakmeter
11:08:50amiconnThe peakmeter will get a bit more self-contained, by only exposing the peak_meter_draw_get_btn() function, with an additional parameter (the timeout)
11:09:05amiconnThat requires changes in wps.c and wps-display.c
11:11:53 Join tvelocity [0] (
11:12:11amiconnThese peakmeter changes should also reduce the lcd transfer interferences in the radio screen
11:12:41kurzhaarrocker*grin* The radio screen is something I never saw
11:13:40 Join courtc_ [0] (
11:14:04 Quit courtc (Read error: 110 (Connection timed out))
11:15:17amiconnCurrently the whole radio screen redraws at every peakmeter refresh...
11:15:18kurzhaarrockerWould those levels be: 1) peak_meter_draw_get_btn(0) draws once, gets a button and returns immediately, 2) peak_meter_draw_get_btn(<some time>) and disk is spinning -> no busy polling, 3) timeout is set & disk isn't spinning -> busy polling ?
11:21:13amiconnThe draw once might be a good extension
11:21:46amiconnHowever, I want 3 performance levels (1) low-cpu, used for recording when the mpeg threadsaves data
11:21:56amiconn...peeking once per redraw
11:22:37amiconn(2) normal performance, peeking once per tick, like the wps currently does. Used for playback.
11:23:01amiconn(3) high performance, used for recording, busy-polling with just yield() in-between
11:23:42kurzhaarrockerseems reasonable
11:24:37amiconnThe draw once will probably be automatic if the timeout is zero
11:25:06amiconn...only thing necessary would be to change the loopto foot-controlled
11:25:45kurzhaarrockerSomething else I always wanted to implement but never did: a numerical wps tag that displays current / max peak. But then you'd need a button to reset the max peak...
11:26:38 Join muesli- [0] (
11:27:43kurzhaarrockermuesli? Hm. I need some milk and a spoon.
11:28:07thegeekmake a timeout for that max peak kurzhaarrocker
11:28:48kurzhaarrockerin combination with an auto-reset for each song / recording this might do the most important trick, thegeek
11:29:14amiconnThere is a timeout
11:29:30muesli-kurzhaarrocker get your hair dyed ;)
11:31:34amiconnI plan to expose the pcm peak calculation function on iriver to the plugin api after the freeze
11:31:36 Join [IDC]Dragon [0] (
11:31:38 Nick solex is now known as milk-and-spoon (
11:32:00amiconnThis should allow to make oscillo*.rock and vu_meter.rock work on iriver
11:32:28*[IDC]Dragon looks for Stockholm people
11:32:48kurzhaarrockerThe timeout by itself isn't very useful. If you want to know the peak of something you must be able to select the region you're interested in.
11:33:08[IDC]Dragona friend needs a hotel room 3.-4. sept.
11:33:25amiconnhi Jörg :)
11:33:46[IDC]DragonLinusN, B4gder, do you read?
11:34:06amiconnkurzhaarrocker: Reset on track change, perhaps when the timeout is set to eternal
11:34:12[IDC]Dragonstockholm seems packed with a congress or such
11:34:27LinusNso no available rooms?
11:34:29[IDC]DragonI thought maybe the "natives" have an idea
11:34:31muesli-did you try
11:34:42 Join Moos [0] (
11:34:44muesli-do it :D
11:34:58kurzhaarrocker... and a reset whenever you ffwd. Maybe definable by an option, amiconn
11:35:04muesli-bonyour madame
11:35:17Moosmonsieur please :)
11:35:18amiconnNooo, not another option.... ;)
11:35:36LinusN[IDC]Dragon: does it have to be in stockholm city, or can it be outside?
11:35:36 Nick milk-and-spoon is now known as solex (
11:35:56LinusNand does it have to be dirt-cheap?
11:36:03muesli-[IDC]Dragon general speaking they provide cheap accomodations
11:36:05[IDC]Dragonno dirt, please ;-)
11:36:13muesli-dunno if its in your taste ;)
11:36:20[IDC]Dragonmuesli: nothing available
11:36:27muesli-no dirt in hostels...very difficult ;)
11:37:09muesli-Moos ;)
11:38:13[IDC]Dragonthey are desperate enough for anything, I guess
11:39:34muesli-but it cant be worst than in oz
11:40:19muesli-reckon you would get a clean hostel in stockholm
11:40:37B4gderI found another hotel-finder service at but that's only in swedish
11:40:57[IDC]DragonLinusN: he studied that already
11:42:58B4gdertry "Scandic Malmen", it showed available rooms to me
11:44:21B4gderhopefully not everything is in swedish
11:44:58LinusN1 room, 2 adults
11:45:14*[IDC]Dragon tries to figure out "vuxna" and "barn"
11:45:22B4gderadults and children
11:45:23LinusNadults, kids
11:46:10[IDC]Dragonthat link was too deep
11:46:13 Join Zagor [0] (
11:46:45B4gderits probably easier to just call them ;-)
11:49:04muesli-i just call to say i love yooou ;)
11:49:14 Quit kurzhaarrocker (Remote closed the connection)
11:49:28*ashridah flees
11:54:42 Join hicks [0] (
11:56:55 Join CyberCr33p [0] (n=chris@
11:57:07 Part CyberCr33p ("Leaving")
12:09:53 Join preglow [0] (
12:11:07preglowLinusN: remove line selector option? why?
12:11:30preglowahh, as long as bar will be the only option, it's all good
12:11:37LinusNbecause we want to reduce the number of unnecessary options
12:11:46LinusNand save some code space
12:12:06preglowyes, agreed
12:12:09preglowthe bar is just fine
12:15:04preglowwill it happen before release?
12:15:56preglowenglish.lang needs some work on Cap Usage, it's a jumbled mix right now
12:16:56*amiconn wants an interim way of disabling voice strings for archos, as localisation v2 seems to take quite some more time...
12:17:37 Join ep0ch [0] (n=ep0ch@
12:18:32ep0cham i going deaf or has playback on ihp just got a bit quiter recently? :s
12:19:06ep0choh i might be coming down with a cold, that could be it...
12:25:50B4gderit could also be the ep0ch detection filter we added
12:28:24Zagoryay, it works!
12:29:00Mooswhat? :)
12:29:48Zagorthe ep0ch filter :)
12:30:07Moosah ok
12:31:56 Quit muesli- (Read error: 104 (Connection reset by peer))
12:42:22***Saving seen data "./dancer.seen"
12:46:11solexI noticed that rockbox hangs when I use an image tag in my .wps file without a trailing '|'
12:46:18solexIs this considered a bug?
12:48:40solexcan anybdy confirm this?
12:49:08amiconnYes, and rockbox hangs when loading a .cfg while music is playing
12:49:22amiconn(as reported yesterday)
12:50:38solexamiconn: doesn't hang here
12:51:46[IDC]DragonB4gder: "Scandic Malmen" booked, thanks
12:51:57B4gderpleased to help
12:52:16amiconnHmmgot that consistently. Maybe it only happens when some setting really changes, need to test which one
12:52:46amiconnBut archos work currently has priority, preparing the release...
12:53:20solexthat's ok. I'm hoping for a fast release cause there are cool features in the queue :)
12:53:26[IDC]Dragontomorrow I'll "harvest" the rombox dailies
12:54:27[IDC]Dragonnice that I don't have to build them all by myself ;-)
12:54:41amiconn[IDC]Dragon: Two questions for you
12:55:23amiconn(1) Could you test whether the peakmeter in the radio screen still works on Ondio fm with my latest changes? I'm quite sure it does...
12:56:09[IDC]Dragonis there one? I thought this was removed because of interference with too much cpu activity
12:56:27amiconn(2) We will need to extend the voice file format with a field to indicate the codec used. (preswapped mp3 for archos, and ltiple codecs (non-swapped)) for iriver
12:56:43amiconnAny suggestions on this?
12:56:58[IDC]Dragon(2) the file format is flexible, iirc
12:57:20amiconnKeep in minf that Bagder is working on localisation v2, which will also need changes in the voice file format
12:57:31[IDC]Dragoncontains its own header size, has a version number, iirc
12:57:35*amiconn looks at Bagder ;)
12:57:50[IDC]Dragonhow do you do that in IRC?
12:57:52B4gderyes, it is "work in progress"
13:00:01[IDC]Dragonamiconn: look at "struct voicefile" in talk.c
13:03:02amiconnLooks fairly extensible, however, it seems the code doesn't check the version
13:03:35[IDC]Dragonno need to, while not having more than one ;-)
13:10:21*LinusN just killed the line selector option
13:10:56Maxime`too bad :p
13:11:00LinusNdamn, i forgot to bump the settings version
13:11:09 Join ep0ch_ [0] (n=ep0ch@
13:16:46LinusNholy moses, this page is really out-of-date:
13:28:00solexdoes anybody use it at all?
13:28:31B4gdermost people probably just save a config using rockbox
13:28:36B4gderand edit that if they feel a need
13:28:49 Quit ep0ch (Read error: 110 (Connection timed out))
13:29:20solexand this saved file is heavily commented.
13:29:36solexif the possible values aren't abvious
13:33:03LinusNso, should we remove it?
13:33:17 Quit tvelocity ("Leaving")
13:33:26preglowi say yea
13:33:27 Quit webguest39 ("CGI:IRC (EOF)")
13:33:46Zagorme too
13:33:58*LinusN sharpens his axe
13:33:59preglowit's too much of a bother to update it when it's not even necessary
13:34:12solexI wouldn't miss it. It would probably easier to generate this information from the code.
13:34:22solexjust like the config dump.
13:36:50LinusNdead and buried
13:43:26*Febs Febs was getting a cup of coffee and is too late.
13:43:56FebsI was going to ask you to wait 5 minutes to see if there was anything I should copy to here:
13:44:12*Febs wonders why he called himself "Febs Febs"
13:44:18MoosLinus: still red build
13:44:32B4gderFebs: you have echo enabled? .-P
13:44:45MoosLinus: time to oops commit :)
13:45:03LinusNcoming up
13:53:50preglowso, any bugfixing work that i can do?
13:54:18*ashridah mutters
13:54:21ashridahdamned lecturer
13:55:18LinusNpreglow: the cpu boost bug in the iriver playback would be nice
13:55:36preglowcpu remains boosted after stopping playback?
13:55:52preglowalright, nice excuse for learning that part of rockbox anyway
13:55:59preglowi've got an hour to try
14:02:42solexGood luck!
14:04:46amiconnLinusN: Would adjusting the IRAM sizes for iriver violate the feature freeze?
14:05:13amiconnIf not, we could allow some more IRAM for codecs, and preglow could perhaps do some magic for mpeg layer 1+2
14:05:23preglowoh yes
14:05:54LinusNas long as you are positive that it won't wreck the archos builds, go ahead
14:06:15Slasheramiconn: Hmm, how much more? there are not much iram left anyway
14:07:17LinusNthe mainstack can be considerably smaller
14:08:14 Quit edx (Read error: 110 (Connection timed out))
14:09:44preglowoh yes
14:09:52preglow16k should be enough, yes?
14:09:59preglowif so, that would be a considerable amount for the codecs
14:11:03preglowi'm also going to start tinkering with stuffing all cache critical functions in a segment of their own
14:13:02preglowit's a really small change
14:13:05preglowand it helped libmad
14:14:03FebsI've updated the manual for .cfg files:
14:15:41LinusNFebs: good
14:16:14LinusNFebs: you can also add a tip, that you can remove any setting you want from the .cfg file
14:16:29LinusNso you keep only those you are interested in
14:17:06LinusNto create, for instance, a flat.cfg setting that resets the bass/treble settings
14:17:17FebsGood idea. I'll do that right now.
14:19:23preglowwhy does rockbox only see a changed rockbox.iriver when you're on the root?
14:19:39preglowdoes it search for rockbox.iriver relative to the current dir, or what?
14:21:11LinusNit looks in the dir cache
14:22:01preglowcouldn't we just as well do a real check? the disk is always spinning when you're exiting usb mode anyway
14:22:12preglowat least i've never had any os put it to sleep
14:23:04LinusNsure, go ahead
14:23:16LinusN(submit a patch, i mean)
14:23:33preglowwill try
14:27:09 Quit B4gder ("Lämnar")
14:30:37preglowi also want a more advanced plugin format, so we dont have to hard code extension and path
14:30:42preglowgive me infinite time, please
14:31:19LinusNadvanced plugin format?
14:34:16amiconnSlasher: Reduce the main stack to 8KB as on archos, then we have an extra 24KB
14:34:18preglowwell, so we can extract information about a codec from the codec file itself
14:34:38amiconnShould allow to split IRAM 50-50, i.e. 48KB for core and 48KB for codecs/plugins
14:34:39preglowwe also need the codecs to load their own data in the future, so we'll probably need it anyway
14:35:12amiconnThe main stack was increased that much back then because of rockboy, but rockboy doesn't hog the stack that much anymore
14:35:22preglowand if it does, it should be considered a bug
14:37:07preglowSlasher: is it just me, or does the iriver firmware buffer files faster?
14:37:45FebsI updated the config file part of the manual with Linus' suggestion. Question: I remember reading that there was a bug that crashed Rockbox if a new .cfg file were selected during playback. Has that been fixed?
14:39:05Slasherpreglow: iriver firmware does not use as much buffer as rockbox does
14:39:14Slasherit will buffer at maximum 3 files at time
14:41:30preglowahh, yes
14:41:32preglowthat explains a lot
14:41:39preglowwhat a frigging waste
14:42:16 Quit courtc_ (Read error: 110 (Connection timed out))
14:42:24***Saving seen data "./dancer.seen"
14:42:48LinusNFebs: no, that hasn't been fixed, to my knowledge
14:43:39preglowgotta help a friend move now, it seems
14:54:28 Quit Febs (" Like's GUI? Then try HydraIRC -> <-")
15:10:49LinusNtime to go
15:10:52 Part LinusN
15:17:53 Join courtc [0] (
15:22:50 Quit ashridah ("Leaving")
15:28:39 Join kurzhaarrocker [0] (
15:29:26 Quit kurzhaarrocker (Remote closed the connection)
15:29:45[IDC]DragonLinusN forgot to bump the cfg version
15:30:15[IDC]Dragonor, remove the option without shifting the other
15:33:51 Join muesli- [0] (
15:41:13 Join kurzhaarrocker [0] (
15:43:24amiconn[IDC]Dragon: He post-fixed the non-bumped config version...
15:43:32kurzhaarrockeramiconn: If you're in search of options of the peak meter to kick out: I don't think the maximum is relevant for anybody. It's important that you can adjust the minimum, but the maximum can be tied to 100% / 0 dbfs.
15:48:28 Join edx [0] (
15:54:06 Join webguest89 [0] (
15:54:25amiconnkurzhaarrocker: makes sense...
15:55:05amiconnI'll also look into the db calculation, wanting to change 2 things
15:55:48amiconn(1) calc_db() should return negative values based on 0 dB, removing the need to calculate the constant every time
15:56:01amiconn(2) The recursive reverse function is ugly...
15:56:08webguest89Any chance of making the rating/playcount WPS tag work like the playmode does after today's commit? (such that I can go %?rr<|bad|bad|bad|okay|okay|okay|good|good|perfect> or somesuch)
15:56:12kurzhaarrockerthe binary search for reverse calculating from db is quite a joke :)
15:57:17amiconnAlso, the linear interpolation isn't that good, perhaps I'll find something better
15:57:31amiconnI do have some ideas, but I don't want to inflate the function too much...
15:58:59amiconnAt least the reverse could use linear interpolation as well. If it uses the same segmentation as the forward calculation, roundoff errors from back & forth calculation should be small
16:00:04amiconnWe already regained almost 2KB with today's removal of the cursor selection by Linus :)
16:00:36kurzhaarrockeriirc roundoffs from back & forth are not acceptable because it is used to increment / decrement in the peak meter menu.
16:00:37amiconn...which by its own regained >1.3 KB
16:00:54amiconnYes I know.
16:01:10Maxime`hm, what's the thing with the cursor? we can't chose inverted anymore?
16:01:26amiconnIt's always inverted now on bitmap displays
16:06:07 Quit kurzhaarrocker (Remote closed the connection)
16:08:51 Join DangerousDan [0] (
16:13:42amiconn[IDC]Dragon: Do you read?
16:17:30 Quit webguest89 ("CGI:IRC")
16:17:57 Quit muesli- (Read error: 110 (Connection timed out))
16:18:03 Quit AliasCoffee (".")
16:22:24[IDC]Dragonamiconn: now I do
16:23:21 Join pillo [0] (
16:24:01pillohello everyone
16:24:41amiconn[IDC]Dragon: Is there a special reason for the button assignment in the Ondio recording screen being so odd? Any objections to changing them?
16:24:53 Quit crwl (Read error: 110 (Connection timed out))
16:24:58[IDC]DragonI don't remember
16:25:06pilloquick question, if I get a SYS_POWEROFF event in the button queue in a plugin, what do I have to return when exiting the plugin? PLUGIN_OK or some other value?
16:25:47amiconnIt seems that in order to put play/pause on the RIGHT button the assignment of the level adjustment became very unintuitive
16:25:59 Quit Bagder (Read error: 110 (Connection timed out))
16:26:29[IDC]Dragonpillo: call the default handler for everything you're not interested in
16:26:43[IDC]Dragon(hope I remembered that correct)
16:27:14amiconnHowever, record/pause is MODE in the radio screen as well, so imho nothing speaks against doing that in the recording screen too
16:27:26amiconn...regaining the logical layout of the level adjustment
16:27:40[IDC]DragonI don't like the Mode recording of the FM screen
16:27:45pillo[IDC]Dragon: that's how it is now. but I wanted to handle that event.
16:28:09[IDC]Dragonit was a quick hack
16:28:20pillo[IDC]Dragon: how do I call the default handler and exit the plugin simultaneously?
16:28:41amiconn[IDC]Dragon: Why?
16:28:53 Join Hansmaulwurf [0] (
16:29:05amiconnIirc the archos firmware also uses something different than RIGHT for recording
16:29:25pillo[IDC]Dragon: first call the handler and then exit? but does the SYS_POWEROFF handler ever return? :)
16:29:25amiconnRIGHT is only used as start for playback
16:29:41amiconnpillo: What do you want to do?
16:29:43Hansmaulwurflittle Question: what is the standart font on rockbox?
16:30:09pilloamiconn: save the settings in the text viewer plugin when you shutdown
16:30:09amiconnAnd yes, the default handler doesn't return on SYS_POWEROFF except something goes very wrong
16:30:11 Join crwl [0] (
16:30:14Hansmaulwurfi cant find it and my WPS was for the original font
16:30:30amiconnpillo: Do it in the callback of the default handler
16:31:00amiconnThat is what this callback is for - it gets called by the default handler on those events that lead to exiting the plugin
16:31:11amiconn(USB connected and poweroff)
16:31:42[IDC]Dragonamiconn: mode should bring the FM presets, because this is more frequently used than recording and does no "harm"
16:32:03amiconnHmm, but everything else is taken...
16:32:04pilloamiconn: uh-oh... I didn't even know such a callback existed... shame on me :o
16:32:34amiconn...and putting the menu on short MODE and recording on long MODE is also not very intuitive
16:32:35pilloamiconn: the text viewer does simply return when SYS_USB_CONNECTED is received in the button queue atm
16:32:44[IDC]Dragonit's just been easier this way, I was abusing the preset F-Button text for menu entries
16:33:35[IDC]Dragonimho, recording has to go into the quick menu, instead of presets
16:33:51[IDC]Dragon(on platforms with not enough buttons)
16:34:13amiconnQuick menu?
16:34:25amiconnWhere the heck do we have a quick menu on Ondio?
16:34:41[IDC]Dragonin the FM screen, iirc
16:35:05amiconnLeft/Left+Repeat - tune down
16:35:14amiconnRight/Right+Repeat - tune up
16:35:24amiconnUp/Up+Repeat - volume up
16:35:34amiconnDown/Down+Repeat - volume down
16:35:42amiconnMode - record
16:35:49amiconnMode+Repeat - Radio menu
16:35:59amiconnOff - Leave with radio stopped
16:36:07amiconnOff+Repeat - Leave with radio playing
16:36:17amiconnAnything I missed here?
16:36:23[IDC]Dragon#define FM_MENU (BUTTON_MENU | BUTTON_REPEAT)
16:36:31amiconnYes, the radio menu
16:36:43amiconnAll platforms have this
16:37:01[IDC]Dragonbut the Ondio needs extra entries here
16:37:10pilloamiconn: nevermind, I found out- thanks!
16:37:13[IDC]Dragonfor which didn't fit on buttons
16:38:08[IDC]Dragonand such entries need strings
16:38:52[IDC]Dragonso I've put the preset in there, recycling the F-button string
16:38:59amiconnYes I know that
16:39:50amiconnThe bad thing with burying the recording feature in a menu is that you can't start the recording quickly if you hear your favouite sonf in the radio
16:40:03[IDC]Dragondoesn't solve the recording screen dilemma, I know
16:40:22amiconnYes, that too
16:40:25[IDC]DragonI was just saying I'm not happy with the FM screen either
16:40:38[IDC]Dragonbefore we copy that style
16:40:50amiconnAt least it would be consistent
16:41:04[IDC]DragonI happen to press recording unintended all the time
16:41:06amiconnThe level adjustment in the recording screen is really counter-intuitive
16:41:41amiconnThe values are oriented horizontally, so the natural method for in-/decreasing is right/left
16:42:14amiconnHowever, with right I unwantedly started the recording several times
16:42:26***Saving seen data "./dancer.seen"
16:42:50amiconnThe levels are adjusted with up/down, and the selection of the value to adjust which would be intuitively doen with up/down is done with mode...
16:42:50[IDC]DragonI find up/down locical, like volume
16:43:11amiconnNot if you have a bar that in-/decreases horizontally
16:43:35[IDC]Dragonlet's tilt the bar ;-)
16:44:09amiconnStill, this would then imply to select the value with left/right
16:44:17amiconn...and adjust with up/down
16:46:39amiconnPerhaps a MODE double-click would help against accidental recording in the fm screen, and still allow fast enough start of recording?
16:47:12amiconnNot necessary in the recording screen I think
16:48:16*amiconn ordered an Ondio FM
16:49:11 Part pillo
16:53:19 Join webguest62 [0] (
16:53:27webguest62Hello folks
16:55:43amiconn[IDC]Dragon: Any opinion on the doubleclick idea?
16:56:09amiconnPerhaps with an added visual feedback
16:56:51amiconnAnother problem you might have read about in the logs: The recording beep *does* (often) disturb the recording, although I don't understand why :(
17:00:20webguest62Slasher: are you around for explain me quickly one thing about caching?
17:01:10webguest62i did read the yesterday log
17:01:42webguest62you planed to will do a new caching design no?
17:01:57*amiconn can only say that he doesn't like to see dir caching
17:02:12webguest62I don't understand well what is it?
17:02:39webguest62hey amiconn: can you explain me please what do he want to do?
17:05:22webguest62what is this tree cachin fonction?
17:05:46 Join keule [0] (
17:08:06[IDC]Dragondear cgi, pleade don't swallow my posting line :-((
17:08:32[IDC]Dragontoday, cgiirc doesn't like me
17:08:56 Quit keule (Client Quit)
17:09:00[IDC]Dragonamiconn: since when do we have double-clicks?
17:09:09[IDC]Dragonamiconn: double-click? do we have that anywhere else?
17:09:16webguest62anyone please can help me to understand please
17:09:23[IDC]Dragon(that was the missing line)
17:09:35[IDC]Dragonit crossed the late ones!
17:09:44[IDC]Dragon later
17:09:55webguest62what is the file tree caching
17:11:41amiconn[IDC]Dragon: There is one other place where we have double click - the recorder v1 shutdown
17:11:52[IDC]Dragonwebguest62: if you don't understand it, the implementation details are most likely not your concern
17:12:04[IDC]Dragonah, yes
17:12:21[IDC]Dragontimes "manually", I guess?
17:12:36amiconnI think so
17:12:41webguest62ok thanks for "help" :)
17:13:07[IDC]Dragonfrankly, UI design is not my thing
17:13:32webguest62Slasher when you will read the logs please answer me, i read the logs thanks by advance
17:13:34[IDC]Dragonand it suffers from the generalisation like we do
17:14:10amiconn[IDC]Dragon: I know that double-clicking isn't common in rockbox, but I don't see another way to allow for somewhat fast recording in the radio screen while at the same time preventing accidental recording
17:14:32amiconnHow is that done on the fmr, btw?
17:14:52Slasherwebguest62: hi
17:14:59[IDC]Dragonit has the abundance of 3 F-keys :-)
17:15:10webguest62Slasher: Hello
17:15:22Slasherwebguest62: the caching already works on my player, now directory browsing is lightning fast no matter if hd is spinning or not
17:15:38amiconn[IDC]Dragon: Hmm, yes, I think PLAY is used for recording there, and for nothing else
17:15:59[IDC]Dragonwhich is another extra key
17:16:28[IDC]Dragonthe thingis just that the mode-key is so "crowded"
17:17:06[IDC]Dragonit seems common to press it when looking for options
17:17:23amiconn...but then you have to press it long
17:17:30[IDC]Dragonit is the only non-direction key at the front...
17:17:54[IDC]Dragonslip off when pressing, and you record
17:18:02amiconnyes, exactly
17:18:39amiconnThat's why I suggest the double-click. Rockbox would only start recording if it detects two *short* presses within the time window.
17:18:57amiconnEverything else would drop the first click, and execute whatever is pressed second
17:18:58[IDC]DragonI'm sure double-clicks have already been discussed widely
17:19:30amiconnThey aren't easy to implement at the driver level, that's why there are only used rarely (I think)
17:19:33[IDC]Dragonthey could be generated by the driver layer already, if becoming popular
17:19:56t0masyou just need to save the tickcount at the last press
17:20:26t0masthen if the key is the same as the last key, and the tickcount - oldcount is < the set value, you have a "double click"
17:20:27amiconnYes, but I would prefer visual feedback from the first click, like the double-click shutdown does
17:20:38 Part ep0ch_
17:20:44t0masno problem
17:20:50t0maswe send the first click event anyway?
17:20:58t0masso the software can make some visual note on that
17:21:04amiconnDon't forget that we also have release events which need evaluation to decide whether this is a double click
17:21:05t0masand do the actual action on the next one
17:21:21amiconnclick-click is different from click-hold_down
17:21:33[IDC]Dragonyes, naturally a double click is not detected until the 2nd downpress within a timeout period
17:22:04[IDC]Dragonwhich would make normal single presses very sluggish
17:22:12amiconnt0mas: Yes, but if we have to react on the first click anyway, we can do the double-click detection in the upper layer
17:22:23t0masyes, we can
17:22:46amiconnIf they're used rarely, this is imho the better way
17:22:59[IDC]DragonANFSCD: you ordered an OndioFM, where?
17:23:17t0masamiconn: why not do it in upper layer now... and if they get used a lot, change it
17:23:20[IDC]Dragon(And Now For Something Completely Different)
17:23:22amiconnIf they become more popular, we can always move them to the driver layer
17:23:31t0massame tought :)
17:25:33[IDC]Dragonamiconn: then you perhaps want the backlight I saved for you
17:27:14amiconn[IDC]Dragon: I ordered one via Amazon marketplace. Second try already; first dealer told me they were all sold out
17:27:48amiconnSame day this offer was removed from Amazon. The other dealer's offer is also gone now, perhaps I am lucky this time..
17:28:07[IDC]Dragonlet's hope
17:34:05 Quit webguest62 ("CGI:IRC")
17:46:00amiconn[IDC]Dragon: Any opinion what to do with the recording beep?
17:55:04 Join PatentProblems [0] (
17:55:48PatentProblemsDoes anyone have a view on how the Creative Zen interface patent will affect Rockbox?
17:58:34 Join ep0ch_ [0] (n=ep0ch@
17:59:20 Join amiconn_ [0] (
17:59:28PatentProblemsDoes anyone have a view on how the Creative Zen interface patent will affect Rockbox?
17:59:34[IDC]Dragonah, the beep
18:00:01[IDC]Dragonsaw that, perhaps it's I2C "overload"?
18:01:34[IDC]Dragonlike, we keep the chip so busy that it can't properly perform recording
18:04:43 Join Sucka [0] (
18:08:36Hansmaulwurf <−−−−- :lol:
18:08:44Hansmaulwurfwie gemein ^^
18:16:03 Quit amiconn (Read error: 110 (Connection timed out))
18:16:04 Nick amiconn_ is now known as amiconn (
18:17:21 Quit PatentProblems ("CGI:IRC")
18:23:17 Join webguest79 [0] (
18:24:47amiconn[IDC]Dragon: yup
18:25:20 Join Lear [0] (
18:27:21webguest79hello Lear
18:27:42webguest79any news about APE tags ? :)
18:27:54Learwhat about them?
18:28:46webguest79this is one patch in tracker
18:29:00Learyou mean the one about adding ape support for mp3 files?
18:29:06webguest79maybe you could add them :)
18:29:25Learnot sure if it should be added...
18:29:37webguest79yeah APE tag support for Rockbox
18:30:03webguest79what's problem
18:30:09LearI made a patch like that a couple of years ago and it was rejected then... :)
18:30:20Maxime`"feature freeze" maybe?
18:30:37webguest79APE tag is not really a new feature
18:30:46webguest79just improvments
18:31:05webguest79Lear: what the problem did for
18:31:39webguest79why rejected?
18:31:50amiconn[IDC]Dragon ?
18:31:52Lear"too esoteric" I believe...
18:31:53webguest79hardware limitations?
18:31:57webguest79a ok
18:32:11amiconnImho ape tags are non-standard for mp3 and shouldn't be supported
18:32:56webguest79never implemented
18:33:10amiconnFormats that natively use ape tags are a completely different thing though
18:33:23amiconnAre there any? (any that we can support, that is)
18:34:04Learbut does wavpack have any standard tag format really?
18:34:44webguest79good question
18:35:27Learwait a minute, it does, I remember now...
18:35:48LearThe wavpack exe even has an option for adding tags...
18:36:52webguest79Lear: anything coming soon in your todo list? :)
18:37:02webguest79for Rockbox
18:37:16Learnothing noticable at least.
18:37:35Learmaybe that wav patch, if I can find a few test files at least. :)
18:37:50 Quit hicks (Read error: 110 (Connection timed out))
18:38:12webguest79It would be very cool
18:38:28webguest79want you wav files?
18:39:03webguest79*do you want? excuse my bad english
18:40:06Learthe patch should be okay, but I'd like to have a few of the more uncommon kinds, to make sure it was applied properly. not sure what Audacity can generate.
18:40:53[IDC]Dragongotta go
18:41:09webguest79Lear: oki
18:41:15amiconnThe recording beep issue...
18:41:15[IDC]Dragondid you read my beep comment?
18:41:35amiconnHmm, yes I did now
18:41:42amiconnOverlooked it before :/
18:42:02 Join bagawk [0] (i=1000@unaffiliated/bagawk)
18:42:04[IDC]Dragonbut I really gotta go, cu
18:42:09amiconnAny idea what to do about this?
18:42:14bagawkhey [IDC]Dragon
18:42:15amiconnHmm, bye
18:42:21 Quit [IDC]Dragon ("CGI:IRC")
18:42:30***Saving seen data "./dancer.seen"
18:42:31bagawkamiconn, Do you still have those photos of the white backlight around?
18:44:00amiconnThe old url should even be still valid is dead
18:44:13amiconn(if my apache server is running)
18:44:41bagawkConnection refused
18:45:22amiconnThen my apache isn't running. Can't change that now, first have to get home to do this
18:45:34 Join criis [0] (
18:46:54 Join webguest05 [0] (
18:47:06 Quit webguest79 ("CGI:IRC")
18:47:07 Join linuxstb [0] (
18:47:29linuxstbLear: What does your WAV patch do? Is it online anywhere?
18:47:45LearNot mine, the one in the patch tracker
18:47:56*linuxstb visits patch tracker
18:52:20SlasherHmm, now the dir cache works great and the file browser is faster than ever before ;) i still have to make it handle special cases (on the fly cache update) and cache rebuilding after usb connection
18:52:30linuxstbThat patch almost makes WAV qualify as a codec. It seems good after a brief look.
18:52:41linuxstbLear: Are you going to commit it?
18:52:59LearI was thinking about it, after the feature freeze.
18:53:09ep0ch_Slasher: commit commit! :)
18:53:24Slasherep0ch_: not before the feature freeze is over :)
18:53:25LearI plan to clean up metadata.c a bit, so I might take this patch as well... :)
18:53:26 Join TCK- [0] (
18:53:39ep0ch_slasher: no one will notice ;)
18:53:46Slasherhehe :D
18:54:13ep0ch_i thought the feature freeze was just for archos anyway?
18:54:22thegeekI love you Slasher:)
18:54:52LearSlasher: does it need much memory?
18:54:52amiconnSlasher: I wonder how you plan to deal with all those special cases...
18:54:52linuxstbLear: What are your plans for metadata.c ? I was also planning a little work on that - after Slasher makes a change to playback.c
18:54:57Slasherep0ch_: at least any commit that could in theory affect the archos code as well should not be committed. And this code might have some effect in theory
18:55:08Slasheramiconn: I already have some plans :)
18:55:26amiconnI still severely dislike the dircache idea...
18:55:30linuxstbWe need some way to distinguish different codecs with the same file extension - e.g. ALAC/AAC (.m4a) and the various filetypes inside .ogg
18:55:36Slasheramiconn: I will reverse a few kilobytes for the on the fly cache updating. If that space is not enough, then the cache will be switched completely off
18:56:20ep0ch_amiconn: what are your reasons for not wanting it?
18:56:33amiconnAdding a whole layer of high complexity just for a really minor speed improvement in corner cases contradicts one of the main principles of rockbox
18:56:44Learlinuxstb: primary goal is to generalize some of the tag parsing, so that ape and vorbis tags share more code. But once I got started, I did some more, mainly cosmetic changes.
18:56:52ep0ch_if it saves battery and makes the user experience better then what could be bad with it
18:56:52amiconnI don't see the big speedup resulting from caching all dirs
18:57:06Slasheramiconn: i think that is a major improvement from a user view
18:57:12amiconnUsually you browse only a few dirs then start playing music
18:57:13ep0ch_but the hd has to sping up which takes agesssss
18:57:24LearWhen I checked a while ago, it cut code size by almost 900 bytes. (I've made a few more changes since though.)
18:57:31amiconn...and to the contrary, caching will take considerable time
18:57:45Slasheramiconn: it doesn't affect boot up performance at all
18:57:48amiconnThat's one thing I really really hate about the iriver firmware
18:57:52Slasherwe can load the cache directly from disk
18:57:54linuxstbLear: Is it ready to be committed? I have some changes I want to make, but it sounds like you should commit first.
18:57:56amiconnSo how do you cache?
18:58:05ep0ch_if its transparent to the user it will be good
18:58:12Slasheramiconn: caching should be done only after usb connection or when user requests that
18:58:20amiconnThen you need to cache icrementally, and that doesn't give a speedup if the dir is not yet cached
18:58:28amiconnAnd after boot?
18:58:30Slasherin fact we could do almost always completely transparent caching
18:58:44ep0ch_but when you navigate out of a directory that parent dirs are cached
18:58:51amiconnThe whole disk contents may well change without rockbox noticing...
18:58:57Learlinuxstb: I think so. I haven't quite decided if I really should do some of the changes though.
18:58:59Slasheramiconn: after boot we will load the previous cache, user will not even notice that
18:59:00 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
18:59:10amiconnSlasher: I will notice for sure
18:59:20amiconnI use bootloader USB very often
18:59:21Slasherand if we like, we can start transparent cache rebuilding at the same time (user can start playing music normally etc)
18:59:41thegeekI dont see the problem
18:59:45thegeekif you dont like it, disable it
19:00:00thegeekI usually browse a lot of dirs
19:00:06thegeekso I would very much like it
19:00:09amiconnthegeek: If the player starts, rockbox can't assume anything about the disk contents
19:00:18linuxstbLear: If you want a second opinion, email me your new metadata.c file before you commit it.
19:00:18thegeekI get that,)
19:00:33thegeekI'm not saying it's easy to implement
19:00:48Slasheramiconn: then it should be possible to start without cache before the transparent rebuilding is complete (if user configures it that way)
19:00:53thegeekjust that if it can be done, with a little overhead, it would certainly be worth it for me
19:01:48amiconnSlasher: The only thing that might work is transparent caching of dirs already visited. However, this still has the problem that a plugin might change the contents of just about any dir...
19:02:15amiconnEven now we have the problem of rockbox not noticing that
19:02:34Slasheramiconn: mkdir, rmdir, open (when creating new file) will cause the cache live update
19:02:37amiconn(though fixing that is simple, we just have to redeclare the meaning of a plugin return code)
19:02:40Slasherso that shouldn't be a problem
19:02:56amiconnrename, delete,...
19:03:22Slasherand the cache is used only when calling opendir_cached or readdir_cached
19:03:40Slasher(on archos they are redefined to opendir and readdir only)
19:03:44amiconnHmm, yet another layer of complexity...
19:04:21amiconnIf you do it that way, the extra caching in tree.c is redundant ... except for database mode .... etc
19:04:56amiconnNext interesting question is how much memory the cache will take, and what happens if the cache gets filled up
19:05:30Slasheramiconn: now it takes ~340 KiB (i have a very large collection of files on the 40 GB drive)
19:06:09Slasheramiconn: if the cached gets filled up on the fly, it will be switched off..
19:06:09amiconnHmm, how many files?
19:06:23Slasherhmm, just a moment (i fill run find /iriver | wc -l)
19:07:04Slasherand the files have quite long filenames also
19:07:07amiconnRemember that one FAT32 dir can have up to 32*65536 = 2M entries
19:07:32Slasheryes, we should have some memory limit to the cache..
19:07:47Slasherif that is exceed, we cannot enabled the cache
19:07:50amiconnWe don't have dynamic memory...
19:07:54Slasher5493 files
19:08:20Slasheramiconn: that's true, but something like 50-100 extra KiB for cache reverse _should_ be enough
19:08:35Slasherif it's not, the cache would be rebuild on next reboot
19:08:36bagawkwhat does the piped wc do?
19:08:51ep0ch_count lines
19:08:55Slasherbagawk: it will calculate the line count that find outputs
19:10:22 Quit TCK (Read error: 110 (Connection timed out))
19:10:52 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:13:27*amiconn is still not convinced that dir caching is a good idea
19:14:03ZagorI'm strongly convinced it's a bad idea
19:14:30Slasherhmm, but if it's an option it shouldn't hurt anybody..
19:14:41Zagorit hurts the code
19:14:50Zagormake it larger and more complex
19:15:33Slasheris it so critical with iriver because we have plenty of memory to implement all nice features?
19:16:01Slasherit shouldn't make the code so complex..
19:16:44Slasheronly few calls to cache update from mkdir, remove, etc.
19:17:14 Quit linuxstb ("CGI:IRC (EOF)")
19:18:52 Part criis
19:20:20ep0ch_surely there is nothing wrong with trying out a good idea though
19:26:41Zagorwell, I don't think it's a good idea... :-) in fact I don't see the point at all.
19:29:53bagawkI do not see it as a good idea also
19:39:47 Quit edx ()
19:40:37Mooswhich are the disavanteges?
19:40:42Mooshi HCl
19:41:33Mooswhy is'nt a good idea?
19:41:59MoosIf the user can choose to enable or disable this feature?
19:47:25 Quit webguest05 ("CGI:IRC (EOF)")
20:04:28amiconnbagawk: webserver is running now.">
20:04:28 Quit Maxime` (Read error: 104 (Connection reset by peer))
20:05:19 Join Maxime [0] (
20:10:54ZagorMoos: turn the question around. what are the advantages? why is it a good idea?
20:15:07 Join edx [0] (
20:19:47 Quit Maxime ("Format, souhaitez moi bonchance :d")
20:26:06 Quit christel (Remote closed the connection)
20:33:24 Join christel [0] (i=christel@freenode/staff/christel)
20:40:36SlasherZagor: for user it should be only an improvement: very fast directory browser because we will never need to use the hd and thus no need to wait for the hd to spin up
20:41:15Slasherit will also save some power if user just wants to browse the dir tree
20:42:33***Saving seen data "./dancer.seen"
20:47:35Moosand a boot a bit faster, no? :)
20:47:52Mooshey t0mas
20:47:58t0masit won't boot faster
20:48:05t0masbut dirbrowsing will be faster
20:48:28SlasherMoos: user wont notice any decrease on boot time
20:48:45Slasherthe caching is not done the same way iriver firmware does
20:48:47Mooswhat do you think you, about this for other point of view?
20:49:02MoosSlasher: ok
20:49:20Moosi noticed this :)
20:49:30Moosabout this future feature?
20:49:47MoosZagor and amiconn don't think is it a good idea
20:50:15t0masI haven't read it
20:51:58*amiconn still wonders why dir caching should speed up dir browsing so much that it would be worth the complexity
20:52:43amiconnWithout dir caching, there is exactly *one* spinup that delays browsing - at the start of your browsing action
20:53:08amiconnIf browsing continuously, the disk won't spin down, so there are no significant delays
20:53:11*t0mas doesn't care
20:53:23t0mastime for a shower
20:53:25Slasheramiconn: the dir browser will be faster with caching even if the hd is already spinning
20:53:35Slasherusers should find it more comfortable
20:54:03amiconnThe word "significant" is significant here
20:54:58Slasherhmm.. anyway i think many people would like that feature, especially if it doesn't have the disadvantages that the iriver feature has
20:55:22Moosit's true :)
20:55:31Slasherand the code will not be very complex
20:57:19Moosamiconn: what are the disavantages for you?
20:58:19amiconnMoos: The disadvantages for me are (1) introducing code complexity for (2) no real advantage
20:58:36thegeekimho it would be an advantage
20:58:57thegeekI am frequently annoyed by the hd spinup delay when browsing dirs
20:59:13Moosamiconn: for the crazy browser is it avantage
20:59:28amiconnMoos: It isn't
20:59:35 Join LinusN [0] (
20:59:46Moosand Slasher said don't introduced complexity in the code
20:59:48amiconnIf frequent spinups annoy you, you've set the spindown delay too low
20:59:55MoosBonsoir Linus
21:00:02LinusNone major point is: how will the cache be kept coherent?
21:00:24LinusNhow can rockbox guarantee that the cache is up-to-date?
21:00:33SlasherLinusN: i have some ideas for this (just trying them at the moment)
21:00:41thegeekthat is imho an implementation problem
21:00:52LinusNthegeek: of course it is
21:00:53Slasherone is that mkdir, rmdir, etc. file operations will inform the cache and cache does a live update
21:01:03LinusNSlasher: that is not a problem
21:01:05thegeekofcourse, if it proves that the disadvantages are greater than the advantages, I do agree that it is not worth it
21:01:06amiconnthegeek: Yes, and it is the main and real source of complexity
21:01:08Slasherit will keep some reversed space (for example 50-100 KiB) for that updating
21:01:11LinusNthe problem is usb mode
21:01:18Slasherif that space is consumed, then the cache will be disabled
21:01:26amiconnLinusN: USB mode *and* reboot
21:01:28SlasherLinusN: after usb mode, we can do transparent cache update
21:01:31thegeekI'm just presenting the issue from an average users viewpoint
21:01:35Slashersame after reboot
21:01:38thegeekif the browser is faster, that is good
21:01:39amiconnYou'll never know what ran between 2 rockbox sessions
21:01:45Slasheronly after refresh is complete, we will enable the cache
21:01:52amiconnIt could be bootloader USB mode, iriver fw USB mode...
21:01:53LinusNSlasher: so it will scan the disk in the background?
21:01:55thegeekif it means that it will be buggy, the drop it
21:01:58Slasherand the updating will be really transparent to user
21:02:12LinusNthat will eat a lot of battery
21:02:15thegeekvbut there are still advantages to having it
21:02:40SlasherLinusN: it's true but user can disable the caching.. and after boot-up we _could_ load the cache directly from disk without updating
21:02:56LinusNno we canät
21:02:59Slasherbut then we must assume that the user has not uploaded any new files to the disk
21:03:12LinusNwe can't assume that
21:03:19Slasherit's true..
21:03:28bagawkHey LinusN
21:03:30Slasherso we would need to refresh it anyway
21:03:42LinusNadding such a cache opens up a huge can of worms
21:04:14LinusNwhere we attempt to keep a mirror of the hard disk structure
21:05:18Slasherbut currently only filetree.c uses the cache (opendir_cached, readdir_cached), and open() etc. would never use it
21:05:32Slasherthat will reduce the risk of failures
21:06:52LinusNopen(9 and friends still have to update the cache
21:07:48LinusNgotta go
21:07:52 Part LinusN
21:08:10 Join paugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
21:14:05 Quit edx ()
21:23:02 Quit paugh ("Leaving")
21:23:39 Quit ep0ch_ ("Trillian (")
21:24:58 Join edx [0] (
21:27:38 Join Maxime [0] (
21:29:28 Quit bagawk ("Leaving")
21:35:25 Part Zagor ("Client exiting")
21:36:15 Quit courtc (Read error: 110 (Connection timed out))
21:37:22 Join Zagor [0] (i=foobar@pdpc/supporter/sustaining/Zagor)
21:38:27Zagorthe point is, it will NOT be instant. it will only be instant the second time you access the same dir, which is quite rare to do actually. or if you preload dirs you will slow everything else down and waste a lot of battery.
21:40:09amiconnAgreed 100%
21:40:31 Join _pill [0] (
21:40:33 Quit pill ("changing servers")
21:40:35 Nick _pill is now known as pill (
21:42:08 Join Wett [0] (
21:43:15 Join webguest47 [0] (
21:43:47 Quit Maxime ()
21:43:48 Join webguest51 [0] (
21:44:09webguest47can rockbox be compiled on linux?
21:44:29Zagorwebguest47: yes
21:46:27 Join Maxime [0] (
21:47:30 Quit Maxime (Client Quit)
21:47:58 Quit webguest47 (Client Quit)
21:48:00 Join webguest47 [0] (
21:48:11webguest47are there instructions anywhere for compiling on linux?
21:48:32 Quit pill ("changing servers")
21:48:41Learrtfw ;)
21:49:03 Join pill [0] (
21:49:11Zagor, see section "For Developers"
21:49:49 Quit webguest51 ("CGI:IRC")
21:50:58 Join Maxime [0] (
21:51:25webguest47 thanx
21:53:53 Quit Maxime (Client Quit)
21:55:30Slasherhmm, does the fat have any timestamp or counting number that will be changed when the disk was updated?
21:59:29Wetthm, I have a little question about lcd_blit, someone ?
21:59:48 Join Maxime [0] (
22:03:14 Join matsl [0] (
22:07:26 Quit Maxime (Read error: 104 (Connection reset by peer))
22:09:11 Quit Zagor ("Client exiting")
22:09:38 Join webguest07 [0] (
22:12:43 Quit Lear ("Chatzilla [Firefox 1.0+/undefined]")
22:16:38webguest47i get this error when compiling binutils: ../../../binutils-2.16/libiberty/fibheap.c:285: error: `LONG_MIN' undeclared (first use in this function)
22:19:04 Join linuxstb [0] (
22:19:36 Join Bagder [0] (
22:23:33linuxstbwebguest47: I've never had that problem. Do you have /usr/include/limits.h on your machine?
22:24:28 Quit Hansmaulwurf ("( :: NoNameScript 3.81 :: )")
22:25:36 Join Maxime [0] (
22:28:33 Quit Maxime (Client Quit)
22:28:49linuxstbWhat was the "configure" line you typed?
22:28:50 Join xen` [0] (
22:29:46webguest47 ../../binutils-2.16/configure −−target=m68k-elf −−prefix=/home/adam/rockboxbuild
22:31:51linuxstbThat looks perfect. I don't know what to suggest. Personally, I use the CVS version of binutils - but would be surprised if that's the problem.
22:34:11linuxstbWhen you typed "make" for the first time, did the configure script display "yes" in the line "checking for limits.h... yes"
22:35:14linuxstbSorry, got to go.
22:42:34***Saving seen data "./dancer.seen"
22:42:34 Quit webguest47 ("CGI:IRC")
22:44:44 Join Maxime [0] (
22:52:21 Quit dpassen1 ()
22:55:32 Join LinusN [0] (
23:00:08MoosRe Linus :)
23:00:21LinusNi think calling rolo_load() in UIE() is a little risky
23:00:39LinusNbetter reboot
23:01:03LinusNMoos: hi
23:01:08amiconnYes, it may hang (and I already experienced this, on archos)
23:03:16LinusNi'll change it and also add a hard drive shutdown
23:04:00LinusNperhaps also backlight off...
23:06:00amiconnLinusN: I have a problem with your peakmeter performance setting idea :/
23:06:05 Quit matsl (Remote closed the connection)
23:06:20amiconnOn archos this would imply a callback from firmware to application...
23:06:30amiconn(mpeg.c calling peakmeter.c)
23:06:45amiconnHowever, I like the idea...
23:06:51LinusNi don't even remember what my idea was... :-)
23:07:24amiconnYour idea was that the playback/recording code should actively set the peakmeter performance according to requirements
23:07:46amiconnFor recording this means the mpeg thread has to control it (actually only partly)
23:08:58 Join XavierGr [0] (
23:09:34LinusNanother reason for mpeg.c to be in apps/
23:10:09amiconnHmm :/
23:10:33amiconnThat's a huge task impossible to complete before release :(
23:11:33XavierGrHi all...
23:13:22XavierGrLinusN did you managed to see the jpeg patch I submitted?
23:14:36 Quit Maxime (Read error: 104 (Connection reset by peer))
23:15:03 Join Maxime [0] (
23:15:15LinusNXavierGr: yes, i've seen it
23:15:30XavierGrWhat's your opinion?
23:15:45LinusNwell, i have seen it, but i haven't reviewed it
23:15:45XavierGrIt works very well for me...
23:15:53XavierGroh I see...
23:16:09LinusNmy priorities are at the release right now
23:16:34*amiconn wonders how this tiny fix in mpeg.c works for preventing the NULL pointer accesses...
23:16:52XavierGrSo there is no chance of commiting this to the release, right?
23:19:46 Quit Maxime (Client Quit)
23:19:55LinusNXavierGr: no
23:22:08XavierGrThough it would be good for archos users too, anyway I hope that you will implement it after you examine it and after the release.
23:22:37LinusNwe'll have a look
23:23:18 Quit Bagder ("Off to search for that connect-resetting peer guy!")
23:23:31*amiconn tests with memguard active....
23:23:53amiconnLinusN: I'd want to enable memguard by default...
23:25:03LinusNamiconn: do that if you dare :-)
23:25:15 Join Maxime [0] (
23:25:36amiconnWell, it should help finding bugs (especially when it gets implemented for iriver as well)
23:27:35LinusNbut it might be annoying
23:28:15LinusNfor example, the recent null access wasn't that harmful
23:28:59 Join webguest90 [0] (
23:29:12amiconnNo, but it was a bug that went undetected for quite some time because memguard isn't enabled by default, and nobody bothered to enable it and test...
23:29:23webguest90any idea what this error is when i try to compile:
23:29:25webguest90src/it/itread.c: In function `it_load_sigdata':
23:29:56LinusNwhat are you compiling?
23:30:10webguest90hang on let me get on a real irc client so i can cut and paste
23:30:32amiconnLinusN: Perhaps we should have a global #define that decides whether memguard is enabled by default?
23:30:33 Nick StrathAFK is now known as Strath (
23:30:40 Join _MadMan_ [0] (
23:30:58amiconnThis could then be set by configure, or manyally as EXTRA_DEFINE
23:30:59_MadMan_src/it/itread.c: In function `it_load_sigdata':
23:30:59_MadMan_src/it/itread.c:1161: error: insn does not satisfy its constraints:
23:30:59_MadMan_(insn 2275 1148 1149 77 (nil) (set (reg:QI 8 %a0)
23:30:59DBUGEnqueued KICK _MadMan_
23:30:59_MadMan_ (mem/s/j:QI (plus:SI (reg/f:SI 8 %a0 [353])
23:30:59_MadMan_ (const_int -288 [0xfffffee0])) [0 mididata S1 A8])) 37 {*} (nil)
23:31:00***Alert Mode level 1
23:31:00_MadMan_ (nil))
23:31:01LinusNamiconn: a configure option?
23:31:02_MadMan_src/it/itread.c:1161: internal compiler error: in reload_cse_simplify_operands, at reload1.c:8365
23:31:04_MadMan_Please submit a full bug report,
23:31:06_MadMan_with preprocessed source if appropriate.
23:31:08_MadMan_See <URL:> for instructions.
23:31:10_MadMan_make[4]: *** [/home/adam/rockbox/build/apps/codecs/dumb/itread.o] Error 1
23:31:18LinusN_MadMan_: use another version of gcc
23:31:20 Quit webguest90 (Client Quit)
23:31:38_MadMan_i'm using the one i built
23:31:46_MadMan_per the cross compiler instructions
23:31:55amiconn...or perhaps in autoconf.h
23:31:56 Join Bagder [0] (
23:32:31amiconn_MadMan_: Then you didn't follow the instructions completely
23:33:25LinusN_MadMan_: the instructions are somewhat ambiguous
23:33:36LinusNbut they say for coldfire: " Get gcc 3.4.X"
23:34:36_MadMan_yeah the page is confusing
23:35:04_MadMan_thanks lemme try again
23:35:11_MadMan_so you don't need 3.3.4?
23:35:54 Join Tali [0] (
23:36:08Talihello hello
23:36:18LinusNhi hi
23:36:30 Nick Tali is now known as Talim (
23:36:34LinusN_MadMan_: no, only for archos
23:36:50Talimim still hopeful for my iriver h320 :(
23:37:14 Join midk [0] (
23:37:32 Part _MadMan_
23:37:48Talimid like to be able to play FLAC :D
23:38:03midkrofl, did anyone see this?">
23:38:12midkor maybe someone here makes it, dunno. :)
23:38:27midkquite embarrassing. :)
23:38:29Bagderrasher makes it
23:38:56midkBagder, hm, can't remember seeing him around much..
23:39:11Bagdernot recently perhaps, but he's around a lot
23:40:29BagderI had an interesting experience a few minutes ago
23:40:35BagderI ripped a CD
23:40:53Bagderand my player got all weird and wouldn't spin down when completed
23:40:57Bagderwouldn't eject
23:41:01***Alert Mode OFF
23:41:09Bagderjust woooshing all the time
23:41:41Bagderand when I *rebooted*, my linux got stuck when probing/whatever with the IDE bus on the unit
23:41:59Bagderhad to power off/on again completely
23:42:23midkworked afterwards?
23:42:33Bagdernow all is good again
23:42:39midkhrm, weird. :)
23:46:41amiconnLinusN: Fixed up the build instructions a bit :)
23:53:01t0masgood night :)
23:54:15 Quit Talim ()

Previous day | Next day