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 2004-11-15

00:04:10***Saving seen data "./dancer.seen"
00:12:04 Join LinusN [0] (
00:12:41amiconnhi LinusN
00:25:04Digital007are you sure rockbox will work on the iriver?
00:26:10Digital007and would u get back to the iriver f/w if things went pear-shaped?
00:26:18Digital007and would the remote be supported as well?
00:27:32Digital007Linus do u think its possible to record from the FM tuner? (doesn't matter if WAV or MP3)
00:28:00LinusNhowever, the hard drive may disturb the reception
00:30:06Digital007Linus ive noticed something wiv the iriver
00:30:20Digital007the more you turn up the volume the less effective is the bass setting
00:30:24Digital007the archos wasn't like that
00:31:31Digital007but then again the iriver has a "software" bass setting, the Archos was hardware
00:31:33LinusNi can imagine that
00:31:55LinusNit is done to avoid distortion
00:32:13LinusNrockbox does that on the Player
00:32:22LinusNand should do it on the recorder
00:32:29Digital007i wonder if iriver really have stopped making the H1xx series
00:32:52LinusNi think they have, and that it was a stupid idea
00:32:59Digital007i agree
00:33:10Digital007i dont like their new plastic devices with colour screens
00:33:19Stryke`H120 was hardware-wise the best player of this generation
00:33:30Digital007i have an H120
00:33:49Digital007i think its beacuse they want to concentrate on their colour-screen devices
00:34:19Digital007but put it this way, if they stop making the H120, it will make it a very desirable item due to its rarity
00:34:22Digital007unlike the ipod
00:34:58Digital007However their is no mention on their website about the H1xx series being discontinued
00:35:12LinusNof course not
00:35:22LinusNthey want tp sell out the remaining stock
00:35:37Stryke`have they been real good at communicating with the userbase, in any regard?
00:35:45Digital007the shufflebug remains
00:35:50Digital007which really pisses me off
00:36:28Stryke`its the lack of the promised features which piss me off
00:36:42Stryke`i don't use shuffle as much as many, i listen to albums usually
00:36:44Digital007Mind you, its still a much better machine than the archos
00:36:55Digital007and the support of drag n drop is a real boon
00:37:09Digital007Personally i really hate colour screens on MP3 players
00:37:20Stryke`i love drag n drop, software-less transfer is the greatest
00:37:25Bagderdrag n drop?
00:37:30Digital007in windows explorer
00:37:33LinusNMass Storage
00:37:57Digital007i think a colour screen on an mp3 player is completely superflous
00:37:57Bagderusb_storage is my friend
00:37:59Stryke`LinusN: if a tag database is utilized, will it require a program on the host PC to generate it, or will the player do the work?
00:38:08LinusNhost pc
00:39:00Bagdera plugin could be written to write the db too
00:39:05Bagderit would just be a lot slower
00:39:46Digital007I think only the H110 is being discontinued, the 120 and 140 are still being made
00:42:56 Quit AciD (Read error: 104 (Connection reset by peer))
00:44:32Bagderfunnily, the h340 is cheaper than the h140
00:44:52Digital007yes bagder but the h340 is cheap plastic
00:45:16Stryke`and significantly thicker, right?
00:46:45Digital007roll on rockbox for iriver
01:01:57 Quit Digital007 ("CGI:IRC (Ping timeout)")
01:05:42LinusNBagder: found the repeat bug
01:05:50LinusNnot all that easy to fix though...
01:06:06LinusNand it's a really ancient bug
01:08:35amiconnLinusN: Repeat bug?
01:09:44LinusNfiles in a playlist can be played twice under certain circumstances
01:10:27amiconnUh? I never observed this....
01:10:39LinusNvery few people have
01:10:44amiconnYou mean 1060759?
01:16:07amiconnHmm, really strange. You said you spotted the cause?
01:16:53LinusNit happens when there are several tracks loaded in memory, and one of the playlist entries was invalid
01:17:51LinusNthen new_file() loads the wrong track when it has passed the "skip point"
01:18:32LinusNwatch my upcoming commit for details :-)
01:20:50amiconnThat explains why I never observed it, there are 2 points against it: (1) I never had playlists with invalid tracks, because I almost never save playlists. (2) It is very unlikely with a 2 MB box to have several tracks in memory (but the reporter of this bug has the 8 MB mod)
01:21:24LinusNit was the 8mb mod that was the mainclue
01:21:27LinusNman clue
01:24:00amiconnBtw: What to do with non-reproducable bugs in the tracker? (I think of 1063452)
01:25:16LinusNamiconn: easy. that's probably because he hasn't updated it to the new format
01:26:03amiconnI agree. Still there is the report...
01:27:11amiconn...usually we don't set the solved bugs to "fixed" until the next release. What about such "invalid" bugs?
01:27:32LinusNwell, tell him to update the font
01:28:12LinusNyou know..."unzip the *entire* file...
01:34:37LinusNwow, that bug has been there since before the dinosaurs... :-)
01:34:45LinusN(the repeat bug)
01:37:02amiconnSo it's time to squash it now. The iRiver will have an even larger playback buffer...
01:38:00LinusNi have committed a fix
01:38:16LinusNbut we might want to look over the whole playlist/tag handling
01:38:45amiconnYes. There still are those NULL pointer accesses...
01:38:58LinusNamong others
01:41:57 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
01:47:27amiconnBug #1064280 seems to be a design "bug" of win_crox0c itself. Instead of using the Win1251 code page (like the other wincrox?c fonts do), it tries to resemble the latin alphabet
01:49:52amiconnBut I wonder how some old version could have worked!? Was the .bdf exchanged at some point?
01:50:02LinusNi have no idea
01:50:10LinusNi have to go to sleep now
01:50:25LinusNcu tomorrow
01:50:34 Part LinusN
02:04:14***Saving seen data "./dancer.seen"
02:08:56 Join Tang [0] (~chatzilla@
02:09:18 Quit Tang (Client Quit)
02:29:51 Join amiconn_ [0] (
02:35:01 Quit amiconn (Nick collision from services.)
02:35:01 Nick amiconn_ is now known as amiconn (
02:45:04 Join Tang [0] (~chatzilla@
02:45:13 Part amiconn
02:45:14 Quit Tang (Read error: 104 (Connection reset by peer))
03:13:45 Part scott666_
04:04:18***Saving seen data "./dancer.seen"
05:22:16 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
06:04:21***No seen item changed, no save performed.
06:10:32 Quit izzy ("leaving")
06:43:07 Join matsl [0] (
06:54:27 Join HoseHead [0] (
06:58:14 Quit HoseHead (Client Quit)
07:10:11 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
07:24:53 Join midk [0] (
07:30:24 Join LinusN [0] (
07:37:19 Quit matsl (Remote closed the connection)
07:47:37 Join oxygen77 [0] (
08:04:24***Saving seen data "./dancer.seen"
08:06:54 Join AciD [0] (
08:11:03 Join amiconn [0] (
08:14:28 Join Chu [0] (
08:14:57 Nick Chu is now known as Undercover (
08:16:23UndercoverDoes Nikkelitous come in here?
08:16:36midknever seen him
08:18:59 Quit Undercover (Client Quit)
08:20:34amiconnGood morning
08:23:49 Quit AciD (Read error: 104 (Connection reset by peer))
08:24:09 Join AciD` [0] (
08:26:58LinusNamiconn: lots of reports on empty recordings
08:27:26LinusNor no files generated
08:30:13amiconnDo you mean the one one the ml?
08:31:40LinusNthere's one in the forum as well
08:34:25amiconnHmm, I can't find it?
08:35:27amiconnBtw: I found that I did not throw away the alternative recording transfer routines I made for some of my recording tests - I made a patch from them
08:37:38 Join Phoe [0] (~5409103a@
08:38:51Phoehello, could anyone posibly recant to me the button presses to get to the contrast control menu in rockbox? My display has suddenly undergone a white out and I cannot see the text well enough to reconfigure
08:39:11LinusNrockbox version?
08:39:21LinusNjukebox model?
08:39:52Phoe2.3. FM Recorder
08:40:53amiconnLinusN: I thought that was about recording level only (I even posted in that thread). Furthermore, this looks like a different problem (panic, no file at all). Unfortunately I cannot check this, as I have no Ondio FMR...
08:46:17PhoeIts actually the old style Jukebox 20Gb FM Recorder, I'm sorry I didn't realise they'd put out so many new models with similar names. I've not looked at their site for years..
08:52:10LinusNamiconn: "...and two instances of no recording file generated (but no error indication)."
08:53:46LinusNPhoe: (from the file browser) F1,down,down,right,down,down,right,down,down,right,down,down,down,right
08:54:01LinusNthen you use up/down to adjust
08:54:14amiconnLinusN: Yes. Imho that's a different thing than an empty recording (zero byte file or file with only the header, as observed with the recording bug fixed in August)
08:59:14Phoethankyou very much everyone. I was dreading the commute to work without my music.
08:59:37PhoeI suppose this might be indicitive of a deeper problem, I need it on contrast 60 to be "normal"
08:59:52LinusNanother approach is to remove the battery for a few seconds, which clears the settings
09:00:16amiconn...and reloads them from the config sector!?
09:00:19LinusNso the default value is way off then?
09:00:31LinusNamiconn: are they still doubled on disk...?
09:00:38Phoeyes, it happened suddenly yesterday morning, up to that point it was fine
09:01:06amiconnLinusN: I think so.
09:01:14PhoeI'll just have to hope that whatever caused it doesn't suddenly correct itself leaving me with a pitch black display ^^
09:01:40amiconn(and imho that makes sense, at least on the v1)
09:02:00LinusN(why not on v2/fm?)
09:02:08Phoein anycase, thank you all very much, and goodbye. Time to hit the train.
09:02:11 Quit Phoe ("CGI:IRC")
09:02:26LinusNamiconn: ah, the shutdown
09:02:44amiconnLinusN: Because on the v2/fm, you don't (regularly) change batteries, but some users do on the v1
09:19:41 Quit AciD` (Read error: 104 (Connection reset by peer))
09:30:17BagderLinusN: my generated makefile doesn't run make in tools
09:30:51Bagderok, it is done from firmware/Makefile
09:34:15LinusNexactly, that's why i changed that file
09:34:45LinusNBagder: what if i set CC= instead of CC=gcc
09:35:37LinusNdidn't work... :-(
09:36:09Bagderwell, I bet no one is using anything but gcc anyway
09:36:17Bagderand if they do, they can run make in tools first
09:36:26LinusNcc works as well
09:36:33 Join AciD` [0] (
09:40:55LinusNBagder: saw my mpeg.c commit?
09:41:55Bagdernice to have that one nailed
09:56:36 Quit oxygen77 ("Cho")
09:56:40 Join oxygen77 [0] (
09:56:53LinusNhehe, the bug was introduced in 2002, version 1.4 :-)
09:59:16 Join Zagor [242] (
09:59:23LinusNmorning Zagor
10:03:08LinusNi've been thinking about the wps context menu
10:03:35LinusNi think a separate wps-plugins.txt could be a neat idea
10:03:49LinusNexpecially if it's not included in the release zip
10:04:15LinusNthen the user can create it, and it will be persistent between releases
10:04:27***Saving seen data "./dancer.seen"
10:05:06LinusNand it could even be updated from the "browse plugins" screen
10:05:23LinusN"Add to WPS menu"
10:05:29LinusN"Add to XXX menu"
10:05:54*Zagor tries to recall last week's discussion...
10:07:02LinusNwe were discussing if we should extend viewers.txt with context menu info
10:07:38LinusNto implement "add to favorites" in a good manner
10:08:45Bagderthe problem with _not_ providing such a file in the default package is that most people won't find out how to do it
10:08:52Bagderand thus gets a worse experience
10:09:02LinusNworse than what?
10:09:15Bagderworse than us providing a fine default
10:09:22LinusNi know
10:09:40Bagderhow about we support two file names, so if the user one is made, we load the default one?
10:09:49Bagderisn't made
10:09:55LinusNfir enough
10:10:28LinusNformat suggestion:
10:10:47LinusN<plugin file name>:Menu text
10:11:10LinusNgives us a translation problem
10:12:05LinusNeach "official" plugin could have a LANG entry
10:12:17LinusNgood for voice as well
10:12:48Bagderwe don't have translation for plugins anywhere
10:12:59Zagori don't want to separate between official and unofficial plugins
10:13:05LinusNi know
10:13:06Bagderso we should rather fix that once and for all, and then we should have the menu item included in that fix
10:14:00LinusNas it is now, the blind folks have problems with the "open with" function
10:15:32Bagdervoice support for plugins is even harder than language support
10:16:20LinusNBagder: voicing/translating the plugin names is not the same thing as translating/voicing the plugins themselves
10:16:35Bagderbut closely related
10:16:56LinusNyes, but we can solve the "open with" problem easily
10:17:17Bagderwe can actually solve most of these problems that way
10:17:42Bagderby stuffing everything in the official lang file
10:17:50LinusNnot very convenient though
10:18:32LinusNso how do we solve the wps-plugins.txt "problem"?
10:23:15Bagderwithout translation for now, I'd say
10:23:51LinusNso the default wps-plugins.txt is in english
10:23:59LinusNfair enough
10:24:17LinusNehum, "wps-rocks.txt", of course
10:24:43LinusNand wps-user-rocks.txt?
10:25:13Bagderor wps-rocks-user.txt
10:25:41LinusNis this a concept that is general enough?
10:26:02LinusNi mean we might want a similar thing for other screens
10:26:32LinusNand those are added to the end of the wps context menu?
10:27:25LinusNthat's my idea at least
10:27:31*LinusN goes to work
10:27:34Bagdersounds fine
10:29:55amiconnI wonder how this could be voiced... For proper language support, it is imho necessary to have a <plugin>-<language>.lng file and a <plugin>-<language>.voice file. Not easily maintainable, though...
10:32:00Bagderit would make a million files
10:33:38LinusNyes, if there are a million plugins :-)
10:34:16Bagderwe have 24 languages
10:34:29Bagderso that systems makes 48 files per plugin
10:34:45LinusNthere could be a million .lang files, but only 24 .lng and .voice files
10:35:06LinusNwith shared .lng ang .voice files
10:35:07Bagderthen <plugin> wouldn't be part of the name I guess
10:35:17Bagderyes, with a shared system it could
10:35:35Bagderthe downside with that is that plugins don't get as independent
10:35:55Bagderbut I guess we need to compromise
10:35:58 Join Lynx_ [0] (HydraIRC@
10:36:25LinusNyes, in this case it's hard to have the cake and eat it
10:36:48amiconnIf you allow a user-defined wps-rocks.txt file, it is *very* hard to voice the menu items at all
10:36:51LinusNor, independent plugins could have their own files
10:37:08LinusNamiconn: not necessarily
10:37:39LinusNif each plugin had a defined name in .lang, the second argument in the file could be omitted
10:38:12LinusNso you just add the plugin file name, and it is correctly named in the menu
10:38:19LinusNand voiced
10:43:46amiconnSo I'd vote to only have the plugin name in the config file, not with an optional menu text
10:44:17amiconnWe have to keep the voice file size in mind, avoid adding too much.
10:44:23LinusNthat was my idea, but Bagder extended this to have all plugin strings there as well
10:44:53LinusNmaybe even adding the plugin names would be too much, i don't know
10:45:23amiconnThe default should be hard coded into rockbox, like with the default wps, default lng etc
10:47:33amiconn(voice) Switching to 12 kHz (possible by using -B 64, see MasLimitations) gives us some more room, but requires a small change in the talk routines
10:48:19amiconnI already "requested" this from Jörg: The mas seems to swallow the last split-second of sound if the dma stops.
10:48:52amiconnThis gets worse with lower bitrates, so I suggested scheduling the "pause" clip before stopping if the queue runs empty
10:50:19LinusNyeah, i remember that
10:55:04 Join kurzhaarrocker [0] (
10:57:25LinusNkurzhaarrocker: have you looked into the pause problem in the trigger code?
10:57:49 Join [IDC]Dragon [0] (
10:57:59*kurzhaarrocker blushes
10:57:59kurzhaarrockernot yet - I had to prepare the drumset for recordings this weekend
10:59:46kurzhaarrockerI still haven't made up my mind wether pause mode should inhibit the trigger to stop the recording.
11:02:28Zagorguys, I need to restart apache. those of you on the web irc client need to log in again.
11:03:58Bagderany svn experiments started?
11:04:07Zagorno, other stuff
11:04:39LinusN"stuff" - sounds reassuring :-)
11:05:07 Quit [IDC]Dragon ("CGI:IRC")
11:05:13 Join [IDC]Dragon [0] (
11:05:20kurzhaarrockerAre there any well known converters to convert a cvs repository into svn?
11:05:28Bagderat least two
11:05:43Bagderand our cvs repo is very easily converted
11:06:17kurzhaarrockerand that keeps the history info accessible I assume?
11:06:30Bagderkeeps the entire history and everything, yes
11:07:31[IDC]Dragonwhat's the benefit of svn in our case?
11:07:46Bagderaccess over HTTP
11:07:54Bagderfile renaming
11:07:56kurzhaarrockersvn is sexy
11:07:59Bagderatomic commits
11:08:07kurzhaarrockerfile moving
11:08:37[IDC]Dragonhttp is cool
11:08:47kurzhaarrockera big plus of svn: LinusN and Bagder want it
11:09:16[IDC]DragonI liked the turtoise cvs client on Windows
11:09:28Bagderthere's a tortoisesvn too
11:09:33LinusNtortoise works with svn too
11:09:35Lynx_svn means subversion?
11:09:53*Bagder runs off
11:10:00[IDC]Dragonah graet
11:10:04amiconnLinusN: Fortunately there is svn for cygwin too :)
11:10:13[IDC]DragonLynx_: yes
11:11:16amiconnLinusN: Is it possible to convert a local cvs working copy into an svn working copy?
11:13:10kurzhaarrockerI assume in the worst case you had to make a patch, checkout from svn and apply that patch.
11:13:54amiconncumbersome :(
11:14:20kurzhaarrocker<- knows hardly anything about svn, there might be other ways
11:15:45[IDC]Dragonapache alive?
11:16:42*[IDC]Dragon doesn't see any more irc, except in the log
11:16:57 Part [IDC]Dragon
11:17:27 Join Bagder_ [0] (
11:18:15 Join [IDC]Dragon [0] (~d90a3255@
11:19:04[IDC]Dragonok, I see the lines popping up again
11:19:47[IDC]Dragonfunny, I was typing "into the dark"
11:21:42[IDC]Dragonamiconn: I'll do the "silent ending" for the voice, just had no time yet
11:23:17LinusN[IDC]Dragon: i looked at radio_init()
11:23:31LinusNit puts the radio in low power mode
11:23:43LinusNhow about calling radio_stop() instead?
11:24:19LinusNso the radio is in a defined state
11:25:06*[IDC]Dragon looks...
11:26:58[IDC]Dragonseems equivalent, yes
11:28:54LinusNyes, lunch
11:29:13[IDC]Dragonnow that you remind me, I recall there may be a flaw with the low power mode
11:29:32[IDC]DragonI measured higher current even after exiting from radio
11:29:49 Quit AciD` (Read error: 104 (Connection reset by peer))
11:30:15[IDC]Dragon~8 mA extra, most likely the tuner
11:30:48amiconn[IDC]Dragon: Did you see (or even try) the new keyboard already?
11:31:26 Quit Bagder_ ("Leaving")
11:31:55[IDC]Dragonamiconn: saw it only in the log, an hour ago
11:32:25[IDC]Dragonhave not been able to appreciate it, yet
11:33:03LinusN[IDC]Dragon: might be that the CPU inputs don't like to be floating
11:34:37[IDC]Dragonwhich input?
11:35:33[IDC]Dragonaren't they only input during i2c read?
11:35:46LinusNwhat is the default state of those pins?
11:36:02[IDC]Dragonoutput, for LCD, iirc
11:36:17*[IDC]Dragon is away
11:36:18LinusNah yes
11:46:36 Join AciD` [0] (
12:04:29***Saving seen data "./dancer.seen"
12:29:56 Quit midk (Read error: 104 (Connection reset by peer))
12:29:57 Join midk_ [0] (
13:06:43 Quit AciD` (Read error: 104 (Connection reset by peer))
13:25:08 Join AciD [0] (
13:34:10 Quit AciD (Read error: 104 (Connection reset by peer))
13:50:32 Join AciD [0] (
14:04:31***Saving seen data "./dancer.seen"
14:31:19 Join elinenbe [0] (~elinenbe_@
14:50:00 Join ripnetUK [0] (
14:50:31ripnetUKis the web log of this irc channel working? nothing in it since 1136?
14:50:46ripnetUKthats a yes :)
14:51:46*Bagder pokes logbot with a stick
14:52:03Bagderit logs fine
14:52:28ripnetUKyeah, i checked the log and my comment got in... everyone was just quiet for a few hours
14:52:32Zagorahh, boring it shows in the log :)
15:25:19 Join HoseHead [0] (
15:25:26LinusNToms Hardware reviews the Neuros II player, commenting the size of the player:
15:25:30LinusN"Of course, if you do jog with the Neuros, at least you can use the Neuros player as a self defense bludgeoning weapon!"
15:34:54HoseHeadAnyone know if there's a way of getting 320kbps songs to play without hanging every 30 seconds while it spins up and rebuffers? (other than setting the hdd spindown at like 35s...)
15:35:32Zagorit should never hang like that. what model are you using?
15:36:52LinusNHoseHead: try fiddling with the anti-skip buffer
15:37:04LinusNbut as Zagor said, it should not happen at all
15:38:34 Join webguest98 [0] (
15:40:25 Quit webguest98 (Client Quit)
15:50:51 Quit HoseHead ("CGI:IRC (Ping timeout)")
16:03:03Zagorgreat feedback
16:03:52kurzhaarrockermaybe he's still waiting for playback to continue?
16:04:33***Saving seen data "./dancer.seen"
16:07:23 Part LinusN
16:07:54 Quit midk_ (Read error: 54 (Connection reset by peer))
16:46:41 Join methangas [0] (
17:02:50 Join mecraw_ [0] (~lmarlow@
17:10:57 Part Zagor
17:30:51 Part kurzhaarrocker
18:00:25 Join BlueBandicoot [0] (
18:01:16 Quit BlueBandicoot (Client Quit)
18:04:36***Saving seen data "./dancer.seen"
18:16:11 Quit [IDC]Dragon ("CGI:IRC")
18:38:53 Join matsl [0] (
18:54:29 Quit AciD (Read error: 54 (Connection reset by peer))
18:54:55 Join AciD [0] (
18:54:57 Join einhirn [0] (
19:07:26 Join Stryke` [0] (~Chairman8@
19:18:19 Quit einhirn (Read error: 104 (Connection reset by peer))
19:22:21 Quit ripnetUK ()
19:57:48 Join einhirn [0] (
20:04:38***Saving seen data "./dancer.seen"
20:11:39 Quit Lynx_ (" Like's GUI? Then try HydraIRC -> <-")
20:31:44 Quit oxygen77 ("Tcho")
20:31:50 Join oxygen77 [0] (
20:47:03 Join [IDC]Dragon [0] (
21:14:45 Quit elinenbe (" HydraIRC -> <- \o/")
21:45:38 Join scott666_ [0] (
21:54:43 Quit scott666_ ("i'll be back...eventually...")
22:04:42***Saving seen data "./dancer.seen"
22:11:47 Quit methangas (" Try HydraIRC -> <-")
22:37:11 Quit oxygen77 ("Cho")
22:41:43 Join LinusN [0] (
22:42:47Bagderevening LinusN
22:47:11 Quit Stryke` (Read error: 104 (Connection reset by peer))
22:47:19LinusNhola Bagder
22:47:41 Join Stryke` [0] (
22:48:48 Join [av]bani [0] (
22:50:40*[av]bani pokes linusn
22:51:09[av]baniafaict the ihp1xx cpu will never be fast enough to do any kind of realtime ogg encoding
22:51:27[av]baniand noone has ever written an asm encoder anyway
22:51:50[av]baniprobabyl the best can be hoped for is mp3 or wav capture, then after-the-fact compression
22:52:16LinusNour plan is wav and flac, and maybe later mp3
22:52:17[av]banieg capture to wav, then compress while youre on the way home
22:52:54[av]baniand less-than-realtime would allow one to do ogg in the ihp _now_ as opposed to much later
22:53:40[av]baniwell considering an asm decoder is hard enough on the coldfire, i seriously doubt even an asm _encoder_ is possible
22:54:13[av]baniat best it would be a seriously crippled one, with lots of shortcuts and not so good quality
22:54:48[av]banibtw i guess you're not a 68k expert? (given your movem.l comment in irc)
22:55:22[av]baniin one of the irc logs you 'discovered' to use movem.l for task switching
22:55:49[av]baniwhich i would interpret to mean youre not very versed in 68k asm
22:55:52LinusNi never said i discovered it
22:56:07[av]baniwell more like, you figured out you could use it for that
22:56:18LinusNi said that we could settle with one movem, since we didn't need to store the EMAC registers
22:56:21*Bagder knows LinusN has done his fair share of 68k asm
22:56:44[av]banido the emac regs have a dirty flag?
22:57:11LinusNi don't know, but i doubt it
22:57:22[av]baniyou could make emac functions or macros which keep a dirty flag, and then use that
22:57:41[av]baniunless youre not going to use emac at all
22:57:52LinusNwe will use it in the codec's
22:58:17LinusNbut that will only be in one thread, so there won't be a need to save the regs
22:58:55[av]banii take it gcc doesnt support emac at all
22:59:42[av]bania bunch of macros then?
22:59:53LinusNyes, inline asm
23:00:46[av]banii guess you know about using movem.l for fast bzero() / memset()
23:01:29[av]baniand memcpy :)
23:02:04[av]banihmm you found usb detect, i assume it was one of the gpio pins
23:04:33[av]baniwhats the highest priority atm, eg stumbling blocks
23:04:49LinusNata driver is my next project
23:05:08[av]banino stock code can be nicked from somewhere else? :)
23:05:41LinusNwe have a working ata driver
23:05:56LinusNin rockbox
23:06:04LinusNi just have to port it to the iriver
23:06:39[av]banigaaah you guys really considering svn?
23:06:58LinusNwhy not?
23:07:05[av]baniit has .. issues
23:07:15LinusNsuch as?
23:07:19[av]banii use cvs and svn for multiple projects
23:07:22Bagderand cvs doesn't?
23:07:29[av]baniless than svn
23:08:04[av]baniwhen svn fails (in a multitude of obscure ways) it gives completely nonsensical error messages
23:08:49[av]banialso, the much touted branching is very problematic
23:09:49[av]baniif you branch from a branch, then blow away the parent branch (a common thing to do), it makes merging back to trunk very hard / impossible
23:10:20[av]banicvs has its problems, but i havent found that svn actually solves anything
23:10:48Bagderthen you are lucky
23:10:59[av]baniit also tries to be 'cvs-ish' but when it differs, it differs in obscure and sometimes undocumented ways
23:11:22[av]baniwhat are you trying to _solve_ that svn does that cvs does not?
23:11:38Bagderone little thing: access over HTTP
23:12:10BagderI could name more things
23:12:27[av]baniwell having used both, i have to say if thats your only reason then thats a huge mistake
23:12:32[av]banibut its your mistake to make i guess
23:12:42Bagderit is not the only reason
23:12:48[av]banithe other?
23:12:49Bagderyou asked for reasons, I named one
23:13:34Bagderrenames/moves and atomic commits
23:13:48Bagderit may have its downsides
23:13:56Bagderit is still the sane way to go
23:13:59[av]banii dont like svn's reliance on apache, given apache's less than stellar history of security
23:14:02Bagderthe bugs should be fixed
23:14:17Bagderit doesn't rely on apache if you don't want to
23:14:26[av]banii know, but everyone does do it that way
23:14:40[av]baniits hard not to use apache, if you want access via http
23:14:44[av]baniits very tied to it
23:15:39[av]banirenames/moves is pretty rare, in all projects ive dealt with
23:15:53Bagderyes, because cvs doesn't do them
23:16:03[av]banibecause you dont usually do that in code...
23:16:16[av]baniat least any of the projects i've worked on, cvs or not
23:17:01LinusNif cvs had let us rename and move files, the rockbox source would have looked a lot different
23:17:04[av]baniif you are renaming/moving often enough that you require svn over cvs, imo something is wrong with your developers
23:17:14BagderI disagree
23:17:20Bagderour code evolve
23:17:30Bagderand so does our file tree
23:17:59LinusNwe had a different idea about the file tree four years ago than we have now
23:18:23[av]baniwell, svn tries to be cvs-ish but when it diverges, there is often no documentation about it
23:18:29[av]baniand when it fails, theres no documentation either
23:18:35[av]baniyou have to guess
23:18:47Bagderit is in heavy development
23:19:02[av]baniyes, and thats not something you want to be using for your rcs is it?
23:19:09Bagderwhy not?
23:19:15Bagderit makes the bugs get solvde
23:19:18[av]baniif you have to juggle svn bugs along with your own...
23:19:30[av]banithey dont though, or at least they apparently arent considered high priority
23:19:40LinusNif it helps the svn development, yes
23:19:44BagderI disagree there too
23:19:52[av]banithe impression i get is that the svn devs either dont think they are an issue
23:20:01[av]banior subconsciously avoid doing things that make svn break
23:20:40[av]banibut i amange to break it all the time :()
23:21:16Bagder... and I do have commit access in the project ;-)
23:21:30[av]baniwell imo having used both is
23:21:59[av]banicvs is clunky but it works,
23:22:12[av]banisvn is fragile, when it breaks it is very frustrating
23:23:02[av]baniits very pretty otherwise, like a glass chandelier
23:23:30[av]banibranches are 'neat' but fail in non obvious ways
23:24:13LinusNand what do the developers respond to your bug reports?
23:24:14[av]banimove/rename is neato, but you can delete/readd in cvs (though you do lose the history, which is a given)
23:25:27[av]baniwell the svn devs i talk to seem rather cranky
23:26:42LinusNthat's bad
23:28:00[av]banibut there are some non obvious things about svn for cvs users
23:29:30[av]baninormally you do 'cvs update' to see if you have anything in your tree which is M odified or ? should be added
23:29:36[av]banior A scheduled to add
23:30:26[av]baniin svn, svn update is only for grabbing updates, and 'svn status' takes over that function
23:30:33[av]baniwhich is distinctly different from 'cvs status'
23:31:33[av]banialso you really never ever want to merge from trunk into a branch, it will cause problems merging back into trunk later :<
23:31:42[av]baniyou only want to merge back to trunk
23:32:45[av]baniin cvs you have a simple .cvsignore file, in svn you have a clunky interface for that called properties
23:38:24[av]banilets just say that in having used both, i have not found that svn _solved_ anything
23:39:53[av]banicvs might be clunky but it is well understood and widely tested and rarely _breaks_
23:40:07[av]banisvn is still in development, and rather feels that way
23:40:18LinusNi see
23:40:52[av]banimaybe in time it will become mature but it doesnt feel that way now :()
23:45:12[av]baniif you are thinking of rearranging your tree with 4 year old files then i assume the history of those files is not super critical
23:45:32[av]banithat is, you dont have any urgent reason to revert to older revisions
23:45:49[av]baniso add/delete should not be a major issue
23:46:13[av]baniits ugly,but doable
23:47:33Bagderthe files are not that old
23:47:38Bagderthe hiearchy is
23:48:32Bagdertime to sleep
23:48:38LinusNnite Bagder
23:48:47LinusNi'd better sleep too
23:48:56LinusNcu around
23:49:09 Part LinusN
23:49:15 Part [av]bani
23:52:59 Quit [IDC]Dragon ()

Previous day | Next day