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 2005-04-24

00:02:12rasheroh jeez, they're all using the same bloody implementation
00:04:24preglowlisp :///
00:05:08rasherBagder: that looks nicely standalone
00:05:26rasherpreglow: I have a tcl version as well
00:05:49t0masgood night
00:05:54 Quit t0mas ("bye")
00:06:45amiconnrasher: There is a crc32 implementation somewhere in the archos flash plugins
00:07:20amiconnActually, there are 2: One space-saver in the flash plugins, doing a nibble at at time. This one is relatively slow
00:08:03amiconnThe second is in the flash image creation tool. This one does a full byte at a time. Somewhat bigger lookup table, but also faster
00:08:24rashersounds like this one more or less
00:08:40amiconn1 KB lookup table
00:09:49rashermake_firmware.c ?
00:10:54rasheroh there
00:10:58amiconnyup. CalcCRC32
00:15:15 Quit stevenm ("Leaving")
00:17:05 Join noclue2 [0] (
00:18:38 Quit Ka_ (Read error: 54 (Connection reset by peer))
00:20:07preglowtcl :///
00:20:34 Join Ka_ [0] (
00:22:35***Saving seen data "./dancer.seen"
00:23:03amiconnI started hacking the MMC driver to support variable blocksizes... and I just found *why* the 2 GB MMC uses 1024 bytes/block instead of the usual 512
00:24:05amiconnThe reason is - it has to. It is impossible to encode a block count above 2^21 with the scheme used in the CSD block
00:24:31amiconnBut 2 GB with 512 bytes/block would be almost 2^22 blocks
00:30:22 Quit johntb ("Leaving")
00:33:31 Quit ehntoo (Read error: 110 (Connection timed out))
00:36:37rasheramiconn: what would "UINT8" and "UINT32" equal to?
00:36:39rasherchar and int?
00:36:43HClunsigned char
00:36:45HClunsigned int
00:36:50rasherclose :)
00:37:03amiconnUINT32 == unsigned long, to make sure
00:37:41HClmost of the time sizeof(long)==sizeof(int), on x86 anyways
00:38:02rasherwell we're dealing with all sorts of wacky archs here though :)
00:38:09HClyup :)
00:44:03rasherhm, it actually produces something that vaguely looks like a crc
00:44:26rashersome files return negative values though
00:44:31rasherwhich I wasn't expecting
00:44:45rasheror maybe %d isn't suitable for feeding with an unsigned long
00:44:50rasherfor snprintf
00:46:46rasherit's.. instant on iriver
00:46:54rasherI'd probably need to loop it a while
00:46:58rashernot now">
00:54:15HClyes, it isn't
00:54:17HCluse %u
00:54:18HClfor unsigned
00:54:20HClor %x
00:54:23HClfor hexadecimal
00:54:46HClnice, i'll look at it tomorrow
00:54:48HClrather sleepy now
00:55:25 Part rasher
01:05:32 Join bippy [0] (
01:05:33bippyowt worth testing ?
01:09:46MoosCamarogood night all
01:10:01 Quit bippy (Client Quit)
01:10:10 Part MoosCamaro
01:17:58 Join Hornet_ [0] (
01:26:47 Quit crwl (Read error: 60 (Operation timed out))
01:30:33 Join Hornet__ [0] (
01:36:12 Quit Hornet (Read error: 110 (Connection timed out))
01:40:20 Quit preglow ("leaving")
01:45:27 Quit Ka_ (Read error: 110 (Connection timed out))
01:47:40 Quit Hornet_ (Read error: 110 (Connection timed out))
02:12:52 Quit Sucka ("a bird in the bush is worth two in your house")
02:14:28 Quit Shagnar ("( :: NoNameScript 3.81 :: )")
02:22:39***Saving seen data "./dancer.seen"
02:31:43 Join Strath [0] (
02:39:22 Join DMJC [0] (
02:46:42 Join stevenm [0] (
02:47:10stevenmAnybody want to explain to me why the PIC C compiler take every line of inline ASM I write and replicate it FIFTEEN TIMES ?!
02:47:26*stevenm beats pic c in the head with an uglystick
02:48:55stevenmthis friggin code looks like it's fallen out of the ugly tree, hit every ugly branch along the way, and then got run over by the ugly truck
02:50:51stevenmwhy are there 15 GOTOs. Strung together ? ?
02:55:26 Part stevenm ("This compiler makes baby jesus cry")
03:38:49 Join ashridah [0] (
03:56:59 Join Ka_ [0] (
04:06:12 Join QT_ [0] (as@area51.users.madwifi)
04:09:09 Quit DMJC ("Leaving")
04:17:11 Quit QT (Read error: 110 (Connection timed out))
04:22:40***Saving seen data "./dancer.seen"
04:33:13 Quit cYmen_ (Client Quit)
05:07:04 Quit gromit` (Remote closed the connection)
05:10:59 Quit Strath ("Client closed")
05:44:27 Join Strath [0] (
06:18:15 Join stevenm [0] (
06:22:44***Saving seen data "./dancer.seen"
07:07:49 Join lostlogic [0] (
07:44:07 Quit stevenm ("Leaving")
07:47:00 Quit XShocK (Read error: 104 (Connection reset by peer))
08:22:48***Saving seen data "./dancer.seen"
08:44:20 Join Harpy [0] (
08:44:57 Quit edx (Read error: 110 (Connection timed out))
09:04:37 Quit ashridah ("Leaving")
09:05:38 Join ashridah [0] (
09:30:40 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
09:47:13 Join edx [0] (
10:13:33 Quit Ka_ (Read error: 110 (Connection timed out))
10:14:41 Join Ka_ [0] (
10:22:52***Saving seen data "./dancer.seen"
10:51:17 Join t0mas [0] (
10:56:10 Nick QT_ is now known as QT (as@area51.users.madwifi)
11:03:47HCl*sighs and wonders what goes wrong*
11:04:09ashridah<HCl> let me check my notes
11:07:10HClokay, i think i might've found it
11:07:19ashridahheh. it's a webcomic. one of the guys is an inventor. things keep blowing up on him. when they do, his typical phrase when someone asks him if something's supposed to be on fire is 'let me check my notes'
11:08:45HCl :P
11:10:04HClone day i'll want to rewrite dbtree.c so its easier to add things to the database structure without having to search the entire code for all the places with the sizes of the structures..
11:13:36HClwatashi wa baka desu
11:13:51HCldisappointing myself by trying to browse the id3 db immediately without rebooting to load the new code P
11:14:08 Join muesli- [0] (
11:15:35HClheh, you fix one thing, and you break the other O.o.
11:15:37muesli-g'day mates
11:16:02*ashridah looks at muesli-
11:16:09ashridahyou're Australian?
11:16:54ashridahguess not.
11:17:37muesli-have been in oz last october ;) since i am "g'day" coined ;)
11:17:55ashridahrighto then
11:19:34muesli-yepp :D
11:19:47muesli-where do you come from?
11:20:23ashridahAustralia. born and bred. sadly. :(
11:20:49ashridahspent most of my life in Melbourne.
11:21:26muesli-didnt made it to money was running out ;-(
11:22:02ashridahyou didn't miss too much. the weather's been pretty crappy this year.
11:23:09muesli-"good" to previous plans were to fly over to tasmania and going back to mainland in mb
11:24:35HCli need someone who's worked on the database code..
11:24:40HClwould that be Bagder ?
11:24:48*HCl goes to check the cvs..
11:25:27muesli-wth are cvs???
11:25:31HClzagor, ok.
11:25:36HClconcurrent versioning system.
11:25:40HClits where rockbox' code is stored.
11:26:14HCli need zagor to tell me how it would be possible for songs4album to malfunction, when "allsongs" works fine..
11:26:42*HCl goes to run his database tester on the database he has on his iriver..
11:29:31 Quit amiconn (" l8rz")
11:31:25HClnope, the database is fine..
11:40:19*HCl stares at the simulator.
11:40:38HClunsupported database version...... 82...!? O.o.
11:46:45 Quit t0mas ()
11:49:28HClah.. something goes wrong with endian detecting on linux..
11:51:54 Join rasher [0] (
11:53:46rasherchanged the %d to %x and added a rb->close
11:54:00rasherlet me loop it a bunch and add a timer
11:54:03HCl :)
11:54:13HCli'm working on getting the database to work...
11:54:32HClartists and album names work...
11:54:35HCland browsing them too
11:54:39HClbrowsing all songs works..
11:54:48HClbut browsing songs in an album is broken... for unknown reasons
11:55:12HClbut i got the simulator to work, should help.
11:55:32rasherwould moving the crc_table outside the function change anything?
11:56:21HClbut you could place it in iram
11:56:31HClhow big is the table?
11:56:43rasherI think amiconn said 1k
11:56:52rasherI'll try without it first
11:58:23HClall songs by artist works too, i just noticed that
12:00:46 Quit _Lucretia_ ("Leaving")
12:04:59rasherwell it's certainly not the crc'ing that's going to take long
12:07:22rasherlooping it 10000 times takes 32 seconds
12:07:49HClthen if we can make it unique enough, it might be a good idea to search for files in our database by hash, rather than filename
12:09:56HClthis is odd... it goes to songs4artist..
12:10:47HClthats wrong, it should be songs4album
12:12:16HCl if (newextra >= artiststart)
12:12:16HCl c->currtable = songs4artist;
12:12:16HCl else
12:12:16HCl c->currtable = songs4album;
12:12:24HClanother check for the old database format
12:12:26HClin the middle of the code
12:12:57HClright, after i fix this, everything should work fine.
12:14:35HCl :)
12:15:45HClit does :)
12:15:53rasher13 seconds at 120MHz ^_^
12:16:04HClnice :)
12:16:31rashergotta run now
12:19:32 Join MoosCamaro [0] (
12:19:38*HCl fixes his tabs
12:19:46MoosCamarohi all
12:22:54***Saving seen data "./dancer.seen"
12:26:21 Quit muesli- (Read error: 113 (No route to host))
12:30:04HCli love it when you can change something from 0% to "done" on the wiki :)
12:44:27HCli need to fix my id3 tags...
12:45:57*HCl wants a new browse mode too, left/right being hop to next letter in the alfabet, and select being enter, and stop being leave
12:53:51 Join muz [0] (
12:54:28muzHCl: you know all this database stuff ur doin... i was wonderin if any of it involves being able to build the database without a pc?
12:56:42HClnot yet.
12:56:58HClon the archos, it will be pretty much impossible. its possible on iriver, technically.
12:57:54muzcool, btw i like ur idea about using left and right to go down a letter in the alphabet
12:58:03 Join amiconn [0] (
12:58:22HClmmm, need to know what other people think, i know i have too many artists and albums to do without that at least.
12:59:40muza friend of mine has an archos unit which functions as an external drive, and whenever you copy new mp3's to it the unit itself builds the database with those new songs (i think its the gmini400)
13:00:44HClwell, the biggest problem is that our database is heavily sorted at the moment.
13:01:28HClsorting has to happen in ram
13:01:38HClarchos recorders and players have extremely limited ram..
13:01:57HClthe 32mb of iriver should be enough
13:02:05amiconnBuilding the db is definitely possible on the archos
13:02:09HClofcourse we can not use mergesort
13:02:15HCland use sorts like bubblesort
13:02:31amiconnI mean, we _can_ use mergesort
13:02:32muzok i dont understand that one bit
13:02:53HClamiconn: how so?
13:02:54muzkeep chattin
13:03:05HClisn't mergesort too big to fit in ram
13:03:09HClsince you have to have all entries in ram
13:03:37amiconnIn fact, the thing what mergesort is about is when you have to sort a file which doesn't fit in ram at once
13:04:02HClwell.. its actually about splitting the search problem into bits in order to achieve n*log n complexity...
13:04:27amiconnMaybe we're talking about different things here
13:04:34HClbut i guess you can run mergesort on disk
13:04:36HClrather than memory.
13:04:56HClmergesort = splitting the array in half, then mergesorting the left half, mergesorting the right half
13:05:00HClthen merging the two sorted halves
13:05:44HClwhere an array of size 1 is already sorted
13:05:55HClits an recursive algorhythm...
13:05:57amiconn(1) Load first <ram-size>block in memory, (2) do whatever sorting on it, (3) write block to disk (e.g. separate file 1)
13:06:15amiconnRepeat until no more data, writing to separate file 2...n
13:06:25HClohyea. i forgot that you can mergesort on any amount of blocks
13:06:29HClnot just split them in two
13:06:36HClbut in any number.
13:07:00HClanyways, creating the database on player is not something that has a high priority for me..
13:07:22amiconn(4) Open all intermediates, read first entry of each, write the one that is first-in sequence and read again from the intermediate you just used the entry from.
13:07:24HClif you have the old database on the player, it should be much less hard on the player anyways, since it can use data of past runs
13:07:35amiconnRepeat until all intermediates are EOF
13:08:10HCli'd still want to do it in ram on the iriver though
13:08:25amiconnTo speed it up, the reading of intermediates can of course be done in chunks of (ramsize/no_of_intermediates)
13:08:54HClwell, i'm not gonna implement it for now, heh.
13:09:01HCli'm wondering
13:09:06amiconnYou see, I already thought about generating the database on target...
13:09:15HClare there any objections if i upgrade the cvs to the new database format?
13:09:23HClit can do everything the old one can
13:09:26HCland works fine
13:09:30amiconnI'd wait for Zagor for that
13:09:36HCli'm not in a hurry
13:09:44amiconnHe's the database man
13:09:52HClyea, i noticed by checking the cvs
13:10:08*HCl prods people around
13:10:11amiconnDid you try to compile for archos? I'd be interested in the code size change
13:10:12HClanyone awake who has an iriver?
13:10:17HClhold on
13:10:22HClcode size change is minimal, really.
13:10:30HCllet me build for archos
13:10:48HCli don't have an archos toolchain..
13:11:00HClcan you check the code size change?
13:11:05HClthere's a patch against current cvs
13:11:16amiconnWhich source files does that involve?
13:11:23HClonly dbtree.c
13:12:02amiconnHmm. A unified diff would be preferred
13:12:15HCli made a context one, how can i do an unified one ?
13:12:17HClcvs diff -u ?
13:12:52HCli don't like unified..
13:12:53HClbut okay.
13:12:57muzhey HCl ill try out ur new database if u want on my iriver
13:13:35muzshall i just use ?
13:13:45HClmuz: yup :)
13:14:18HCli put a unified version on it.
13:14:24HCli personally always make context ones.
13:14:30HClcause they're much more human-readable
13:15:17muzerrr when i booted it it said unsupported database version
13:15:30HClyou need to get the
13:15:32amiconnMust be a personal preference thing... I can read unified diffs easier than contextual ones...
13:15:32HClfrom the same location
13:15:37HCland update it to the new format
13:15:39muzlol oops
13:15:53HClamiconn: *nods*
13:16:56muzhow do i update it to the new format
13:17:10muzshall i run makedb.bat?
13:17:33HClperl.exe −−path /
13:17:42HClunless you altered makedb to run songdbv2
13:17:45muzcrap im running windows
13:17:47HClor renamed songdbv2 to
13:17:54HClget the perl.exe
13:17:56HClof the tagdatabase wiki
13:17:58HClwith the dll
13:18:03muzk i think i know wat to do
13:19:19HClit should react exactly the same as the old format
13:19:34HCland there's no code that utilizes the new data stored in it yet
13:20:55 Join matsl [0] (
13:22:21muzis it bad not to safely remove the iriver?
13:23:07*HCl sees a stray temp[100] that he needs to remove.
13:23:11*HCl removes.
13:25:06 Quit Ka_ (Read error: 60 (Operation timed out))
13:25:25 Quit muz ("CGI:IRC")
13:30:35 Join cYmen [0] (
13:35:01 Join t0mas [0] (
13:36:20 Join Ka_ [0] (
13:36:20t0masmy neighbor's son (dutch: buurjongen) just died
13:36:40t0masI talked to him just yesterday... just fine... alive..
13:36:40HClwhat happened? :(
13:36:51t0masgot hit by a hockystick that afternoon
13:37:01HClwow o.o.
13:37:11t0masthis morning passed away when he got up... partents weren't at home
13:37:28t0mascalled us... my mother helped him back into bed
13:37:33t0masdailed 112
13:37:41HCl :(
13:37:49MoosCamaroall my condolece
13:37:50t0maspolice searched his parents
13:38:00t0masand before they found his parents he was just dead :X
13:38:24HClhow did he get hit with the hockeystick..? where..?
13:38:36t0masnormal game...
13:38:51t0mas(dutch: net onder 1 rib, in z'n lever...)
13:39:01t0mastnx MoosCamaro
13:39:10t0masHCl: hij had een leverbloeding..
13:39:15HClah :(
13:39:24t0masmaar gister had ie nog niks
13:39:39HClyea, its something you don't easily detect
13:39:45HCli didn't know you can die from that though..
13:40:00t0masyesterday he stepped by... having nothing but a little pain in his side... (thinking it was a bruce)
13:40:11t0masdrank something... and this morning he's just dead
13:40:38HCl :(
13:40:48t0masHCl: buikholte vol bloed, longen in elkaar geklapt... ze hebben nog geprobeerd om te reanimeren enz.. onderweg naar t ziekenhuis :(
13:42:56HCl :(
13:43:39t0masand the next disaster was my girlfriend...
13:44:10t0masa friend of me was here... and I asked him to tell her I wasn't comming with them this afternoon
13:44:43t0masso he told her: "Tomas's neighbor is dead" and she didn't really understand... and freaked thinking me or my sister was dead
13:45:03HCl :x
13:45:07HClpoor girl.
13:45:21HCli don't understand how you can misunderstand that though :X
13:45:34t0mascellphone.. crying friend telling it
13:45:56HCli guess its just one of those days :X
13:46:03HCl :/
13:46:20HCldid your neighbour have any siblings? :(
13:46:33t0massiblings? (dutch?)
13:46:44HClzussen of broers..
13:46:49t0masoooh ok
13:47:01t0masyes, he has a little broter.. 6 yo and a sister 4
13:47:08HCl :/
13:47:11HClhow old was he?
13:47:22HCl :/
13:47:36t0masmy mother picked up his brother/sister
13:47:57HClhow are they?
13:48:02t0mastotally scared...
13:48:06HCl :(
13:48:08t0masdidn't understand what was happening
13:48:38t0masand then their parents came here... just crying... as they were told at the same time Job was dead...
13:48:47HCl :(
13:48:56t0masand the little boy asked "Als Job dood gaat, weet hij dan wel hoe dat moet?"
13:49:08t0masthat was just too eh... zielig
13:49:17HCl :x...
13:49:25HClif they don't understand
13:49:30HClit will be easier for them to cope with it..
13:49:46HCland the parents having other children helps too... at least he wasn't their only child..
13:50:22t0masbut they do understand what dead is...
13:50:29HCl :/
13:50:37HCl*sighs deeply*
13:50:42t0masonly I don't think they understand they're never going to see Job again
13:51:24t0masman... I don't even really got it myself the first hours
13:51:51HClwell, you have to remember the good times you shared, and if you have a religion, use that a bit for comfort..
13:51:54t0masit's just to weird... someone just "disappearing"
13:52:03HCl :/
13:52:22HClyea, i know..
13:55:08HClamiconn: whats the codesize difference?
13:55:18HClamiconn: be sure to remove that temp[100]
13:55:22HCl :/
13:55:30*t0mas is away: gonna walk a little round...
13:55:34t0mastnx for just listening...
13:55:41HClno thanks..
13:55:49 Join muesli- [0] (
13:55:49HClits the least i can do :/
13:56:06*HCl knows what its like to lose someone to death, unfortunately :/
13:58:30*HCl is gonna find some distraction by watching f1
13:59:12MoosCamaroyea the San Marino GP
14:00:11Hornet__I know what it's like to lose someone too.. I know nobody can really make you feel any better by feeling sorry for you or any of that, believe me I know..
14:00:19amiconnHCl: +128 bytes :)
14:00:23MoosCamaroHCl: look the frenchs car Renault
14:00:24muesli-schu is in hopeless position ./
14:00:31t0masyeah... well it's ok to have some people listen...
14:00:34Hornet__But for what it's worth, I'm sure he was a good kid
14:00:51Hornet__And it really is a hard thing to accept
14:00:52t0masHhe was a fabulous hocky player...
14:01:02Hornet__I wish you and his family the best of luck
14:01:18Hornet__Sometimes the only thing you can do is remember.. never forget
14:01:28MoosCamarothe life is sometimes very difficult
14:02:37MoosCamarotime cure sometimes
14:03:05Hornet__After 8 years it still brings tears to my eyes
14:03:43Hornet__When you lose someone close, even time doesn't take away the pain that replaced their life
14:03:51Hornet__Especially if it's not a natural death..
14:04:01Hornet__The person I lost was shot to death, murdered in cold blood.
14:04:01amiconnHCl: songdbv2.plk produces some strange console output
14:04:10HClamiconn: strange?
14:04:23amiconnWide character in print at tools/ line 325.
14:04:24HClamiconn: what does it say thats odd..?
14:04:26amiconn15 times
14:04:31Hornet__But.. life goes on. We just have to do our best.
14:04:33HClodd... i don't have that o.o
14:04:37*Hornet__ goes to bed.
14:04:40HClline 325?
14:04:47t0masHornet__: beiing murdered is just the workst thing that can ever happen
14:04:59amiconn15 is just the number of songs I have on my Ondio
14:05:22HCl325 is the line of the dumpint thing..
14:05:33HClmaybe its your perl...?
14:05:37HCli don't get those errors...
14:05:43HClnot on linux, not on windows..
14:06:05amiconn$ perl -v
14:06:11amiconnThis is perl, v5.8.6 built for cygwin-thread-multi-64int
14:06:25HClcan you try the original
14:06:28HClsee if it has it too? doesn't do this
14:06:53amiconnJust checked
14:07:27HCli'm confused :(
14:07:31HCllet me ask my local perl guru
14:09:18amiconnThat's indeed strange. The dumpint() function is identical
14:09:18 Quit muesli- (Read error: 54 (Connection reset by peer))
14:11:07amiconnThe only cause I can imagine is that the number is >= 2^32 in one of your calls
14:11:45HCldumpint writes it in bigendian
14:12:07HClmy perl guru is offering a better way to write an raw int to a file
14:12:15HClbut we need to know which endianness it is
14:12:25amiconnWe need big endian
14:12:39amiconnAll our current targets are big endian
14:14:07HClprint DB pack "N", $int;
14:14:11HClreplacement for our dumpint
14:14:20HClhe's better at perl :P
14:14:59HCltry replacing dumpint with print DB pack "N", $num;
14:15:00HCl ?
14:15:04HClsee if that helps
14:18:04HClyou're probably right though...
14:18:17HClit would give that error if the number was > 2^32
14:18:24*HCl goes to check his dumpints
14:18:34amiconnI found it
14:18:37amiconnline 508
14:19:10amiconn(unsigned)(-1) is >2^32 for a 64-bit-int perl
14:19:36HClthat pack line should fix it then.
14:20:29HClor you can just & 0xFF
14:20:55HClthat line in dumpint
14:21:12HClthe pack is probably cleaner.
14:21:24HClsince its especially to write and read binary formats
14:21:34HClin 32bits
14:22:58***Saving seen data "./dancer.seen"
14:23:22amiconnSame goes for dumpshort, I just googled a bit
14:23:32HCldumpshort ands them properly
14:23:49HClso it doesn't have that problem
14:24:07amiconnYes.. but imho print DB pack "n", $num should be better optimised
14:24:18HCli don't know how to do shorts that way though
14:24:22HCl "n" is short?
14:24:52 Quit edx (Read error: 104 (Connection reset by peer))
14:24:55 Join edx__ [0] (
14:25:33 Quit edx__ (Client Quit)
14:25:40amiconnHCl: yes. See
14:29:28amiconnSeems to work (though I can't check with a larger database; only have my Ondio with me)
14:30:02amiconnwork == both the "print pack"-changed script and the target code
14:35:40HClokies :)
14:35:57HCllet me modify them, i'll change it in cvs too
14:37:02 Join preglow [0] (
14:38:19HClhmm... i want to alter the script so that if an album doesn't have an album tag, it takes the name of the directory instead
14:43:40HClin fact, i think its malfunctioning when you have multiple songs without an album..
14:44:51preglowso, you got things up and jumping again?
14:45:19MoosCamarohi preglow
14:48:49 Join xen` [0] (
14:49:33HClpreglow: nah, the new format is working, it does everything the old format did
14:49:38HClincluding bugs in the old format :P
14:50:03preglowwouldn't want people going homesick on the old format, now
14:50:24HCl :P it doesn't make a difference, really
14:50:32HClyou won't notice it having a different version of the database
14:50:38HCland i haven't written any code that utilizes the new info yet
14:50:47HCli should start on runtime info
14:50:53HClbut i suddenly got struck by lazyness
14:51:32*HCl prods rasher
14:51:37HClrasher: hows the crc thing?
14:51:47HClis there any code that checks for duplicate hashes?
14:53:09amiconnHCl: database size increased by about 30% for my really small testcase
14:53:15amiconn(1 album, 15 songs)
14:53:24HClsounds plausable.
14:53:33HClit stores the genre now too
14:53:59HClas well as having a fixed filenamelength thats determined like the length of the other fields
14:54:14amiconnThat's probably strange: Max genre length : 8 (Metal)
14:54:23HClit aligns it to 32bits
14:54:28amiconnah ok
14:54:39HCllike all fields, or at least, all fields should be aligned
14:58:23HCli don't quite understand why they should be aligned though.
14:58:26HClcan anyone explain that?
14:58:58amiconnThe number fields must be aligned
14:59:06HClhmm.. why..?
14:59:14HCli think i might have an alignment problem somewhere..
14:59:43HCl0000720: 0000 0000 0000 0003 41ff ffff ff
14:59:44amiconnReading a short/long from an unaligned address (1) causes slowdown on coldfire (2) crashes on SH1
15:00:07HCli might have an alignment bug in my somewhere
15:00:13HCli'll need someone with an archos to test it..
15:00:21HClcause i don't think that last -1 is aligned at all.
15:01:01HClits the end of a fileentry
15:01:03amiconnIt didn't crash here, although my db might be aligned by coincidence
15:01:04HClin my test database.
15:01:20HClyou can throw it in
15:01:22HCli can check
15:01:27amiconnSH1 crashes with CPUAdrErr on unaligned accesses
15:01:46HClit might very well be a bug in the alignment of the filename length
15:01:49HCllet me check that..
15:02:05HClsince we don't read the ints after it yet.
15:03:31HCli'm dumb.
15:03:36HCli know where it misaligns.
15:03:45HClhold on, let me change that.
15:04:45HClwill we get problems reading shorts then too?
15:04:50preglowyou're using a number based genre system? :/
15:04:55HClpreglow: no no.
15:05:13preglowalmighty jesus be praised
15:05:15HClamiconn: do we have to align on 16bits or 32bits?
15:05:18amiconnShorts must be 16 bit aligned at least
15:05:27amiconnLongs must be 32 bit aligned
15:06:04HCli'll just converts all my shorts to longs.
15:06:16HCloh wait. unless i use two shorts.
15:06:19HClwhich i do.
15:06:30HCli'll convert my rundbdirty flag to 32bits
15:08:09HClits my rundbdirty flag, i made it 8 bits
15:13:54HClupdated the patches and the perl script
15:14:33HCl0000720: 0000 0000 0000 0000 0000 0344 ffff ffff ...........D....
15:14:36HCllooks much better :)
15:17:28HCli broke it though ;x *goes to fix and sighs*
15:19:33HClhm.. or not..
15:19:38*HCl wonders..
15:23:18*HCl broke something :/
15:23:52HClhow the hell could it break anything :(
15:30:22 Join muesli- [0] (
15:32:06MoosCamarovery beautiful duel in F1 Alonso Vs Shumi
15:33:13 Nick Lynx_awy is now known as Lynx_ (HydraIRC@
15:33:14MoosCamaroferrari lose in his land by Renault
15:41:12 Quit muesli- (Read error: 60 (Operation timed out))
15:48:14 Part amiconn
15:48:45*HCl prods stuff.
15:50:24HClah, nothing is broken, heh. aside from what i mentioned earlier
15:50:30HClat least i improved my database checker xD
15:56:23*HCl works on making the script default to directory and filename if there aren't album tags / title tags
15:57:26HCl15:54 < Juerd> HCl: my ($base) = $filename =~ /(.*)(?:\.[^.\s]+)?$/;
15:57:31HClxD @ my local perl guru
15:59:22preglowthere is a basename module
15:59:29preglowi propose you use that instead
16:01:16HClwill that work standalone?
16:01:22HClour current perlscript doesn't need any modules
16:01:25HCli was hoping to keep it like that
16:02:46HClthats much better
16:02:48HCldoes anyone object
16:02:50preglowusing modules is not a bad thing
16:02:57HClwhen i make the in the cvs
16:02:58preglowas long as they're standard
16:03:04preglowperl has quite a number of standard modules
16:03:15HCldefault to the directory name and the filename base if the file doesn't have those tags?
16:03:45preglowthey're all listed on that site, use them diligently unless you like reinventing the wheel :P
16:03:50HClwell i don't have a suffixlist yet...
16:03:59HCli'll leave that open on cvs if someone wants to fix it :X
16:04:17preglowit's optional
16:04:28preglowyou don't have to specify anything but the filename
16:04:28 Quit Ka_ ("* poof *")
16:05:45HClit is..?
16:06:27preglowwell, the docs say so, and i've used it several times with much luck
16:06:44HClCan't locate File/ in @INC (@INC contains: .) at line 10.
16:06:47HClBEGIN failed−−compilation aborted at line 10.
16:06:55HClgood enough reason for me to not use it ;x
16:08:11preglowyou're doing something wrong, or the perl install is retarded
16:08:51 Quit ashridah ("sleep")
16:09:23preglowmight also be an old perl, i don't know when that module became standard
16:09:32Lynx_HCl: the old script was used by some people with just perl.exe downloaded from the wiki page, they would not have any module. but it should be possible to put the module into the one perl file, i think the old makdb has a module included
16:10:03preglowoh, it's been in perl from perl 5, it seems
16:10:14HClLynx_: yea, i use it like that.
16:10:26preglowget a decent perl install
16:10:33HCleither way, i think including the module when we already have a line that works is pointless..
16:10:40preglowok by me
16:10:43preglowit's less obfuscated
16:10:50preglowunless you really think regex is beautiful
16:11:04HClnot at all xD
16:11:06preglowi don't, and i really like regex :P
16:11:06HCllike i said.
16:11:22HCli leave it open to anyone on the cvs to fix it if they want to
16:11:28HCljust gonna test the script
16:11:29preglowsure, but you really should get a better perl install
16:11:31HClbefore comitting
16:11:43HCli'm not gonna install a windows perl install on my iriver xD
16:12:00preglowajhahha, you have perl on your iriver????
16:12:31Lynx_i always wanted perl.rock
16:12:31preglowi hope this new db shit will be written in c
16:12:41preglowLynx_: i'm very sure that can be arranged...
16:12:42HCland yes.
16:12:46HCllike most people with rockbox.
16:12:49HCli have perl.exe on my iriver
16:12:52HClthat i can run easily
16:12:55HClfrom any windows installation
16:12:58HClto update the database
16:13:02Lynx_preglow: i doubt that ;)
16:13:15preglowLynx_: why? perl is very portable
16:13:33preglowit'd require large chunks of the mp3 buffer, but hell, it should be possible :P
16:13:51HClannoying. its skipping my directory without id3tags entirely.
16:14:00HClshould it do that at all?
16:14:42*HCl goes to find an id3tag editor to remove the songtitle..
16:15:01 Join Ka_ [0] (
16:15:09preglowMP3::Info is included verbatim in the file
16:15:39preglowoh well
16:15:45preglowsomeone go embed perl
16:16:28preglowHCl: you plan to use the same code for updating on the host and on the device?
16:16:39Lynx_preglow: well, a basic version of perl maybe...
16:16:48HClpreglow: ofcoursenot.
16:17:03preglowwhy, that would be optimal
16:17:14preglowthen you can sync the db from the player as well
16:18:37HClwe'll have c code to do that one day. not perl.
16:18:48preglowwhy not the same c code for both platforms, then?
16:18:56HClbecause we don't have any c code yet.
16:19:01HClto build the database.
16:19:01preglowno, not yet
16:19:12preglowbut i was asking if that is your plan
16:19:20HClif its faster
16:19:30HClthe c code will most likely be designed for limited resources
16:19:51preglowwell, yeah, but it'll still have to do exactly the same thing
16:20:21Lynx_the bad thing about the tag db is that you need more or less perfect tagging of the files to make browsing there useful ;)
16:20:29preglowof course, heh
16:22:29Lynx_the c version will have to get the tag data from an mp3, is that difficult do write?
16:22:52preglowid3v1 is shit simple and very bad
16:22:57preglowid3v2 can be very convoluted
16:22:59***Saving seen data "./dancer.seen"
16:23:46HClokay, my perl guru's basename replacement didn't work.
16:24:08preglowthe basename module wasn't coded for fun
16:24:12HCli know
16:24:57HClcan i just add the basename module to the end of the script?
16:25:10HClif not, i'll just keep the extension on it for now.
16:26:01preglowinstead of making this huge bloated pl file containing all tons of modules, wouldn't just bundling them with perl.exe be better?
16:26:12HClhe fixed it :P
16:26:21HClwith apologies that he hadn't coded in a while
16:26:40HCl ($base = $f) =~ s/\.\w+$//;
16:26:41preglowi use perl too little these days myself
16:26:43HClthat works
16:26:49HCl :)
16:27:10HClMax artist length : 12 (K's Choice)
16:27:10HClMax album length : 16 (Cocoon Crash)
16:27:10HClMax song length : 24 (05. Butterflies instead)
16:27:12HClhows that?
16:27:20HClalbum was <no album tag> before
16:27:26HCland i removed the title tag of that song
16:27:30HCl :)
16:27:33HClmuch better, no?
16:27:42preglowthat'll keep directory name as well, btw
16:27:58HClyes, the directory name is already seperated
16:27:59HClfrom the file
16:28:10HCllet me commit this..
16:28:40preglowit'll fail if there's a stray . and the end of the file as well :P
16:29:14 Join asdsd_ [0] (
16:30:28HClno it won't.
16:30:31HCllook at my example file
16:30:35HClit has a . ..
16:30:38HClbut okay
16:30:45HCli already committed it
16:30:51HClfeel free to improve it :x
16:30:56preglowfiles very seldom have stray dots
16:31:01preglowisn't even legal in some filesystems
16:31:06HCli know
16:35:36HClit really shouldn't have to rebuild when you alter stuff that nothing depends on ;x
16:37:16Lynx_fixes the stray . ;)
16:37:42HClfeel free to update that in cvs, do test it though
16:38:04t0masHCl: stray dots => dutch?
16:38:16HClpunten zomaar ergens in een regel
16:38:20t0masah ok ?
16:38:41Lynx_HCl: naw, i doubt it's important
16:38:55t0mashm... I have some files with that... I think
16:39:14t0masseen people doing it... tracknr.artist.title.mp3
16:40:01Lynx_t0mas: that'll work, it only removes the last . and stuff after that
16:40:18t0masah ok, what was the problem then?
16:40:30preglowHCl: stray dots at the END of the filname
16:40:34preglowHCl: like hello.mp3.
16:40:56preglowlike i said, not a really big problem
16:41:08preglowignoring it is pretty safe
16:41:10HClfiles like that won't even get recognised as playable
16:41:14HClwill they?
16:41:22preglowprobably not
16:45:07HClits finally reading my k's choice albums correctly
16:45:17HClwith my change in the default name for albums
16:49:33t0massee mailinglist
16:49:49HClgot an url link?
16:49:51HCli got the digest option
16:49:57HCli'm not gonna receive it till tomorrow
16:50:45HCldid he restart his player?
16:50:56*HCl goes to try
16:53:02HClmaybe the new dumpint is malfunctioning for him?
16:53:34*HCl gets claimed by his cat
16:53:34 Quit Ka_ (Read error: 104 (Connection reset by peer))
16:55:06 Join Ka_ [0] (
17:01:37*HCl fixes a tiny bug if there are files in the root of the player without an directory
17:05:53HClhm thats pretty cool.
17:06:14HCli have this album of top 715 songs, and every artist has that album in his list
17:06:25HClof the artists that made at least 1 of the songs
17:06:40HClbut aside from that, i don't see what that person is talking about ;X
17:06:45HClit works better than ever :3
17:07:14HCli need some people with an archos
17:08:03Lynx_HCl: I'm right here ;)
17:08:39HClLynx_: mmm?
17:08:45Lynx_i have an archos...
17:08:48HClah great.
17:08:53HCli have a very simple request
17:08:59HClcheck out and install latest cvs
17:09:04HCland the latest
17:09:09HClsee if it works for you
17:10:50HCli want to get someone to confirm that data on disk actually has to be 32bit aligned as well as data in memory.
17:10:55HClit seems odd to me..
17:11:04preglowdata on disk doesn't have to be aligned?
17:11:07preglowdata in memory has to
17:11:12HClyea, thats what i was thinking.
17:11:20HClthe database is very strict on aligning all data though.
17:11:23preglowbut you often just load the entire struct from disk to memoruy
17:11:23HClit confuses me slightly
17:11:25preglowmemory <-
17:11:47HClthen it doesn't have to be aligned, only the struct in memory will have to be.
17:12:00preglowif you just read it into the struct, verbatim
17:12:02preglowthen it has to be
17:12:14preglowif you read element by element, then no
17:14:28preglowand since reading the entire struct from disk to memory in one go is by far the fastest
17:14:33preglowi guess that's what they dio
17:14:40HCli was mostly curious
17:14:45HClthe new database is completely aligned
17:16:12Lynx_looks good...
17:17:10Lynx_seems so, yes
17:18:03HClokay well i don't understand the problem them
17:18:15HClhe might have had all his files in his root without album tags..
17:18:19HCli just fixed that bug.
17:19:44Lynx_i didn't put the −−strip option, what does that do?
17:20:20HClstrip a path
17:20:24HCllike if you have a mountpoint
17:20:35HCldo you have linux?
17:20:43HClit might be useful for someone to try with the strip option
17:25:57Lynx_something is not right...
17:26:46Lynx_the db seems messed up.
17:27:00HClin what way?
17:27:02Lynx_what should i see when i turn on db browsing?
17:27:13Lynx_the artist, song, album screen?
17:27:14HCl4 options
17:27:25HCland search
17:27:43Lynx_well, i have seemingly random parts of songs and folders, way less than are on the player
17:28:11HClcan you send me your database?
17:28:31Lynx_in the first screen after turning on db browsing. seems like some stuff it cut off at the beginning and other errors
17:29:14HCli don't have a database checker for the old format though, i'm just gonna try to load yours on my iriver
17:30:37HCllate response
17:30:39HCltry again?
17:31:14HClnot working
17:31:18HClare you behind a firewall?
17:31:20Lynx_hmm, firewall issue i guess
17:31:32HCltry throwing it on my ftp
17:32:17Lynx_ah. whichever.
17:32:25HClthat works too
17:32:40 Quit Ka_ (Read error: 110 (Connection timed out))
17:34:04Lynx_umm, i used the bleeding edge build, is that ok btw?
17:35:00HClyou actually have files in your root directory without tags?
17:35:24HClyour database seems fine to me.. where does it go wrong?
17:35:58Lynx_i my have untagged files in root
17:36:10HClyup, i could tell, heh
17:38:18HClcan you tell me what goes wrong where?
17:38:36Lynx_actually i have only one untagged song in root
17:39:02HClyea, maybe you just have files with <Unbekannt> as album tag
17:39:22Lynx_works now...
17:39:46Lynx_I just rebooted the player. But i had done that before...
17:42:43Lynx_hmm, weird. Btw, characters like are replaced by symbols in the perl output. Are they also in the db?
17:44:11HCli think markun was working on unicode support
17:44:44Lynx_cool. i just noticed was replaced by -
17:44:46preglowuse utf8; :>
17:48:45Lynx_HCl: i figured it out, i accidentally used an old db with the new rockbox ver
17:49:01HClthats still odd.....
17:49:16HClthere haven't been major changes in cvs concerning the db
17:50:10Lynx_does the db have to be on the player at boottime?
17:50:21HClyou need to reboot when you update it.
17:50:26HClbit unhandy
17:50:28HClbut thats the way it is
17:50:29HClat the moment
17:51:07HClso no problems with the database as it is in current cvs then?
17:51:59Lynx_no, doesn't seem so
17:52:12Lynx_and the is actually ok on the player
17:53:35Lynx_but it seems to sort uppercase before lowercase. So AZ is before Aa. Is that on purpose?
17:53:57HClthats natural, and yes, i consider it a bug.
17:54:06HClbut it was that way with the original database too.
17:54:59Lynx_i just noticed it. I guess it's good if people can not find their DJ Bobo songs ;)
17:55:15HClthe original firmware does it wrong too.
17:55:36Lynx_Is that problem located in the perl script?
17:55:52HClor actually.
17:55:53HClyes o.o.
18:02:07t0maswhen I need a timestamp in C++
18:02:33t0masand it returns a number..
18:02:39t0maswhat kind of number is it?
18:02:41t0masdouble? int? long?
18:02:46t0masah, own type?
18:03:28t0mastime_t == unixtime ?
18:03:36HClpretty much, its just an 32bits type
18:05:20 Join Ka_ [0] (
18:05:56HClits just an int like
18:10:32Lynx_HCl: the sorting is easy to fix. I changed it and it seems to work...
18:10:47HClLynx_: whats the change?
18:11:50t0masok, next question...
18:12:00t0masis there a simple way to calculate md5?
18:12:08HClmd5sum <file>
18:12:15t0masin C
18:12:22preglowthere are free routines on the net
18:12:27t0masyeah, found some
18:12:40t0masbut they're all a little bloated for my idea...
18:12:49preglowthat works
18:13:04HClmd5sum is slow.
18:13:07preglowbloated? md5 is a pretty complex algo
18:13:15HClthe old crc32 is more suitable for speed
18:13:18HClbut less reliable
18:13:42preglowwont hash collisions be pretty disastrous for you?
18:14:02HClbut so is speed.
18:14:03Lynx_HCl: change the "sort keys %x" to "sort {uc($a) cmp uc($b)} keys %x" in the parts where it prints to DB
18:14:28HClwill try
18:15:10Lynx_thats's sort {uc $entries{$a}->{'TITLE'} cmp uc $entries{$b}->{'TITLE'}} keys %entries for the song table
18:15:23Lynx_just uc for upper caseing
18:15:47 Join edx [0] (
18:16:22rasherpreglow: if we up the hash to 1kb, I'm pretty sure collissions are HIGHLY unlikely
18:17:03HClLynx_: yea
18:17:13HClLynx_: got cvs access?
18:17:18HClLynx_: if you do, commit that
18:17:21HCli'll commit it
18:19:48preglowi'd really hash a bit more than that
18:23:03***Saving seen data "./dancer.seen"
18:23:47rasherdoesn't rockbox read id3 already?
18:23:58rasherre: c version of
18:24:04HCldirectly from the file, yes, not from the database.
18:24:05preglowrasher: nothing bad up with breezy these days? think i'll get myself an upgrade now
18:24:18Lynx_HCl: did you leave out the filename table on purpose?
18:24:18rasherbeen a while since I booted it tbqh
18:24:23HClLynx_: yes.
18:24:30HClfilenames *should* be sorted normally.
18:24:35HClits needed for the binary search.
18:24:49HCl(though database v1 doesn't do binary search on filenames)
18:25:17Lynx_wow, i actually contributed something to rockbox except testing and uninformed advice ;)
18:25:39HCl :P
18:25:46HClif you want, i'll add your realname to credits
18:26:03Lynx_naw, thanks :)
18:41:56 Quit einhirn (Read error: 54 (Connection reset by peer))
18:49:25 Join einhirn [0] (
19:12:07 Join amiconn [0] (
19:15:39HClhey amiconn
19:15:51HClyou have an archos right?
19:16:02HClcan you test the current on cvs with latest cvs?
19:16:08preglowhe sure as hell doesn't have an h1x0 :-)
19:16:26HCl ;p thats mean preglow
19:16:45amiconnpreglow: grr :-(
19:17:03amiconnI just broke the filesystem on my Ondio
19:17:11amiconnNeed to reinstall rockbox & files
19:17:27amiconnSeems it wasn't rockbox' fault though
19:17:58amiconnI played a bit with the archos firmware and the 2 GB card. Seems the card confused the original fw
19:18:13 Nick Lynx_ is now known as Lynx_awy (HydraIRC@
19:18:19amiconnIt overwrote large parts of the card with 0-bytes every 32 bytes...
19:18:30amiconns/card/internal flash/
19:19:04amiconnHCl: You're right about modules - should be able to run standalone
19:19:18 Join bipple [0] (
19:19:31bippleANything new worth testing ?
19:19:38HCli made a lot of bugfixes on the in cvs
19:19:41amiconnMost windows user won't have a full perl installation
19:19:41preglowbipple: cvs logs should tell you that
19:19:50HClamiconn: yea, including me
19:20:02preglowmost MANLY windows users should have :P
19:20:02bippleDont give me all this, just a simple yes/no answer is the jobby lads
19:20:13amiconnHCl: Yeah, noticed that. However, you should change dumpshort too
19:20:30HCli did.
19:20:43HClthe songdb in cvs simply doesn't have a dumpshort
19:20:50HClthats only in songdbv2
19:21:16bipplenowt worth testing then?
19:21:52amiconnHCl: Ah, didn't check
19:22:08bipplelooked in CVS, dont know what it means anyway
19:22:28rasherHCl: added thoughts on hash function
19:22:40preglowbipple: just fixes and small stuff
19:22:50*HCl checks
19:22:52bippleok preggers
19:23:19 Quit bipple (Client Quit)
19:25:16HCli actually have a whole album without tags, heh.
19:25:23preglowi've got several
19:25:35preglowi used to strip tags from my stuff a long time ago
19:25:37rasherwell, that wouldn't be a problem
19:25:43preglowi actually wrote tools to do so
19:25:45rasherempty tags would be though
19:26:25rasherfind -iname "*mp3" -exec id3v2 -D "{}" ";" # :D
19:26:51rasher(WILL delete all tags!)
19:26:54preglowdidn't use linux in those days
19:26:59HCl :P
19:27:07rasherI love being able to do such things like that
19:27:49MoosCamarohi preglow, scuse me, have you got news about mp3 sounds please?
19:28:04preglowlinus is working on that
19:28:21MoosCamarook thanx
19:28:32MoosCamaroi believe you worked on it
19:28:54preglowno, i'm just working on the codecs themselves
19:28:56preglownot playback
19:29:07HClcoffee sounds like a good idea
19:29:17preglowthink i'll make a cuppa
19:29:30MoosCamarook, and there is progress?
19:29:54preglowno, i don't have time to work on it
19:30:04preglowbut mp3 is good enough to be used now
19:30:36MoosCamaroa ok
19:31:23 Quit xen` (Read error: 110 (Connection timed out))
19:31:30MoosCamaroi know you are a very good coder:)
19:31:48MoosCamaroyou're precious for this project
19:32:05MoosCamaroothers too
19:32:25preglowwhy, thanks
19:32:40MoosCamarofor informations
19:32:59MoosCamaroi don't like ask a lot
19:34:24rasherWe're used to it :)
19:39:42*HCl hmmz.
19:41:32amiconnHCl: I did use the −−strip option, since I am on cygwin
19:42:19amiconntools/ −−path /cygdrive/g −−strip /cygdrive/g
19:46:28HClamiconn: good
19:46:34HCldoes it work properly?
19:46:53amiconnDidn't try your latest version yet. I'll do that
19:47:00HClyea, you should
19:47:09HClsomeone reported a problem with it
19:47:15HClbut i don't think i changed anything vital
19:53:25amiconnSeems to work (again, with my limited testset
19:53:46 Join crwl [0] (
19:54:48HClcan't fit more on it? :X
19:54:55amiconn128 MB
19:55:09amiconnI'm working on 2 GB MMC support, you know...
19:55:15HCl :)
19:55:29HCli forgot you hadn't gotten it to work yet
19:56:24amiconnComplex change - I need to deal with splitting blocks into sectors
19:57:39rasherlet me try.. just what's in cvs?
19:58:02amiconngtg again
19:58:05 Part amiconn
20:07:05*rasher runs
20:09:15*HCl slaps v2 malfunctioning and hopes he'll have the same problem in v1...
20:09:36rasheruh.. something's wrong
20:09:56rasherArtist->Album is wrong
20:10:00rasherthe links don't fit
20:10:01HClhow so?
20:10:06HClwhat do you mean?
20:10:25HClif an album contains a song made by the artist
20:10:29HClthe album will be in the list
20:10:33rasherwell when you enter an artist... I get random things
20:10:40HClhave you rebooted?
20:10:51rashereven other artists o.O
20:10:55HClcan you send me your database?
20:10:56HCl x.x;
20:10:59HCli'll be back later >.<;
20:11:24HClsend it when i'm back
20:23:07***Saving seen data "./dancer.seen"
20:29:41rasher <−− best. username. ever.
20:40:15noclue2nice find
20:40:54preglowi can't even view it
21:24:25 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
21:37:37HClrasher: still around?
21:42:23 Quit matsl (Remote closed the connection)
21:45:41HClsome thing is definately bugged for "all songs of artist..
21:53:22*HCl wonders whether there are any regular developers around at all..
21:53:37preglowdo i count? :P
21:54:41HClyes :P
21:54:49HClyou had an iriver right
21:54:53HClyes you do.
21:55:09HCli'm just gonna need some people to help me test what exactly broke the original v1 database
21:55:28HClcause i can't really believe that it were my bugfixes that broke it..
21:55:50HCland i need to know what exactly is broken..
21:55:57preglowif it requires a large database of id3 tagged music or much time, i can't help you :V
21:56:14HClwell, if you at least have *some* id3 tagged, it should work.
21:56:20HCli dunno, how much time do you have? o.oo.
21:56:25HClohno! i'm growing extra eyes
21:56:30*HCl removes one o.o
21:57:39preglowit only supports id3 tags at the moment?
21:58:22HCli think so? i haven't written the original
21:58:26preglowman, that .pl file is going to look nice when ogg tags, ape tags and flac tags have been added as well
21:59:27preglowgimma sec and i might help
21:59:44HCl :P sarcasm, i suspect
21:59:57HClpretty much, i need you to find stuff that looks / is broken
22:00:50HCli'm hoping i'll just have to check/rewrite songs4artist, and that thats the only thing thats broken
22:04:29preglowi have no id3 tagged music that i'm familiar with
22:06:48HClthanks anyways..
22:06:52preglowi use vorbis more or less exclusively
22:07:03preglowonly mp3s i have are randomly downloaded tidbits, heh
22:07:09HCli see...
22:07:23HClstill you work on libmad, not tremor
22:07:25HClwell, i don't mind :P
22:07:36preglowlibmad was commited first, heh
22:07:38 Quit Harpy (Read error: 60 (Operation timed out))
22:08:26HCli don't suppose you'd be able to tell me how to check out a certain version of from cvs?
22:08:54HCli need 1.19..
22:12:10 Quit Rick (Read error: 54 (Connection reset by peer))
22:13:14 Join Rick [0] (
22:14:25 Join muesli- [0] (
22:15:02preglowcvs co whaterever.c -r 1.19 ?
22:16:05 Quit Hornet__ ("Leaving")
22:16:47HCldoesn't work :/
22:19:29HClinstead of checkout
22:19:55preglowok, never did it myself
22:20:03HClsame :)
22:20:12HCljust figured it out with google
22:21:19HClnow i'll find out whether i broke it or whether rasher broke it
22:22:20t0mascvs annotate is useful for that
22:22:38HClwhat does it do?
22:22:43t0masit tells you who changed a line the last time
22:22:52t0masso if you find the bug... just check who edited that :)
22:23:10***Saving seen data "./dancer.seen"
22:25:23HClwith rashers version its even more broken.
22:25:43HClnow to revert to zagors version
22:28:02HCli have a hunch this one will work semi-okay, aside from no albums with multiple artists are possible
22:28:06HClwhich is not what we want either.
22:28:44HClthe current is better. but it needs code adjusted in dbtree.c since we can't trust the artist tag in the album anymore to point to the artist of our song
22:31:02HClits definately rasher's fix for multiple artists per album thats broken it all
22:31:10HCli agree with rasher that its needed though..
22:31:20HClneeds a code adjustment..
22:32:53 Join Psy^Dead [0] (
22:36:05Psy^Deadhi hcl
22:36:16 Join amiconn [0] (
22:36:18HCli need to talk to zagor about how multi artist albums could be properly implemented.
22:36:21HClhey amiconn
22:36:34Psy^Deadzagor doesnt seem to be here...
22:36:46HCli found out that the current tagdb errors are caused by rashers changes to make multi artist albums possible..
22:36:56HClPsy^Dead: hoping he'll read the log
22:36:58Psy^Deaddamn that rasher! :P
22:37:10HCli highly agree with him that such albums should be possible
22:37:14Psy^Deadme too
22:37:24amiconnHCl: Zagor doesn't do that afair
22:37:29HClthe current actually does it pretty well
22:37:31HClamiconn: what do you mean?
22:37:38amiconnreading the backlog
22:37:42HCloh. right.
22:37:45HClah well.
22:37:52t0masghehe... I normally search it for Tomas or t0mas :)
22:37:54amiconnLinus most often does
22:39:39 Nick Psy^Dead is now known as nobby-BRB (
22:39:46*HCl changes digest mode on the mailinglist
22:39:53HClat least then i'll be able to respond normally
22:40:15t0masjust sort all rockbox mailinglist and cvs mail into 1 dir...
22:42:12*HCl adds a warning to the tagdatabase wiki
22:42:25HCli need to prod both rasher and zagor about multiple artist albums
22:42:29HClits not originally my bug
22:42:32HClbut it has to get fixed
22:44:59*t0mas is going to watch some tv... and sleep
22:45:40HClmy default-to-directory for albumnames if not present isn't perfect either..
22:46:15 Quit muesli- (Read error: 60 (Operation timed out))
22:46:17HClamiconn: did you work on a lot?
22:46:26amiconnNot at all
22:46:32HCljust wondering.
22:46:43HCli think i'll just go and relax a bit then, since there's not much i can do
22:47:03HClzagor is the only one who knows exactly how the v1 database works and what should all get changed to get multiple-artist-per-album support
22:47:03 Nick nobby-BRB is now known as nobby (
22:47:39HClrasher made a good attempt, i bugfixed it a little, it works better. but its not good yet
22:47:40amiconncvs says: bagder, zagor, rasher and hcl did
22:47:51HClah. then it was Bagder
22:48:33nobbymushroom mushroom!
22:48:54HCl :P
22:49:19nobbysomeone should make a eggbot to do that every 10 times someone says badger
22:49:34preglowi can whip one up in perl in a jiffy
22:49:43nobby>:D do it!
22:50:16preglowhaha, nah
22:50:21nobbyawww ok
22:50:42preglowi'm wary of distractions
22:50:55nobbywhat you working on now?
22:51:13preglowschool stuff
22:51:23nobbyditch that :P
22:51:43preglowi want to, but that means me suddenly getting another year of school
22:51:57preglownot very much leeway in the fifth year
22:52:15 Join XShocK [0] (
22:59:49 Join Yono [0] (
23:28:54 Join amiconn_ [0] (
23:29:09 Quit amiconn (Nick collision from services.)
23:29:10 Nick amiconn_ is now known as amiconn (
23:30:22HClhey XShocK
23:30:24HCland amiconn
23:30:28HClXShocK: whats new?
23:35:48 Join keith [0] (~keith@
23:43:57amiconnThis doesn't werk :(
23:44:20ferenczyI registered the #xclef_mt-500 player's channel, if somebody interests... (more info in topic)
23:44:34 Join hile [0] (
23:44:46 Quit hile (Client Quit)
23:45:06 Join hile [0] (
23:48:39HClamiconn: mmc? :/
23:49:27amiconnMy new code works for the default block size (which is of course necessary), but not yet for 1024 bytes (for which I wrote it)
23:49:38HCl :x
23:49:41HCli'd look at it
23:49:45HClbut i got my hands full already
23:49:48HCland no testcase for mmc
23:51:38XShocKHCl: hi
23:51:55XShocKi wasn't here
23:51:55HClhow goes?
23:52:16XShocKpretty well
23:52:23preglowferenczy: anything new happened on that front?
23:53:51XShocKHCl: what about rockboy dynarec? in what stage is it?
23:54:11ferenczypreglow> how do you mean that?
23:54:11 Join _n2_ [0] (
23:54:14HClum. stalled till rockboy assembly cpu core gets worked on
23:54:21preglowferenczy: what, if any porting progress has happened
23:54:23HClno one is working on rockboy at the moment
23:54:54preglowmmm... asm core...
23:55:37HClinterested preglow? :P
23:55:44HCli think codecs are a higher priority
23:55:46preglowin a nutshell, yes
23:55:48preglowbut codecs first
23:55:59HClthe good thing of a rockboy assembly core.
23:55:59preglowand codec work will take me a long time :/
23:56:02HClis that we already have one
23:56:04ferenczypreqlow> I'm disassembling original FW of the player...
23:56:05HClin x86 assembly
23:56:13preglowHCl: url?
23:56:27HClits supplied in the gnuboy release, i can toss you the gnuboy tgz
23:56:33preglowi'll fix it
23:56:41preglowjust wondered if it was non-standard or anything
23:56:46HClnah, its standard :)
23:56:48HClonly, x86
23:57:40preglowthe site is acutally down
23:57:43preglowso if you've got the tgz
23:57:46HClhold on
23:58:42HClalso, it has some other functions in x86 assembly
23:58:46HClthat are still C in rockboy

Previous day | Next day