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 2002-09-06

00:12:58 Join BoD[] [0] (
00:13:03BoD[]mblelop !
00:21:30 Quit Jet8810 (Read error: 60 (Operation timed out))
00:26:50***Saving seen data "./dancer.seen"
00:35:53 Quit hardeep ("Leaving")
00:41:15 Quit GomGom ("bozo_fr: je suis avec toi")
00:59:56 Quit datazone-work ("I like core dumps")
01:03:57 Join lithermon [0] (
01:06:15 Nick _seb_ is now known as seb-away (
01:15:43 Join pimlottc [0] (
01:30:57 Join Jbell73 [0] (
01:44:17pimlottcno stat in rockbox is there
02:06:37PsycoXulpressing stop to exit the menu also stops playback with it
02:09:22PsycoXulin current cvs
02:10:25PsycoXulsame for exiting id3 info screen
02:10:39PsycoXuland when's the id3v2 thing getting fixed
02:11:34PsycoXulnone of my files tagged by the 'id3v2' utility show their v2 tags in rockbox because they're for some reason bigger than the buffer...
02:26:52***Saving seen data "./dancer.seen"
02:32:32pimlottcI'd like to make a request to the fat32 guy for a stat function
02:40:53 Quit elinenbe (Read error: 104 (Connection reset by peer))
02:45:48 Join elinenbe [0] (
02:52:58pimlottcand as long as I am talking to the void
02:53:32pimlottcthe build scripts between simulator and real binary has some different behavior
02:53:50pimlottcfor binary I can add a .c file to apps and it is compiled and linked in
02:53:58pimlottcfor simulator it is not and I'm not sure how to add it
03:21:06 Quit lithermon (Connection timed out)
03:32:44pimlottcand as long as I am talking to the void
03:33:06PsycoXuland repeating yourself into the void
03:35:27 Join newone [0] (
03:38:07pimlottcthe repeat
03:38:49pimlottcthe button accel is nice
03:44:45pimlottchandy for the slidebar I'm making
03:46:56pimlottcsimplifies things immensely
03:57:04pimlottcold lcd charset is so lame!
04:00:41elinenbewhat is old LCD charset?
04:00:56elinenbeI want the button accel to get even faster, but I am unable to do it.
04:01:10pimlottcit's on the notes page
04:01:21elinenbeI use LCDPROP
04:01:29pimlottcolder players have a different character set which also only allows for 4 custom characters
04:01:36pimlottcand it's hard to do much with 4 custom characters
04:01:44elinenbei like button acceleration
04:01:50elinenbedoes anyone have a recorder?
04:01:52pimlottcif there more characters in the rom that I could use for the slidebar then it would help
04:02:10pimlottcbut a lot of the characters are random graphics that aren't much use at all
04:03:42pimlottcI gotta go brainstorm on this...
04:11:00 Join RipnetUK [0] (
04:16:52 Nick seb-away is now known as _seb_ (
04:26:53***Saving seen data "./dancer.seen"
04:46:15 Quit edx|sleep (Read error: 110 (Connection timed out))
05:16:51 Nick _seb_ is now known as seb-sleep (
06:04:21 Quit newone ()
06:18:15 Join edx|sleep [0] (
06:26:57***Saving seen data "./dancer.seen"
06:29:23 Nick edx|sleep is now known as edx (
06:38:17 Nick dwihno is now known as dwihno|gone (dwihno@Bald067.Baldakinen.Umea.SE)
07:01:32 Quit edx ()
07:08:06 Join merwin [0] (
07:08:18 Quit merwin (Client Quit)
07:08:22 Join merwin [0] (
07:10:03 Quit BoD[] (Read error: 54 (Connection reset by peer))
07:14:11 Join LinusN [0] (
07:14:32LinusNelinenbe: i think you repeat acceleration starts out too slow
07:15:53merwinheya linus
07:17:37merwinLinusN: i'm trying to figure out how to use the mpeg_val2phys() function... i'm doing loadable EQ and since bass can either end at 12 or 15 depending on the mas, I don't know quite how to use mpeg_val2phys
07:17:41Mode"#rockbox +o LinusN " by ChanServ (ChanServ@services.)
07:18:39LinusNmerwin: you should use mpeg_sound_max() to find out the max
07:19:00merwinhow about bass & treble?
07:19:18LinusNmerwin: you should use mpeg_sound_max(SOUND_BASS) to find out the max
07:19:22LinusNof the bass
07:19:35LinusNand mpeg_sound_max(SOUND_TREBLE)
07:19:58pimlottcdo we have any idea what the proportion of pre-4.50 and post-4.50 players is
07:20:05merwinpimlottc: do you ever sleep?
07:20:14LinusNpimlottc: you mean how many there are out there?
07:20:28pimlottclinusn - yeah like the % of each type for players
07:20:38LinusNi honestly don't know
07:20:51LinusNbut i assume that there are far more post-4.50
07:20:54pimlottchaving 4 characters and that wacky charset is such a pain
07:21:04merwinLinusN: Also, on the players, when you press stop to exit the menus it stops the song also. I'm not quite sure how to fix that (other than trial and error). Figured you may know it since it's been a problem lately
07:21:04LinusNi rarely meet pre-4.50 people
07:21:16pimlottcI'd like to keep thing uniform between pre and post but I'm wondering if it's really worth it
07:21:39pimlottcmerwin - bah and you people... I've been here all for hours talking to myself :P
07:22:02merwinpimlottc: tree icons can't be kept uniform.... without more dynamic icon loading code since there's more than 4 icons
07:22:12merwinwhy the hell am i listening to country?
07:22:31pimlottcmerwin - you mean file type icons?
07:22:33merwinbugs bunny... much better
07:22:35merwinpimlottc: yeah
07:22:40pimlottcmerwin - um we only display 2 at once
07:22:53merwinpimlottc: yeah, but it loads all of the icons at once
07:22:57LinusNpimlottc: so what is it that you want to keep uniform? the charset?
07:23:01pimlottcmerwin - but there's no reason it has to
07:23:14LinusNmerwin: it doesn't load all ocons at once
07:23:27merwinpimlottc: that's what i'm saying :)
07:23:28LinusNonly the ones needed, in tree.c
07:23:41pimlottclinusn - I am making a slidebar for options.
07:24:14LinusNyou need two chars for that
07:24:39pimlottcyou can do it with few but you can do it better with more
07:24:56LinusNbetter, as in good-looking?
07:25:17pimlottcbetter looking and more functional
07:25:23pimlottcI assume you are thinking of −−-+−−- style
07:26:28LinusNno, i am tninking or a simple bar
07:26:39LinusNor ###
07:26:52LinusNwith pixel resolition
07:27:20LinusNwhat did you have in mind?
07:27:22pimlottcthere's many improvements you can make on it
07:27:59pimlottclike for balance you might want more like
07:28:06pimlottc[ ===| ]
07:28:45merwinLinusN: What happens when you try to put, say, a bass level of 20... will it just set the highest available?
07:28:51pimlottcor even more like [==−−|====]
07:29:02pimlottc _____
07:29:05pimlottc( )
07:29:06LinusNmerwin: at the moment, i think it will go ballistic
07:29:08pimlottc ~~~~~
07:29:16pimlottcthings that start to really look nice
07:30:24pimlottcthe optimal way I want to do it needs 6 characters - 2 ends, the current pos char, filled char, empty char and zero point
07:30:42LinusNok, 5 then
07:30:51LinusNempty char is 0x20 (space)
07:31:03pimlottcnod you can do that
07:31:22pimlottcit's nicer as an oval-like hollow shape
07:31:38pimlottcof course part of the problem is I'm a perfectionist
07:31:43LinusNi think we can ise two models
07:31:45pimlottcI have spent all night with this :)
07:32:03LinusNwith a simpler for the oldplayers(tm)
07:32:30pimlottcfor some styles you can use rom characters for things like filled characters, just that the pre charset does't have many good candidates
07:34:09LinusNmerwin: i fixed the STOP key issue
07:34:21LinusNpimlottc: true
07:35:20pimlottclike you could use the [] in pre for ends but the way they are positioned you end up wasting 2 chars for brackets 2 pixels wide
07:37:23LinusNwell, you can't have everything :-)
07:37:39pimlottcyeah I know
07:37:50pimlottcjust can't decide
07:38:37pimlottcend markers are nice to show the size but they lessen the number of available pixels for the bar
07:39:01pimlottcbut without end markers you have to use non-blank empty spaces to show the maximum bar size
07:39:36LinusNstill, with 4 characters, you can have 2 end markers, one filled horizontal line and one to show the cursor
07:40:05merwingrr... "subscripted value is neither array nor pointer"
07:40:05merwin global_settings.bass = (mpeg_sound_max[SOUND_BASS]-atoi(buf_val+1));
07:40:38LinusNmpeg_sound_max is a function, not an array
07:40:52*merwin slaps merwin around a bit with a large trout
07:40:57merwini deserve that
07:40:57pimlottclinusn - assuming your zero-point is at the end of a charcell
07:41:13pimlottcalthough for balance control the zero-point would be best in the middle
07:41:37LinusNaha, i see
07:42:07*LinusN goes to fill his coffee cup
07:42:31pimlottcalright I should go to sleep
07:42:39pimlottcbut first laundry to finish
08:00:31merwinCustom EQ done, now docs
08:06:35LinusNmerwin: k00l!
08:07:34 Join coutta [0] (
08:09:12LinusNcoutta: that 200-minute file, how big is it?
08:12:36couttaLinusN - just added to the bug report
08:12:43LinusNi saw it
08:12:54LinusN220Mbytes is not long...
08:13:05LinusNi wonder what it might be...
08:13:25LinusNdoes it start from the absolute beginning, or just somewhere in the beginning?
08:13:44couttato be honest, resume has never worked on files this big. starts from the absolute beginning
08:14:20 Join Bagder [0] (
08:14:21LinusNhow about a 100Mb file?
08:14:24LinusNor 50?
08:14:28LinusNwhere is the limit?
08:14:57merwinmorning bagder
08:15:06LinusNBagder: the daily build still hasn't games and "screensavers"
08:15:40Bagderthe define issue still is as we discussed previously
08:15:58Bagderconfigure update doesn't fix that
08:16:14LinusNi don't get it
08:16:17Bagderbut I re-ran configure in the recorder-build dir yday
08:16:30Bagderthe defines USE_GAMES and USE_SCREENSAVERS are not set
08:16:41Bagderand imho, we should not demand them
08:16:46LinusNwhat exactly does "configure update" do?
08:16:59BagderLinusN: it updates the makefile
08:17:01LinusNit says "setting defaults"
08:17:05Bagdersupposedly with the same settings
08:17:31LinusN"getting defaults from it"
08:17:36Bagderthe games/screensavers flags were not added "correctly"
08:17:42LinusNaha, now i see
08:17:43 Join rwood [0] (
08:18:17couttaLinusN:largest file I have is 32MB (other than the 220MB)
08:18:20merwinhmm... my cvs commit isn't working
08:18:23merwinor add for that matter
08:18:24LinusNBagder: did you run configure manually?
08:18:31BagderLinusN: yes yday
08:18:41LinusNthen it should set the variables
08:18:41Bagderso today's build should work
08:19:05Bagderbut we should fix the variable issue
08:19:13LinusNcoutta: and the 32Mb file works?
08:19:19LinusNBagder: yes
08:19:27Bagder'configure update' should make the build remain as before with the same options
08:19:33couttaLinusN: yes
08:19:35LinusNi realize that
08:20:47merwinthere, committed
08:22:28LinusNcoutta: does fast-forward work?
08:22:34LinusNall the way?
08:23:13couttaLinusN: before or after the resume ?
08:23:26LinusNin general
08:23:33PsycoXulso how about auto-loading of some special .eq file in the current dir when you start tracks there, and mp3filename.eq for specific songs?
08:24:04couttaLinusN: yes, it works perfectly since you guys fixed the 86 minute limit
08:24:26merwinPsycoXul: bah :)
08:25:02merwinPsycoXul: loads default.eq when you enter the folder... but what if you have a song called default.mp3?
08:25:47PsycoXulthen maybe it should load default.mp3.eq? or something? heh
08:26:22merwinPsycoXul: i create something, and you just keep WHINING and WHINING :)
08:26:31PsycoXuli'm not whining!
08:26:40PsycoXuli'm just seeing greater potential :p
08:26:58***Saving seen data "./dancer.seen"
08:28:21LinusNhehe, now i need to find a 200Mb VBR file somewhere... :-)
08:28:43LinusNis it an audio book?
08:29:34couttaLinusN: it's a recording of a radio program
08:30:10 Nick coutta is now known as coutta|away (
08:31:07merwinLinusN: if i was at home with my dsl i would help
08:31:22merwinLinusN: i've got the audio track for "Song Of The South" in vbr mp3 :)
08:36:28merwinWe need a "Reset Config Defaults" option.
08:36:32merwinin the menus
08:36:47coutta|awayLinusN: I can send it to you DCC if you wish. I am away from my desk (soon) for an extended period
08:36:54LinusNmerwin: make one
08:36:58 Join hardeep [0] (
08:37:11LinusNcoutta|away: i'll try with my 73Mb file first...
08:37:15merwinLinusN: should be easy, it already does this when a new version is encountered
08:37:16LinusNand i can't do dcc
08:37:41LinusNmerwin: just call settings:reset()
08:37:42Bagderyou can dcc it to me
08:37:44LinusNmerwin: just call settings_reset()
08:37:44hardeepIs this the resume bug?
08:37:49LinusNhardeep: yes
08:37:53Bagderand I can send it to Linus
08:38:13hardeepLinusN: I think I know the cause... the AJBR files are VBR but don't have a toc
08:38:21merwinLinusN: wow, easy enough
08:38:27LinusNwhy would the TOC be a problem?
08:38:40hardeepLinusN: the time display is wrong... but I bet the song is starting at the right location
08:38:53LinusNit is the file position that is saved
08:38:57LinusNcoutta|away: is this true?
08:39:10LinusNor does it play from the beginning?
08:43:00LinusNok, my 73Mb CBR file worked
08:46:16LinusNwell done Bagder
08:46:46LinusNi don't like elinenbe's key repeat feature, it starts with a too slow rate
08:47:27Bagderwon't that feature collide with the ffw feature?
08:47:37LinusNand it makes the fast-forward run like the wind! :-)
08:47:39Bagderthe ffw accel
08:48:05rwoodmerwin: any problem with best buy trading-in your unit with the sick drive
08:49:12merwinrwood: nope
08:49:25merwinwill this work?:
08:49:25merwin while(button = button_get(false));
08:49:45Bagderyes, that'll eat all existing key events
08:50:25merwinBagder: ack... i want to wait until the user presses a key and then eval button... so while(!button
08:50:46Bagderno button_get(true) waits until a key is pressed
08:51:05merwinso i really only need button = button_get(true);
08:53:18LinusNi assume it is the "Are you sure?" question?
08:53:31merwinLinusN: yup
08:53:45LinusNi'm a mind-reader, just like Bagder!
08:54:00Bagderyou do good my apprentice
08:54:30merwinit doesnt like this line
08:54:31merwin settings:reset();
08:54:43merwinsettings_menu.c:260: warning: suggest parentheses around assignment used as trut
08:54:44merwinh value
08:54:44merwinsettings_menu.c:261: warning: implicit declaration of function `reset'
08:54:58merwini fixed the truth part
08:55:23hardeepit's "settings_reset()" isn't it?
08:55:30merwinwell tehn
08:55:49merwini didn't think we were doing oop :P
08:56:04LinusNmerwin: you missed my correction
08:56:24merwinLinusN: oh
08:57:42merwinjust testing, then commiting
08:59:14merwinwoo, works
09:00:31merwincommmitted, just like i should be
09:02:25merwinI found the perfect thing today... I was on a long 6 hour drive with bad radio and only a tape player. So i decided to take out my rockbox, plugged in my earbud headphones, and had nonstop tunes all the way down the road
09:03:20hardeepmerwin: i believe that's illegal... :)
09:03:30adi|homeit is..
09:03:35merwinhardeep: not if it's only one earbud, right?
09:03:46merwinLinusN: i think your or my commit broke the build
09:03:55hardeepmerwin: i think even one...
09:03:59merwinLinusN: and i think it was yours
09:04:08merwinhardeep: why? having a hands-free headset is legal?
09:04:38hardeepmerwin: a better solution... use a cassette adaper (archos sells one but there are others) or an fm transmitter
09:04:54merwinhardeep: it's a rental, i didn't want to spend the $15 on a casette adapter
09:06:12rwoodi tried several fm transmitters - couldn't find one that didn't drift
09:06:52hardeeprwood: i've been using the iRock from radio shack... it's not great but it works (i only have a cd player in my car :( )
09:07:14hardeepmerwin: it looks like you're right... you can't drive with headphone that cover *both* ears
09:07:15merwinrwood: you gotta get the kind that plugs inbetween your deck and the real antenna in the back... and comes out front with a plug
09:07:24merwinhardeep: Really?
09:07:37hardeepmerwin: that's what it says on california dmv site
09:08:01merwini wonder about illinois, wisconson, and minnesota? heh, went into all of them
09:08:31rwoodmerwin: i'm getting a new head-unit with line-in - best option
09:08:47merwinrwood: i got a JVC mp3 playing deck with a line-in :-)
09:08:57merwinrwood: damn good deck in itself also
09:09:18rwoodmerwin: which one - i've been looking at the 909 - would rather have black face
09:09:41BagderLinusN: you tried the new credits scroll?
09:09:41merwinnot exactly sure... it's the one that best buy sells
09:10:09merwinBagder: i like the new scroll... tried it in the sim
09:10:25Bagderit is faster on the target
09:10:39Bagderat least faster than the x11 sim
09:11:01merwinI love it... Iraq says they will freely let the US check out all their arms depots, and we still say we're going to start a war with them :)
09:11:10merwinBush is a fuckin' idiot IMHO
09:11:20merwinnot to start a political riot or anything...
09:12:20bobTHCbush is a muppet...
09:13:08merwinbobTHC: I think he is the muppet called Animal
09:13:35bobTHCdehind him the only thing counting $$$....
09:13:53*merwin needs to leave... i have to call my cell phone to see where I hid it
09:14:01merwini also need sleep
09:14:04merwing'nite all
09:14:06bobTHCbush said that a good war is better than a good economic program !!
09:14:17bobTHCit's really an ass hole!!!
09:14:18Bagdernight merwin
09:14:25bobTHCgood night!
09:14:37 Quit merwin ()
09:17:42Bagdernow, why does the build go red?
09:27:19LinusNi am looking into it, very strange
09:27:48Bagderyes indeed
09:28:39LinusNdo we include . in the -I path?
09:28:55LinusNfeatures.h is a bad filename in that case
09:29:19BagderI can't see any -I.
09:29:31LinusNin types.h:
09:29:33LinusN#include <features.h>
09:29:36Bagderbut yes
09:29:42Bagderthat must be it
09:30:04Bagderhehe, it isn't -I.
09:30:12Bagderit is -I[appsdir]
09:30:17LinusNwhich is .
09:30:26Bagderno, not when sim builds
09:30:44Bagdernever mind, I'll change file name I guess
09:31:47Bagderoptions sounds better
09:32:44LinusNi bet options.h is a system include too :-)
09:33:10Bagderbut orb sounds rather safe
09:33:21Bagder"ORBit: A CORBA v2.2 ORB"
09:33:51LinusNi guess CORBA won't be in until Rockbox 1.8
09:34:52Bagderwith options.h it builds fine
09:37:42LinusNtime to upgrade the memory on my machine. cu after the reboot. wish me luck!
09:37:58Bagdergood luck!
09:38:28*LinusN puts his antistatic wrist-wrap on
09:38:39*LinusN holds his breath
09:38:39hardeepmay the force be with you
09:38:48 Part LinusN
09:41:07 Join matsl [0] (
09:42:09bobTHCthe little java machine is a real futur possibility or just a big dream ?
09:42:37BagderbobTHC: you mean on target?
09:42:46Bagderits not a dream
09:42:51Bagderit is a nightmare ;-)
09:43:21BagderI guess it could work, but in practise it would be unbearable
09:43:25Bagderand use all the ram
09:44:57bobTHCrolo jvm boot java or audio (it's really far from the original goal @all)
09:45:37bobTHCrolo boot jvm or audio
09:46:04bobTHCi don't know why i think too much on nothing good .... sorry
09:46:25Bagderno worries
09:48:36rwoodhow about rolo address book - rolo rockbox?
09:49:04 Quit hardeep ("Leaving")
09:49:08bobTHCRockbox Vmanager ?
09:49:29rwoodrockbox/simple pda
09:50:51bobTHCwith t9 and txt viewer it's not a multimedia PDA but nor a "simple pda" ;) for Ver 2.x
09:51:55rwoodyou're way ahead of me
09:53:23bobTHCfor 1 good idea i have 200 bad ideas ;)
09:53:41bobTHCisn't it badger ?
09:53:48rwoodpretty good ratio
09:54:15BagderI have this idea-filter built-in into my irc client ;-)
09:55:05Bagdernight rwood
09:55:09 Quit rwood ()
09:55:56 Join LinusN [0] (
09:56:03LinusNwilma! i'm hooome!
09:56:38Mode"#rockbox +o LinusN " by ChanServ (ChanServ@services.)
09:56:48LinusN512Mb now
09:57:12bobTHCddr or sd ?
09:57:40bobTHCouch light speed ram ;)
09:58:12adi|homethats funny bobTHC
09:58:18adi|homei have a similar ratio
09:58:52adi|home2 good idea, 100 mediocre, 50 really bad, with 1 phenominal idea every 750 ideas or so...
10:01:08bobTHCit's more near the reality... but i'm i'm feeling good this morning.. ;)
10:06:57adi|homeanyone happen to have any Apoctalyptica?
10:11:24LinusNwhat is that?
10:12:48HesIn file included from recorder/sokoban.c:21:
10:12:48Hesoptions.h:24:1: warning: "USE_SCREENSAVERS" redefined
10:12:48Hes<command line>:1:1: warning: this is the location of the previous definition
10:12:48DBUGEnqueued KICK Hes
10:12:48Hesoptions.h:28:1: warning: "USE_GAMES" redefined
10:13:00Hesfrom current cvs
10:13:29Bagderrerun your configure
10:13:35Bagderbut I'll fix that too
10:13:58Hesoops, forgot to update tools
10:14:17Hessorry 8-)
10:14:36adi|homeLinusN: its guys using classical instruments to play metallica music :)
10:17:10bobTHC"Atmospheric" hard rock style adi ?
10:17:18Synthe4 cellos
10:17:50Bagder46 people counted, 4 bad votes, 134 used
10:17:50BagderPos Design # Points 1st vote 2nd vote 3rd vote
10:17:55Bagder 1: 6 36 8 6 0
10:17:55Bagder 2: 24 26 7 1 3
10:17:55Bagder 3: 7 21 5 2 2
10:17:55DBUGEnqueued KICK Bagder
10:17:55Bagder 4: 25 21 3 4 4
10:19:06Bagder26 designs have got votes
10:20:26matslbagder: is my blank vote cunted as bad?
10:20:34Bagderit isn't
10:20:39Bagderit's just ignored
10:20:53Bagderotherwise it would've said 47 people ;-)
10:20:54LinusNmatsl: we always ignore you :-)
10:21:30adi|homeIM wining!!! im winning!!!
10:21:35matslSo I have noticed.
10:21:49adi|homei don't know how..
10:21:52adi|homebut thats cool...
10:22:34bobTHCthe 6 is really a classic :(
10:22:47LinusNi like 7 the best
10:22:53 Join mecraw12 [0] (~lmarlow@
10:25:25bobTHClinus : swedish friendship rulez !
10:25:37mecraw12drunk... drunk... drunk...
10:27:01bobTHChigh , high , high ;)
10:27:03***Saving seen data "./dancer.seen"
10:28:55CtcpIgnored 2 channel CTCP requests in 5 minutes and 7 seconds at the last flood
10:28:55*mecraw12 hiccups
10:37:07 Join saruk [0] (
10:37:40 Nick saruk is now known as quelsaruk (
10:44:16 Join Tsd [0] (Tsd@
10:48:17 Join Zagor [0] (
10:48:54Bagderhey Z
10:50:25LinusNHi Zagor
10:50:38LinusN"16 sectors per request"
10:50:39quelsarukhi Zagor
10:51:04LinusN16 sectors
10:51:09LinusNi want 256
10:51:15LinusNlame HD
10:51:25Zagorhaha. get a new hd.
10:51:36LinusNsame as yours
10:51:49Zagoryou get 256 per request, you just don't get 256 buffered
10:51:51LinusNwhat did you get?
10:51:55Zagor16 here too
10:52:10LinusNaha, it's the size of the FIFO?
10:52:24Zagoryes. the debugf text is misleading.
10:53:23LinusNZagor: have you tried the key repeat accel?
10:53:43Zagoris it added?
10:53:55LinusNelinenbe added one tonight
10:54:07LinusNit starts with a very slow rate, imho
10:54:18LinusNtoo slow
10:54:31LinusNand it really speeds up the FFW!!! :-)
10:54:39Zagorwe need to differentiate between player and rec for this. players need slower repeat since they have smaller screen.
10:54:54Bagderso how is this gonna work in the ffw case?
10:55:04LinusNdon't know...
10:55:36LinusNcan i put the lcd framebuffer back to DRAM?
10:55:46Zagorfine with me
10:56:01LinusNthe iram is full when i build debug builds
10:56:06BagderLinusN: could you time it and see how biff diff it is
10:56:20bobTHCon the original HD only 512k of cache isn't it ?
10:56:31ZagorbobTHC: yes
10:57:21bobTHCon the ibm travelstar i have 2 Meg... it's not a problem with the new request improvement ?
10:57:42LinusNbobTHC: the opposite
10:58:17bobTHCi think it's better but it's just a question of compatibity ..
10:58:26Zagorthe drives report how many sectors they can use
10:58:31Zagorwe read that setting and use it
10:58:48Bagderwhat do you guys think of the credits scroll speed?
10:58:55LinusNit flickeres
10:59:01ZagorBagder: i only tried it in simulator yet
10:59:12bobTHCit's so good! i love that, it's autoadapative .....
10:59:30BagderZagor: it's faster on target
11:00:12Bagderit now moves two pixels per "move"
11:00:21BagderI could make it move one and sleep a little shorter
11:00:42LinusNthat would be better, i think
11:01:37LinusNBagder: i think bouncer is too slow
11:01:51LinusNand it starts out funny, the letters "jump"
11:02:11LinusNthe last "o" and "x" make a large jump
11:02:20Bagderah that one
11:02:26BagderI noticed that too
11:06:41LinusNi want "snow" in, and "boxes"+"blank" out
11:07:14Bagderdr Z said he'd do it
11:07:19Zagori'll be off buying some furniture today, i'll fix that later
11:09:49Bagdersanke-sinus test coming up ;-)
11:10:20LinusNBagder: i want to change parameters with the arrow buttons
11:10:26Bagdersure thing
11:13:02Bagdersanke is darned cool ;-)
11:18:56 Nick coutta|away is now known as coutta (
11:19:06LinusNi have tried different intervals for the repeat accel
11:19:06couttaI am embarassed
11:19:21LinusNit's not very useful, imho
11:19:45 Quit fragglet (Read error: 60 (Operation timed out))
11:19:47LinusNif it doesn't start slow, it will only be able to advance one step
11:19:51couttaI have reported an incorrect bug - it should be that it does not have the correct start time upon resume
11:20:13LinusNcoutta: yes
11:20:33LinusNis your file recorded with the jukebox?
11:20:43LinusNthen it has no Table Of Contents
11:20:50LinusNno VBR header
11:21:03LinusNand the resume can't find the correct time
11:21:38LinusNnote to self: make sure to create a VBR header in the Rockbox recorder
11:22:08LinusNcoutta: there are tools to recreate VBR headers
11:23:43LinusNelinenbe's repeat acceleration starts with a repeat rate of 8 instead of the previous 4
11:24:07LinusNand it accelerates up to 2, which is the maximum the tree scroller can handle
11:25:23LinusNif we start at the (i my opinion better) original value of 4, it will only accelerate two steps
11:25:47LinusNi just tried that
11:25:54LinusNit is good, i think
11:26:14LinusNit gives a little extra "kick" in the dir scroller
11:26:36LinusNnow i like it better
11:26:44couttanow that i know there's no VBR header i'll adjust my brain to cope
11:27:06LinusNyou guys should try to change REPEAT_INTERVAL_START to 4, to see if you agree with me
11:30:08 Part LinusN
11:34:34 Part coutta
11:43:41elinenbewhat does this mean?
11:43:41elinenbecvs server: Updating .
11:43:41elinenbecvs server: failed to create lock directory for `/cvsroot/rockbox' (/cvsroot/roc
11:43:41DBUGEnqueued KICK elinenbe
11:43:41elinenbekbox/#cvs.lock): Permission denied
11:43:41elinenbecvs server: failed to obtain dir lock in repository `/cvsroot/rockbox'
11:43:41***Alert Mode level 1
11:43:41elinenbecvs [server aborted]: read lock failed - giving up
11:46:14Zagorit means i have revoked your commit access, as I said I would. you must use anonymous CVS.
11:46:53elinenbewhy is that?
11:47:11Zagorbecause you consistently fail to follow my instructions and requests.
11:49:05 Join morgan [0] (
11:49:53morganhey, does anyone here know how the new ROLO feature is supposed to work? I have one the most recent 9-5-2002 daily build and can't find it...
11:51:23morganand incidentally, is someone working on the recording stuff yet? I'd like to lend a hand if it'd help...
11:51:59Zagormorgan: switch off mp3 filter, go to a *.mod file, press play.
11:53:42***Alert Mode OFF
11:55:45 Join alkorr [0] (
11:56:28 Part alkorr
11:56:38morganah, or .ajz for recorder, thanks
12:05:36 Nick elinenbe is now known as elinenbe|sleep (
12:05:36DBUGEnqueued KICK elinenbe|sleep
12:05:57Zagorelinenbe and I have sorted out our mutual misunderstandings and he's back in the team
12:06:32bobTHCahhh good news
12:07:16 Quit mecraw12 (
12:07:16 Quit RipnetUK (
12:07:16 Quit Hadaka (
12:07:16 Quit adi|home (
12:07:37NJoinadi|home [0] (
12:07:37NJoinmecraw12 [0] (~lmarlow@
12:07:37NJoinRipnetUK [0] (
12:07:37NJoinHadaka [0] (
12:07:37Mode"#rockbox +o adi|home " by
12:08:07bobTHCteamwork is hard but so efficient and good for all ppl....
12:08:09 Quit Tsd (Read error: 104 (Connection reset by peer))
12:09:15elinenbe|sleepI had a few good shots on Zagor, but he ended up knockin me out.
12:09:50elinenbe|sleepdoes anyone else get the "disk: NULL" message when they plug in their USB?
12:09:51bobTHCmake love not war ! ;)
12:10:04 Join Tsd [0] (Tsd@
12:11:19 Quit Hadaka (
12:11:19 Quit mecraw12 (
12:11:19 Quit adi|home (
12:11:19 Quit RipnetUK (
12:11:38NJoinadi|home [0] (
12:11:38NJoinmecraw12 [0] (~lmarlow@
12:11:38NJoinRipnetUK [0] (
12:11:38NJoinHadaka [0] (
12:11:38Mode"#rockbox +o adi|home " by
12:12:54bobTHCouch, servers spliting
12:16:52bobTHCinstalling security patch and come back
12:16:54 Part bobTHC
12:21:25Zagori'm going shopping. see you later!
12:21:27 Nick Zagor is now known as Zagor|away (
12:24:47 Nick quelsaruk is now known as quel|out (
12:27:04***Saving seen data "./dancer.seen"
12:37:39 Join notch [0] (
12:53:24 Quit RipnetUK ()
13:01:54 Quit Hes (Read error: 60 (Operation timed out))
13:07:18 Join fragglet [0] (
13:07:44 Quit Hadaka (
13:07:44 Quit mecraw12 (
13:07:44 Quit adi|home (
13:08:01NJoinadi|home [0] (
13:08:01NJoinmecraw12 [0] (~lmarlow@
13:08:01NJoinHadaka [0] (
13:08:01Mode"#rockbox +o adi|home " by
13:14:47 Join bobTHC [0] (
13:20:32 Part bobTHC
13:25:42 Join Snorlax [0] (
13:26:45 Join bobTHC [0] (
13:29:50 Join idefx [0] (
13:31:24idefxi'm trying to implement boyer-moore, but, it may need a dynamic array, isn't it ?? which is not possible ?? or is there a solution to do it without dynamic array ?
13:32:19Bagderyou need to init the array first, yes, why is that a problem?
13:33:38idefxBagder: but the array may be dynamicaly made, isn't it ?? i mean, i can't know, during the compilation time, how big it should be, and take a bigger one may take too much place. Or may be have i got a bad idea of the pre-scanning stuff ?
13:34:19Bagderwell, I can't see how it ever needs to be larger than 255
13:34:20Bagdercan you?
13:35:46idefxBagder: what would you put in this array ?
13:36:21Bagderthe thing with BM is to figure out how far you can advance ahead from the current index
13:36:47Bagderso I'd probably store how many bytes you can move
13:38:34idefxor how many bytes you should move ? my idea was to store next position of each char of the patern, would be nice, but, neeed to know the number of char in the pattern.
13:39:00Bagderand you don't?
13:39:37idefxi mean, the number of different char, wich can't be known before entering this char.
13:40:35Bagderyou of course make the array when you know the whole string to search for
13:41:02 Nick seb-sleep is now known as seb-school (
13:41:58idefxwell, i'm getting lost... the idea is : how to make an array containing the position of next char, for example, in blabla, from the last char, the next b is 2 char further, the next a is 3... ?
13:42:07 Join lohap [0] (
13:42:33lohaphi all just wondering why the newest recworder build is missing of the dialy build page
13:43:20Bagderidefx: I don't understand your problem at all
13:44:19lohapthe newest build for the recorder is missing 20020906
13:44:31Bagderlohap: well, what would you think is the reason?
13:44:42lohapi have no idea
13:45:30lohapi just like to download the newest version every day and today they aint one
13:45:33Bagderthe daily build builds the latest from cvs
13:45:39Bagderit happens that it fails
13:46:19lohaplol some now i have to download the source and compile myself barr
13:46:20matslidefx: I found some alternatives: ;-)
13:46:27idefxBagder: Well, i'm on the last char of blabla, the letter corresponding in the text is a b, so, i have to move 2 position right. But, i have to know how many position i should move, which takes to much time to be tested each time, i have to make an arreay containing the position of the next b, for example. As i explain you, i'm finding a solution in fact... Make 2 array, one containing the position of the next letter, and one for the first a
13:47:25idefxmatsl: nice ;) but I wan't to make my Boyer-Moore.
13:47:35idefxBagder: it doesn't sound clear this way..
13:48:01 Quit lohap ()
13:48:34idefxBagder : i explain my idea: for blabla, there will be a 255 char array, saying: 0,1,2 at position of a,l and b. which are the first position of each of this char.
13:48:43idefxfrom the end ;)
13:49:04matslidefx: There was some simpler variants in the URL that had good practical performance. But if you want to do Boyer-Moore I guess you have to ;-)
13:49:40idefxand one more char containing: [0,0,0,3,3,3]
13:49:48idefxi mean, one more char array.
13:50:55idefxmatsl: this is now a challenge for me ;)
13:52:07idefxwell, let's go coding ;)
13:52:55matslidefx: Don't get stuck on this.
13:53:33idefxdon't worry
13:53:42bobTHChis nick expain is mind
13:54:00bobTHCfixed ideas ;)
13:56:55 Quit Snorlax ("gittar ny!")
13:59:51 Join Hes [0] (
14:03:45Bagderveeeeery useful bounce improvement coming up soon
14:05:13pimlottchey bagder
14:05:40Bagderhey ho
14:06:05pimlottcgot a problem with the build scripts
14:06:22Bagderwhat happens?
14:06:53pimlottcif you are compiling for a firmware, you can create a new .c file in apps and it gets picked up and linked in automatically
14:06:53 Join LinusN [0] (
14:07:01pimlottcbut if you compile for simulator, it doesn't
14:07:05BagderLinusN: now go play with the new bounce ;-)
14:07:23Bagderpimlottc: that's correct, they need to be put in the simulator makefile manually
14:07:43pimlottcit should be the same way
14:07:57pimlottcany reason why not
14:08:31Bagder"should" ?
14:08:36Bagderthen please feel free to fix
14:08:52pimlottcI would but I'm not very familiar with makefiles
14:10:22Bagderwell, there was no obvious solution to me so I made it this way
14:10:38pimlottcwhy can't it do whatever it does normally
14:11:10Bagderlearn makefiles, then you'll know
14:11:30pimlottcc'mon now
14:11:56Bagderit just seems unnecessary to explain why when you don't grasp the makefile concepts
14:12:15Bagderbesides, I'm sure it is fixable
14:12:18Bagderwith some effort
14:21:33LinusNBagder: i love it!
14:21:49CtcpIgnored 2 channel CTCP requests in 21 minutes and 55 seconds at the last flood
14:21:49*Bagder bows
14:22:20LinusNi want to be able to save the settings! :-)
14:22:32Bagder.bns files ;-)
14:22:48LinusNin ever directory
14:23:01LinusNand i want it to move to the music
14:23:07Bagderexactly my thought
14:23:12Bagdersomeone will suggest that soon ;-)
14:23:25LinusNyeah, a mainiac!
14:25:30LinusNBagder: have you fixed the fixed font code?
14:26:17BagderI noticed that I had fooled myself a bit
14:26:27BagderI didn't use the font I thought I did
14:27:06***Saving seen data "./dancer.seen"
14:27:17BagderI talked with Zagor about it, and we've pretty much agreed on that it might be smarted to use .bdf fonts all the way
14:27:44LinusNwint one bdf font built-in?
14:28:07Bagdernah, the built-in would need to be a custom format somehow
14:28:19Bagderbut it could be useful to avoid the .ajf format
14:28:34LinusNyou mean that the font files should be BDF?
14:28:52Bagderit takes some more parsing
14:29:06LinusNthat sounds very good
14:29:19LinusNless work for the average user
14:29:33Bagdereasier to just grab a font from somewhere
14:29:36LinusNwhen are you done with that?
14:30:50LinusNi moved the frame buffer back to DRAM
14:30:59LinusNtime to look for new IRAm candidates
14:31:10LinusNlike the str**** functions maybe?
14:31:23Bagderhm, yes that might be a good idea
14:31:35LinusNthe LCD functions are obvious candidates
14:31:49Bagderright, but some of them are there already
14:32:54matslAnybody doing something on text input?
14:33:11BagderI don't think so
14:33:30matslwith idefx search stuff coming we need it
14:34:01BagderI don't ;-)
14:34:05*Bagder grins
14:35:23matslWell I have a naive implementation of text input but don't know it I should commit it since the search stuff might be a bit outside the main development.
14:38:20 Quit Hadaka (
14:38:20 Quit mecraw12 (
14:38:20 Quit adi|home (
14:38:48NJoinadi|home [0] (
14:38:48NJoinmecraw12 [0] (~lmarlow@
14:38:48NJoinHadaka [0] (
14:38:48Mode"#rockbox +o adi|home " by
14:45:48matslHave to check if the suns still out there. cu
14:46:02 Quit matsl ("Liece")
14:46:14Bagderhey, there are webcams for that ;-)
14:55:55pimlottcI think someone has a text input patch
14:58:33pimlottcso where's this iram?
14:58:59Bagderon the sh itself
14:59:21pimlottcoh the cpu
14:59:37pimlottchow much is there? 16, 32k?
15:00:43Bagderso only a selected few functions go in there
15:03:10LinusNmemcpy() is there, for example
15:03:38pimlottc4k on both record and player?
15:07:33 Join coutta [0] (
15:07:45LinusNpimlottc: yes, 4k on both
15:07:59 Join sylvr [0] (
15:08:01 Part coutta
15:13:17LinusNwhat should we do when playing/resuming VBR files without TOC?
15:13:33LinusNestimate as good as we can?
15:13:45LinusNor regenerate the TOC (takes time)
15:14:02Bagderwe should estimate
15:14:07Bagderthe toc is the xing header isn't it?
15:14:40 Quit Tsd ()
15:14:43Bagderone option is of course to waste some more resume space
15:15:07LinusNfor a toc?
15:15:16Bagderno, for a resumed-time
15:15:44LinusNwhat happens to the time information when you FFW?
15:15:54LinusNit must be wrong after a FFW
15:15:54BagderI don't know
15:16:01LinusNwithout a toc, you are lost
15:16:04Bagderfor a non-toc one, indeed
15:16:42LinusNrecreating the toc with an external tool is very much recommended
15:17:10LinusNthe file just screams: "Toca Me"! (Fragma joke)
15:17:19*LinusN hides
15:34:48LinusNI have always wanted an easy way to convert my vinyl records to MP3
15:35:02LinusNhere is the solution:
15:36:43Bagdermuch easier than other solutions
15:36:50Bagderjust put them in your scanner
15:37:45LinusNi just love it
15:37:57 Join langhaarrocker [0] (
15:39:48 Nick elinenbe|sleep is now known as elinenbe (
15:39:48DBUGEnqueued KICK elinenbe
15:40:16langhaarrockerDo you think the sh1 compiler is intelligent enough to optimize modulo operations? Eg. it finds out when the modulo operation is used with a constant of base 2?
15:40:53Bagderpossibly when optimizing at some level
15:41:16LinusNcompile and check the assembly output
15:41:39elinenbeLinusN: I changed the key acceleration to start at 4. Also, with the new key acceleration you may need to change the ff/rw skip size and ff/rw acceleration to give nice results.
15:42:06elinenbeLinusN: I use 2sec skip with 2x accel every 3 seconds.
15:42:21langhaarrockerLinusN: I'm a java programmer and don't fiddle with assembler... ;)
15:44:15langhaarrockerDo I have to install extra tools to disassemble or is it included somewhere in the sh1 tool chain?
15:44:18elinenbeAnyone: every time I commect with USB and then dump on a new firmware I get "Disk: NULL"
15:44:47langhaarrockerelinenbe: I had this once and then never again.
15:45:16elinenbelanghaarocker: I get it EVERY time.
15:45:23Bagderlanghaarrocker: it's just a gcc option
15:46:34langhaarrockerelinenbe: Maybe my box is different: I exchanged the harddrive somewhen. Since then I only have to hit the on button and my recorder starts. With the old drive I had to hold down the button for a while.
15:46:58langhaarrockerBagder: Ok, I'll try to fiddle with assembly tonight.
15:47:03LinusNelinenbe: since when?
15:47:10Bagderlanghaarrocker: use -save-temps and you'll get the assembly file on disk
15:47:47elinenbeLinusN: yesterday
15:47:59 Join datazone-work [0] (
15:48:09LinusNgcc replaced my (x % 8) with (x & 7)
15:48:41langhaarrockerLinusN: cool! Thus I don't have to alter my code!
15:48:45elinenbeLinusN: If I ever just plug in the USB and then unplug it I get that.
15:48:55LinusNelinenbe: model?
15:48:57elinenbeLinusN: I don't even have to modify the file system
15:49:02elinenbeLinusN: recorder 6k
15:49:18LinusNbut it worked the day before yesterday?
15:49:35elinenbeI never have seen this before.
15:49:40langhaarrockerMaybe it's something about disk spinup before boot
15:49:50 Quit morgan ()
15:50:05elinenbedisk spindown is at 20 seconds... if that matters.
15:50:11LinusNwhat was the last build that worked?
15:50:18elinenbeah.... I don't know.
15:50:51elinenbeSometime before yesterday when I was doing the button acceleration thing. I had to reboot everytime I loaded a firmware onto it.
15:51:21langhaarrockerelinenbe: Normal boot is fine?
15:51:29elinenbeprobably before I get yesterday's CVS commits (in the morning here) EST. (I am 6 hours behind you)
15:51:35elinenbeNormal boot is fine everytime.
15:51:49elinenbeit is just when I plug in the USB and then unplug it.
15:52:04elinenbeI can plug it in, and then unplug it 3 seconds later, and I get it.
15:52:39langhaarrockerDo you use linux?
15:53:11elinenbebut I use linux too.
15:53:57langhaarrockerDon't you use the "disconnect hardware" button (or however it was called) before unplugging?
15:54:23langhaarrockerI can't disconnect my drive within 3 seconds or WinXP will complain.l
15:55:21langhaarrocker<boss alert!! must work!!>
15:55:22 Part Bagder
15:55:23pimlottchow long is the voting going to last
15:55:23 Part langhaarrocker
16:02:41LinusNelinenbe: it is probably the ATA driver changes that cause this. can you revert to version 1.45 of ata.c and retry?
16:03:07LinusNby the way, the modulo optimization will of course only work on unsigned types
16:03:07elinenbeI will do it.
16:07:21 Join Phil [0] (
16:08:37 Part LinusN
16:09:55 Join LinusN [0] (
16:10:02LinusNdamn tunnel
16:10:21Phil@ the end of the tunnel sometimes is light...
16:11:33elinenbeLinusN: that fixed it.
16:11:34LinusNyeah, fibre
16:11:39LinusNelinenbe: really?
16:11:43elinenbeLinusN: sorry I took so long, but I was in the bathroom.
16:11:57elinenbeLinusN: yup works perfectly now... every time!
16:12:11LinusNok, i'll investigate
16:12:18LinusNwhat error code did you get?
16:13:23elinenbe"disk: NULL"
16:13:47elinenbebut one time I got something like "disk -1" I am nut sure −− I do not remember exactly
16:15:45LinusNor ata: -1 perhaps?
16:17:15elinenbethat was it.
16:22:36Philis comparing current_tick before and after executing some code a good way to measure it's duration?
16:24:25Hesyeah, but beware - it warps at times
16:24:54PhilWhat does 'warp' mean?
16:25:13PhilOverflow and restart from 0?
16:25:39HesNot very often though 8-)
16:26:21PhilStrange. It's declared as long. Wasn't also a long value used to count milSecs from 1.1.1970?
16:27:08***Saving seen data "./dancer.seen"
16:27:20LinusNno, secs
16:27:52PhilOk. Thats 1000 times better!.)
16:27:57LinusNand if you use the TIME_AFTER() macro you don't have to care about the wrap
16:28:16LinusNlook at how sleep() is implemented in kernel.c
16:28:29elinenbelater all.
16:28:35 Nick elinenbe is now known as elinenbe|laundry (
16:28:35DBUGEnqueued KICK elinenbe|laundry
16:29:37PhilLinusN: that's a bit too much hardcore c for a java dude like me. Never saw anything like __attribute__ before.
16:30:14LinusNdon't care about that, look at how the TIME_AFTER() macri is used
16:30:30LinusNBTW, the attribute thing is to put the code in (fast) internal RAM
16:30:32Philok sorry, I only looked at the declaration.
16:32:48LinusNPhil: remember that the tick counts 1/100 seconds
16:33:04LinusNso if your code executes faster than that you will not be helped by the tick
16:35:41PhilI count 3 - 7 ticks -> I'm on the safe side (unfortunately)
16:36:01LinusNwhat is it you are doing?
16:37:31PhilIt's the time a game cycle in the new wormlet takes. I want to get it constant and adjust the button_get_w_tmo parameter accordingly.
16:38:15LinusNwhat is consuming the most time?
16:38:26Philhit detection
16:40:05LinusNdo you need to optimize it?
16:40:10PhilIt was O(wormlength) and now I reduced it to O(number of worm bends). That saved a lot.
16:40:49PhilIt's not critical for the game but I wanted to reduce the waste of cpu time.
16:42:40LinusNi haven't played it for a while
16:43:03LinusNis it still so that you can have an argh places right in front of you when you eat?
16:43:22Philyes. Should I fix that?
16:43:32PhilI'll do.
16:43:52LinusNthe player should have a chance to avoid it
16:44:08LinusNotherwise you can lose just because of bad luck
16:44:38PhilNow the real problem is that all the contrahend worms (with artificial stupidity) eat all the food in front of your nose... :)
16:44:40 Join edx [0] (
16:44:52LinusNhi edx
16:55:36bobTHCbye all good weekend !
16:55:47 Part bobTHC
16:59:20 Nick Zagor|away is now known as Zagor (
17:00:24 Part Phil
17:04:45LinusNhey Zagor
17:05:13LinusNthe new ata.c broke the USB on recorder6 for some reason
17:09:14Zagorreally? interesting...
17:09:24LinusNdisk: NULL
17:09:36LinusNand sometimes ata: -x (he doesn't remember)
17:09:37Zagorah, right. what elinenbe reported. so that was the ata?
17:09:54LinusNhe reverted to ata.c 1.45 and the problem was gone
17:10:28Zagorhmm. maybe we shouldn't use READ MULTIPLE then. it doesn't really give much of a performance boost anyway.
17:10:46LinusNit may not be READ_MULTIPLE at all
17:10:56LinusNyou did lots of other changes
17:11:10LinusNit may be a simple compare of the return code
17:11:13Zagornothing relevant. not intentionally, anyway.
17:11:22LinusNif ata_xxxx() == -1
17:11:28LinusNor something
17:12:17LinusNand you added identify()
17:12:40Zagorbut that's not ran after usb
17:13:29LinusNusb runs ata_init()
17:13:41Zagoryes but not those portions of it
17:13:57Zagorif (!initialized)
17:14:29Zagordisk_init doesn't check the ata_read return code at all. that's bad
17:16:41Zagoractually, us using of READ_MULTIPLE shouldn't affect usb. we don't change any disk settings or anything like that. you can mix READ_MULTIPLE and READ_SECTORS freely.
17:16:47Zagorso it must be something else
17:17:18Zagoran interesting thing is that the 'return -2' on line 294 previously returned 0. i found that odd.
17:17:42LinusNyeah, in write()
17:18:29 Nick dwihno|gone is now known as dwihno (dwihno@Bald067.Baldakinen.Umea.SE)
17:18:48LinusNZagor: we still have a suspicious return code, on line 314
17:18:50dwihnoHello Computer!
17:19:05LinusNlittle endian eh?
17:19:13ZagorLinusN: ah, missed that one
17:19:14dwihno (divx5)
17:19:32dwihnoLinusN: I'm the red indian, thank you very much! :)
17:20:17LinusNZagor: have you checked if the compiler optimizes the sector*SECTOR_SIZE multiplication?
17:20:25LinusNin the loop
17:20:44dwihnoBtw, has RockBox gone through some kind of optimization the last day? It felt a lot faster today!
17:20:52LinusNdwihno: it sure has
17:21:05LinusNDRAM burst mode
17:21:10dwihnoIt was a major speed boost when switching between tracks
17:21:14dwihnowhich means?
17:21:20LinusN3 times faster RAM
17:21:30CtcpIgnored 2 channel CTCP requests in 48 minutes and 47 seconds at the last flood
17:21:30*dwihno feels the speed!
17:21:35dwihnoGreat work! :D
17:21:49dwihnoThose optimizations made me happy today :)
17:22:02dwihnoI always do a CVS update and build my rockbox before going to work! :)
17:22:35dwihnoAnything new on the font loading code?
17:23:07 Join Phil__ [0] (
17:23:42ZagorLinusN: that line is one calculation for every 4096 word copies. "optimising" it makes no difference
17:24:09*dwihno saw some sheep yesterday
17:24:11dwihnoand some goats
17:24:14dwihnoand some horses
17:24:19dwihnoand some hamsters
17:24:26dwihnoand some smċgoats
17:24:34dwihnobut they were afraid of me :(
17:24:57LinusNZagor: on line 237?
17:25:13Phil__dwihno: Humans are more clever: you don't even see them...
17:25:24ZagorLinusN: yes
17:25:56LinusNi don't get it
17:26:13idefxHey, my BM seems to work... but it's more dirty than dwihno's horses ass-holes...
17:26:15LinusNeven if sectors == 1, the multiplication would be performed 256 times
17:26:30Zagornormal transfers do 16 sectors in one go. that's 4096 words. the comparison is calculated before the loop and then put in a register for comparison.
17:28:41notchmmmnnn sheeeep....
17:33:56LinusNZagor: that was my question, does the compiler optimize that loop?
17:34:08LinusNand you say it does, so i believe you
17:34:30Zagorah, you mean the loop? i thought you meant the actual calculation.
17:35:03Zagoractually, i haven't checked. but it's a normal 'for' loop, i would be very surprised if the comparison value is not held in a register (unless the value is changed in the loop, of course)
17:35:37LinusNI have seen stupid compilers before, but i guess this optimization is pretty trivial
17:36:19Zagoryou are right, we should check. alan used a shift instead of multiplication, but alan is always optimizing everything ;-)
17:36:28LinusNthe dynamic memory discussion has fueled up in the mailing list...
17:36:36Zagorooh, fun fun
17:37:50LinusNactually, the scratch buffer isn't such a bad idea...
17:38:08Zagorwhat do they want it for?
17:38:19*Zagor is too lazy to read ;)
17:38:37LinusNlarge disk loading buffer for string searching in playlists
17:39:16dwihnobtw, is there some special reason the time display still is being displayed by default in the wps?
17:39:24Zagoreh? we use a 16KB buffer for playlists and read 2500+ lines/sec. bigger buffers don't help.
17:39:36Zagordwihno: just 'cause nobody fixed it yet :-)
17:39:41LinusNZagor: not *that* buffer
17:39:43dwihnoZagor: Me fix, me mail you patch! :D
17:39:50Zagordwihno: goodie
17:40:05dwihnowhen is there going to be a devcon btw? :)
17:40:17ZagorLinusN: i know we're not talking about that buffer. but it's a good comparison. the disk is the limit here, not the ram buffer.
17:40:20LinusNdwihno: some day
17:40:35LinusNZagor: maybe not for the string search
17:40:49LinusNi don't know, i haven't seen his code
17:42:14ZagorLinusN: well I say as always: show me the need and of course I will consider it
17:42:55Phil__Zagor: I hope you talk differently with you lover...
17:43:02ZagorPhil__: hehe
17:43:18idefxsomeone talking about my code ?
17:43:24Zagorshe's got wifi. she's happy. ;-)
17:43:43Zagoridefx: well, related anyway
17:43:43dwihnoZagor: In what way are you and Daniel related?
17:43:48Zagordwihno: brothers
17:43:52dwihnoCool :)
17:43:57dwihnoThe code brothers :D
17:44:08dwihnoIf you are the police, he is the villain? :)
17:44:17idefxZagor: i'm coding now, so, don't expect me to follow the discussion ;)
17:44:19Phil__What's a wifi?
17:44:43ZagorPhil__: wireless networking
17:52:03LinusNZagor: the multiplication is optimized into a shift
17:52:32LinusNstill, it is done for every loop
17:52:54Zagorthat's bad. so we should pre-calculate it.
17:53:01LinusNi'll fix
17:53:49Zagorthis john wood fellow is pretty annoying
17:54:23LinusNhe sure doesn't give up
17:56:12idefxdamn, i give up for today: i have to rewrite my search_string code, because this way, it isnot optimised for BM at all... I'm gonna get crazy...
17:56:26notchsomeone should dynamically allocate john wood off the list....
17:57:57LinusNZagor: yeah, i managed to shave off a few instructions...
17:58:06idefxSo, i go... Bye
17:58:08notchI can feel 50 spamm to allocate/to not allocate meassages coming... uggh....
17:58:09 Quit idefx ("[BX] Gary Coleman uses BitchX. Whatchoo talkin bout foo?")
17:59:45 Quit notch ()
18:00:23 Join TotMacher [0] (
18:01:02 Quit TotMacher (Client Quit)
18:01:04 Join TotMacher [0] (
18:01:07 Quit TotMacher (Client Quit)
18:02:59 Join TotMacher [0] (
18:04:29 Quit TotMacher (Client Quit)
18:04:49 Join TotMacher [0] (
18:05:32LinusNTotMacher: hard to decide? :-)
18:05:50TotMacheriam testing mirc with auto-perform
18:05:57TotMacheri connect to 3 networks
18:06:00TotMacherthat rox
18:06:28dwihnoSo mirc finally has multi-server support?
18:06:41dwihnoand it only took, like, 5 years? :D *haha*
18:07:24LinusNthere is an xchat port for windows, if you want to try a real IRC client
18:07:27dwihnopimlottc: hey, be careful you evil doer! :)
18:07:35TotMacherhehe :)
18:07:41pimlottcdwihno - hey you had it coming
18:07:47pimlottcdwihno - good choice btw :)
18:07:53dwihnopimlottc: vielen dank
18:08:01dwihnopimlottc: i en rosa helikopter...
18:08:29pimlottcdwihno - english?
18:08:30LinusN...skall jag flyga hem till dig...
18:08:57dwihnoLinus - the dance master
18:08:59Phil__Neederlands versta ik niet.
18:09:12dwihnoBut I do prefer Caramell
18:09:20ZagorLinusN: unfortunately it made no difference to the disk benchmark :-)
18:10:19pimlottcdarn, no swedish on babelfish
18:10:24LinusNZagor: :-)
18:12:45dwihnoRosa helikopter = pink helicopter
18:13:16pimlottcooooooh that explains _everything_
18:13:28 Quit TotMacher ()
18:13:39 Join TotMacher [0] (
18:13:49dwihnoIt's a kickass tune!
18:13:54pimlottcso you are a pink helicopter?
18:14:12*pimlottc grins
18:14:30LinusNgotta run now
18:14:33LinusNbye all!
18:14:43dwihnobye Linus
18:14:58pimlottcall I know is one word in swedish
18:14:59LinusNdwihno: are you coming to Snaxx?
18:15:11dwihnoLinusN: When is that?
18:15:23dwihnopimlottc: fjortis is a good word too! :)
18:15:26pimlottc(and also 'ikea')
18:15:46LinusNdwihno: sept 18
18:16:07pimlottcdwihno - what's that (fjord-related?)
18:16:22 Part LinusN
18:16:38dwihnopimlottc: it's just a term for underdeveloped small kids
18:17:32 Join Tsd [0] (Tsd@
18:22:14Phil__Swedish language allways looks like joke told by a rubber duck to me. When seeing those words I somehow expect pronouncing them invovles a lot of squeeking.
18:27:09***Saving seen data "./dancer.seen"
18:38:37Phil__No, that didn't squeek.
18:38:47dwihnowhat squeak?
18:39:00Phil__That's what I meant.
18:39:08Phil__No squeek
18:43:37Phil__Do you squeek swedish?
18:43:37Phil__Erm - speak ;)
18:46:21 Part Phil__
18:46:45elinenbe|laundryZagor: you here?
18:46:53 Nick elinenbe|laundry is now known as elinenbe (
18:46:53DBUGEnqueued KICK elinenbe
18:52:26elinenbeI got the "disk: NULL" error and the "ata -1" error on the latest ata.c
18:52:42elinenbewell, not the one committed about an hour ago
18:52:48elinenbeon v. 1.46
18:52:55elinenbe1.45 however is okay
18:53:15elinenbeit only happens when I put it into USB mode and then take it out of USB mode
18:53:17Zagorelinenbe: that's quite interesting
18:54:06Zagorahh, i have an idea
18:54:14Zagordo you use linux or windows for usb?
18:54:46elinenbebut if I plug it in, and then unplug it without doing anything it happened.
18:55:04elinenbeeven if I plugged it in for 1 sec. −− before windows even rocognized it.
18:55:15elinenbeas long as the recorder shows the USB icon.
18:55:36Zagormaybe the driver did work before the UI showed it
18:55:48Zagormy theory is it changed the multisector setting on the disk
18:55:58Zagorcan you dcc?
18:56:22elinenbemaybe try it :) −− I am behind a firewall, but I may be able to DCC
18:56:24Zagori'll send you a test ajbrec.ajz, ok?
18:57:11elinenbeI can DCC.
18:57:28Zagorno, wait. that won't work!
18:57:31dwihnoI can DCC too! :) What a joy!
18:57:45Zagori'm silly...
18:57:47dwihnoZagor: You and Bagder were a part of Horizon?! You guys are trés cool! :D
18:57:58CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
18:57:58*dwihno is with the famous grouse! ;D
18:57:59elinenbewhat is Horizon?
18:58:09Zagoran old C64 demo group
18:58:35dwihnoZagor: will you sign a 5.12" floppy for me? :)
18:58:52Zagorlinus was in horizon too
18:58:57*dwihno will come to the Snaxxx just to collect autographsw
18:59:17Zagorwe've been programming together since the 80s :-)
18:59:32elinenbeyou got a .ajz for me to test?
19:00:26Zagorelinenbe: i'm working in it. the fix I sent you won't work
19:04:38Zagorsending a new version
19:05:46Zagorelinenbe: problems?
19:06:03elinenbeyou send?
19:06:23Zagori'm sending now...
19:06:26Zagori'll try again
19:07:51TotMacherwhat about a enque function for songs ?
19:08:24ZagorTotMacher: not until fat32 write is done
19:08:29TotMacherthat i can choose a song, that will defenitively played
19:08:56TotMacherno playlist
19:09:06TotMacherlike the queue playmode in stock firmware
19:09:17TotMacherso that i can choose the next song
19:09:29 Quit elinenbe (Read error: 104 (Connection reset by peer))
19:09:43TotMachera it wont care about shuffle or normal play
19:10:18 Join elinenbe [0] (
19:10:34elinenbeZagor: it seems to work fine
19:11:30Zagorelinenbe: nice!
19:11:41Zagorthen my theory was correct
19:11:44elinenbeso, whatever you did makes it work ouak.
19:11:57Zagorok, i'll check it in then
19:12:02elinenbewhat is your theory?
19:12:16Zagori'm resetting the multisector value after usb, and then it works
19:12:48elinenbeoh −− obviously.
19:13:34Zagorso it seems the win driver sets it to 1 or something
19:16:43 Quit edx (Read error: 54 (Connection reset by peer))
19:17:19 Join edx [0] (
19:18:33 Nick edx is now known as edx|dead (
19:19:54 Nick edx|dead is now known as edx (
19:27:59 Nick edx is now known as edx|eating (
19:28:52 Join LinusN [0] (
19:29:04LinusNZagor: ah, you found elinenbe's bug!
19:32:57Zagorodd though, that the driver changes this parameter
19:34:16LinusNthat's interesting, yes
19:34:30LinusNand, i hate those dynamic memory discussions...
19:35:29Zagorignore them. i do.
19:35:36Zagoralmost, anyway :-)
19:35:48LinusNbut i still see their point
19:36:26LinusNit would however be extremely complicated to handle dynamic memory in the mp3 thread *SHRUG*
19:36:43LinusNi'm about to say the famous words:
19:36:47Zagorit sure would. if they want it, they write it.
19:36:54LinusNSHOW ME THE CODE!
19:36:57Zagorbeat you to it :-)
19:38:42LinusNbut still, we have a lot of issues regarding the memory handling
19:38:58LinusNi mean, the xxx-file limit and stuff
19:39:06*Zagor slaps LinusN
19:39:11Zagordon't mention that :-)
19:39:38LinusNso it won't hurt at least spending some spare brain cells thinking about it...
19:40:54Zagoryou're obsessed. write some recording code, or something :-)
19:41:22elinenbewhen LinusN feels bad about something... he takes action
19:42:14Zagorwell, he feels bad about some really odd things sometimes ;)
19:46:56 Join merwin [0] (
19:50:00LinusNi feel bad about john wood not getting it
19:54:17merwinLinusN: hmm?
19:56:58 Nick edx|eating is now known as edx (
19:59:22*merwin does not have the will to do the 1 hour of work that needs to be done today
19:59:52 Quit Tsd ()
20:27:11***Saving seen data "./dancer.seen"
20:30:37 Quit merwin ()
20:36:51 Quit edx (Read error: 54 (Connection reset by peer))
20:37:07 Join edx [0] (
20:38:52 Quit quel|out ("KVIrc 3.0.0-beta1 "Eve's Avatar"")
20:43:37 Part LinusN
20:48:32 Quit TotMacher ()
21:24:39 Quit datazone-work ("I like core dumps")
21:51:40miahhrm, keylock hates me
21:52:01miahi walk down the street with my archos in my pocket, with keylock on, and randomly it will skip to the next song, as if keylock wasnt on
21:52:15miahi have no idea what button combination is getting hit to cause it.. but its annoying :/
21:53:08HesIt might be that there is an HD error which causes it to skip
21:53:23HesI have lots of those while driving (with keylock on)
21:53:32Heswhich I don't have with archos firmware or older rockbox
22:03:49 Nick edx is now known as edx|tv (
22:06:40miahnah, there isnt hd error
22:06:54miahbecause if i stop, and start playing that directory again while holding the archos it plays through the songs properly
22:11:02Hesthat doesn't rule out errors...
22:11:12Hesthe disks do get errors while walking/shaking/jogging
22:11:34Hesand currently rockbox does not retry - it gives up on the first error and skips to the next track
22:11:55dwihnowell, using the unit while jogging doesn't sound that good :)
22:14:20miahi dont jog, i just walk
22:14:28Hesi don't even walk, i drive
22:14:34miahif it just skips to the next track on a error, it could get an error while walking
22:14:34dwihnoMr comfortable
22:14:42miahi could see that
22:14:57miahparking in boston is not an easy thing, so i take bus's and subway alot
22:15:04dwihnoboston gurka
22:15:06miahand walk even more
22:15:37Hesbike is easy to park ( 8-)
22:16:42dwihnothe bike looks dangerous
22:19:42miahif you park your bike here
22:19:46miahyou come back and pieces of it are gone
22:19:58miahseat, wheels, anything people can manage to take
22:20:01dwihnoyou arm it with claymore mines
22:20:08miahyea, you pretty much have to
22:20:20dwihnoare there no ways to prevent ppl from stealing parts?
22:22:12miahnot really
22:22:29miahi've seen bike thieves lock their chains to bikes
22:22:36miahso when you come to get it you cant
22:22:45miahand then they go back later and remove both chains
22:22:47dwihnoWhat a sick world we live in
22:22:52miahmaster keys and such
22:23:06miahits messed up
22:23:15dwihnoThat's the price you pay, to live in the U.S.A (rhyme!)
22:23:26dwihnomaster keys? I thought they were nonexistant on the newer locks
22:26:16HesSomeone tried to steal mine last summer
22:26:31Hesdidn't succeed though... I got a new chain & lock
22:26:50Hessome drunk or drug user probably, not a professional
22:27:14dwihnoWhat equipment does the pro's use?
22:27:16***Saving seen data "./dancer.seen"
22:28:13elinenbecar jacks
22:28:20elinenbeand freon −− freeze it and then break it.
22:28:23elinenbewith a hammer
22:28:26elinenbeI am a pro :D
22:29:18dwihnocar jacks?
22:29:28dwihnoyou mean the things you use to raise a car to switch a tire?
22:29:29elinenbeput a car jack in a Ulock (one from a small car like a VW)
22:29:37elinenbeand then crank it like you ar lifting a car
22:29:42elinenbelock blows apart
22:30:02dwihnowelcome to the insecure world
22:30:10dwihnoif you want to keep it - keep it with you 24/7
23:07:15 Join hardeep [0] (
23:08:23Zagorhi hardeep
23:18:17 Quit Synthe ("Connection Lost - Excess Blood")
23:20:33 Join Synthe [0] (
23:50:01 Nick edx|tv is now known as edx|sleep (
23:58:55 Quit hardeep ("Leaving")

Previous day | Next day