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

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

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

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

#rockbox log for 2008-02-05

00:04:01 Join Gr34t3st [0] (
00:06:11 Part Gr34t3st
00:11:14 Join LambdaCalculus37 [0] (
00:12:06DerPapstgood night at all :-)
00:12:12 Quit DerPapst ("So Long And Thanks For All The Fish!")
00:13:30 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
00:13:39 Quit mewshi ("Konversation terminated!")
00:20:57 Part linuxstb
00:24:24 Quit ender` (" They could trim up the framework by treating System.DBNull the same as String.Empty, ala Oracle-style...")
00:24:45 Join LambdaCalculus [0] (n=1800dac6@gateway/web/cgi-irc/
00:25:41 Quit LambdaCalculus (Client Quit)
00:27:42 Join LambdaCalcuclus3 [0] (
00:27:43 Quit LambdaCalculus37 (Read error: 110 (Connection timed out))
00:27:45 Nick LambdaCalcuclus3 is now known as LambdaCalculus37 (
00:27:48 Quit Jon-Kha (Remote closed the connection)
00:28:13LambdaCalculus37GodEater: Here?
00:29:07LambdaCalculus37linuxstb_: Around?
00:30:04linuxstb_LambdaCalculus37: Just about.
00:30:30LambdaCalculus37linuxstb_: I have both Rio Karma CDs.
00:30:53LambdaCalculus37Doesn't look good, though. :(
00:31:16linuxstb_I not that surprised...
00:31:26LambdaCalculus37There's no source diffs on either CD.
00:32:28 Quit MethoS- (Remote closed the connection)
00:32:40linuxstb_Does any of the documentation mention ecos/
00:33:08LambdaCalculus37Let me look again.
00:34:34 Join Gr34t3st [0] (
00:34:58Gr34t3stcan anyone help me with a 30g iPod video with rockbox?
00:35:27scorche|wnot really....unless you were planning on telling us what issue you are having :)
00:35:38LambdaCalculus37linuxstb_: I don't see any mention of either eCos OR RHEPL.
00:35:51Gr34t3stMy girlfriend can't charge her iPod with rockbox now.
00:36:06linuxstb_Here are some interesting comments about ecos -
00:36:08Gr34t3stWhen she plugs in her iPod to a charging dock or to her computer it doesn't connect and charge.
00:37:10scorche|wGr34t3st: well, it shouldnt connect, but it still should charge
00:37:29Gr34t3stwhen she plugs it in it doesn't do either. It stays off.
00:38:48Gr34t3stwe would simple uninstall it but she can't even turn it on.
00:39:11linuxstb_Hold MENU+SELECT together for a few seconds to reset it. Sometimes ipods don't like to turn themselves on.
00:39:53LambdaCalculus37linuxstb_: According to RHEPL, "The license does not require users to release the source code of any applications that are developed with eCos."
00:40:18 Quit EspeonEefi ("さよなら")
00:40:26Gr34t3stYou're a life saver. She was getting ready to kill me.
00:40:34linuxstb_LambdaCalculus37: I know. But it requires them to release the source code to modified parts of ecos itself to both RedHat and anyone who they distribute a binary to.
00:40:35Gr34t3stThanks a ton man. :D kudos
00:41:25linuxstb_LambdaCalculus37: And all they did was to give that patch to RedHat - not to users...
00:41:33LambdaCalculus37linuxstb_: Yes, you're right! Next sentence: "However, if anybody makes any changes to code covered by the eCos license, or writes new files derived in any way from eCos code, then we believe that the entire user community should have the opportunity to benefit from this."
00:42:01LambdaCalculus37Well, GodEater did try to ask on the eCos discussion list, but to no avail.
00:42:34linuxstb_I know - he asked for help in tracking down that patch, and I guess no-one knew anything about it...
00:42:57linuxstb_But I think it's definitely worth trying to contact - they're still distributing firmware binaries.
00:43:28 Join japc [0] (
00:43:32 Quit ompaul (Client Quit)
00:43:51linuxstb_Or not... Seems rioaudio,com is empty...
00:44:16LambdaCalculus37linuxstb: They switched to, methinks.
00:44:45linuxstb_That site doesn't work either... worked a week or two ago.
00:45:05 Nick fxb is now known as fxb__ (
00:45:09 Quit spiorf (Remote closed the connection)
00:45:29linuxstb_Ah yes, that site works.
00:45:40LambdaCalculus37Back later... dinner.
00:47:24 Join cool_walking_ [0] (
00:53:12 Quit OlivierBorowski (Remote closed the connection)
00:54:27 Quit conando (Read error: 110 (Connection timed out))
00:56:15 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
00:59:31 Quit kugel ("ChatZilla 0.9.80 [Firefox]")
01:00:01 Quit Gr34t3st ("Copywight 2006 Elmer Fudd. All wights wesewved.")
01:03:33 Join roolku [0] (
01:04:00 Join Gr34t3st [0] (
01:04:53Gr34t3stWhy do some of the wps files act like rockbox's original wps?
01:05:04***Saving seen data "./dancer.seen"
01:05:48 Quit hcs ("Leaving.")
01:06:53krazykitGr34t3st, they're either out of date or require unsupported patches
01:07:33 Join MethoS- [0] (
01:10:28 Quit MethoS- (Remote closed the connection)
01:10:54 Quit Gr34t3st ("IceChat - Chillin with the Best of em")
01:12:52 Join z35 [0] (
01:13:11 Quit obo ("bye")
01:20:24 Quit Axio (Remote closed the connection)
01:23:32 Quit roolku ()
01:23:50 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
01:29:45 Join cool_walking_200 [0] (
01:38:54LambdaCalculus37linuxstb: No general contact emails for Rio.
01:43:30 Quit lee-qid (Read error: 110 (Connection timed out))
01:44:39 Quit Nico_P (Remote closed the connection)
01:46:04 Quit cool_walking_ (Read error: 110 (Connection timed out))
01:46:31 Quit JdGordon|w (" ajax IRC Client")
01:46:41*LambdaCalculus37 is wondering where to go to find out about eCos source diffs
01:50:50 Nick cool_walking_200 is now known as cool_walking_ (
01:52:02 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
01:52:33saratogado we want spec sheets for anything samsung makes?
01:52:58 Nick stewball`ghost is now known as stewball`sleep (n=WTFOMGBB@
01:54:24 Part pixelma
01:54:52LambdaCalculus37saratoga: Anything, you say? ;)
01:55:25saratogawe'll see
01:57:54*amiconn doesn't understand why his linux box became so slow when building rockbox
01:58:09amiconnMore than 4 minutes for an H300 build :(
01:59:22amiconnKernel 2.6.24 problem?
02:01:33amiconnHmm, or ccache effect :\
02:16:35 Quit saratoga ("CGI:IRC (EOF)")
02:17:01 Quit BigBambi (Remote closed the connection)
02:18:14tessaraktthe original SanDisk firmware is crap
02:21:27 Part LambdaCalculus37
02:22:41tessarakteven the German translation is crappy ...
02:23:12 Quit XavierGr (Read error: 110 (Connection timed out))
02:24:28 Join aliask [0] (n=chatzill@rockbox/developer/aliask)
02:24:57 Join ctaylorr [0] (
02:31:36 Quit Horscht (Read error: 110 (Connection timed out))
02:31:59 Join Horscht [0] (
02:32:04 Join DerDome [0] (
02:32:20 Join andrew___ [0] (
02:36:00 Quit scorche|w ("CGI:IRC (EOF)")
02:37:31 Quit Thundercloud (Remote closed the connection)
02:37:55 Quit BitTorment (Remote closed the connection)
02:42:03 Quit GodEater (Read error: 110 (Connection timed out))
02:46:20 Join Febs [0] (
02:50:26 Quit andrew__ (Read error: 110 (Connection timed out))
02:55:13 Join Thundercloud [0] (
02:58:47 Quit tessarakt ("Client exiting")
03:03:38 Join FOAD_ [0] (
03:05:05***Saving seen data "./dancer.seen"
03:10:18 Quit Thundercloud (Remote closed the connection)
03:17:55 Quit DerDome ("Leaving.")
03:20:24 Quit FOAD (Read error: 110 (Connection timed out))
03:20:24 Nick FOAD_ is now known as FOAD (
03:20:49Febssoap, I posted a thread on Head-Fi to try to find testers for the 11 Color/Photo patches Buschel is trying to test.
03:31:04 Join Geeky123_ [0] (
03:35:30 Quit linuxstb (Read error: 113 (No route to host))
03:35:42 Quit Geeky123_ (Client Quit)
03:35:55 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
03:39:13soapcool Febs
03:48:11 Quit japc (Remote closed the connection)
03:58:28 Quit aliask ("ChatZilla 0.9.80 [Firefox]")
04:06:05 Join webguest55 [0] (n=458deaa8@gateway/web/cgi-irc/
04:07:12 Quit ctaylorr (Read error: 110 (Connection timed out))
04:08:09 Quit webguest55 (Client Quit)
04:11:24 Join Jon-Kha [0] (
04:14:09 Join Falco98 [0] (
04:17:42 Join psycho_maniac [0] (
04:30:19 Join jcollie [0] (
04:33:09 Join Mit [0] (
04:33:09 Quit PaulJam (".")
04:33:57MitHi I have a question about the rock box, and it's support for a 60 gig Ipod Photo...
04:34:11Mithave I done something wrong when I installed it?
04:34:25Mitcause every time I try to use the nice themes for rockbox the images look all corrupted
04:34:38Mitand for that matter the rockbox logo at boot time looks corrupted too
04:42:11 Part Mit
04:44:00 Join Evilnick [0] (
04:47:18 Quit hcs (Read error: 110 (Connection timed out))
04:52:19 Quit aneka ("KVIrc 3.2.4 Anomalies")
05:02:45 Join joseph_ [0] (
05:02:51 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:03:07 Join miepchen^schlaf [0] (
05:04:43 Quit Evilnick ("ChatZilla 0.9.80 [Firefox]")
05:05:09***Saving seen data "./dancer.seen"
05:05:30joseph_Hi, My ipod (5g) has recently "lost" all songs. Luckily the files were not damaged and restored by gtkpod, but it seems the file system (vfat) has been damaged. Can anyone offer any word of advice - particularily: do I need to remove rockbox before I attempt a "restore" via iTunes?
05:06:37Febsjoseph_: no. The iTunes restore will wipe out Rockbox anyway.
05:06:58joseph_any advice on how to possibly fix the vfat?
05:07:11FebsWhat OS?
05:07:17joseph_linux of course
05:07:40joseph_but if the best tool is a windoz i will find a box
05:08:05FebsI'm (mostly) a Windows user, so I would do a chkdsk /f /r.
05:09:01FebsHave you tried fsck in Linux?
05:09:04joseph_I run fsck.vfat -n -l /dev/sdc1, which I assume is similar (-n for just report, no fix). it said something like
05:09:18joseph_FSINFO sector has bad magic number(s):
05:09:18joseph_ Offset 0: 0xc161d252 != expected 0x41615252
05:09:18joseph_ Offset 484: 0xe141f272 != expected 0x61417272
05:09:50joseph_I am not sure if letting it fix that would damage the ipod?
05:10:15FebsIf you're going to restore in iTunes anyway, you have nothing to lose by trying.
05:10:47joseph_yes but I rather not, as restoring was painfull :)
05:11:33joseph_I mean, no matter how bad I muck up the file system restore is possible?
05:12:04joseph_I am asking here because as rockbox developers you might know ...
05:15:21joseph_thanks anyway
05:15:27 Quit joseph_ (Remote closed the connection)
05:16:19 Quit markun (Read error: 104 (Connection reset by peer))
05:22:19 Join rocko [0] (i=rocko@
05:22:32rockoI'm an avid user of rockbox
05:22:46rockolately, I've been having a few problems relating to my ipod color
05:22:58rockoit beeps when it goes into regular ipod firmware
05:22:59 Quit Falco98 ("ChatZilla 0.9.80 [Firefox]")
05:23:21rockoand when I turn the thing off after using 10% of my battery, when I turn it back onit says there's 0%
05:24:25rockoI don't know what to do
05:31:18 Join crzyboyster [0] (n=6210f01d@gateway/web/cgi-irc/
05:33:03crzyboysterAnyone know what in the world this is supposed to be? >
05:39:03Horschta remote addon for players?
05:39:49HorschtI have no clue
05:40:02 Quit joshin (Read error: 110 (Connection timed out))
05:40:04Horschtanyways, I am off
05:40:09 Quit Horscht ("")
05:40:41JdGordonthe up and down are much smaller than the << and >> arrorws :(
05:41:51 Quit crzyboyster ("CGI:IRC (EOF)")
05:43:52JdGordonbah, stupid large font messing with my head
05:43:59JdGordonanyone game to test this patch?
05:57:02 Join JdGordon_ [0] (
05:57:02 Quit JdGordon (Read error: 104 (Connection reset by peer))
05:57:10 Quit JdGordon_ (
05:57:10 Quit miepchen^schlaf (
05:57:10 Quit linuxstb (
05:57:10 Quit z35 (
05:57:10 Quit jhulst (
05:57:10 Quit Karti (
05:57:10 Quit gromit`` (
05:57:10 Quit DaCapn (
05:57:10 Quit SirFunk_ (
05:57:10 Quit Bagder (
05:57:10 Quit Lynx_ (
05:57:10 Quit cagnulein (
05:57:10 Quit lids (
05:57:10 Quit ze (
05:57:10 Quit pabs (
05:57:10 Quit [omni] (
05:57:39 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
05:59:12 Quit DogBoy (
05:59:12 Quit simonrvn (
05:59:12 Quit qweru (
05:59:12 Quit scorche|sh (
05:59:12 Quit johnh51 (
05:59:12 Quit tchan (
05:59:12 Quit BHSPitMonkey (
06:00:47NJoinDogBoy [0] (
06:00:47NJoinsimonrvn [0] (i=simon@unaffiliated/simonrvn)
06:00:47NJoinqweru [0] (
06:00:47NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
06:00:47NJoinjohnh51 [0] (
06:00:47NJointchan [0] (n=tchan@lunar-linux/developer/tchan)
06:00:47NJoinBHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
06:01:38NJoinJdGordon_ [0] (
06:01:38NJoinmiepchen^schlaf [0] (
06:01:38NJoinlinuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
06:01:38NJoinz35 [0] (
06:01:38NJoinKarti [0] (n=jim@
06:01:38NJoinjhulst [0] (n=jhulst@unaffiliated/jhulst)
06:01:38NJoingromit`` [0] (
06:01:38NJoinDaCapn [0] (
06:01:38NJoinSirFunk_ [0] (
06:01:38 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
06:01:38NJoinLynx_ [0] (
06:01:38NJoincagnulein [0] (n=Cagnulei@
06:01:38NJoinlids [0] (
06:01:38NJoinze [0] (
06:01:38NJoinpabs [0] (
06:01:38NJoin[omni] [0] (
06:02:23 Join perrikwp [0] (n=982165ed@gateway/web/cgi-irc/
06:03:35JdGordon_and there bacl
06:03:56JdGordon_ok, whos going to test my patch? or ill just commit it
06:03:58 Join lids_ [0] (
06:05:42jhMikeSwhat change?
06:06:20JdGordon_quickscreen vith viewports
06:06:30JdGordon_and customizable items
06:06:42rasherdoes it kill SYSFONT_ lang entries?
06:07:35rasherPreferably in all languages including them..
06:07:53JdGordon_I actually forgot about them, but ill make sure to remove them
06:08:02JdGordon_that might just push it closer to acceptable red :)
06:12:17JdGordon_ill commit that once i fix the langs.. so testers... :)
06:14:58 Join toffe82 [0] (
06:14:59JdGordon_rasher: do i set the unused sysfonts to deprecated?
06:15:59rasherJdGordon_: I believe so, yes
06:16:32JdGordon_bugger.. thats gonna be a bitch to do :)
06:16:58JdGordon_LANG_SYSFONT_OFF is still used.. but this patch deprecates about a dozen strings
06:18:15rasherThis is why xml and a "langfile tool" would be nice
06:18:32rasherthe xml only being needed because it makes the tool quicker to write
06:18:57rashersomething along the lines of `langtool -deprecate LANG_SYSFONT_OFF *.lang`
06:19:40JdGordon_yes, that would be sweet
06:19:45JdGordon_shame we dont have that :(
06:20:13 Quit lids (Connection timed out)
06:20:55rasherYou're more than welcome to pick up FS #8233
06:21:33JdGordon_16 strings gone
06:21:40 Nick JdGordon_ is now known as JdGordon (
06:26:47*JdGordon wonders if it would be ok to to jump the lang version because of this and do it the easy way
06:27:13 Join midgey [0] (
06:27:31JdGordon1 down, 15 to go
06:28:15JdGordonaparently lang editors dont understand why the _SYSFONT_ are needed
06:28:34rasherIt really should have been in the desc field for each and every one
06:29:54 Quit tvelocity ("Αποχώρησε")
06:30:37 Nick lids_ is now known as lids (
06:32:35 Join hannesd_ [0] (
06:33:04 Quit Karti (Remote closed the connection)
06:37:45JdGordonfuck.... kate is a crap editor.. its fubaring the lang files
06:39:37rockoI talk to this girl Kate
06:39:47rockoI'm fairly in love with her.
06:46:32JdGordongonna be a nasty red... :p
06:48:59 Quit hannesd (Read error: 101 (Network is unreachable))
06:48:59 Nick hannesd_ is now known as hannesd (
06:54:46 Quit midgey ()
07:05:11***Saving seen data "./dancer.seen"
07:09:41 Join karashata [0] (
07:10:56psycho_maniachaha only read rockons comments about him being in love with "kate" haha then read above that
07:23:28 Quit perrikwp ("CGI:IRC (EOF)")
07:27:50 Join ch4os_ [0] (n=ch4os@unaffiliated/ch4os/x-059673)
07:32:04 Join joecool|wired [0] (i=joecool@no-sources/joecool)
07:33:32joecool|wiredlostlogic: ping
07:45:53MattAndrewI'm designing some hardware to run rockbox. Anyone know if it's a Bad Idea to completely power down the ram after the player has been put in pause mode and the backlight timeout has expired? Basically, this would mean Rockbox would have to boot up again every time you let your music player sit for 2 minutes in pause mode. Good idea/bad idea?
07:47:04hcssounds like the sort of thing that should be reserved for a "sleep mode"
07:47:20hcsor, why not just show down entirely?
07:48:10MattAndrewyeah, it is shutting down entirely. I just wanna know if Rockbox can be set to go back to where it was in the middle of playing a song when it boots back up.
07:48:27hcsyeah, it can
07:49:00MattAndrewand the time it takes to boot up to that point would be how long?
07:50:27 Join Adam [0] (n=2093c0aa@gateway/web/cgi-irc/
07:50:51hcsI guess it would depend on how fast your hardware can boot rockbox, I don't know.
07:52:02AdamCan neone help me out?I got a sansa e260 and want to put rockbox on it. I downloaded the current build, but i can't find the installer that goes with the software. What do I do?
07:52:32scorcheAdam: have a look at the manual
07:52:51AdamIs there a complete zip package that i can d/l that has the current build, installer, rockdoom, themes, fonts, etc. all in one package??
07:53:15scorcheAdam: please look at it :)
07:53:37 Part toffe82
07:57:31 Join Adam2 [0] (n=2093c0aa@gateway/web/cgi-irc/
07:57:55 Quit Adam ("CGI:IRC (Ping timeout)")
07:58:30Adam2K. I just d/l the rockbox utility but it only has archos and iriver players listed on the install. WTF. No sansa
07:59:45 Join GodEater [50] (n=bryan@rockbox/staff/GodEater)
08:01:03Adam2I downloaded the rockbox utility and it only has the archos and iRiver mp3 players on the install list, no apple, sansa, etc.
08:01:14scorchewhat install list?
08:02:21Adam2The screen that says please select the player you'd like to install rockbox on
08:02:46scorchewhat did you download?...i see them all just fine
08:03:19 Join WafflesDr [0] (
08:03:54Adam2I musta got an older version of the installer cuz I found what I was looking 4.
08:04:54Adam2Now it's saying in the progress screen that the mount point is wrong when I start the install
08:05:06scorchethen change it
08:05:54 Quit Adam2 ("CGI:IRC (EOF)")
08:06:04MattAndrewhcs: okay, good point. This is going to be a 180MHz arm9. I can't find out how fast the arm in the sansa is or I'd compare it.
08:06:21LloreanWhy do we have a customizable quicksceen?
08:06:38scorcheMattAndrew: 80MHz boosted
08:07:01GodEaterLlorean: the reason starts with Jd and ends with Gordon
08:07:36LloreanGodEater: Shouldn't this have been in the patcher? Was it discussed and I missed it?
08:07:43Lloreanerr, tracker?
08:07:54GodEaterI think it got discussed for about ten minutes yesterday
08:08:00GodEaterthat's discussed in ""
08:08:21GodEaterI'm officially handing in the support towel
08:08:25scorcheMattAndrew: you can see how fast it is running at any one time (and change it) through System > Debug > CPU Frequency
08:08:59MattAndrewscorche: ooo. nice feature.
08:09:20LloreanGodEater: He also added a top item, which probably makes the screen unexitable on some targets
08:09:25 Join matsl [0] (
08:09:26 Quit psycho_maniac (" its bed time")
08:09:27LloreanIf a user sets a top item.
08:09:42GodEateryeah, I think amiconn told him that at the time too
08:09:46MattAndrewAccording to portalplayer's pdf about the pp5024, it has 2 arm7tdmi's that can run up to 100MHz each. So, this would be about as fast as that.
08:09:47scorchehe did...
08:10:00LloreanJdGordon: Why on Earth did you commit that?
08:10:19scorcheMattAndrew: well, we dont utilize both cores yet and we only run up to 80MHz boosted
08:10:45*scorche goes and has a look at the delta..
08:11:08LloreanQuite the opposite of green
08:11:44scorcheand a bit o red
08:11:59scorche(in the scoreboard)
08:14:41amiconnLlorean, GodEater: The top item doesn't make the screen unexitable. It is just impossible to use the top item the real quick way of operating the quickscreen (which has been broken for months and months anyway)
08:15:01Lloreanamiconn: Some players depend on the "Up" button to get out of the screen, don't they?
08:15:06LloreanSpecifically, iPods?
08:15:14MattAndrewscorche: I'm curious: can you power down the unused coprocessor?
08:15:23*GodEater goes to buy coffee
08:16:00 Join Rob222241 [0] (
08:16:20amiconnLlorean: Ah, you're thinking about the ipod... that wasn't what was discussed yesterday though
08:16:48LloreanWell, is what he committed not enabled for iPods then?
08:17:39 Nick joecool|wired is now known as joecool (i=joecool@no-sources/joecool)
08:17:39 Quit joecool ("Leaving")
08:18:49amiconnscorche: The failed build is caused by an unstable build server...
08:19:03amiconnThe delta row colour is nasty though :(
08:19:38LloreanWhy was "customizable" added at the same time as making them viewports?
08:20:00amiconnJdGordon: Why does your quickscreen commit even touch the binsize of the Ondios and the Player?
08:20:08amiconnThat can't be right...
08:20:12LloreanI know JdGordon's been rather against the current choices for quickscreen contents, but I have to admit that I'm kinda pissed off about a fairly major change to an existing system not getting a mail to the list or any warning.
08:20:25amiconnThose targets have no quickscreen
08:22:28amiconnYeah, I also don't understand why the customisation was added.
08:23:34LloreanI hate to sound as cynical as I'm going to, but I imagine just because he thought we were less likely to revert it, even if generally speaking we're against it vocally all the time.
08:23:35amiconnEspecially since hidden customisation features were discussed a few days ago, and the consensus was that hidden options shouldn't exist
08:23:46amiconnAt least I thought it was that... :\
08:23:46 Join joecool [0] (i=joecool@no-sources/joecool)
08:23:58 Quit spr0k3t_ (Read error: 104 (Connection reset by peer))
08:24:37 Join spiorf [0] (
08:24:39 Join spr0k3t_ [0] (
08:25:01*karashata notes a couple things about the recent change to the quickscreen: can't find where to customize it, the font colour defaults to black and not the foreground colour, and scrolling lines in the WPS end up updating over the quickscreen
08:25:03GodEaterI thought we'd always been against menu customisation
08:25:54Lloreankarashata: It wasn't widely tested, I fear. Possibly because that would've resulted in more discussion. =/
08:26:24*karashata nods
08:27:00*amiconn thinks that commit should be reverted for now
08:27:02WafflesDrHey, I just tried to install RockBox on a Sansa e260. Autodetect did not work so I had to manually set it. It said the installation was complete, but the only thing I noticed was RockboxUtility.exe and RockboxUtility.ini on the player. Now I tried to run it again and it says it does not detect any sansa players. I do not have the Sansa Update Utility installed and I tried running it as Administrator, it still won't work.
08:27:04*GodEater votes for a revert until it gets more testing
08:27:16amiconnSeveral things can't be right
08:27:17LloreanGodEater: I'm not against *very* finite customization options. For example, two or three choices for the short-press on record, and a limited pre-set choice of options for the quickscreen, choosable by a menu interface.
08:27:19*karashata agrees with amiconn and GodEater
08:27:28*Llorean thirds the revert suggestion.
08:27:46GodEatermotion carried ?
08:29:54GodEaterWafflesDr: what version is the stock firmware ?
08:30:07 Join conando [0] (
08:30:20WafflesDrlet me check real quick, I know its 1.xx, not sure the details
08:30:31 Join markun [50] (n=markun@rockbox/developer/markun)
08:30:45GodEaterwell if it's 1.xx anything it should get detected
08:31:30WafflesDrfirmware 01.01.11A
08:32:04GodEaterand you're sure it's in MSC mode, not MTP ?
08:32:19scorcheMattAndrew: yes...the second core cna be put to sleep
08:32:26WafflesDryup, its MSC
08:32:46WafflesDrMSC is where you can see all the files, hidden and everything right? I changed it over before I did anything to the player.
08:32:46*GodEater exhausts his knowledge of sansa installs
08:32:51scorcheamiconn: yeah...i didnt bother to give it more than a passing glance...did the dishes..
08:32:55amiconnJdGordon: ping....
08:33:19 Quit Rob2222 (Read error: 110 (Connection timed out))
08:35:35 Join joecool_ [0] (i=joecool@gateway/gpg-tor/key-0xA7BF8828)
08:36:06WafflesDroh you know what
08:36:19 Quit joecool (Nick collision from services.)
08:36:22 Nick joecool_ is now known as joecool (i=joecool@gateway/gpg-tor/key-0xA7BF8828)
08:36:24WafflesDrI deleted the RockboxUtility.exe and .ini that ended up on the player (i don't think i put them there) and now its installin
08:36:29 Nick joecool is now known as joecool|sleep (i=joecool@gateway/gpg-tor/key-0xA7BF8828)
08:38:01 Quit joecool|sleep (Client Quit)
08:38:04amiconnJdGordon: Uhm, you changed the copyright line of an existing source file.....
08:38:27 Join homielowe [0] (
08:40:28 Join joecool [0] (i=joecool@no-sources/joecool)
08:43:55 Join linuxstb_ [0] (n=chatzill@rockbox/developer/linuxstb)
08:44:57 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:45:28pondlifeJdGordon: +1K for the blimmin' quickscreen??
08:46:23*GodEater welcomes pondlife late to the party
08:46:34*pondlife reads the party logs
08:46:42 Join ender` [0] (
08:46:46pondlifeHey, I have to sleep too, y'know
08:46:57 Quit ol_schoola (Read error: 104 (Connection reset by peer))
08:46:57GodEaterdon't we all /
08:46:59*scorche hands pondlife an "angry liquor"
08:47:09*pondlife drinks and reverts
08:48:05amiconnThe quickscreen works even worse than the old one with a large font on archos
08:48:22karashatasomething else I just discovered, the quickscreen isn't properly updating the repeat mode when switching it
08:48:28pondlifeIf you're going to have a customisable quickscreen, I'd much prefer a binsize reduction
08:48:40amiconnThe right option hides its value
08:48:43 Join ol_schoola [0] (
08:49:30*GodEater proposes "Customisable quickscreen" as "Most unpopular commit to Rockbox of 2008"
08:49:35amiconnWhy are the left & right option not at the same y position - that's one of the things which should be dead easy with viewports. far.
08:49:44amiconnThe icons are also missing
08:49:46GodEaterpondlife: goes without saying ;)
08:49:48pondlifeGodEater: Beating the keyclick into second place??
08:50:09GodEaterbeating it so far into secondplace it's not even worth considering for the award
08:50:43*pondlife notices the support towel rack has an extra towel on it
08:50:50karashataack, somehow I managed to trigger a restart of my currently playing track from within the quickscreen and now my WPS messed up...
08:50:57GodEaterbtw - did you give up with my VM ?
08:51:12pondlifeBut not had any time to proceed
08:51:23karashatathough going to the main menu then back to the WPS fixed that easy enough
08:51:23pondlifeGotta go now, might have time this weekend
08:51:24GodEaterjust wondered if you'd gone with rasher's instead
08:51:29GodEaterok - later!
08:51:29 Quit linuxstb_ (Read error: 104 (Connection reset by peer))
08:51:38amiconnSlight correction: the icons are there if there's enough room. But the text postions are broken, and the decision whether it has to scroll is broken too
08:52:02 Nick joecool is now known as joecool|sleep (i=joecool@no-sources/joecool)
08:54:01JdGordon.. and so it begins :)
08:54:33*scorche sees daniel walk into the lions den
08:54:52JdGordonicons arnt missing...
08:55:17JdGordonand left and right not being level is becuase its harder to read when they are
08:56:06lostlogicjoecool|sleep: should I know you? reping tomorrow if so
08:56:54scorcheJdGordon: i think a good first thing to address is why was this simply committed with a minimum of discussion, no task, etc?
08:56:56pondlifeJdGordon: If you can get it back to (a) non-customisable and (b) a green delta, then do so ;p
08:57:17JdGordon(a) easy.. (b) blame viewports
08:57:50pondlifeHmm, in which case maybe the viewport implementation needs more work?
08:57:50Lloreanpondlife: And (c) actually working? There seem to be a lot of things broken about it judging by the people already speaking up
08:58:05pondlifeI've not got as far as downloading it yet.
08:58:20JdGordonworks fine here... even tested with different fonts in the rec sim..
08:58:21pondlifeBut +1K for a cosmetic change is poor value, IMHO
08:58:48JdGordonlike i said.. blame vp on that
08:59:17pondlifeIn which case do it without using them...
08:59:39 Quit scorche (Nick collision from services.)
08:59:42 Quit advlaptop2019 (Read error: 104 (Connection reset by peer))
08:59:47 Join advlaptop2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
08:59:48JdGordonyou'll notice it doesnt use sysfont anymore? try doing that without viewports..
08:59:52LloreanJdGordon: Well read karashata's comments from just a bit earlier
09:00:00JdGordoncant... drunk
09:00:08lostlogicJdGordon: you had one beer, you're just sleepy
09:00:10 Join scorche [0] (i=Blah@rockbox/administrator/scorche)
09:00:11lostlogichave a coke and fix it.
09:00:20pondlifeBut better still, let's see if we can use this opportunity to improve the memory efficiency of viewports?
09:00:39pondlifeI mean, if it's going to be big red deltas each time we add one....
09:00:51GodEaterpondlife: "we" or "linuxstb your implementation gets a D-, redo from start"
09:01:18pondlife"we" - meaning everyone but me
09:01:34 Nick advlaptop2019_ is now known as advlaptop2019 (n=advcomp2@unaffiliated/advcomp2019)
09:01:43GodEaterWRT delta size, isn't this the first time VP has been commited though ?
09:01:57pondlifeAh, I thought that had already gone in.
09:01:58GodEateri.e. later things using it wouldn't have such enormous redness
09:02:09pondlifeI may have misunderstood
09:02:34GodEateryou know what they say about assumption...
09:02:35amiconnJdGordon: As it is now, the diferent y positions make it harder, even impossible to read with large fonts on small screens
09:02:37 Join petur [50] (n=petur@rockbox/developer/petur)
09:02:47pondlifeThe log message "use viewports" made me think it was using an already committed feature.
09:03:01amiconnAnd they wouldn't be hard to read on level if there would be a spacing between
09:03:05GodEaterpondlife: I may have missed the commit too - in which case I could be wrong :)
09:03:16 Quit cool_walking_ ("Miranda IM! Smaller, Faster, Easier.")
09:03:18JdGordonthere is spacing between them
09:03:24GodEaterI thought linuxstb had just commited the driver portion
09:03:24amiconnI think the icons should be placed into the centre if there's not enough room in the outside
09:03:26GodEaternot the userspace bits
09:03:45JdGordon_something_ had to be first
09:04:06*karashata notes that the quickscreen takes around 5 or 6 seconds to exit back to the WPS after pressing the button to leave it...
09:04:20pondlifeGodEater: 8th Jan has - "Add the viewport functions to the screens API..."
09:04:21JdGordonand check the logs.. i said from the beginign all viewport comits would be red
09:04:47amiconnThe bottom option doesn't use the single-line switch you proposed yourself... and the left & right options aren't deciding properly whether they have to scroll, and if they scroll, their placement is wrong (at least the left one), overwriting the icon
09:05:12amiconnAnd what about the quickscreenless targets??
09:05:14JdGordonhow many lines are viewable with your font?
09:05:15***Saving seen data "./dancer.seen"
09:05:28GodEaterpondlife: I take it all back then
09:05:31karashataalso, the viewport for the quickscreen doesn't cover the whole WPS screen, there's a line across the top, I'm guessing about 8 pixels thick, that doesn't get covered
09:05:48GodEaterkarashata: for the status bar perhaps ?
09:06:02amiconnThat is, 4 lines with the font where everything breaks (Nimbus-14), and 7 with the font where it breaks partially (Rockfont-8)
09:06:08karashatanone of my WPSes use the status bar though, so it looks bad
09:06:09amiconnI have the statusbar enabled of course
09:06:43JdGordonamiconn: change quickscreen.c:66 to >=3 instead of ==4 and it should be ok
09:06:53JdGordonstatusbar was always shown with the old code
09:07:12JdGordonslight delay is because settings_apply() is always called now if an item is changed
09:07:19JdGordonwtf with the colours
09:07:27JdGordonand wps scroll will be fixed in a sec
09:07:31 Join broom [0] (n=c27f0812@gateway/web/cgi-irc/
09:07:42JdGordonget_foreground() seems to be broekn
09:08:36broomHello. Why is the parameter add_files in tagcache.c:check_dir() declared as int? Shouldn't it be bool (for better readability)?
09:12:11 Quit joecool|sleep (Remote closed the connection)
09:12:59 Join EspeonEefi [0] (i=espeonee@STRATTON-FIVE-SEVENTY-NINE.MIT.EDU)
09:14:16karashatahmm.. the repeat mode in the quickscreen changes from off to all when changing modes, but not from all to one or any of the other modes, and when it's switched back to off again it changes back to off
09:14:52karashatathe mode does change though, the quickscreen's just not showing the mode changing
09:15:43 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
09:16:37 Quit ol_schoola ()
09:17:11 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
09:17:29 Quit linuxstb (Nick collision from services.)
09:17:39 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
09:19:21karashatathe show files option doesn't change at all when changing what to show
09:19:43JdGordonI cant reproduce that at all... all 4 optinos work fine
09:20:05karashataI only have three options
09:20:18 Quit EspeonEefi ("さよなら")
09:20:19JdGordonthe defaults? they all work fine here
09:20:23linuxstbpondlife: I would agree - if the current viewport implementation is going to result in significant red everywhere they're used, then something's wrong.... But this patch doesn't seem to be a good example (it's including other changes at the same time) - and the binsize delta for the Player/Ondio (without quickscreen) are bigger than some targets with quickscreen (e.g. other Archos devices).... So my guess would be that the colour handli
09:20:24linuxstbng should be changed.
09:20:55JdGordonthere are no other changes
09:21:35amiconnImo the icons should be placed into the centre at all times (iirc it even was that way before multi-screen support)
09:21:46*karashata goes to test the quickscreen with the default theme to see what happens
09:22:04amiconnAnd we probably want a compile-time option to select between 3-item and 4-item quickscreens
09:22:25JdGordonthe extra item makes SFA difference.... just set it to none if you dont want it
09:22:36 Quit Seed (Nick collision from services.)
09:22:36JdGordonand how are the icons not showing in the center?
09:22:42 Join Seedy [0] (
09:22:49 Join advlaptop2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
09:22:51 Quit advlaptop2019 (Read error: 104 (Connection reset by peer))
09:22:53 Quit JdGordon (Remote closed the connection)
09:23:13 Join JdGordon [0] (
09:23:39karashatahmm, show files will show all and supported, but not the other two, repeat shows off and all, but not one, shuffle, or A-B, shuffle shows both no and yes fine
09:23:43LloreanJdGordon: you can't set it to "none" if we end up removing the as-yet-not-fully-discussed customization.
09:23:45karashatawith the default theme
09:24:08LloreanJdGordon: Which, you still haven't explained just committing without more discussion, since it's ostensibly against the overall intent.
09:24:38amiconnJdGordon: A 4-way quickscreen doesn't make sense on ipods, as then the screen can't be exited anymore (in a logical way). And on recorders, it makes little sense because of the limited space
09:25:01amiconnAnd having the decision between 3 and 4 options dynamic increases code size
09:25:09 Join knittl [0] (
09:25:18*karashata still wonders where the customizable part comes in, he can't find anywhere to customize anything
09:25:45GodEaterkarashata: read the commit message....
09:25:55*karashata goes to check
09:26:02 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:26:17 Nick advlaptop2019_ is now known as advlaptop2019 (n=advcomp2@unaffiliated/advcomp2019)
09:26:35amiconnThat is, the ipods *could* have a 4th option on Select, the same way as the recorders could have it on Play (that'd even work for quickmode)
09:26:51JdGordonthe only wtf bug is the colours... linuxstb spot my mistake?
09:27:00amiconnBtw, quickmode still doesn't work, and leaving the F3 quickscreen is also broken
09:27:25amiconn(these things were broken before your commit, I know)
09:27:26*JdGordon checks log.... nope... pretty sure it mentions nothing about fixing quickmode
09:27:38*linuxstb isn't against the idea of configurable menus, as long as a) It is only adding additional ways to access items (which is what the quickscreen is anyway); b) it's easily configurable (i.e. there's a GUI).
09:28:12scorchei would weigh in more, but i would like to see a better delta
09:28:55 Join Insectoid [0] (
09:29:02 Join EspeonEefi [0] (i=espeonee@STRATTON-FIVE-SEVENTY-NINE.MIT.EDU)
09:29:48InsectoidWith latest build on Ipod Video 32 mb, if the voice file is not current with english.lang, playback locks up the player. The voice is also apparently not current on the voice build page, rendering the player inoperable.
09:30:13 Quit broom ("CGI:IRC")
09:31:13JdGordonamiconn: how about screenshots so the tweaking that you say needs to happen can happen?
09:32:03 Quit jcollie (
09:32:03 Quit n17ikh|Lappy (
09:32:03 Quit roxfan (
09:32:03 Quit HellDragon (
09:32:03 Quit dancek (
09:32:03 Quit Ave (
09:32:03 Quit Tuplanolla (
09:34:22pixelmaJdGordon: did I miss an explanation why targets without quickscreen (Player, Ondios) got such a big red delta too?
09:34:55NJoindancek [0] (
09:35:02JdGordonyou obvviously missed my commits also
09:36:57linuxstbJdGordon: I'm not sure why that wouldn't work, but fg_pattern and bg_pattern also need to be set for greyscale targets. Also, wouldn't it be better to write a function to populate a full-screen viewport struct with defaults?
09:37:30NJoinroxfan [0] (
09:38:01JdGordoniirc only colour targets have fg/bg colours in the settings atm and yeah, slipped my mind
09:38:22linuxstbIt doesn't matter if they're customisable - they still need to be set to the defaults.
09:41:15*pondlife thinks anything customisable should have a GUI
09:41:37pondlifeNo need for a text editor (unless you want to do that)
09:41:51Lloreanpondlife: Themes being the exceptin.
09:42:05pondlifeWPS, you mean?
09:42:21LloreanWe don't have a GUI for creating WPS, Iconsets, settings icontypes to file types, setting file colors.
09:42:44*linuxstb would love to see a plugin to do that...
09:42:45pondlifeIconsets should be selectable via a GUI.
09:42:50Lloreanlinuxstb: I would too, actually
09:42:59Lloreanpondlife: They're selectable, just not creatable.
09:43:26LloreanActually, I'm not sure if you can select a .icons file directly, or need a .cfg for it.
09:43:41 Join Buschel [0] (
09:44:03JdGordonyou cant
09:44:44JdGordonbtw.. you wanted a small delta.. a gui for this would be an extra 1k probaly... ill do it in a plugin
09:46:17*pondlife would like to see the whole settings menu moved to a plugin...
09:46:48pondlife..but I know that's not a popular idea.
09:47:10 Quit WafflesDr ()
09:47:12*JdGordon suggested that over a year ago
09:47:38NJoinHellDragon [0] (i=jd@unaffiliated/helldragon)
09:48:27 Join Rob2222 [0] (
09:48:56 Quit matsl (Read error: 110 (Connection timed out))
09:49:10 Join CaptainSquid [0] (
09:49:20LloreanJdGordon: And it was discussed and rejected. I'm in favour of it too.
09:49:38LloreanThe key though is that it was discussed. Rather than just pushed through with less than a day's notice, and no chance for comments...
09:49:43NJoinn17ikh|Lappy [0] (
09:50:09NJoinjcollie [0] (
09:50:09 Quit Rob222241 (Read error: 104 (Connection reset by peer))
09:50:51JdGordon4th time lucky maybe?
09:51:06pixelmaoh dear, already put on MajorChanges...
09:51:32JdGordoncourse... otherwise people will complain they have no idea how to change the setting
09:51:33LloreanOkay, seriously, what the hell.
09:51:56pixelmaI thought it is still being discussed?
09:52:03LloreanNo offense, but frankly this is bullshit. The project has long been "no customizable menus", and you intentionally avoid discussion and commit it?
09:52:28pondlifeRevert revert revert
09:52:28LloreanSVN isn't your private playground, and you've pretty much been ignoring the fact that we've asked why you didn't wait to discuss it more.
09:53:36LloreanNot only that but it seems you were enough in a hurry that you pushed in a fairly buggy patch since you were more interested in getting it in, than providing it for testing first.
09:53:57 Quit Buschel ()
09:55:15pondlifeI find the "green fetish" comment a little offensive too. It's not a fetish, it's a fundamental.
09:56:12pixelma"quiet the masses" isn't much better...
09:57:04LloreanSeriously, I mean, putting it immediately on MajorChanges when you knew it was unpopular among at least a moderate sized portion of the devs seems like a bid to get the users on your side so it's harder to revert.
09:57:17*pondlife wonders if SVN should have a tube attached, and require a breath sample for the commit command... ;p
09:57:40 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
09:57:42LloreanI mean, I gave you a chance to offer an explanation, but since you don't seem interested, I'm going to come out and say it really feels like you're just trying to force it on the project without discussion because you didn't feel like it'd get in legitimately.
09:58:14scorchepondlife: only to enforce the Ballmer Peak, surely
09:58:42 Join S [0] (
09:58:44JdGordonim taking scorche's advice and reply tomoorow after a good night sleep
09:58:47 Quit JdGordon ("zzz")
09:58:58LloreanSo, who wants to do the honors of reverting it until tomorrow?
09:59:05 Quit conando (Read error: 104 (Connection reset by peer))
09:59:17pondlifeI would but I'm at work right now
09:59:31 Join tvelocity [0] (n=tony@
10:00:52scorcheanyone else jumping, or shall i boot up the VM?
10:01:32*petur has no svn access at work
10:01:47*scorche boots it up
10:01:57LloreanI can do it, but it'd be possibly the first commit by phone
10:02:13*petur edited the wiki
10:04:21Lloreanscorche: Am I doing it or you?
10:05:07scorchei have it booted up and getting ready to svn a clean revision and will do it if it is truly wanted...or we could wait a bit for cooler heads
10:05:24LloreanHe can re-commit it tomorrow.
10:05:41LloreanBut for the moment, it needs discussion.
10:06:00GodEaterwouldn't it be an idea for it to a) be working, and b) be agreed first ?
10:06:04LloreanI'll take the heat on it.
10:06:15LloreanGodEater: As in, revert the change until it at least works?
10:06:23Lloreanscorche: I'll do the commit. He already hates me anyway
10:06:40linuxstbAt the very least, patches should be working _before_ committing...
10:06:53scorcheyou sure?...i am in the middle of going through svn
10:06:54GodEatercan't resist this any longer :
10:07:04Lloreanscorche: I've already reverted the changes locally
10:07:10LloreanI was just waiting to see if you were ahead of me.
10:07:17*scorche shrugs
10:07:25LloreanAnd I'm good at taking hear
10:07:39*scorche is a neutralish type :)
10:07:56 Join Jeton [0] (n=4f7ebca7@gateway/web/cgi-irc/
10:09:00LloreanOr I could be unable to remember my SVN password. Again. >_>
10:09:14*Llorean feels like a bit of an idiot.
10:09:14JetonI'm a little confused. How do i use the QuickScreen that was commited, on my Sansa e200. Which button is assigned to it?
10:09:38GodEaterJeton: you don't for the time being - we're reverting it
10:09:39LloreanThere we go
10:09:42LloreanI remembered it.
10:09:45XavierGrsee users already want to try it :P
10:09:57Jetonoh, lol. OK :)
10:09:57*Llorean hopes he did this right
10:10:02markunwhich part of the commit will be reverted?
10:10:14markunI like using the UI font in the quick menu
10:10:19GodEaterJeton: but as to "which button" it will be the same button that the quickscreen has always used
10:10:39scorchemarkun: at this point it seems more like an all-or-nothing
10:10:45Jetonwhich was holding the REC button. (?)
10:10:52GodEaterJeton: it's in the manual...
10:11:00scorcheLlorean: all 3?
10:11:01Lloreanmarkun: That'll make its way back in.
10:11:08Lloreanscorche: Huh?
10:11:14scorche3 commits
10:11:14*Jeton checks the manual
10:11:20LloreanWhat about 3 commits?
10:11:38scorchewhat did you revert?...he did 3 commits
10:11:43LloreanIt should revert all of them
10:11:48XavierGrwell IMHO except the buggy part that some mentioned (didn't try it), the idea can be favorable
10:12:20Lloreanscorche: I separated Mike's change, reverted JD's, reapplied mikes, and used that.
10:12:33scorcheXavierGr: it seems that it is more the way it was put across
10:13:34*Llorean tried to go for a very diplomatic commit message
10:14:25XavierGrwell I didn't like the hit and run tactic but also I don't like disputes between people...
10:14:49scorchei think it was batter to revert now and discuss when cooler heads prevail
10:14:57scorcheberrter in point? :)
10:15:36 Join Buschel [0] (
10:15:39pixelmamarkun: btw. I don't understand the database.unignore. Don't know the code but doesn't that slow the process down for everyone who uses database.ignore because now it needs to look in every subdirectory when it could be skipped before? And can't you have the same effect if you are putting database.ignore in your subfolders?
10:15:54XavierGrwhere are the pacificist swedes anyway? :P
10:17:52LloreanOkay, looks like I didn't break anything.
10:18:34markunpixelma: if you have dircache off it could have quite and impact. With dircache on you will not notice a difference.
10:18:39 Nick S is now known as conando (
10:18:41 Quit amiconn (Nick collision from services.)
10:18:47 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
10:19:13markunyou could get the same result with a lot of database.ignore files, but this way I think you can do it a lot simpler
10:19:18pixelmastill don't see a need for it...
10:19:38markunwell, I have a lot of folders with audio files which I don't want in my database
10:19:43pixelmaor you could organise your files better
10:20:12markunI think they are organized quite well
10:20:45markunI have a 'audio' folder, with 'podcasts', 'albums
10:20:45 Quit Jeton ("CGI:IRC (EOF)")
10:20:59markun', 'misc', 'SPC' etc
10:21:23markunand also in the root I have a 'test' folder whith files which used to crash rockbox or for test plugins
10:22:08markunthis way I just need 2 files (ignore in the root, and unignore in albums) and don't need to keep adding ignore files when I add folders with audio files somewhere.
10:23:22markunpixelma: anyway, did you notice any problems yourself or are you just worried (which I can also understand)? I could revert it and keep using it in my own build (as I did before).
10:26:19 Quit TMM (Remote closed the connection)
10:26:33 Join broom [0] (n=c27f0811@gateway/web/cgi-irc/
10:27:15pixelmahaven't noticed problems, I just couldn't understand it (not sure I know more now). I don't use the database much and database.ignore even less (only by way of the included one in .rockbox)
10:27:57 Quit linuxstb (Read error: 113 (No route to host))
10:27:58broommarkun, pixelma: what about having a ignore-recursive file which would avoid scanning sub folders?
10:28:35 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:28:43broomA simple ignore would ignore the folder but then proceed with the subfolders
10:28:45markunbroom: could also be done ;)
10:29:10markunbut keeping the unignore file, right?
10:29:24broommarkun: but then users will ask why the unignore files doesn't have effect! :-)
10:29:38markunyes, it's a bit tricky
10:29:50broommarkun: yes, two types of ignore, one unignore.
10:30:21markunI think I'll wait a bit before I make more changes to it until we come up with something nice and logical
10:30:46pixelmabroom: that sounds even more complicated...
10:30:54broomSo more users are asked to tell their opinion about it.
10:31:28Lloreanmarkun: Can you do database.ignore->database.unignore->database.ignore nesting?
10:31:34markunLlorean: yes
10:31:40LloreanThen I'm happy with the existing system.
10:31:51broompixelma: that's a way to combine few (un)ignore files (markun) with quick scans without dircache (pixelma)
10:31:59LloreanIt's basically a way to flip the "database" switch on and off at different depths, and that's good enough for me, even if I'll never use it.
10:32:53markunLlorean: I think the problem that pixelma has with it is that the whole tree is always scanned (like it was before the database.ignore files existed)
10:33:00broommarkun: and make the func parameter bool (instead of int) :-)
10:33:30linuxstbmarkun: I'm also not sure about your .unignore commit - for the reason you just said... e.g. if you have the 32000 file High Voltage Sid Collection on your disk, you don't want the database to scan it.
10:33:37Lloreanmarkun: There's an easy solution: database.force.ignore, tells the database that there are no unignore files beneath it. :-P
10:34:15linuxstbOr as pixelma said, organise files differently ;)
10:34:18markunlinuxstb: yes, for non dircache users it will not be very nice.
10:34:34broomLlorean: exactly my thought, just another file name
10:34:39markunbut I think dircache should be on by default on the targets where it makes sense
10:35:04 Quit karashata (" HydraIRC -> <- The professional IRC Client :D")
10:35:16markunlinuxstb: does the high voltage sid collection have many sub dirs?
10:35:17pixelmamarkun: that's a quite personal choice...
10:35:30Lloreanmarkun: HVSC has a LOT of subdirs.
10:35:37markunpixelma: sure, but I think it would be a better default.
10:35:52Lloreanmarkun: I think it still has some significant bugs to work out
10:36:01LloreanNot initializing with very large numbers of files, etc.
10:36:21LloreanBut I'd agree, once dircache works "well enough", it should probably be enabled for disk-based SWcodec targets.
10:36:22markunLlorean: good point
10:36:26pixelma...and on a few targets where you can have the most music (no SIDs and the like though) are the Archos HDD based players and there is no dircache)
10:36:55markunpixelma: so?
10:37:38 Join matsl [0] (
10:38:07broommarkun: so the need to avoid subdirs scans?
10:38:53markunah, tagcache but not dircache... sorry
10:39:16markunpixelma: would you be willing to do a speed test on an archos?
10:39:28markunor do you only have an ondio?
10:39:30 Join Mathiasdm [0] (
10:39:42pixelmaonly Ondio
10:39:49markunsomeone else?
10:40:14 Join cool_walking_ [0] (
10:42:59amiconnI admit that I don't really get the purpose of unignore files... I do like the ignore feature though
10:43:00 Quit Llorean (Read error: 104 (Connection reset by peer))
10:43:54amiconnIt's especially handy for my .mod file folder (I'm running the MOD player patch), because the mods provide kinda broken metadata
10:45:19amiconnmarkun: Using dircache is *very* subjective... e.g. I don't want it on X5 (too much RAM waste - the cf iaudios only have 16MB), and also not on Mini (makes no sense, as the mini's hdd spins up in less than 500ms)
10:47:58 Quit jhulst ("Konversation terminated!")
10:49:18markunamiconn: "where it makes sense""
10:51:53 Quit broom ("CGI:IRC")
10:52:24 Join ender [0] (
10:52:43amiconnYeah, and that definition is rather subjective
10:53:44 Quit mokkurkalve ("leaving")
10:55:52markunChosing good defaults is mostly a subjective thing. Still I think it's important.
10:56:26 Quit CaptainSquid ("Miranda IM!")
11:05:16***Saving seen data "./dancer.seen"
11:06:04 Quit ender` (Read error: 110 (Connection timed out))
11:07:58pixelmamarkun: speed impact on my c250 with a 1GB microSD (both almost completely filled, so 3GB in total), obviously no dircache but flash based - and the only database.ignore in .rockbox: 5 seconds difference (10 seconds with r16213, 15s with r16214).
11:08:25markunquite a lot
11:09:57markunI'll try to come up with a nice solution, otherwise I don't mind reverting it.
11:16:38 Nick daurn is now known as daurnimator (i=daurn@unaffiliated/daurnimator)
11:18:54 Join Arathis [0] (
11:20:47 Join CaptainSquid [0] (
11:33:00 Join BitTorment [0] (
11:38:21 Join Av3 [0] (
11:38:51 Join ender` [0] (
11:44:05 Join devurandom [0] (
11:45:22markunhi devurandom
11:45:55devurandomCould it be that there are quite some problems involved with the database and skipping of tracks? (RockBox from 02.02. on iPod Nano)
11:47:02devurandomIt seems as if it just displays the next track, but still plays the old and then skips to a completely different track...
11:48:50amiconnBagder: The problem on the ml (error in system.c) is due to compiling an ages-old rockbox source with gcc 4. And it's pretty obvious that he's trying to build for archos player
11:52:34*LinusN thinks "ignore" and "unignore" are silly names, why not "exclude" and "include" instead?
11:52:42 Quit ender (Read error: 110 (Connection timed out))
11:54:24markunLinusN: yes, sounds better
11:56:37markunbut can we still change it?
11:57:14pixelmadevurandom: I also have a weird problem with skipping tracks on my Sansa (in certain folders, no database) - I'm still trying to find out what determines a "working" and a non-working folder. Did you use iTunes to copy the music to your Nano? Just thinking it might explain the randomness for you, it's not that random here...
11:57:50devurandomI purged my iPod after switching to rockbox.
11:58:34pixelmahmm, no idea then. Maybe it's related, maybe not. :\
11:59:20devurandomFunny enough, I somehow booted to orginal firmware...
11:59:27devurandomI thought that got replaced by RockBox...
12:00:49 Join BigBambi [0] (n=86ceaf37@rockbox/staff/BigBambi)
12:00:54pixelmathe original firmware is not deleted and even can't be because you have to use the original firmware for USB connection. You can dual boot and AFAIK the manual describes how...
12:02:12BigBambiIndeed it does
12:02:17pixelma^ this is only true for portalplayer targets
12:02:48pixelma(the USB connection part)
12:03:32devurandomI thought the connection screen was black/white after installing RB... (While it was coloured with Apple's.)
12:03:39devurandomProbably was just some impression then.
12:03:50BigBambiThat is Apple's emergency disk mode
12:03:59BigBambiRockbox reboots the iPod to it
12:04:00cool_walking_LinusN: It's not clear what the precedence is there. It may seem like you have to database.include a folder to get it to initially be included.
12:04:49 Quit Jon-Kha (Remote closed the connection)
12:05:05devurandomSo I at least know now how to evade this... ;)
12:05:13BigBambidevurandom: On portalplayer targets (as pixelma says), rockbox does not have its own USB mode - you have been using Apple's all along.
12:05:28devurandomBtw: It also boots the original firmware when you shutdown your PC while RB is connected. ;)
12:05:54devurandom... Fully boots the orig fw. Not just some diskmode.
12:07:04devurandomHad that 2 times till now, so I guess it was not just some "random feature", but probably can be reproduced.
12:07:19BigBambiI don't know, I don't own an iPod
12:08:00BigBambiBut I believe since the start of our own USB stack went in to SVN, there have been a few oddities with the automatically rebooting iPods
12:15:29cool_walking_LinusN: Also, it's not clear what should happen if both "exclude" and "include" files are present in the same directory, whereas it seems obvious (to me at least) that "unignore" should take overrule "ignore".
12:18:26cool_walking_LinusN: Also, once people are already using a protocol, it's debatable whether it's a good idea to change it e.g. HTTP "referer".
12:19:31 Join agm3nt [0] (
12:20:32 Join mf0102 [0] (n=michi@
12:20:47 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:21:27 Quit homielowe (Read error: 110 (Connection timed out))
12:21:45Nico_PjhMikeS: congrats on the S progress! I had no idea you were working on it :)
12:23:25 Join axionix_ [0] (
12:26:00Buscheldoes anybody know, if the high battery consumption (long time HDD-access while buffering) with 5G 60/80GB has been fixed with revision 16197?
12:26:08BigBambiNico_P: Oooh, I missed S progress! In the logs?
12:26:24Nico_PBigBambi: in SVN!
12:27:22Nico_Papparently jhMikeS didn't mention it here today or yesterday
12:27:28BigBambiAh, it went on and off the front page too quickly for me
12:27:34*pixelma wants to point Nico_P to ...
12:27:37BigBambiExciting :)
12:28:03Nico_Ppixelma: thanks
12:28:08Nico_PI'll look into it
12:28:19pixelmathank you :)
12:29:02*Nico_P noticed a bunch of new tracker entries about playack bugs :(
12:29:20Nico_Pseems I still have lots of work to do
12:29:47Nico_Pwas there any new information about the playback glitches?
12:29:52BigBambiSilly you taking it on
12:30:03BigBambi(but don't stop!)
12:31:41pixelmaNico_P: I have found out a bit about skipping the 3rd song in a folder (but only consistent symptoms, unfortunately not the cause), I'm still looking but running out of ideas a bit
12:32:16Nico_Pdid jhMikeS' priority_yield commit help a bit?
12:32:23pixelmahaven't posted it yet because of that
12:32:59pixelmaon my Sansa I could only notice that the unresponsiveness during buffer is gone
12:37:39Nico_Ppixelma: when does the sliding_puzzle crash occur?
12:38:26pixelmawithout an album art bmp or when music is not playing and your trying to change the mode (which will try album art mode first)
12:38:37 Quit axionix (Read error: 110 (Connection timed out))
12:38:52Nico_Pshould be a rather easy fix
12:41:16pixelmabtw. there were also appearance problems with sliding_puzzle, e.g. on the Nano it looked like for a while. That's already fixed and improved though...
12:41:59Nico_Pouch... I merely committed stripwax's patch though
12:42:10pixelmaimproved appearance that is
12:44:21Nico_Pah, I'd like to point out that the AA crash isn't my code
12:44:49 Join Jon-Kha [0] (
12:46:01Nico_Phmm, on my sim it tries to open 'archos/.rockbox/albumart/(null)-(null).bmp', but doesn't crash
12:46:38amiconndevurandom: The rockbox bootloader boots into the OF if hold is enabled. That's intended and documented behaviour
12:54:27 Join Zagor [242] (
12:56:30Nico_Ppixelma: I just committed a fix
12:56:39 Quit Weiss (Remote closed the connection)
12:56:58 Join Buschel_ [0] (
12:57:09 Join roxfan2 [0] (
13:03:48 Join Febs_ [0] (
13:04:22 Quit Buschel (Read error: 110 (Connection timed out))
13:04:24 Quit Febs_ (Client Quit)
13:04:27 Join gevaerts [0] (
13:04:56gevaertsDoes anyone here know something about logf ? I can't get it to work on my sansa c250.
13:05:19***Saving seen data "./dancer.seen"
13:06:43amiconnNico_P: The "No file" test looks kinda odd...
13:07:20Nico_Pamiconn: it's what the id3 info contains when there is no file loaded.
13:07:54Nico_PIIRC audio_current_track does that
13:09:26 Quit roxfan (Read error: 110 (Connection timed out))
13:12:43pixelmawhat if someone changes this string?
13:13:41Nico_Pthat someone should grep the code to see if it's used anywhere else :)
13:14:08Nico_Pbut maybe a #define would be better, or even doing that at the WPS level
13:14:13amiconnUser-visible strings should be localised. Checking such string usually isn't a good idea
13:14:42 Nick Buschel_ is now known as Buschel (
13:15:03Nico_PI agree. I'll try to come up with a more elegant solution
13:16:02amiconnI know that the problem is not in the album art code - but the album art code exposes it
13:16:43amiconnI also know that it can be annoying to try to fix one problem, only to find that the problem is related to another problem in a different module, which is related to....
13:16:51 Join DerDome [0] (
13:16:56devurandomamiconn: So if I shutdown my PC while my iPod is locked, it will always boot into Apple's fw?
13:19:41pixelmaNico_P: but at least it doesn't seem to crash anymore (tried a e200 sim) :)
13:20:50 Quit BigBambi ("CGI:IRC")
13:21:33 Quit Febs (Read error: 110 (Connection timed out))
13:25:05 Quit cool_walking_ (Read error: 104 (Connection reset by peer))
13:25:41 Join ender [0] (
13:34:34 Quit Arathis ("Bye, bye")
13:37:44 Quit jcollie ("Ex-Chat")
13:38:05gevaertsIs building with debug enabled supposed to work ? I get apps/main.c:583: undefined reference to `debug_init'
13:38:06 Join Aevum [0] (
13:39:43 Quit ender` (Read error: 110 (Connection timed out))
13:40:53 Part Christo1
13:46:19Nico_Pgevaerts: not on all targets
13:49:00gevaertsNico_P: I'm trying to work out why the read on Zagor's USB storage driver doesn't work, so I'll need either DEBUG or logf(). Do you have any hints ?
13:50:24Nico_Plogf should be ok
13:50:52Zagorgevaerts: I'd suggest a good USB snooper too.
13:51:25gevaertsZagor: I have access to one at work, so that won't be a problem.
13:52:47gevaertsAm I supposed to do more than enable logf in the build and choose either logf or logfdump in the debug menu to get some output ?
13:53:04GodEaterthink you need to uncomment some #defines too
13:53:18GodEaterin whichever file you want info from
13:53:36 Join Weiss [0] (
13:53:36Nico_Por add #define LOGF_ENABLE
13:54:45gevaertsOK. That's ptobably what I was missing.
13:56:08 Join K4rP4D [0] (
13:58:09 Quit Insectoid (Read error: 110 (Connection timed out))
13:58:12 Join Insectoid [0] (
13:58:14gevaertsThanks. I get some logging now.
13:59:04gevaertsWhere should I report results (assuming I get some) ? In FS #7962, or should I open a new one ?
13:59:17GodEater7962 I should think
14:00:09gevaertsThe "problem" with 7962 is that it is closed.
14:00:23ZagorI'd say open a new one. 7962 is closed which makes it a bit odd to start adding patches and/or comments there.
14:00:38GodEateroh ok
14:01:00gevaertsZagor: ok. I'll do that once I have something to report.
14:03:56 Join karashata [0] (
14:04:40pondlifeNico_P: Hope you had a nice break
14:04:53 Quit ComradeHaz ("leaving")
14:04:54Nico_Ppondlife: very nice, thanks :)
14:05:03pondlifeI suspect many of the reported bugs will share the same cause
14:05:16pondlifeThe playlist index being misused in some places maybe?
14:05:27pondlifeOr in many places
14:06:00pondlife is nicely reproducible in the sim
14:06:30pondlifeAlthough I've no idea what's causing it still!
14:06:54GodEateryour gdb stopped working pondlife ? :)
14:07:03Nico_Pyeah I encountered it too
14:07:05pondlifeNah, my boss started working :/
14:07:26pondlifeI used to get lunchtimes ...
14:08:02 Join Thundercloud [0] (
14:08:31GodEaterpondlife: it's a bad sign when you start referring to your g/f as "my boss"
14:08:41pondlifeHaha, no.
14:09:08pondlifeThat's mrs pondlife to you
14:09:25 Join midgey [0] (
14:09:41GodEaterwhatever - I'm not calling her "boss".
14:09:44GodEaternot even for you
14:18:31 Quit DerDome ("Leaving.")
14:25:35 Join PaulJam [0] (
14:25:43 Join ol_schoola [0] (
14:32:30 Quit karashata (" HydraIRC -> <- Organize your IRC")
14:32:40 Join DerDome [0] (
14:34:51gevaertsThis is interesting : if I limit the transfer to 64 bytes it works, but not wih 128 bytes or more
14:35:08GodEaterI think that's what Zagor found
14:35:19 Join crazy_bus [0] (
14:36:05Zagorbut I didn't get reliable function even with 64 bytes
14:36:14gevaertsAccording to my tracer, it actually sends 95 bytes
14:36:35 Quit DerDome (Client Quit)
14:37:10crazy_busis there anyway to get rockbox to save a bookmark when the player is plugged into a usb port. As the player sometimes boots straight into the original firmware without saving the bookmark and sometimes does nothing
14:37:37 Join jcollie [0] (n=jcollie@
14:37:56GodEatercrazy_bus: sounds like that would be a good feature
14:38:52 Join DerDome [0] (
14:41:06crazy_busI was told that rockbox would shut down and load the of when usb was plugged in unless the select button was pressed at the same time. But sometimes it does it and sometimes not. Is there a specific set of conditions for rockbox to restart sometimes and sometimes just charge?
14:43:37 Quit knittl (Read error: 104 (Connection reset by peer))
14:43:51 Join knittl [0] (
14:44:04GodEatercrazy_bus: which player ?
14:44:26GodEaterall the DAPS I own work properly when holding down whichever button it is to avoid the USB connection
14:44:36GodEaterbut I don't own an e200, so I couldn't comment
14:45:03krazykitthe usb-restart is hit-and-miss on portalplayers, it seems.
14:45:18krazykitif you want to JUST charge, hold select when you plug in the cable
14:45:35 Quit DerDome ("Leaving.")
14:45:47GodEaterkrazykit: it's not hit and miss on my ipod
14:45:55crazy_busis there anyway to change that. So if you want to just restart you press restart?
14:46:20crazy_bus*press select
14:46:31 Quit midgey ()
14:46:38GodEatercrazy_bus: you'd have to make yourself a custom build
14:46:57krazykitGodEater, really? sounds like you're one of the lucky ones. my sansas never restart regularly and there's always someone complaining about a similar issue
14:47:52crazy_busthanks for your answers. One more question. Do some mp3's not work well with bookmarks. I had one that was over a hour and you sometimes had to try it three times for the bookmark to take it to the correct time
14:47:53GodEaterkrazykit: the issue about it restarting reliably seems more to do with the host OS
14:48:07pixelmait's hit and miss per build on my c200, either it's working with a certain build or not
14:48:11krazykiti've heard that too, but i still have reliability issues in linux
14:48:22GodEaterI only have issues with it in Windows
14:48:28GodEaterlinux has been 100% reliable so far
14:48:32 Join DerDome [0] (
14:52:33Nico_PGodEater: hi! how have the playback skipping isses been lately?
14:53:10GodEaterSince I stopped using Album art I don't get any skipping at all
14:53:21GodEaterbut I suspect that's just hiding the issue
14:54:41GodEaterbuffering and dircache are still fighting though
14:54:52Nico_Pwas that before or after jhMikeS' commit?
14:55:24GodEaterwhat, that I stopped using AA ?
14:55:25 Part LinusN
14:55:42GodEaterbecause his commit didn't make much difference
14:57:03 Quit Aevum (Read error: 104 (Connection reset by peer))
14:57:05GodEatermy plan would be to basically stop dircache from doing anything at all until buffering has filled available memory
14:57:15GodEaterbut I don't know how easy that would be to accomplish
14:57:36GodEaterwhen dircache / buffering aren't fighting over the disk, dircache finishes building in 14 seconds
14:58:01GodEaterwhen they're both competing though, I get non-stop disk activity for 3 to 4 minutes
14:58:32GodEaterand for some reason, as soon as you use an AA theme in that case, playback starts to stutter
14:58:42GodEateralthough I really don't understand why - the image is surely only loaded once
14:59:58Nico_Peven if several tracks use the same pic it will be loaded several times
15:01:21soapnew builds up Buschel
15:01:33GodEaterNico_P: that's, um, suboptimal ?
15:02:22GodEaterhard to improve ?
15:02:28 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
15:02:39 Quit LambdaCalculus37 (Client Quit)
15:02:41petursame goes when playing the same song several times, it is also buffered several times, no?
15:02:49Nico_Ppetur: yes
15:03:00 Join ryant [0] (n=47a85de2@gateway/web/cgi-irc/
15:03:06 Quit crazy_bus ("ChatZilla 0.9.80 [Firefox]")
15:03:08Nico_PI tried smoething some time ago but but there were new issues
15:03:08 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
15:03:41Nico_Plike the buffer getting filled with metadata if there was one track repeated (the audio would be reused, but not the metadata)
15:03:42gevaertsZagor: do you have a datasheet for the usb controller in the c200 ?
15:03:56ryantis it possible to combine two builds of rockbox
15:04:05ryantie. like take the features of one and put it with the other
15:04:19Nico_Pryant: not without compiling one yourself
15:04:32ryantyeah, darnit
15:04:44ryanti tried copy over the files, and combining the uhhh what's that file called
15:05:08ryantit looks like it tells rockbox where everything is
15:05:19Zagorgevaerts: we have docs for what we *think* is the usb controller, namely the same as in the Motorola iMX31
15:05:23***Saving seen data "./dancer.seen"
15:05:51ryantthanks, Nico_P
15:06:02GodEaterNico_P: this MoB is a bit of a nightmare eh ? :)
15:06:19ryantviewers.config... that's the one
15:06:36Nico_PGodEater: hehe, more playback.c than MoB actually ;)
15:06:38ryanti couldn't mod that and add the application files, could i?
15:06:49GodEaterNico_P: of course, silly me
15:07:13GodEaterryant: no you couldn't
15:08:10Zagorgevaerts: I can't find it in wiki, grab it at
15:08:30gevaertsZagor: yes. I'm now looking there.
15:08:50 Quit Insectoid (Read error: 110 (Connection timed out))
15:08:55Zagorthe usb module is documented in the big man pdf (~2MB)
15:10:00 Join Insectoid [0] (
15:10:25Zagoruh, sorry no. it's in the 14MB reference manual.
15:11:22Zagorchapter 32
15:11:49gevaertsZagor: I have it now. I'll try to figure something out later. Time to get back to some real work now.
15:12:13 Quit gevaerts ("Leaving")
15:14:06ryantgetting stardict on my 5g > murder > deleting files to make room for cygwin and compiling rockbox
15:15:33 Join joshin [0] (n=joshin@VDSL-130-13-144-158.PHNX.QWEST.NET)
15:19:12Nico_PGodEater: have you tested jhMikeS' S work?
15:19:18GodEaternot so far
15:19:28Nico_Pon my S the bootloader doesn't get past kernel_init
15:19:28GodEaterI'm not even sure what he did ? :)
15:19:39Nico_PIIUC he managed to get interrupts working
15:21:29 Quit DerDome ("Leaving.")
15:22:14GodEaterthat's excellent news
15:22:32 Join nicktastique [0] (n=nick@unaffiliated/nicktastic)
15:22:39GodEaterdo I need to build both bootloader and rockbox? Or is the bootloader I have there currently sufficient ?
15:22:52Nico_Pyou need to rebuild the BL
15:23:02*GodEater goes to do that
15:27:41Nico_Pthe main binary doesn't build
15:27:59GodEaterjust transferring the bootloader over now
15:28:23Nico_Pon my unit the BL shows the version number then freezes
15:30:00GodEaternot sure sendfirm is working
15:30:39GodEateris there a way to get the BL to pause on the version screen ?
15:31:22*GodEater supposes that removing rockbox.gigabeat might accomplish that
15:31:32 Join n1s [0] (
15:31:34Nico_Pyou could try adding a sleep() call
15:31:51GodEaterbut I'm not convinced the sendfirm has worked
15:31:52GodEaterI get errors
15:32:32 Join DerDome [0] (
15:32:39Nico_Pme too, but the file still gets sent
15:32:57 Join aliask [0] (n=chatzill@rockbox/developer/aliask)
15:35:39Nico_PGodEater: is there too much output for you to have time to read it?
15:36:03 Part at0m|c ("Clear Channel Event")
15:36:19*n1s is excited to see gigabeast progress :D
15:37:21GodEaterNico_P: it goes by to fast, and is replaced by the rockbox logo too quickly
15:38:15aliaskNico_P: Just built then, I'm freezing at in kernel_init too
15:38:15Nico_PGodEater: then try adding a sleep call. for me it caused a freeze, but at least you'll be able to confirm that sendfirm works
15:38:34GodEaterNico_P: I've just deleted the rockbox.gigabeat instead
15:38:52Nico_PGodEater: ah yes, good idea :)
15:39:00Nico_Pversion number should end with 2 and not 1
15:39:09GodEaterthen it didn't work
15:39:10Nico_Paliask: good to know I'm not the only one
15:39:23GodEaterI'm still on the 00000001 version
15:40:22Nico_Pwhat errors does sendfirm show?
15:40:43GodEaterI got it to work now
15:40:56GodEaterI get the freeze straight after it prints the bootloader version number
15:44:46Nico_Pfunny that jhMikeS got a different behaviour
15:45:01aliaskOr maybe he forgot to commit something
15:46:07GodEaterwhat did he get then ?
15:46:26Nico_P"Now it progresses to ATA init and fails with -11 but without crashing or hanging."
15:46:38aliaskThe problem is in tick_start somewhere
15:47:01Nico_PI don't know what he meant by "Replace all bootloaders"
15:47:14GodEaternot have a dual boot ?
15:47:44aliaskI think it just meant that the changes would only be seen if you update the bootloader
15:49:30 Quit Jon-Kha (Remote closed the connection)
15:50:00 Quit tvelocity ("Αποχώρησε")
15:50:00GodEaterwell it definitely hangs here =/
15:50:53GodEaterNico_P: what model do you have? 30 or 60 ?
15:51:20 Join Jon-Kha [0] (
15:51:22 Quit ryant ("CGI:IRC (EOF)")
15:51:30GodEaterwell it's not a model difference then :)
15:53:54 Quit Mathiasdm ("Yuuw!")
15:54:03Nico_Pno. do you know which model jhMikeS has?
15:54:15aliaskI have an s60
15:54:27Nico_Paliask: any idea what's wrong?
15:54:38aliaskTracing it now
15:56:24Nico_PI have to go do some shopping
15:56:38aliaskI'll be asleep in 10 minutes though
15:56:50aliaskWork tomorrow morning :(
15:57:16 Join desowin [0] (n=desowin@atheme/developer/desowin)
15:57:45Nico_Paliask: ok, no rush anyway
15:57:59Nico_PI'll tell jhMikeS about the freeze when I see him
15:57:59 Join lee-qid [0] (
15:58:27aliaskI'm so pleased to see the port make progress.
15:58:35Nico_Pme too :)
15:59:37aliaskmcr p15, 0, %0, c7, c0, 4 is what's causing it
15:59:48aliaskTesting to see if it works without that call
15:59:51 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
16:02:38 Join Tuplanolla [0] (n=jani@unaffiliated/tuplanolla)
16:03:00 Quit Rob2222 (Read error: 104 (Connection reset by peer))
16:03:12 Join Rob2222 [0] (
16:04:41 Quit Insectoid (Read error: 110 (Connection timed out))
16:04:59Buschelsoap: thanks for the new builds −− i am too dumb :/
16:05:15 Quit devurandom (Remote closed the connection)
16:05:18 Quit Zagor ("Client exiting")
16:06:22aliaskRemoving the call definitely stops it from crashing, the bootloader runs all the way through. Now testing with a rockbox.gigabeat to boot
16:06:46*GodEater wonders what the mystery call does
16:07:52aliaskFuddles with the mystical cp15 register on the ARM core. No idea what it's supposed to do though :)
16:08:09amiconnIt's not mystical at all. See the arm reference manual
16:08:42*GodEater doesn't own it
16:08:48amiconnAnd btw, cp15 is not a register, it's a coprocessor
16:09:00GodEaterI also don't suffer sufficiently with insomnia to want to read it currently =/
16:09:01 Quit Horscht ("User was distributing pornography on server; system seized by FBI")
16:09:26amiconnGoogle for ddi0100e_arm_arm.pdf
16:11:41 Join shotofadds [0] (n=1485000d@gateway/web/cgi-irc/
16:13:09aliaskLoading the rockbox.gigabeat produces the rockbox logo for a split second then a full screen "ATA error: -11\nPress ON to debug"
16:13:16 Join Insectoid [0] (
16:13:27aliaskSo yes, looks like it does the job once that call is gone.
16:13:51aliaskAnyway, I'm off to sleep. Great work jhMikeS!
16:14:07 Quit aliask ("ChatZilla 0.9.80 [Firefox]")
16:14:40 Quit shotofadds (Client Quit)
16:14:44 Join shotofadds [0] (n=1485000d@gateway/web/cgi-irc/
16:15:05amiconnIiuc, said instruction puts the cpu into low power mode and waits for an interrupt
16:15:21amiconnSo if interrupts are disabled, it will of course freeze the machine
16:16:42 Quit shotofadds (Client Quit)
16:16:42GodEaterI thought that jhMike's code *enabled* interrupts though, or was intended to ?
16:16:48 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
16:18:33Buschelhi horscht
16:18:44HorschtHi Buschel
16:19:02Buschelwhat were the results of your short test yesterday evening?
16:19:23HorschtI posted them on the forums, but soap already pruned them
16:19:29Horschtissue not fixed yet
16:19:53Buschelhmm, did you create a bug report in flyspray?
16:20:09Horschtno, i went to bed straight after :)
16:20:11soapoops - over pruned.
16:20:36Horschtno worries
16:21:01Buschelyep, no problem as we can meet here :)
16:21:20Horschtyeah, and after all you had already noticed the same issue, soap
16:21:38Horschtthat's why you made an old SVN checkout I assume
16:22:31*GodEater wonders where aliask found that line
16:22:44soapHorscht, I /suspected/ the same issue. I swapped out the builds as fast as I could to prevent someone from downloading them and running a full bench only to learn @ the end that the numbers were "crippled"
16:27:49 Join bumbl [0] (
16:30:24 Join perrikwp [0] (
16:31:47 Quit CaptainSquid ("Miranda IM!")
16:31:56 Join Klevi [0] (
16:31:58HorschtBuschel, flyspray seems to already have a bugreport (FS# 8397) on the disc activity issue. I don't think I can help much by "re-reporting" it
16:32:22KleviHi all, long time no see. =)
16:32:40Buschelhorscht: ok
16:33:12GodEaterNico_P: sendfirm appears to only work if there are no usb mass storage devices attached =/
16:33:22GodEaterI can't make it work unless I unplug my ipod first
16:35:13 Nick Av3 is now known as Ave (
16:36:17 Quit petur ("reboot")
16:37:35 Join Febs [0] (n=chatzill@
16:39:31 Join qwedsa [0] (
16:41:58 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
16:43:33 Join kugel [0] (n=kugel@unaffiliated/kugel)
16:44:09 Quit joshn_ (Read error: 110 (Connection timed out))
16:46:09 Join petur [50] (n=petur@rockbox/developer/petur)
16:48:37Buschelhorscht: did you try to check via dialy builds whith which revision this behaviour came in?
16:48:47preglowthe disk thrashing?
16:48:54Buschelpreglow: yes
16:49:02preglowi think it happened with a low-level commit by jhMikeS
16:49:14preglowsome spinlock related commit, i think
16:49:16 Join snipexv [0] (
16:49:30Buschelpreglow: why isn't it reverted then?
16:49:52Buscheluntil the bug is fixed and verified?
16:49:56preglowthat one, i think
16:50:06preglowdon't know
16:50:14preglowi'm not sure, btw, so better ask someone that is :)
16:50:32HorschtI have tried 15998 or something (cant remember the exact rev, but gonna try 16002 next
16:51:03Buschelhorscht: seems like preglow just answered my question −− no need to test
16:52:11HorschtBuschel, he is assuming, I could verify
16:52:39Buschelhorscht: if you've got the time :)
16:53:14Horschtwell, shouldn't take more than an hour :)
16:54:33Horschti'll try 16105 first, though
16:54:58 Nick fxb__ is now known as fxb (
16:55:37preglowme brb
16:55:46 Quit qwedsa_ (Read error: 110 (Connection timed out))
16:55:59GodEaterhow certain are we of the ATA addresses for the gigabeast?
16:56:38Nico_PGodEater: the sendfirm issue is probably down to libmtp
16:56:47GodEaterNico_P: probably :)
16:56:53GodEaterI just thought I'd make you aware
16:57:23 Join joshn_ [0] (
16:57:52Nico_PGodEater: and the line aliask disabled is in kernel_imx31.c:tick_start
16:58:04GodEaterNico_P: yeah I found it
16:58:07GodEaterI disabled it to
16:58:14GodEaterand am trying to work out wth is going on with ATA
16:58:31GodEaterATA_STATUS is coming back with nothing in it currently
16:59:50GodEaterI think it's the master_slave_check which is going wrong currently
17:00:09GodEaterI'm wondering if the ATA controller is so slow we need to wait a bit longer for a result from the status register
17:00:35 Join jgarvey [0] (
17:01:16GodEaterI stuck a sleep(500) in the slave_check so I could read my lcd message with the status register in
17:01:30GodEaterand after it exits, the ATA error returned in main.c is now -32
17:02:39GodEaterwhich I believe suggests that we *do* need a few ticks before issuing a command and checking its result
17:02:41GodEaterthoughts ?
17:03:01Nico_Pwhat do -11 and -32 mean?
17:03:09 Join karashata [0] (
17:05:26***Saving seen data "./dancer.seen"
17:06:58GodEaterit's the return codes from various routines all being added together
17:07:23 Quit stewball`sleep (Read error: 110 (Connection timed out))
17:07:30GodEater-30 something suggests something in check_registers went awry
17:07:47GodEaterwhich is the next call but one to master_slave_check()
17:08:32GodEaterand the one inbetween is only ever called if hard_reset is true, which I suspect it's not in this case
17:12:03GodEaterI'm wondering how to get short pauses inbetween SET_REG calls and reading the results without completely hacking ata.c to bits
17:12:04 Join Aevum [0] (
17:13:31GodEateror for some ATA genius to offer alternative solutions ;)
17:16:41*GodEater wonders where everyone has gone
17:17:05 Join OlivierBorowski [0] (
17:17:13Buschelreal life? ;)
17:17:21Nico_Pno ATA genius here maybe :)
17:17:40GodEateramiconn was here earlier - he qualifies :)
17:19:18Horscht16105 and up has the buffering issue on ipod 5G, 16099 and 16095 are not much better with buffering. 16105 requires ~2:30 minutes to buffer a playlist. 16095 and 9 require ~1:45. all the builds pre 16000 require only ~1:13 to buffer the same playlist
17:19:35Horschti'll run a few more
17:20:43Buschelhopefully someone who knows what to do will read your results and take some action
17:20:50GodEaterHorscht: is that at first powerup, with dircache updating ?
17:21:42soapok, work got canceled for the day. Buschel - anything you need?
17:22:07GodEaterNico_P: looks like something similar to what I want is done for the PP5002
17:22:09Horschtfirst powerup, entering the library, waiting for the HD to shut down, then queue all songs of a certain artist (21 files 140MB or something) and waiting for the HD to shut down again, GodEater
17:22:29Horschttime stopped from queuin to stopping
17:23:29 Quit Insectoid (Read error: 104 (Connection reset by peer))
17:24:35 Join MethoS- [0] (
17:24:58GodEateranyone know what IDE_CFG_STATUS does on the PP5002 ?
17:24:59Nico_PHorscht: thanks for the info... looking up
17:25:26HorschtI'll do more testing, to narrow the rev down at which point buffering issues started to arrise
17:27:13 Quit kugel ("ChatZilla 0.9.80 [Firefox]")
17:27:14Buschelsoap: nothing for now, just hoping the ipod color build works now. as soon as the results are posted I'll post the condensed version.
17:27:29 Join MethoS_mobile [0] (
17:28:42BuschelHorscht: is the buffering time of current svn about ~1:45 or still at ~2:30?
17:28:59Horschtin between
17:29:08Horschtmore towards 2:30, though
17:29:14 Quit bumbl (Read error: 104 (Connection reset by peer))
17:29:23Buschelhmm, so jhMikeS' bugfix didn't work
17:29:39Buschel(rev 16197)
17:31:49 Join Insectoid [0] (
17:31:59lostlogic /win6
17:32:03 Nick Klevi is now known as Klevi_Dj (
17:34:13GodEaterNico_P: I've done a filthy hack to see if I can get the ATA checks to pass :)
17:34:22amiconn2:30 buffering time??
17:34:35GodEaterNico_P: which didn't work :(
17:34:55Buschelhorscht: it's interesting to see what impact the duration of active HDD has. doubled HDD-acitivity in the 5G equals ~5-6mA. maybe it would be interesting to check for reading speed to _lower_ the duration of HDD-activity?...
17:35:03*amiconn gets ~1:15 buffering time (i.e. slooow) on G5.5/80 with peakmeters in wps, and 0:30 without peakmeters
17:35:44amiconnThat's with current svn, and without aa or bitmapped wps of course
17:36:55 Join fehmicans [0] (n=canavar@
17:37:21Horschttill the hd shuts off again
17:37:47Nico_Pamiconn: have you given the idea of sending me LinusN's video some thought?
17:37:53 Join r4v5 [0] (
17:38:03amiconnThat'd have to be decided by LinusN
17:38:16 Join frogskank [0] (
17:38:36Nico_Pobviously, but it also depends on whether you still have work to do on it
17:39:01r4v5Hey, does audio being recorded get passed through the processor on the way to the headphones?
17:39:25r4v5i.e. would it be possible to put in a hook to allow for filtering of the recorded input pre-monitoring
17:40:47preglowon the way to the headphones, that depends
17:40:51preglowi don't think it does on coldfire
17:41:00preglowbut that can be changed, why would you want this, btw?
17:41:17preglowand filtered how?
17:41:23r4v5to allow for filtering of mic noises and/or eqing
17:41:31preglowwhy only for monitoring?
17:41:49r4v5well i could do it to the recorded file but it'd be useful to have it in real time as well.
17:42:06preglowwell, depending on what you're recording to, there's not much cpu left for processing
17:42:14preglowwav is pretty much all that'd allow much dsp to happen at all
17:42:18r4v5ah, that's true.
17:42:27 Quit petur ("*plop*")
17:43:09preglowwell, unless you're on gigabeat or anything fancy
17:43:18r4v5nah, sansa.
17:43:39preglowwell, in theory, that does have a whole extra core we can use for dsp, but that'd need much work
17:44:12r4v5it does?
17:44:15pixelmabut it's not a good recording device anyways
17:44:19preglowbut yeah, i've been toying around with dsp recording options, but i don't do a lot of recording, so doubt i'll work much on it
17:44:30 Quit agm3nt ()
17:44:38r4v5you're going to make me look through hardware specs and the like instead of paying attention in school today.
17:44:49preglowsansa is dual core, yes
17:45:01preglowi don't think the second core does much when recording
17:45:45GodEaterNico_P: reading from the ATA_STATUS register, then waiting, then reading again seems to make a difference
17:46:55r4v5pixelma: how do you mean? the noise level and such?
17:47:44LambdaCalculus37r4v5: Yes, that's one of the reasons.
17:48:22LambdaCalculus37I have a c240 that I bought for voice dictation. I get a hiss whenever I try recording, and adjusting my settings isn't making much of a difference.
17:49:07pixelmaabove all I mean that you can only record at 22.05 kHz (jhMikeS once said 32kHz *might* be possible)
17:50:32r4v5LambdaCalculus37: a hiss or digital crosstalk?
17:51:03Horschtok, update. 1:45 is also the time stock pre 16000 builds take to buffer music. So the issues seems to have been introduced between 16095 and 16105
17:52:11Horschtnote: the actuall buffering of music takes 20seconds vs 1:07minutes respectively
17:52:29Horschtjust the HD still takes some time to eventually power off again
17:53:05LambdaCalculus37r4v5: A hiss.
18:00:58GodEaterNico_P: you still here ?
18:01:34GodEater <−− this is the small change I've made to the master_slave_check()
18:01:48GodEaterwhich for whatever reason causes it to return 0, and move on to the next check
18:02:08GodEaterand therefore return an ATA: -32 error
18:02:12GodEaterinstead of the -11 one
18:02:50GodEaterbut for whatever reason, if I try and move that code into the #define for SET_REG (so it happens all the way through ata.c), it fails to work
18:03:34GodEaterand then I get the -11 error again
18:03:57GodEaterI'm going to have to go home shortly
18:04:07GodEaterso if you feel like playing with it more
18:04:14GodEaterthat's what I've been trying
18:04:43 Join kugel [0] (n=kugel@unaffiliated/kugel)
18:04:47Nico_PI might do a bit of playing, yes
18:05:01GodEatergood luck :)
18:05:23pixelmaLambdaCalculus37: do you have a microSD with music and your c200 at hand?
18:07:22LambdaCalculus37pixelma: Yeah, in my pack. Let me get it out.
18:08:08LambdaCalculus37Okay, got it out.
18:09:36pixelmacould you try creating a root.m3u8 with plugged in microSD (by chosing Playlists > Create Playlist from the menu) and check if the m3u8 contains files for both memories? For some reason I only get a playlist with files from the card (and not even all if I'm not overlooking something)
18:09:58 Join DerPapst [0] (
18:10:13DerPapstmorning all :-)
18:11:02LambdaCalculus37pixelma: root.m3u8 contains everything on my Sansa, both card and internal memory.
18:11:19pixelmahmm... wonder what's up here :\
18:11:33LambdaCalculus37Yours isn't?
18:11:36pixelmabut thanks for testing
18:11:50LambdaCalculus37pixelma: No problem. :)
18:12:35Klevi_Djpixelma: Mine also takes songs from the MSD card and the internal memory
18:12:45Klevi_Dj(as for a playlist)
18:13:00LambdaCalculus37Klevi_Dj: c240 or c250? Just curious.
18:13:25LambdaCalculus37pixelma: c250, right?
18:13:44Klevi_DjAnything as to why the USB connection is rather.. annoying to get to work
18:13:49pixelmaok, it's a bit weird because database and stats don't care and use everything off of both memories
18:14:02Klevi_DjI have to remove my MSD card and turn off the player before it works
18:14:20pixelmaKlevi_Dj: same here
18:14:29Klevi_Djdoing it any other way makes it either not USB connect (only charge) or freeze
18:14:30LambdaCalculus37And here too.
18:15:19LambdaCalculus37Klevi_Dj: It works if you boot into the OF first, then insert the cable. That way you don't have to remove the microSD card first.
18:15:33pixelmachkdsk doesn't find any problems neither on the card nor the internal memory
18:16:43Buschelsoap: new patch out for 4G grey/color/photo, mini 1G/2G
18:20:33 Quit perrikwp (Read error: 110 (Connection timed out))
18:21:28Klevi_DjRefresh Database freezes my player
18:21:39Klevi_Djwhich, after booting the OF for some reason always does
18:23:43soapon it
18:23:54 Part pondlife
18:24:05Buschelsoap: we're getting closer to the end :)
18:27:13Horschtdefine "the end" ;)
18:27:42 Quit joshin (
18:27:42 Quit Nimdae (
18:29:24amiconnNico_P: Could you remind me what happened to your gigabeat?
18:30:25soapso from your testing, Horscht 16064 should be fine as far as buffering / disk thrashing is concerned?
18:31:14Nico_Pamiconn: I accidentally dented the HD
18:31:16NJoinjoshin [0] (n=joshin@VDSL-130-13-144-158.PHNX.QWEST.NET)
18:31:35Horscht16099 is allright. buffers 20 seconds, shuts HD down after 1:45, soap
18:31:55Buschelhorscht: #define end as_no_feedback_for_mini1G2G_is_available_only_PCF_config_is_of_interest
18:32:07Horscht16105 and up buffer 1:07-ish and shut down after 2:30
18:32:13Horschthehe, Buschel
18:32:26Horschti don't understand code
18:33:07 Join bumbl [0] (
18:34:51pixelmaoi, LambdaCalculus37: found the culprit... I had a folder with files in a too deep folder structure on my microSD. Creating the root playlist seems to start with the card and then just stopped there without any notice...
18:34:54jhMikeSNico_P: so some gigabeats don't like the mcr instruction in tick_start? Mine won't start the ticks without it.
18:36:32amiconnHorscht: Huh? I don't really understand the times you're posting...
18:36:35bumblwhere could I look up the progress of the iaudio7 port?
18:36:40Nico_PjhMikeS: that line freezes GodEater's, aliask's and my S
18:36:43amiconnThe disk should shut down immediately after buffering....
18:37:08amiconn(immediately == stop command is sent within the same second)
18:37:15Nico_PjhMikeS: I'm not sure the ticks start though
18:37:20jhMikeSNico_P: odd. unless I do that on mine, the timer never starts
18:37:28jhMikeSNico_P: if it gets that far, they do
18:37:45Horschtwell, the hd shows 20 seconds of intense usage (you can hear the HD sounds as know from PCs)
18:37:50Nico_Pyeah well I rech ATA error -11 too
18:37:52jhMikeSsince sleep() calls are made they'd have to or they'd never return
18:38:14Nico_PjhMikeS: and is it possible to use sleep() in the BL?
18:38:27Horschtafter that, you can only hear the disc spinning, and every once in a while some disc activity, (split second)
18:38:38jhMikeSNico_P: It could be
18:39:16jhMikeSI suppose I only need the AVIC turned on there and to copy the vectors in case of data aborts, etc.
18:39:38LambdaCalculus37pixelma: Ah, that would be it. My card just has a few folders no deeper than 2 levels.
18:39:51Nico_PjhMikeS: but basically you managed to get interrupts working, didn't you?
18:40:07pixelmaLambdaCalculus37: well yes, removing it "fixed" the problem :)
18:40:12Nico_PjhMikeS: well congratulations :)
18:40:13 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
18:40:20jhMikeSgracia :)
18:40:24Nico_Pdid you figure out what ptw and aliask were missing?
18:40:37LambdaCalculus37bumbl: No one is working on the iAUDIO 7 at the moment.
18:41:03pixelmaLambdaCalculus37: not really true, there is one (TMM) working slowly
18:41:34*jhMikeS reads logs in more detail
18:41:50HorschtI am sorry if this is no help, soap
18:42:16soapchill - It's all good.
18:42:31jhMikeSNico_P: If aliask saw the ATA -11 then it's all getting as far as I do and we're all in the same spot
18:42:47*BigBambi is very excited by the gigabeast progress
18:42:56BigBambiAnd I can test with an S60 if required
18:42:57LambdaCalculus37pixelma: I do need to read the logs more often. :P
18:43:04bumblpixelma: thank you
18:43:21*LambdaCalculus37 is planning on getting a Gigabeat and may get an S to help test
18:43:22Nico_PjhMikeS: yes. I meant what they were missing all the time they've been trying to get interrupts working
18:43:22jhMikeSyou must replace both the firmware and bootloader of course
18:43:28LambdaCalculus37If I have the money, of course. :)
18:44:41 Quit snipexv ("Leaving.")
18:45:03jhMikeSNico_P: The timer set wasn't right and lines 33 and 34 of kernel-imx31.c I think might have been significant. I should undo them and see if they actually matters.
18:45:42jhMikeSNico_P: btw, which firmware image is everybody combining the bootloader with? I wonder if I'm using something else.
18:46:19 Part bumbl ("Running Gentoo Linux")
18:46:52Nico_PjhMikeS: MD5 is fb76b0025cd6bd2cc7512620ebf30b81
18:47:54*jhMikeS matches that
18:49:46jhMikeSI wonder how that mcr instruction could lock at that point but not in core_sleep :\
18:50:12 Quit csc` (Read error: 110 (Connection timed out))
18:50:37amiconnIt could if either the interrupts aren't enabled, or the timer isn't running
18:51:21amiconnI guess there are multiple places inbetween where either problem is fixed, so that core_sleep works
18:51:24jhMikeSon mine, the timer doesn't start unless I do that. on others it blocks there. since the sleep function works, the tick must be running
18:52:07amiconnWhy would the timer not start if you don't send the core to sleep, or to the contrary, if you do?
18:52:39*amiconn wonders whether the low-power mode is the *only* effect of that cp15 command, or whether it actually does more
18:52:46jhMikeSI've no idea nor why other get different behavior
18:52:57amiconnE.g. flushing cache, or some store buffer or similar
18:53:11jhMikeSthat command is specifically a "wait for interrupt" function of the coproccessor
18:53:32amiconnMaybe some i/o area isn't properly marked as uncacheable, unbuffered?
18:53:53amiconn(just outputting ideas)
18:54:19 Quit amiconn (" bbs")
18:54:26jhMikeSI wonder if clearing the source before enabling it in the vic might be the reason
18:55:14jhMikeSThe I/O areas should not be in cacheable areas
18:57:53 Join obo [0] (n=obo@rockbox/developer/obo)
18:58:01 Quit Insectoid (Read error: 110 (Connection timed out))
18:58:07 Join Insectoid [0] (
18:58:40 Quit idnar (Read error: 110 (Connection timed out))
18:58:46 Quit Klevi_Dj ("Sleep")
18:58:54 Join idnar [0] (i=mithrand@unaffiliated/idnar)
19:01:07jhMikeSaha...that theory seems to have panned out
19:01:48jhMikeSthe VIC seem to get out of sync
19:04:02lostlogicI have an idea that comes out of a discussion in -community and jonno's commit last evening. the introduction of a dev builds vs. user builds dichotomy wherein rather than the old daily builds system we provide some kind of sign off on commits to become user builds.
19:04:31lostlogicnot sure what the bar should be for signoff, but I do think it's not the best thing to have all of our users running off of whatever we throw at SVN −− it discourages the commit early adn commit often philosophy
19:04:39BigBambitesting and stable :)
19:05:27***Saving seen data "./dancer.seen"
19:06:07lostlogicI mean we've failed in the past to make releases because the bar was too high, but surely we can find a happy medium of simply promoting certain (most?) builds to user builds after the commits contained therein have seen some dev testing
19:06:17NJoinNimdae [0] (
19:07:02BigBambilostlogic: Who would have to agree to signing it off? All comitters? A core few? A majority? Consensus?
19:07:19lostlogicthat's the question −− setting the bar.
19:07:41lostlogicfor one thing it would have to have no red and no more yellow than the compiler bug level
19:08:10 Quit DogBoy (Read error: 113 (No route to host))
19:08:34lostlogicperhaps we could come up with a list of specific test actions to be performed on any new build on each major category of targets and once a committer has signed off on each one of those categories it's promoted
19:11:13rasherlostlogic: everyone who's committed since the last user build?
19:11:32rasher.. To agree to signing off a new user build
19:11:33PaulJamwouldn't that list become very large and time consuming?
19:12:45soapcouldn't "user builds" just be an arbitrary time length behind "dev builds" and if a SVN commit can't get major bugs resolved in said timeframe it gets pulled for more work?
19:13:51lostlogicsoap: yes, that sounds better −− a big red button instead of signoff.
19:14:06Nico_Pwhat about user votes?
19:14:15lostlogicso after JdGordon's commit the people discussing it would press the big red button before it made it to user build status.
19:14:26lostlogicvery good −− very toyota.
19:14:34lostlogicNico_P: no, users shouldn't see it :)
19:15:53Nico_Pmakes sense
19:15:56lostlogicsoap: and that is a lot lighter of a process, very practical. Bagder?
19:15:57Nico_Pbut I'd prefer a release
19:16:42lostlogicNico_P: hehe, I'm neutral on releases −− they add a lot of overhead I think
19:16:54soapyou can get into pissing matches re: "I'll push the button on anything attempting to do X"
19:17:09lostlogicsoap: are our committers that immature?
19:17:20soapsee the UN security council for veto powers gone mad.
19:17:29soapI said "can" not "will" ;)
19:18:11jhMikeSblah, here what's happening, the tick won't actually start counting until the first sleep() call for some reason
19:18:22Nico_PjhMikeS: your commit fixed things for me
19:18:26lostlogicwell I can tell you that Toyota and Amazon both have policies in place that allow any old worker who sees a problem moving toward customer delivery to _stop_ it and it is not abused
19:21:32jhMikeSI still don't know why or how the mcr in thread.c should pass but the one in tick_start wouldn't. :\
19:21:32Nico_PjhMikeS: using sleep() in the BL (right after kernel_init()) made it freeze
19:21:47jhMikeSNico_P there's no ticks started there
19:22:22jhMikeSI suppose they could be
19:22:33DerPapsthmmm... i get problems with the sim today... it's simply freezing.
19:23:53 Join GodEater_ [50] (n=bryan@rockbox/staff/GodEater)
19:24:04DerPapststarted somewhere between r16215 and r16225
19:24:29PaulJammaybe this is a silly question, but what would be the major difference between the idea of delaying the user builds and posting the work as a patch on the tracker for testing first?
19:24:36 Join Domonoky [0] (
19:24:51*GodEater_ would like to know the answer to that too
19:25:07lostlogicPaulJam: for better or worse (better, IMO) a lot of devs ignore the tracker and so don't see things even posted there until they go to SVN
19:25:30lostlogicthe red button and delayed user builds puts a definite time frame on "This _will_ go to the users unless there is an objection"
19:25:42lostlogicrather than "this is kinda in someone's queue for eventual commit"
19:26:12 Join SkaFan [0] (
19:26:31GodEater_I can see a lot of reverted commits happening if this road is taken
19:26:40 Quit soap ()
19:26:54SkaFanhi there, can anyone tell me there i can buy a Sandisk Sansa E260 v1 in germany?
19:27:38Horschtyour best bet would probably be ebay
19:28:20SkaFani have found only used players there... and no information about firmware version :(
19:28:49 Quit Insectoid ()
19:28:57GodEater_well you're only likely to find v1 players that are used - they're not produced any more
19:29:11BigBambiI imagine you would have better luck asking an ebay seller the firmware version than a shop
19:29:35 Quit hannesd ("Client suicide")
19:30:12knittli ordered a v2 yesterday :-$
19:30:19GodEater_lostlogic: the main problem I see with your proposal is that it doesn't take account the "build on commit" system we have going here, which most OSS projects I know of don't have
19:30:29 Join soap [50] (n=soap@rockbox/staff/soap)
19:30:29lostlogicGodEater_: doesn't revert commits −− just stops it from going to users until it's discussed and bug-fixed
19:31:25lostlogicGodEater_: what do you mean? That's a critical part of this system −− build-on-commit lets us test things immediately, and I see the red button perhaps just as a dev tool added to the dev builds page which turns a whole build line red and prevents it from showing up on the downloadable builds page.
19:32:34GodEater_lostlogic: sorry, I assumed you meant the auto-builds wouldn't happen for "dev" builds any more
19:32:40GodEater_crossed wires :)
19:33:50scorche|wlostlogic: on at least one thing, the whole "commit early, commit often" thing did pop up in my mind through all of this..
19:34:48GodEater_doesn't it still require some sort of development roadmap though ?
19:35:33GodEater_otherwise we'll end up with features a and b which are mutually exclusive being fought over right in the svn tree
19:36:02HorschtI think daily builds should be treated as such
19:36:18Horschtbeta, and possibly containing bugs
19:36:32BigBambiHorscht: That is what they currently are
19:37:05Horschtyeah, I don't see why one would want to/need to change that
19:37:21LambdaCalculus37Horscht: There's no good reason to change that.
19:37:35lostlogicI think the problem is t ha tdailiy builds are no more stable than any particular commit
19:37:41lostlogicso we need a way to choose which build gets promoted to that status\
19:37:46lostlogicother than just the last commit before the daily is pulled
19:37:56XavierGrjhMikeS: congrats on your work for the Gigabeast :) Too bad that I am in my exams period and can't help on the process :(
19:38:01lostlogicand that's really all I'm trying to get down to is a way to _choose_ which builds are promoted
19:38:47 Quit karashata (" HydraIRC -> <- The professional IRC Client :D")
19:38:59HorschtI know something. daily builds will be marked daily builds, whereas "recommended" and sorta-stable builds would be called.... dunno... post-beta or something
19:39:47Horschtnah, too abvious
19:39:56GodEater_also correct...
19:40:25jhMikeSXavierGr: Thanks. And good luck on the exams.
19:40:27SkaFan@knittl but rockboy wont run on a v2, or am i wrong?
19:40:36Horschtrockbox wont at all
19:40:57knittlthat's why i'm concerning returning it :(
19:41:07knittlhaven't recieved it yet
19:41:19Horschtthat's why you'll probably have to find a used one on ebay, SkaFan
19:41:33knittli should've ordered it before christmas
19:41:43SkaFani dont do ebay anymore!
19:41:57GodEater_SkaFan: then you probably won't Rockbox either...
19:42:09knittlamazon has used ones too
19:42:13jhMikeSNico_P: I think we need to do fancy stuff so that loading the firmware doesn't overwrite the bl vectors and the bl should be moved higher in memory since it's only at 2MB from DRAM start.
19:42:16Horschtwell, or whatever way of buying used devices you use, SkaFan
19:42:46 Nick roxfan2 is now known as roxfan (
19:42:51Horschtfact of the matter, your chances of getting a new v1 are close to 0
19:43:12Nico_PjhMikeS: what kind of fancy stuff?
19:43:23knittli could donate my v2 to a dev :) and in return he should give me a v1
19:43:43BigBambiyeah right
19:43:48GodEater_dream on
19:43:53BigBambiI don't think you will find any takers
19:44:00 Join DogBoy [0] (
19:44:46Horschtwhat a shame of them, to name the device exactly like the previous, but changing the whole hardware
19:45:30BigBambiWhy? It saves them money, and to the consumer looks the same.
19:45:37GodEater_and apple have done the same
19:45:42BigBambiMost people don't care what is in their DAP
19:45:43jhMikeSNico_P: I'm not perfectly sure but probably switch memory maps and/or have the bootloader copied to a higher address out of the way of any possible fw load.
19:45:51BigBambiIt is annoying for rockboxing of course
19:46:04Horschtapple have done the same?
19:46:14GodEater_of course
19:46:17Horschtwith the nanos?
19:46:25GodEater_and the new Classic
19:46:26rasherThe nano v2 is completely different from the previous ones, yet look quite like it
19:46:34Horschtthe new classic is called classic
19:46:42Horschtit's just from the ipod family
19:46:44GodEater_and yet people still confused it with the video
19:46:45Nico_PjhMikeS: aliask will most probably be much more helpful than me on this kind of stuff
19:46:47jhMikeSThe interrupts will work on the VIC but if there's a fault after loading the firmware it won't go to the correct handler
19:46:55Horschtbecause they don't get it, GodEater
19:47:03LambdaCalculus37GodEater_: It's all because it plays video, that they insist it's a video.
19:47:20Horschtit is clearly named classic, as opossed to Video
19:47:28BigBambiHorscht: The point is to most people, they don't care it is all the same to them, and to Sandisk they save money
19:47:29*jhMikeS is just rambling on about complications due to address conflicts
19:47:34GodEater_jhMikeS: is it not worth me delving further in to the ATA stuff till you get this figured out ?
19:47:42BigBambiFrom their viewpoint Win Win
19:47:50 Quit lids ("Reconnecting")
19:48:10HorschtI don't see how they "save money" by not naming it different
19:48:16jhMikeSGodEater_: go ahead on that one. this shouldn't have anything to do with that problem. it seemt to be failing in master_slave_detect
19:48:16BigBambiThey don't
19:48:25GodEater_Horscht: they don't have to re-advertise it
19:48:33BigBambiThey avoid confusing consumers by naming it the same
19:48:39GodEater_jhMikeS: it fails all the way through the ata.c
19:48:43BigBambiAnd they save money by using cheaper hardware
19:49:05GodEater_master_slave_detect is just the first call to show the symptom
19:49:06jhMikeSGodEater_: Error -11? Doesn't that come from master_slave_detect?
19:49:12LambdaCalculus37The average Joe consumer doesn't care WHAT is in the device, he just wants that iPod name.
19:49:22BigBambiOr Sansa in this case
19:49:26GodEater_jhMikeS: yes - but see the logs from earlier
19:49:32GodEater_when I discussed it with Ni
19:49:44 Join davina [0] (
19:49:50*jhMikeS missed that part
19:50:12GodEater_I managed to get it to pass check_master_slave
19:50:25GodEater_by putting in a sleep call before the read of the ATA_STATUS register
19:50:40GodEater_which then made ata fail with -32
19:51:03jhMikeShow long of a sleep? I tried one tick but never went farther.
19:51:31 Quit dionoea (Remote closed the connection)
19:51:43GodEater_well originally I used a sleep(500) :)
19:51:48GodEater_but I got it down to a sleep(5)
19:52:01knittlwhat language do you use? c?
19:52:04GodEater_I was trying to get SET_REG to use that all the time
19:52:17GodEater_but for some reason it didn't work if I moved the sleep() call in there
19:52:32GodEater_knittl: mostly C, some Assembly
19:52:50knittlok. thanks
19:52:56jhMikeSso check_registers fails.
19:53:17GodEater_jhMikeS: yes, but I think (as I say) this is just a symptom
19:53:20jhMikeSIt's perhaps the case that with the cache enabled, this processor just gets stuff done too quickly
19:53:26GodEater_that's what I think
19:53:36GodEater_it's reading the ata registers too fast
19:53:37 Join gevaerts [0] (
19:53:59GodEater_I was trying to think of a way of slowing it down (hence my games with the SET_REG #define)
19:54:04GodEater_but I couldn't make it work properly
19:54:07jhMikeSthe bootloader runs from physical uncached addresses atm
19:54:43 Join itcheg [0] (
19:54:45GodEater_so doesn't have the same problem accessing the disk you mean ?
19:56:03GodEater_I dragged some lcd code into ata.c, and output the ATA_STATUS register after it had been written to with SET_REG, and it came back as 0x0
19:56:18GodEater_which I assume is why master_slave_check was failing originally
19:57:46jhMikeSGodEater_: yes, the cache is enabled in the bootloader so the only difference should be the physical address execution.
19:58:00GodEater_I noticed we do some magic with SET_REG on the PP5002
19:58:09GodEater_so tried something similar with this
19:58:13GodEater_but couldn't make it fly
20:00:29 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
20:00:45 Quit SkaFan ()
20:05:41jhMikeSany reason memsize is defined as 32 instead of 64? Actually only 63 MB are available until the framebuffer can be moved.
20:06:07*GodEater_ looks blanks
20:08:07*jhMikeS wonders if a bunch of stuff was just simply left undone
20:08:31GodEater_very probably
20:09:34GodEater_I expect defining the memsize was way down the list while interrupts weren't working
20:09:43GodEater_we couldn't get enough code running to run into the limit :)
20:10:44 Quit scorche|w ("CGI:IRC (EOF)")
20:11:07 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
20:12:12jhMikeSI'll just toss the TTB into that 0x80100000 to 0x80200000 address range and use 63MB :)
20:12:44 Join smoking_gnu [0] (n=3e2fbd45@gateway/web/cgi-irc/
20:15:05smoking_gnuHas something like this ever happened to one of you?
20:15:20 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
20:18:33 Join smoking_gnu2 [0] (n=3e2fb5d7@gateway/web/cgi-irc/
20:19:28smoking_gnu2When I try to compile a self-written plugin, I get the following error message: "undefined reference to memcpy" although I don't even use memcpy() anywhere in my plugin!
20:19:36 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
20:20:11 Quit smoking_gnu ("CGI:IRC (Ping timeout)")
20:20:21jhMikeSgcc is probably using it internally
20:20:30smoking_gnu2When I include the MEM_FUNCTION_WRAPPERS it compiles, but my X5 just hangs when I start it (not even the first instruction is carried out)
20:21:18 Quit cagnulein ("Ex-Chat")
20:21:22 Join perrikwp [0] (n=98213f56@gateway/web/cgi-irc/
20:22:00jhMikeSthose need the api pointer to be initialized first. you could provide a trivial one or avoid whatever might required GCC to use it.
20:23:07smoking_gnu2They GET the api pointer: MEM_FUNCTION_WRAPPERS(rb)
20:23:42jhMikeSand the first code you have in plugin_start is "rb = api;" ?
20:23:45smoking_gnu2And I can't avoid anything since I don't know why the compiler even needs it - I don't use it anywhere..
20:24:18 Quit matsl (Read error: 110 (Connection timed out))
20:25:19 Join csc` [0] (n=csc@archlinux/user/csc)
20:25:59 Join matsl [0] (
20:29:29*jhMikeS doesn't know what he's thinking. the bl is starting at 32MB so no load conflict should be a problem. :p
20:30:22Buschelsoap: I should stop making patches... inluded the same typo as before in the patch you've made builds from... the 4G color/phote needs rebuild with an updated patch.
20:30:31*Buschel is hiding away...
20:31:51 Quit DerPapst (Read error: 104 (Connection reset by peer))
20:32:35jhMikeSGodEater_: you could set the dram start from 0x82000000 to 0x02000000 and see if ata_init become a problem because then the bootloader will run from cached DRAM after the remap.
20:32:47Domonokysmoking_gnu2: have you tried you plugin on the sim ?
20:33:06 Quit SacredTerror (Read error: 110 (Connection timed out))
20:33:08jhMikeSGodEater_: in
20:33:48GodEater_jhMikeS: I can't till tomorrow morning unfortunately - my dev environment at home is b0rked
20:33:58GodEater_have to wait till I'm back at work
20:34:06amiconnI wonder whether there are some ata timing regs which just need to be set up properly
20:34:36Nico_PjhMikeS: I'll test it
20:34:51amiconnThe PP5002 set_reg mechanism is more a kludge than a fix. I guess PP goofed there, and got it right in the PP5020 and higher
20:35:30jhMikeSamiconn: the bootloader manages to pass ata_init so why not the second time if not something due to faster execution?
20:36:03jhMikeSthe timings are different when the bl executes it maybe?
20:36:13amiconnThat still doesn't rule out that some timing regs might not be set up correctly
20:36:30 Join robin0800 [0] (
20:36:35amiconnThe first time, uncached execution just slows it down enough so that it doesn't break
20:36:55Nico_PjhMikeS: I just get a frozen windows mobile logo
20:37:10soapBuschel, on it
20:37:37amiconnThe set_reg delay loop became necessary only after I introduced the optimised transfer loop - with plain C transfer it worked without the delay
20:37:41 Quit JdGordon|w (Read error: 104 (Connection reset by peer))
20:37:49*amiconn just read up on lostlogic's proposal
20:38:43amiconnI don't really like that proposal, as it just shifts the problem, leading to less testers of new svn additions
20:39:14amiconnI think the problem is that adding or changing certain things in the core needs to be discussed
20:40:14amiconnSure, some devs don't check the tracker often (I admit I am among those). But the tracker is where discussion of patches should happen
20:40:49amiconnThere's one problem with testing patches though - not sure whether I'm the only one experiencing this...:
20:41:49jhMikeSNico_P: hmmm...double-checked the value 0x02000000 and not 0x020000000 or something :)
20:41:58amiconnOften, my tree is heavily patched with my own work, and I don't want to risk conflicts due to applying another patch. And a full build in a second working copy is time consuming....
20:42:34Nico_PjhMikeS: I'm pretty sure I got it right
20:43:03Nico_PI simply changed the 8 to 0
20:43:08 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
20:43:10*jhMikeS has a couple trees for 1) doing two projects at one 2) testing quick work or patches
20:43:14amiconnAnd that quickscreen commit was not exactly an example of 'commit early, commit often'. It even changed several things in a single commit
20:44:03 Quit robin0800 (" HydraIRC -> <- It'll be on slashdot one day...")
20:44:09amiconn(1) introducing viewports to quickscreen (2) adding a 4th option (3) adding configurability
20:44:29jhMikeSNico_P: that will need a looksee then
20:44:43amiconn(2) and (3) are things which should have been discussed beforehand, and only added after agreeing that they're wanted
20:44:59 Quit kugel (Remote closed the connection)
20:45:03bluebrotherjust from reading the commit message I didn't liked the 4th button.
20:45:13jhMikeSNico_P: wait, 0x82000000 physical should be 0x01000000 virtual since a 1MB section isn't mapped. :p
20:45:30bluebrothera configurable quick screen? Urgh.
20:45:45jhMikeSNico_P: wrong maths
20:46:16amiconnThe 4th button might make sense e.g. on iriver. (a) there's enough space (b) iriver can't use quickmode anyway (c) there's no button collision (like e.g. on ipods)
20:46:35jhMikeSNico_P: try 0x01f00000
20:46:51amiconnBut imho it should be a compile-time option. 3 buttons on recorders and ipods, 4 buttons on irivers, iaudios, gigabeat (?) ...
20:47:23amiconnThe recorders even have that secondary quickscreen, so a 3 button quickscreen is no loss
20:47:44Nico_PjhMikeS: no better
20:47:52amiconnAnd the configurability, hmmmmmrmmmm......
20:48:20*bluebrother wonders if he's still working on the manual in some way ...
20:48:35amiconnThere was some discussion (long ago) regarding a user configurable quickmenu, replacing the old quickscreens
20:48:41jhMikeSNico_P: thanks for checking
20:48:42pixelmaand if a configurable quickscreen at all, I think it needs some documentation at the same time. I admit I don't care about the quickscreen much (got used to not having one on the Ondio)
20:48:55amiconnOf course neither the main menu nor the buttons should be configurable
20:49:26bluebrotheryep, documentation should come the same time.
20:53:26 Join robin0800 [0] (
20:54:02 Quit K4rP4D ("Leaving")
20:55:08 Part robin0800
20:55:26 Join robin0800 [0] (
20:55:51*jhMikeS thinks we should just skip the first 64 bytes of the firmware image when loading it and jump past the vectors area instead of 0x0 and no vector conflicts will occur during load
20:56:03*n1s types into the wrong window...
20:58:17 Quit robin0800 (Client Quit)
20:58:22 Quit linuxstb (Read error: 113 (No route to host))
20:58:26 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:00:10 Join robin0800 [0] (
21:00:27 Quit robin0800 (Client Quit)
21:00:55 Join robin0800 [0] (
21:01:45 Quit robin0800 (Client Quit)
21:01:53 Quit GodEater_ ("There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence.")
21:05:29***Saving seen data "./dancer.seen"
21:06:06 Join robin0800 [0] (
21:06:42 Quit robin0800 (Client Quit)
21:07:17 Join robin0800 [0] (
21:08:02 Join Domonoky_ [0] (n=Domonoky@
21:08:21 Part robin0800
21:08:36 Quit mf0102 ("Verlassend")
21:08:39jhMikeSshoudn't len = filesize(fd) - 8; in load_firmware be len = filesize(fd) - FIRMWARE_OFFSET_FILE_DATA?. if not it will fail if that #define isn't always 8. :\
21:10:55 Join robin0800 [0] (
21:10:55 Quit DerDome ("Leaving.")
21:11:10 Quit robin0800 (Client Quit)
21:12:14 Join robin0800 [0] (
21:12:51 Quit desowin ()
21:12:51 Quit gevaerts ("Leaving")
21:19:37*jhMikeS wonders why te "based on iPL loader" comment never seems to go away
21:20:10 Quit simonrvn (Read error: 104 (Connection reset by peer))
21:20:16 Join simonrvn [0] (i=simon@unaffiliated/simonrvn)
21:20:35 Join TMM [0] (
21:23:40 Join gevaerts [0] (
21:24:59 Quit smoking_gnu2 ("CGI:IRC (EOF)")
21:25:42 Quit Domonoky (Read error: 110 (Connection timed out))
21:25:52n1samiconn, bluebrother: I am in favour of a configurable quickscreen as I don't use any of the options in the current one except setting file view to "all" if settings are reset
21:26:11n1swhich is maybe once every month or so...
21:26:53 Quit Buschel ()
21:27:08 Quit gevaerts (Client Quit)
21:29:10jhMikeSn1s: pretty much the same thing here re: usage
21:29:39n1sjhMikeS: great work on the 'beast :)
21:29:47BigBambiI turn shuffle on and off from time to time when listening to albums or not, but that is it
21:30:08BigBambinls: seconded, ace job jhMikeS
21:31:15bluebrotherwell, at least the guy who commits a customizable quick screen should also commit the manual documentation the same time.
21:32:28bluebrotherI see the quick screen becoming customizeable a problem as people will "see" the no-customizable-menus argument breaking
21:32:39BigBambithere is that
21:32:39*jhMikeS isn't even sure what setting he'd put on a quickscreen since he rarely changes any of them but for debugging
21:32:55BigBambiNo, I'd have to think about it :)
21:33:03BigBambiShuffle I guess + ....
21:34:19 Join DerDome [0] (
21:34:52*JdGordon|w has "folder advance" "shuffle" and "birghtness" set and will probably change shuffle later
21:35:55BigBambiI very rarely change any settings once they are there
21:38:32 Join knitt1 [0] (
21:39:17 Join Mathiasdm [0] (
21:45:12bluebrotherDomonoky_: what's the reason for this EncoderCache thingy? Is creating a new encoder object that slow?
21:46:05 Quit Mathiasdm ("Yuuw!")
21:46:10 Join tessarakt [0] (
21:48:54tessarakt"moin" is Low German for "good [day/evening/morning]"
21:49:30bluebrotherin some areas even "moin moin" ;-)
21:50:18bluebrotherI believe "moin" has even made it to the IRC lingo wiki page ;-)
21:52:03HorschtI always refrain from using "moin" in non-german channels
21:52:04scorche|wmoinmoin wiki? :)
21:52:55bluebrothermaybe we should switch the wiki software to moinmoin? ;)
21:53:05Horschtand, yes, it made it into teh wiki
21:54:36 Quit knittl (Read error: 110 (Connection timed out))
21:55:39 Nick knitt1 is now known as knittl (
21:56:29 Quit LambdaCalculus37 ("CGI:IRC")
21:58:12 Quit bluebrother ("leaving")
21:59:01Domonoky_bluebrother: this cache is because there isnt a parent for those encoder and tts objects, so they wont get automatically deleted, so the cache ensures we dont create the several times :-) *hacky solution*
22:00:08tessaraktscorche: moinmoin wiki sucks
22:00:46 Quit perrikwp ("CGI:IRC (EOF)")
22:01:35 Join linuxstb_ [0] (n=chatzill@rockbox/developer/linuxstb)
22:04:37 Join defukt [0] (n=WTFOMGBB@
22:05:53*amiconn will probably not change the default configuration of the quickscreen, at least not on recorder
22:06:24 Join Llorean [50] (
22:06:28amiconnThere are only 2 options I use occasionally, which are 'file view' and display flip (on F3 on recorders)
22:08:07*amiconn should also quick-check the original quickscreen layout
22:08:38amiconnI'm failry sure the icons were on the inside originally, and only got moved to the outside in the multi-screen transition
22:09:29JdGordon|wso we now have to make sure we dont change anything from 2 years ago with updates?
22:09:36amiconnHaving them inside the text makes adjusting to varying font size easier. Also, the bottom option originally used a single line, and with viewports, could do so again
22:10:06JdGordon|wusing a single line and scrolling is HORRIBLE compared to 2 line
22:10:21amiconnNo, I wanted to check that as I think it both looks better and is easier to handle (provided it uses viewports)
22:10:40amiconnI'm not proposing a single viewport
22:11:03amiconnJust have 2 viewports side-by-side
22:11:59*amiconn should probably draw an illustration of how he imagines quickscreens with viewports
22:15:13 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
22:23:09 Part Loki ("We All Should Wonder Quis custodiet ipsos custodes!")
22:30:43 Join muck [0] (n=4fd3f8c9@gateway/web/cgi-irc/
22:31:10 Join gong [0] (n=4fd3f8c9@gateway/web/cgi-irc/
22:31:11 Quit muck (Client Quit)
22:31:32gongNico_P: here?
22:31:42Nico_Pgong: yes
22:32:11gongNico_P: in your last commit, why do you declare the dot variable once more in the block?
22:32:30gongI mean the line after strcat
22:33:34 Quit knittl (Read error: 110 (Connection timed out))
22:34:08 Quit Domonoky_ (Read error: 104 (Connection reset by peer))
22:34:09Nico_Pgong: good catch, although it doesn't seem to be new to that commit
22:35:57gongNico_P: the var wasn't declared at the top before, that's why I asked. No harm though.
22:36:02 Quit ch4os_ ("Lost terminal")
22:37:17 Quit nicktastique ("Leaving")
22:41:22 Quit gong ("CGI:IRC (EOF)")
22:42:56 Join ging [0] (n=54bd4acf@gateway/web/cgi-irc/
22:43:33 Quit ging (Client Quit)
22:45:35 Join avh22 [0] (
22:46:09 Quit avh22 (Client Quit)
22:47:00 Quit n1s ()
22:47:13 Join axionix [0] (
22:52:52 Join crzyboyster [0] (n=6210f01d@gateway/web/cgi-irc/
22:53:40crzyboysterWhat are target id's?
22:53:46 Join perrikwp [0] (n=982165ed@gateway/web/cgi-irc/
22:54:14Bagderin what context?
22:55:32 Join Falco98 [0] (
22:56:20crzyboysterI'm using (or trying to use) the voicefont.exe app at the bottom of > and it requests a target id in order to make a voice file
22:56:42crzyboysterI need the nano one
22:56:57 Join midgey [0] (
22:57:00 Quit midgey (Read error: 104 (Connection reset by peer))
22:57:31 Join midgey [0] (
22:57:54Bagderthat's just a unique number selected for each specific port Rockbox runs on
22:58:15Bagderset in the configure script, outputted in the generated Makefile
22:58:32Bagderhuh huh?
22:58:54amiconnThat wiki article is outdated....
22:59:05crzyboysterSo the process will no longer work?
22:59:14 Quit perrikwp ("CGI:IRC (Ping timeout)")
22:59:18 Join perrikwp [0] (n=982165ed@gateway/web/cgi-irc/
22:59:30 Quit axionix_ (Read error: 110 (Connection timed out))
22:59:40amiconnVoicefont.exe is still used to build a voice file, but the version attached there is also outdated, and won't work for current rockbox
22:59:46amiconnToday, the build system handles calling voicefont etc
22:59:49scorche|wBagder: so how hard should i push to get on the site? seems quite handy
23:00:25amiconnAll you have to do is to configure a voice build, and then 'make voice'
23:00:32crzyboysterSo what would be the easiest way to make a voice file for a language today?
23:00:36amiconnOf course you need a suitable tts engine installed
23:00:42 Join karashata [0] (
23:00:55crzyboysterI have espeak with the tts language I need
23:01:40crzyboysterWhat more do I need? Can a voice file be generated without a build environment?
23:01:41 Quit jcollie ("Ex-Chat")
23:01:47Nico_Pscorche|w: looks nice indeed
23:02:58amiconncrzyboyster: No, you need a build environment. Although, if you only want to build voice files you don't need the crosscompilers
23:03:50amiconnJust be sure to select the respective simulator in addition to the voice build, then the host gcc (plus sdl) is sufficient
23:04:23amiconnAt least this *should* work - I didn't try it
23:05:15crzyboysterSo the steps at > aren't outdated and would work on cgywin? What other windows build environments are available?
23:05:31***Saving seen data "./dancer.seen"
23:06:11amiconnYes, this method is the current one
23:07:12crzyboysterWow this is a pain in the butt...
23:07:28 Quit davina ("GNU/Linux the free alternative to Windows")
23:07:46amiconnThe note below the next headline ("Using Windows") even says the old method no longer works. I agree that it should be made more visible...
23:08:47 Quit perrikwp ("CGI:IRC (Ping timeout)")
23:08:58crzyboysterDo all windows build environments take up 1gb+?
23:12:01 Quit karashata (" HydraIRC -> <- Wibbly Wobbly IRC")
23:12:16 Join perrikwp [0] (n=982165ed@gateway/web/cgi-irc/
23:12:52amiconnNo, although cygwin installation size can vary a lot depending on what modules you install
23:14:24 Quit fehmicans (Connection timed out)
23:15:34 Quit joshn_ (Read error: 113 (No route to host))
23:15:36*amiconn 's cygwin installation takes ~490 MB (without the crosscompilers), or ~4.1 GB (including crosscompilers and homedir with working copies, build dirs etc)
23:16:11amiconnThat cygwin installation includes a lot of stuff you wouldn't need, e.g. the cygwin X server
23:17:05crzyboysterSo cygwin is just like another windows program that you can run from within windows?
23:17:54crzyboysterAnd are coLinux or vmware any better?
23:18:57 Quit ompaul (Client Quit)
23:18:57amiconnVMware is faster for actual builds. You will need more disk space and ram than with cygwin thoug, and it's less integrated (you'll have to copy files between the VM and the host)
23:19:12scorche|wdepends what you mean by seemed like you wanted a smaller footprint before
23:19:21scorche|wvmware is *much* faster than cygwin
23:19:35amiconnAnd with linux (be it native, in a VM or coLinux) you cannot use SAPI for voice file generation
23:22:18crzyboysterI want a lightweight one that can do the job I want (voice file generation and *maybe* generating builds) and has anyone had experience with coLinux?
23:23:35 Join dionoea [0] (
23:24:48crzyboysterWhat modules in cygwin would I need for this?
23:24:49*amiconn thinks coLinux will behave similar to VMware
23:24:53*preglow thinks the current svn build system is ok, perhaps we want a bigger "DON'T USE THESE IF YOU'RE NOT READY FOR SOME SERIOUS TROUBLE" sign...
23:25:02amiconnMaybe with a little smaller memory footprint
23:25:36 Join ol_schoola_ [0] (
23:26:53 Quit ol_schoola (Read error: 104 (Connection reset by peer))
23:27:45crzyboysterWhat modules for cygwin will be necessary and can they be installed seperately later (after initial installation)?
23:28:10soapreally, the only advantage of coLinux, IMHO, is that it runs on hardware VMware doesn't.
23:28:57crzyboysterI have a somewhat older windows xp computer with 20gb hardrive and maybe 256 mb ram? Will this be ok for cygwin?
23:29:57scorche|wehhhh...not so much...i would go with coLinux at that point
23:32:08soapprocessor and speed?
23:32:49*JdGordon|w prefers coLinux, but its a pita to get setup
23:32:54 Join barrywardell [0] (n=barrywar@
23:33:07crzyboysterI don't know... how would I be able to find that information? And what modules will be necessary for just voice file creation and maybe build compiling on cygwin?
23:33:49crzyboysterNever mind, found that (
23:34:46crzyboysterWill it take over 2gb with just that minus the things that have parentheses after them?
23:35:14*linuxstb_ suggests just installing Linux...
23:35:45JdGordon|whey barrywardell, you have an osx box yeah?
23:35:59crzyboysterUnfortunately, I can't as I have no clue how it works. I will probably stick to cygwin. Thanks for all the info and support! See you all later
23:36:06 Quit crzyboyster ("CGI:IRC")
23:36:17JdGordon|wdid you have to do anything special to get sdl buyilds to compile without warnings?
23:36:37JdGordon|wthey work but i get messags along the line of "blaa.o has no symbols" ?
23:37:00barrywardellI don't remember anything like that, but haven't built a sim in a while
23:37:15barrywardellI'll try one now and see
23:37:20 Quit Rick ("I don't need to be here.")
23:37:23linuxstb_I don't get that either, but I do get lots of warnings about duplicate symbols (stdlib stuff).
23:38:44JdGordon|wranlib: file: /Users/jonathangordon/rockbox/e200-sim/librockbox.a(debug.o) has no symbols
23:38:46barrywardellI do get a "no symbols" message now
23:39:23 Quit matsl ("Leaving")
23:39:24barrywardellI don't think I got it in the past, but I've updated to leopard since then so maybe something has changed
23:40:05JdGordon|wBagder: would that line show up as a build warning?
23:40:30BagderI don't know without checking...
23:41:07linuxstb_JdGordon|w: You're planning on contributing an OSX build server?
23:41:53linuxstb_You don't get the duplicate symbols warnings?
23:42:30linuxstb_Or rather, multiple definitions - e.g. "/usr/bin/ld: warning multiple definitions of symbol _strtok_r"
23:42:38JdGordon|wlemme se... i canceled the build once i got that warning
23:42:51linuxstb_They happen when librockbox.a is being linked.
23:43:32JdGordon|wzzz.. soo much slower than my ain box :p
23:43:45 Quit DerDome ("Leaving.")
23:44:10JdGordon|wyeah, only get those no symbols warnings
23:44:50 Join webguest27 [0] (n=4c10ea95@gateway/web/cgi-irc/
23:45:28 Quit EspeonEefi (Read error: 110 (Connection timed out))
23:45:52 Join Rick [0] (
23:46:27 Quit jgarvey ("Leaving")
23:48:55linuxstb_Hmm, so we could simply not compile debug.c for the sim?
23:49:24JdGordon|wthats not the only one
23:49:30JdGordon|wI get about half a dozen of them
23:50:05linuxstb_Presumably all empty files...
23:50:07webguest27hi i have a question do you guys no of a converter i can use to join 2 or more videos together so i can watch them on rockbox?
23:50:53JdGordon|wyeah, we shuold get out of the habbit of having #if something around the whole .c instead of using sources
23:51:26linuxstb_IIRC there was the same problem with the bitmap libs - so there's now a dummy bmp file to ensure none are empty.
23:52:33barrywardellyeah, I was just thinking about that
23:52:57barrywardellthe linker on osx is a bit weird sometimes
23:54:06barrywardellI committed FS #5767 to fix the problem
23:54:26Lloreanwebguest27: As far as I know, it's legal to just concatenate MPEG2 videos. If they're in the Rockbox format, you just need to join them end to end, assuming Rockbox handles that properly
23:54:36 Quit perrikwp ("CGI:IRC (Ping timeout)")
23:54:54 Join perrikwp [0] (n=982165ed@gateway/web/cgi-irc/
23:56:57PaulJamdoes rockbox read the whole file in order to determine the length of the video? do i do that?
23:57:24 Quit MethoS- (Remote closed the connection)
23:57:40 Join DerDome [0] (
23:58:44LloreanPaulJam: No clue, but if it doesn't support concatenated MPEG-2 files, it's not supporting the spec, and I consider that buggy. ;)

Previous day | Next day