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

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

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

#rockbox log for 2013-08-21

00:00:52jlbiasinipamaury: I solved my compilation problem on gigabeat-fx on g#569
00:00:54fs-bluebotGerrit review #569 at : [touch devices]Disable touch device on softlock/keyhold. by Jean-Louis Biasini (changes/69/569/13)
00:06:34jlbiasinikugel: I updated g#518
00:06:36fs-bluebotGerrit review #518 at : plugins/stopwatch: PLA integration by Jean-Louis Biasini (changes/18/518/5)
00:07:03gevaertsjlbiasini: the setting is now there, but if the new setting is not enabled, all buttons still work if hold is on
00:07:43 Quit Water (Quit: ChatZilla [Firefox 22.0/20130618035212])
00:07:49*gevaerts isn't convinced that the gigabeat wants this setting at all...
00:08:05gevaertsSure, it's technically a touchpad, but rockbox uses it as five buttons
00:10:34jlbiasinigevaerts: the problem has to do with it having a BUTTON_HOLD that need to be change if there is a touchpad
00:10:50jlbiasiniI will still work on that
00:11:13kugelis anyone else wonderign why it's all called touchdev_* instead of just touchpad_*?
00:11:47jlbiasinibecause we want to unify the way touscreen and touchpad get handled on hold
00:12:02gevaertsThat's silly IMHO
00:12:06jlbiasinieither softlocks or holdkeys
00:13:08kugeljlbiasini: don't add english strings to non-english langs
00:13:18kugelthis way they won't appear as untranslated
00:13:25gevaertsTo a user, there is *zero* difference between using a gigabeat F or any target that doesn't have these touch things
00:13:33jlbiasiniI was of the opinion keylocks should stay keyhold (ie disable touch on driver level, but pamaury insisted on having a setting on all touch devices either soft or hold
00:13:59jlbiasinikuel: ok thanks
00:14:24*gevaerts doesn't follow any of this
00:14:43jlbiasinigevearts actually there is: the gigabeat does have a sensitivity_setting
00:15:28gevaertsSo is that a reason to suddenly totally change how hold works on it?
00:15:35gevaertsBecause that's what you're proposing
00:15:43 Quit bertrik (Ping timeout: 256 seconds)
00:16:14jlbiasinino I didn't agree on that but I implement it because pamaury said we have to...
00:16:23gevaertsWhat your patch does right now on the gigabeat F (and probably the X) is adding a setting that makes hold work
00:16:29gevaertsIn other words, it *breaks* hold
00:16:46jlbiasiniI know I want to make it work
00:17:07gevaertsSo what *exactly* is supposed to happen?
00:17:27jlbiasiniActually it is meant to turn holdkey device in softkey device if they have touchpad or touchscreen
00:17:46gevaertsI have no idea what that means in practice
00:19:07jlbiasiniit means that unstead of locking touchpad at driver level (a holdkey has a physical buttun to lock key, that the driver can check directly and silent the touchpad if it's lock)
00:19:27jlbiasiniit should lock it in action.c
00:19:39gevaertsRight, so *nothing* should change for the user?
00:20:00gevaertsSo why the extra setting?
00:21:45jlbiasinithe whole point of changing that lock mecanism is to have a setting that let us choose whether we want to disable the touchpad or not while key are locked to avoid lcd refresh
00:22:05gevaerts*what* lcd refresh?
00:22:59gevaertsOn touchscreen devices I can see the lcd being related, and on softlock devices maybe as well, but how on earth is LCD related to this on devices with a physical hold switch?
00:23:47jlbiasinion gigabeat f there none because the touchpad get silent at driver level. We want it for the fuzeplus. But it was pamaury idea that for consistency reason it should be the same on all touchpad devices
00:24:05jlbiasinieither soft or hold
00:24:21gevaertsThat's just fundamentally *wrong*!
00:25:04gevaertsYou're proposing a setting that (I'm still not sure here) either does *nothing at all*, or *breaks* the hold feature
00:25:05jlbiasiniI won't fight for it because my opinion was to leave all hold target cutting their touchpad at driver level (and less work for me)
00:25:37gevaertsWhere you disable passing the events is *not* the issue
00:25:52jlbiasiniwell when fonctionnal it would just add the choice between disable or locked
00:26:19jlbiasinipamaury: could you help me here?
00:26:27gevaertsAnd what exactly is the difference between inactive buttons and inactive buttins supposed to be?
00:26:45kugeljlbiasini: there's something to consider: hardware hold buttons work anywhere
00:26:51kugelsofthold only in the wps and fms
00:27:38jlbiasiniright +1
00:28:57kugelhow do you unlock the fuze+?
00:29:40gevaertsAs far as I understand it, this setting on the gigabeat F means "Do you want to deactivate buttons by disabling generating the button events, or by ignoring the button events?"
00:30:28jlbiasininot ignoring them, returning keyhold message
00:31:19*gevaerts is very strongly opposed to this change
00:31:22kugeljlbiasini: what is the softlock button (if the touchpad is disabled)?
00:31:30jlbiasinithe main issue is on f+ that those message kill the battery down so we want a setting to kill that
00:31:42jlbiasinipower button
00:31:58gevaertsjlbiasini: those messages ARE NOT THERE on the gigabeat
00:33:17jlbiasinigevaerts: yeah its broken, because my implementation is wrong. I don't have any keyhold target so I cannot test what I do
00:33:39kugelgevaerts: it doesnt seem the patch actually affects the FX
00:33:58gevaertsIf you want to fix softlock, go ahead (I don't have a softlock device, so I don't know the issues there). If you need to change some infrastructure overall to achieve that cleanly, fine, do so. But I don't see ANY reason AT ALL to change *behaviour* on non-softlock devices
00:34:02gevaertskugel: it does
00:34:16kugelbecause it defines HAVE_BUTTON_HOLD and the new logic is within !defiend(HAVE_BUTTON_HOLD)
00:34:34gevaertsWell, that's possibly why it doesn't work as intended
00:34:40kugeloh wait, there's ||
00:34:50gevaertsBut that doesn't change the fact that "work as intended" means "broken" in this case
00:35:28kugeljlbiasini: I agree with gevaerts; a hardware lock button is superior to softlock and your patch means a regression
00:36:23kugelI suggest changing the #ifdefs so that it's not used if HAVE_BUTTON_HOLD is defined (so that you could enable the functionality on the FX by not defining it)
00:37:08pamaurysorry I made myself unclear, the setting is intended for softlock only, I must have been tired when I said so, sorry :( The setting is to choose the behaviour of the touchpad when softlock is on: should it light the screen and display you a nice "button is locked message", or should the touchpad be plain disable
00:37:48jlbiasiniok then let's cut the whole keyhold out
00:38:06pamauryso basically do you want the current behaviour or do you want a phone like behaviour (since virtually all phones do this now)
00:38:31jlbiasinithis is very complicated to explain to pehraps I got confused when I tried to explain the different case
00:39:27pamaurythe complicated comes from the fact that it should apply both to touchpad and touchscreen, at least that's my opinion
00:39:46jlbiasinipamaury: the setting is fine the think is just to have it on softlock and leave the keyhold target do want they do
00:40:10pamaurygevaerts: is that clear and okay to you ?
00:41:23gevaertsI think so
00:41:45jlbiasiniok so It's clear for me to
00:42:28gevaertspamaury: as I said, I don't know much about softlock and its issues, so I don't have any opinions there, but if physical hold switch devices are unaffected (at least in behaviour, if there are under-the-hood changes that's probably fine) I'll leave that to you entirely
00:43:32pamauryok, good; just so you know, the major annoyance of softlock is that each time to touch a touch button (ie very easily), screen lights up, this consumes a lot of battery and is very annoying
00:46:09*gevaerts nods
00:46:27 Quit lebellium (Quit: ChatZilla [Firefox 24.0/20130815180504])
00:46:48jlbiasiniI keep thinking that we should change anything to keyhold target. They cut the touchpad at driver level that's not meant for optimisation. If it was really anoying that could be a limitation to turn around. But the way they work is just fine. Actually I am of the opinion that we don't even need a setting for that. touchdev should be disable on lock IMO
00:47:21jlbiasini*I keep thinking that we should't* :D
00:47:45gevaertsI agree there's no need for a setting there
00:48:22gevaertsAnd yes, ideally the device should be disabled (i.e. powered off) on hold, but I don't know if all devices support that
00:49:17pamauryI agree it should be default behaviour but that would mean changing the current bahviour without a way to go back, some users might find it useful
00:49:52jlbiasinione of the idea of pamaury on the patch is to implement a weak symbol function that make the touchdevice silent and let us redefine it a driver level if we have some power management support
00:49:57pamaurygevaerts: even devices cannot be powered off can filter events in software, so there is no excuse to no implement it ^^
00:50:10pamauryjlbiasini: exactly
00:50:29jlbiasiniactually this is already in the patch
00:50:47gevaertspamaury: presumably that's what they do now...
00:50:54jlbiasiniwe just have to cut the keyhold part
00:51:01pamauryjlbiasini: thanks
00:51:11pamauryok I should go to bed, I'm too tired to say anything useful now
00:51:23*gevaerts too
00:52:09pamauryjlbiasini: I'll have some spare time tomorrow so don't hesitate to poke me when it's ready so I can commit it
00:52:11jlbiasinigevaerts: actually on the fuze+ we get 10% more batt by power down the touchpad
00:52:51jlbiasinipamaury: we cannot commit it before we test the generic touchdev_disable function on touschreen device
00:54:28jlbiasiniI see no reason why it won't works but It's usually a very good criterium for me to know that something actually won't work
00:55:11jlbiasiniAnd I'm going to bed too!! I'll will poke you tomorrow
00:55:34 Quit jlbiasini (Remote host closed the connection)
00:55:48 Quit pamaury (Read error: Connection reset by peer)
01:03:08 Quit ender` (Quit: Power comes from wallsockets -- Starwhite)
01:22:56 Join [Saint] [0] (~saint@rockbox/user/saint)
01:30:04[Saint]IS it too late to point out that softlock can be implemented on touchscreen devices with a touchscreen area?
01:30:34[Saint]I fell off the internet for a while, so I'm not sure if we're still doing this argument or not. :)
01:38:38***Saving seen data "./dancer.seen"
02:01:12 Join wowaname [0] (
02:01:44 Quit madcat1990 (Quit: Leaving)
02:12:28 Quit Raptors (Read error: Connection reset by peer)
02:18:12 Join Raptors [0] (
02:35:51 Quit wowaname (Read error: Connection reset by peer)
02:36:31 Join wowaname [0] (
02:50:55 Quit Raptors (Read error: Connection reset by peer)
02:56:56 Join Raptors [0] (
03:18:30 Join habys [0] (
03:27:45wowanameanyone online?
03:38:39***Saving seen data "./dancer.seen"
03:56:00 Join wow_a_name [0] (~wow@
03:56:08 Quit wowaname (Disconnected by services)
03:56:11 Nick wow_a_name is now known as wowaname (~wow@
04:04:31wowanameis it at all possible to display a progress bar under text?>
04:09:34 Quit kevku (Ping timeout: 245 seconds)
04:20:58JdGordon_wowaname: yes, but not easily
04:21:24wowanamehow hard would it be
04:29:24 Join kevku [0] (~kevku@2001:0:c38c:c38c:3456:2760:3d69:be87)
04:33:02 Quit amiconn (Disconnected by services)
04:33:03 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:33:07 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:33:23 Quit pixelma (Disconnected by services)
04:33:23 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:33:25 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:38:11 Join Substanzlos1 [0] (
04:40:57 Quit Substanzlos (Ping timeout: 276 seconds)
05:03:09 Quit kiwicam (Read error: Connection reset by peer)
05:03:25 Join kiwicam [0] (~quassel@
05:07:25 Quit wowaname (Ping timeout: 256 seconds)
05:18:59 Quit TheSeven (Disconnected by services)
05:19:08 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
05:38:43***Saving seen data "./dancer.seen"
06:05:41 Join Strife89 [0] (
06:18:44 Quit mrtux (Ping timeout: 245 seconds)
06:25:34 Join mrtux [0] (~colin@unaffiliated/mrtux)
06:47:07 Quit fs-bluebot (Ping timeout: 246 seconds)
06:48:47 Quit bluebrother (Ping timeout: 260 seconds)
06:49:56 Join fs-bluebot [0] (
06:50:24 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
06:54:31 Quit fs-bluebot (Ping timeout: 256 seconds)
07:02:22 Join fs-bluebot [0] (
07:04:59 Quit ps-auxw (Ping timeout: 245 seconds)
07:08:53 Join sciopat [0] (
07:11:49 Quit sciopath (Ping timeout: 240 seconds)
07:19:19 Nick SuperBrainAK is now known as DormantBrain (
07:27:01 Quit bluebrother (Ping timeout: 246 seconds)
07:27:29 Quit fs-bluebot (Ping timeout: 245 seconds)
07:28:11 Quit [Saint] (Remote host closed the connection)
07:29:20 Join [Saint] [0] (~saint@rockbox/user/saint)
07:30:19 Join fs-bluebot [0] (
07:34:22 Quit fs-bluebot (Ping timeout: 246 seconds)
07:38:44***Saving seen data "./dancer.seen"
07:39:22 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
07:42:09copper[Saint] "But in the above, if I remove the blanking, same thing...fat grey line in the middle of the art."
07:42:20copperDid you file a bug report for it?
07:42:40 Join fs-bluebot [0] (
08:00:04 Quit fs-bluebot (Ping timeout: 248 seconds)
08:00:57 Quit bluebrother (Disconnected by services)
08:01:03 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
08:03:36 Join fs-bluebot [0] (
08:04:20 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:05:24 Quit bluebrother (Ping timeout: 248 seconds)
08:08:18 Quit fs-bluebot (Ping timeout: 264 seconds)
08:08:27copperDo %Vl() viewports that never get loaded with %Vd() put any kind of strain on Rockbox? CPU / RAM / whatever?
08:15:58 Join fs-bluebot [0] (
08:16:27 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
08:20:13 Quit Scall (Ping timeout: 246 seconds)
08:25:17 Join Scall [0] (~chat@unaffiliated/scall)
08:29:54 Quit Scall (Ping timeout: 264 seconds)
08:31:27 Quit Maxdamantus (Ping timeout: 260 seconds)
08:34:16 Join Scall [0] (~chat@unaffiliated/scall)
08:35:53 Quit Provel (Read error: Connection reset by peer)
08:38:26 Quit kevku (Read error: Connection timed out)
08:39:08 Quit kiwicam (Remote host closed the connection)
08:41:42 Join Provel [0] (
08:44:18 Join kevku [0] (~kevku@2001:0:c38c:c38c:3456:2760:3d69:be87)
08:47:18 Quit Strife89 (Ping timeout: 264 seconds)
08:49:56 Join LinusN [0] (
08:54:16 Join jlbiasini [0] (
09:03:22 Join ender` [0] (
09:13:49 Join Maxdamantus [0] (~Maxdamant@2001:470:f078::dead:beef:cafe)
09:24:09 Join jlbiasini_ [0] (
09:28:05 Quit jlbiasini (Ping timeout: 256 seconds)
09:38:36 Join kiwicam [0] (~quassel@
09:38:46***Saving seen data "./dancer.seen"
09:52:27 Join maruk [0] (
10:04:34 Quit kevku (Ping timeout: 245 seconds)
10:05:00 Join rela [0] (~x@pdpc/supporter/active/rela)
10:11:52 Quit copper (*.net *.split)
10:11:52 Quit advcomp2019_ (*.net *.split)
10:11:52 Quit Mir (*.net *.split)
10:11:52 Quit evilnick (*.net *.split)
10:12:01 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
10:12:12 Join copper [0] (~copper@unaffiliated/copper)
10:12:17 Join advcomp2019_ [0] (
10:12:31 Quit advcomp2019_ (Changing host)
10:12:31 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
10:13:58 Join Mir [0] (
10:14:18 Quit rela (Ping timeout: 263 seconds)
10:19:01 Join kevku [0] (~kevku@2001:0:c38c:c38c:28b5:4a2e:3d69:be29)
10:25:45 Join einhirn [0] (
10:27:54 Quit kevku (Ping timeout: 245 seconds)
10:41:12 Join kevku [0] (~kevku@2001:0:c38c:c38c:3877:f7c3:3d69:bebc)
10:44:18 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:49:38 Join petur [0] (~petur@rockbox/developer/petur)
10:58:20 Join lebellium [0] (
11:33:21 Quit rdn (Remote host closed the connection)
11:35:37 Quit fs-bluebot (Ping timeout: 264 seconds)
11:36:33 Quit bluebrother (Ping timeout: 242 seconds)
11:38:50***Saving seen data "./dancer.seen"
11:40:21 Join fs-bluebot [0] (
11:44:34 Quit fs-bluebot (Ping timeout: 245 seconds)
11:48:51GodEaterthey still exist as a structure in memory, so yes
11:52:45 Join fs-bluebot [0] (
11:53:47 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:54:42 Quit petur (Quit: Nettalk6 -
11:56:22 Quit Scall (Read error: Operation timed out)
11:58:47 Join Scall [0] (~chat@unaffiliated/scall)
12:33:30 Quit kevku (Quit: KVIrc 4.3.1 Aria
13:38:52***Saving seen data "./dancer.seen"
13:39:36 Quit aevin (Remote host closed the connection)
14:06:52 Join petur [0] (~petur@rockbox/developer/petur)
14:41:09 Join amayer [0] (
14:54:28 Quit Raptors (Ping timeout: 248 seconds)
14:58:01 Join Raptors [0] (
15:03:22 Join crose [0] (
15:15:34 Join Rower [0] (
15:15:42copperI don't remember who was working on the theme site
15:16:07copperit occurs to me that the theme pages don't indicate what target the themes are for
15:17:07copperlebellium: I see that you chose not to display the album title on some targets, with the cover artwork taking all the space
15:17:28copperI find it rather strange, since you were vocal about displaying important information
15:18:20copper(I'm browsing the theme site to get new ideas)
15:19:13 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
15:21:15lebelliumsome targets? only one. The Philips SA9200 that nobody uses :) And the album name displays. It's like 3s artist name then 3s album name (or something like that, I don't remember). When the YP-Z5 port is ready (same screen resolution), I'll rework this theme. But on portait themes I want full album art so sometimes it's a bit tricky
15:22:25copperyeah, full album art is tempting
15:23:17lebelliumit really depends on the screen ratio
15:23:22lebelliumfor 240x320 no problem
15:23:36lebelliumbut on 176x220 (E200) I had to cut off the album art
15:23:43lebelliumthe same for 128x160
15:23:58copperdo you actually use those DAPs?
15:24:16lebelliumwhich ones?
15:24:21copperthe tiny ones :D
15:24:25amayercopper: I am probably the one you are reffering to that is working on the themesite. ill add it to my notes about putting the target on searches where the target isnt specified
15:24:32copperamayer: ah, yes
15:25:24copperamayer: also, would you or anyone be opposed to allowing "meta themes" such as mine (i.e. one theme with two or more variants, that share graphical elements)?
15:25:39copperi.e. one zip file with all variants
15:25:50lebelliumI never use the Philips SA9200 because of the touchpad. That's why I won't improve the theme until Rockbox is usable on YP-Z5. I sometimes use the E200 and C200. I never use Fuze+ and I use daily Clip+ and Clip Zip
15:26:41amayerI was thinking about that but I would say one "theme profile" for all the different colors. not all of them in one zip file
15:26:57amayerthat would break the theme downloader in rockboxutil
15:27:19amayerthey all need to be in seperate zip files but i was thinking about putting them all on the same page
15:32:11copperah, yes, I see what you mean
15:32:33coppermostly my problem is that updating 12 themes at once is a drag
15:33:03copperespecially since the uploader insists on asking for confirmation while losing the files that I uploaded via the first form
15:33:16amayercopper: for example you would make a theme page "googley classic" under ipod video and upload all 6 varients to that page
15:33:17copperit's a lot of clicking and fiddling
15:33:32amayercopper: that is one of the main things i am going to fix
15:33:36copperyes that would be nice, especially if I can upload everything at once
15:33:50copperI mean the 6 zip files and screenshot and all
15:34:32copperamayer: maybe you could allow for uploading a second zip file with only screenshots in it, with a standard nomenclature, like wps.png, sbs.png, fms.png, other01.png, other02.png, etc
15:35:01copperso that I could just zip the screenshots and upload a single file for them
15:36:27copperI'm able to script some repetitive tasks on my end, when dealing with the 6 variants
15:38:05amayercopper: thats an interesting idea
15:38:20 Join kevku [0] (~kevku@2001:470:27:773:0:feed:c0f:fee)
15:38:56***Saving seen data "./dancer.seen"
15:39:12copperamayer: I think HTML5 allows for easy uploading of several files at once (see Google, Facebook, etc) but the zip file is an easier solution on your end, I think
15:39:51amayerzip would be capatable with more users browsers
15:40:09amayerif i would code up something in html5 i would also need to code up a fallback
15:40:47copperI have no idea what most people use these days
15:41:16amayerassume: IE 7
15:41:38copperis that the stock browser in Windows Vista or something?
15:41:52Tornemost people are on IE8+ by now
15:42:56amayerTorne: I guess my clients are just behind the times. when ever i dont check my html/css in IE 7 they call and complain because "something doenst look right" but i would guess most of the world is on IE8+
15:43:41Tornewell under 10% of IE users are on IE7
15:43:56Tornethere's 2-3x as many IE6 users :)
15:44:04copperyou only need 1% to complain to make your life miserable ;)
15:44:05Tornebut mostly in china
15:44:21Torneif you don't care about china then IE8 is mostly okay as a minimum
15:44:34Torneif you want it to work in china then you get to code for IE6 ;)
15:50:01amayerwell ill make it compatable with IE 8 then. Ill have to check it at work where i can remote into a windows box. (hard to check in IE when you run 100% linux)
15:50:24amayerbut there is a windows box in a closet at my work that i can remote into
15:50:42Torne7 vs 8 doesn't make a lot of difference for compatibility generally anyway; unfortunately lots of stuff is still broken until 9
15:50:52Torne(or 10)
15:50:54Torne(..or 11)
15:57:44copperis there a keybinding for locking the iPod Video sim?
16:00:41coppersimulating the lock switch
16:01:09coppercapslock sounds like an appropriate key for that
16:02:01 Quit Zagor (Quit: Clint excited)
16:12:46[Saint]Normally I would've pointed to the source, but I guess you wouldn't really appreciate that.
16:13:40[Saint]copper: ^
16:15:56copperof course!
16:16:31[Saint]I don't blame you, eh.
16:16:42copperamayer: also, you could add a download link on the theme page, that's labelled as such (it's not clear at all that one must click on the screenshots)
16:16:45[Saint]I did the same thing 4~5 years ago.
16:16:53copperdid what?
16:17:09amayercopper: I already have that in my notes
16:17:28[Saint]When I looked at the source and saw it was "H" I was just like..."huh, dammit, makes sense...H for Hold". Whoops.
16:18:04copperit doesn't actually lock anything in the sim though
16:18:22[Saint]I had spammed the keyboard back and forth a few times but managed not to hit it, or not notice which key it was, and had to go looking.
16:18:23copperhmmm it does
16:18:25copperjust not the volume
16:18:47[Saint]Hmmm. odd. That /could/ be implemented.
16:18:58[Saint]The sim isn't 100% HW accurate.
16:19:16copperI saw that another theme used the hold switch / button to display alternative information, I figured that makes more sense than "pause"
16:19:28coppersince pausing disrupts playback (obviously)
16:19:37[Saint]Not really.
16:19:45[Saint]Most people have the screen sleep on hold.
16:19:53[Saint]Wasted battery.
16:20:08copperthey can make it go to sleep after 5 seconds
16:20:11[Saint]Pause is way more likely to have the LCD active, IMO.
16:20:15copperlong enough to see the alternative information
16:20:35[Saint]I dunno, I set mine to sleep the LCD immediately on hold.
16:20:40copperyes me too
16:20:57[Saint]hold == I care not for displays :)
16:20:59copperbut that was before I used a mechanism to display alternate information
16:21:06copperalternate / alternative, I don't know
16:21:24copperfrenglish phail
16:21:33[Saint]both work.
16:21:39coppergood to know
16:21:40[Saint]English IS Hard(TM)
16:21:48copperYeah? Try French.
16:22:52copper[Saint]: I suspect users would rather use the hold switch than the pause button, if they're going to use that functionality at all
16:23:12copperand if they don't see it, who cares
16:23:18copperit's not that important
16:23:39[Saint]That's why I <3 touchscreen themes.
16:23:52[Saint]Its so easy to fire an event that does whatever the hell you want it to.
16:25:06[Saint]Toggle display layouts with a touch area? Done. Switch fonts on the fly? Done. You can do some awesome stuff.
16:25:42[Saint]JdGordon_ worked quite tirelessly to provide me with the mechanics needed to do this.
16:27:10[Saint]the key is being able to increment, decrement, set, or query variables from a touch area.
16:29:32[Saint]copper: are your Googly_* themes touch capable?
16:30:16[Saint](bars will automagically generate their own touch areas unless you specify 'notouch', which is a pretty cool feature)
16:37:01copper<−−- young padawan learner
16:38:13[Saint]the easiest way to think about touch areas is with them being like magic blank viewports.
16:38:26[Saint]they work in almost exactly the same way.
16:38:41copperI'm not really interested in that tbh, since I don't have a touch target
16:39:02copperwell I have a smartphone, but I don't use it for music
16:39:13[Saint]the sdl app is pretty cool.
16:39:22[Saint]I use that as a desktop player sometimes.
16:39:34[Saint]themeable desktop rockbox == win.
16:40:46[Saint]its hardly designed well for a desktop player, though. but still kinda cool since you can make it any arbitrary size.
16:48:51 Join rela [0] (~x@pdpc/supporter/active/rela)
16:50:53 Quit jlbiasini_ (Ping timeout: 276 seconds)
17:04:50 Quit bluebrother (Read error: Operation timed out)
17:08:17 Quit rela (Ping timeout: 264 seconds)
17:08:55 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
17:13:13 Join n1s [0] (
17:13:13 Quit n1s (Changing host)
17:13:13 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:26:41 Join petur2 [0] (
17:27:06 Join westwind [0] (~zephyr@
17:28:11 Join jlbiasini_ [0] (
17:29:40 Quit petur (Ping timeout: 248 seconds)
17:31:07 Nick petur2 is now known as petur (
17:31:20 Quit petur (Changing host)
17:31:20 Join petur [0] (~petur@rockbox/developer/petur)
17:37:24westwindI work at a semiconductor company, we designed SoC has a dedicated processor to handle audio decoding and post-processing with 8K+8K caches and 8k+8k TCM. my question is that should I remove the TCM and increase the caches to 16K+16K in next-generation? is it better for audio related processing/
17:37:57Tornedepends :)
17:38:57***Saving seen data "./dancer.seen"
17:38:58Tornethere isn't really a good answer, i don't think. we make use of fast memory when it's available (though we are generally thinking of just internal SRAM rather than TCM), but i'm far from sure that an actual cache wouldn't be better
17:39:22Tornei dunno how much we stuff in iram, either, i would expect more than 8kb
17:39:36Tornesince the on-chip SRAMs we're dealing with are likely much bigger (though not as fast)
17:39:54Tornecache is certainly *easier*
17:40:13Torneeverything benefits from it, without ahving to go and carefully make smart choices about what to put in your TCM for every codec/dsp effect/etc
17:40:53n1scache is a lot easier to use, we tent to stuff as much as we can into the sram as long as it's speed critical. it's abit messy since the requirements of codecs may change based on input and different platforms have varying amounts
17:41:25Tornebeing able to select TCM or cache, and/or lock down cache lines, might be ideal, of course, if your architecture allows :)
17:43:16westwindTorne, We have a software based audio pipeline and it supports for at least 15+ audio format. big instruction size will cause the cache miss heavy. so I have to consider to increase the cache size. I think the caches could be locked down when I really need TCM to optimize a specific audio decoder, am I right/
17:43:18 Join rela [0] (~x@pdpc/supporter/active/rela)
17:43:37Torneif you have cacheline lockdown support then that's generally sufficient
17:44:01Torneif you really must never miss in your inner loop then prefetching and locking it is not really any harder than copying it to TCM
17:44:19Tornesince you aren't gonna be able to keep the same thing in the TCM allt he time anyway
17:44:43Tornenot that i'm an expert on dsp stuff
17:44:51Tornebut i know my way around embedded cache nonsnese ;)
17:45:08westwindI think the TCM or scratch pad is not easy to use.
17:45:24Torneyeah, TCM is a pain if you are using it for multiple purposes
17:45:30Tornesince you effectively are doing overlays then
17:45:37Tornelinking your code to run at different addresses etc
17:45:53Tornewhereas locked down cachelines just require the addition of the code to do the locking, and no other changes
17:46:26TorneTCM might still be faster at runtime though
17:46:39Tornedepends on the exact implementation
17:47:16Torne(when i say faster i mean, like, maybe it's a cycle or two faster) :)
17:47:19 Join saratoga [0] (4160ac54@gateway/web/freenode/ip.
17:47:36saratogaTCM means you use arm?
17:48:42westwindthe other question, do you think a dedicated audio DSP is necessary on a dual core cortex A9 SoC?
17:48:52westwindsaratoga, TCM or scratch-pad
17:49:17westwindsaratoga, Yes, I used to use ARM processor for audio decoding.
17:49:28 Quit kevku (Quit: KVIrc 4.3.1 Aria
17:49:30saratogano dsp is not needed on such a fast processor probably
17:49:30Torneaudio DSP seems likely to be superfluous on an A9
17:49:56Torneunless you are able to keep the A9 in suspend for a long time as a result of decoding on DSP
17:49:59saratogae do everything on maybe 50 mhz on armv4
17:50:34saratogaes if your a9 uses too much power adding a slower arm or dsp core could help
17:50:37Torneif you expect to do a lot of stuff that can be done on DSP and comparitively little that needs the APU there may be a power win
17:50:51Tornebut if you're waking the APU constantly anyway to do other stuff (e.g. talking on radios) then you're not gonna win much/anything
17:51:25saratogaaudio uses so little cpu dsp processors are almost never used for ot these days
17:51:27westwindNo voice, just for audio
17:52:05saratogas this running android?
17:52:23saratogaughhh stupid nexus keyboard glitches
17:52:27westwindbut I heard that the post-processing such as SRS will take a lot of cpu usage.
17:53:09westwindsaratoga, Yes, I suppose the os is android.
17:53:39Tornehaving a DSP core on an android device seems largely pointless, yeah
17:53:42saratogamost effects like that are pretty cheap
17:53:53Torneoffload audio decode to the GPU! :)
17:54:13westwindTorne, GPU is very expensive
17:54:21TorneCheaper than the APU
17:54:36saratogaelechips had one of them, cant rememberr if it was srs, running in a couple kb of ram on a 200 mhz arm9 core
17:55:14saratoga this was 5 years ago
17:55:25saratogaour a9 is unimaginab
17:56:11 Join kevku [0] (~kevku@2001:470:27:773:0:feed:c0f:fee)
17:56:18saratogaunimaginably fast
17:57:38saratogai think including a cortex m4 or similar as an audio processor could be useful to save power perhaps
17:57:47saratogabut only if you exp
17:57:47Tornewestwind: android devices are bascially always doing race-to-completion; powering a DSP constantly may not actually be cheaper than powering the A9 for long enough to fill all your audio buffers and then just shutting it off
17:58:09westwindsaratoga, Yes, but If you owned a whole multi-instances audio pipeline on a ARM9, includes decoding, resample-rate, volume, post-processing downmixing, mixing and I2S/HDMI/SPDIF driver, do you think the arm9 is enough to handle all of those/
17:58:19Torneif you're just doing decoding/effects you don't have any realtime inputs so you can race to completion instead
17:58:33saratogabut only if you expect the user to rarely have the a9 powered and listen to audio at the same time
17:58:35Tornewestwind: probably not, but the A9 is, like, fifty times faster than that :)
17:59:21saratogaont we have that in rockbox?
17:59:33Torneyeah, pretty much.
17:59:54Tornewe do most/all of that, and while it lags if you try and do all of that on a 40MHz ARM7TDMI, it works fine on something like the beast with an ARM11 ;)
18:00:17saratogawe can do aac ,5 band eq, resample and mixing on an 80 mhz arm7tdmi
18:01:06Torneyeah. on an A9 you are more talking about how often you have to wake up at all, rather than the level of utilisation
18:01:19saratogapdif on some devices too but thats a hardware thing so its free to the CPU
18:01:28westwindsaratoga, Torne, Yes, You are the experts, that's why I ask question here.
18:01:44Torneif you only wake up every couple of hundred milliseconds for a moment then the power consumption over time may be *less* than a dedicated always-on DSP doing the same thing
18:01:53Tornedepends how good your silicon is at waking up and going to sleep with low latency :)
18:02:08saratoga audio processing is generally very litte cpu compared to what android expects
18:02:19Tornethat's android's general approach to everything, anyway: go as fast as possible until you've finished all possible work then shut off as much of the hardware as you can
18:02:55saratogaut the question is how power efficient is your system for the work loads you expect
18:04:04saratogawe have seen many arm vendors take the same ip from arm and produce designs that are very different in effciency
18:04:54saratogafor an audio player you will need to deep sleep often while using very little power
18:05:25 Quit rela (Quit: Leaving)
18:05:31saratogaeven the slowest android processors will hardly ever be working for most audio
18:06:30saratogaof course you can screw that up too, ive seen vendors ship codecs with arm optimization disabled for instance
18:06:35westwindsaratoga, If My products are for Set-top-box, dolby ture HD and DTS-HD are needed, how do you think if I should have a DSP?
18:07:10saratogaso if you want to do bad things like that a fast proxessor will help
18:07:32saratogah then power does not matter
18:07:36Torneremmeber that that kind of CPU can already decode pretty advanced *video* codecs with reasonable performance :)
18:08:09 Join Jinx [0] (~Jinx@unaffiliated/jinx)
18:08:11saratogaac3 and regular dts are some of the most primative lossy codecs
18:08:18westwindWe have a VPU
18:08:24saratoga dont know about the newer versions
18:08:28westwindWe have a hardware-based VPU
18:08:44Tornewestwind: right, my point is just that the CPU alone without hardware video decode is probably already capable of HD MPEG4 and SD H264 decoding in realtime
18:08:54TorneEven a crazy advanced audio codec is much less work than that :)
18:09:39saratogado you need to encode them or just decode?
18:09:50Torneif your hardware runs off the mains then i can't imagine ever wanting DSP; no consumer audio workload is going to tax a modern ARM core at all
18:11:21westwindTorne, saratoga, I'm sorry for my bad english, could you please tell me your conclution, do you think a audio dsp is a necessary on dual A9 cortex?
18:11:59westwindTorne, thank you very much
18:12:07westwindsaratoga, How do you think?
18:12:29saratogaprobably not unless you want to do much more then you have mentioned
18:13:15westwindsaratoga, You know, It bother me for a long time
18:13:37saratogajust decoding audio and mixing is so fast your cpu wont notice
18:14:02westwindsaratoga, I get touch with CEVA for a long time
18:14:55westwindsaratoga, Torne, but it's difficute for me to decide buy a DSP
18:15:38westwindsaratoga, I heard many company who use A9 processor don't need a DSP for audio
18:16:02saratogais there some other reason you were considering a dsp or just concern that the main cpu was too slow
18:16:13westwindthey just use ffmpeg for decoding, because of the neon instruction set.
18:16:52westwindsaratoga, No reason, the A9 is a brand-new thing for me
18:16:54saratogawe have reverse engineered many plays and its been a very long time since DSPs were used
18:17:38saratogaarm7tdmi killed audio DSPs for decoding in the late 1990s
18:19:13saratogasandisk, creative and others dropped their dsp systems for arm9 in the mid 2000s
18:19:29westwindsaratoga, Thank a lot!
18:19:33saratogai dont know anyone who still uses them
18:19:42saratogao problem
18:19:50saratogano problem
18:25:43funmanwestwind: where do you work btw? if it's not a secret
18:26:13 Quit saratoga (Ping timeout: 250 seconds)
18:26:39westwindfunman, I'm sorry, I can't say it. because of I mentioned CEVA.
18:27:46funmanthat's ok
18:28:32westwindwhat's the different from #rockbox and #rockbox-community
18:29:18funman#rockbox is for technical talk and #rockbox-community for social chat
18:29:48westwindfunman, I get it, Thanks
18:32:58westwindOn the calculation of Matrix address this part, the dsp may have the advantage
18:34:03westwindFOR 2D ARRAY
18:34:15westwindFOR MDCT
18:35:53westwindand saturated and MAC
18:36:01funmanif you want to decode hundreds of channels at the same time maybe you need a dsp
18:36:33funmanthe cpu can already do saturation and MAC anyway
18:37:13westwindfunman, lol, I'm just thing about the dsp's future, not for me, for the DSP vendor
18:37:50westwindfunman, YES, but ARM9 only has the 32bits saturation but not the 16bits
18:38:07westwindfunman, ARM11 has 16bits saturation
18:38:39funmanA9 is ARM11
18:38:48Tornewestwind: are aware that ARM9 and A9 are different, right?
18:38:51Tornefunman: no it isn't :)
18:39:00westwindTorne, yes
18:39:07*funman lets Torne correct him
18:41:21westwindI think the dsp has advantage for voice but not audio related
18:42:10westwindsuch as Echo Cancellation
18:42:40westwindspeech recognition
18:44:01funmanpossible, echo cancellation doesn't make much sense for rockbox; for speech recognition I don't know if that was ever studied by us
18:44:44westwindfunman, does rockbox have AMR?
18:45:23 Quit maruk (Quit: Leaving.)
18:45:28funmani doubt it
18:46:07 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
18:46:43funmanalthough wmavoice seems to mention it, are they similar codecs?
18:47:41westwindHE-AAC V2 also has speech related thins
18:48:30westwindfunman, total different
18:52:01westwindtorne, saratoga, funman, , Thanks for your expertise. I'm going to bed.
18:56:15 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
18:57:44 Quit westwind ()
19:15:52 Join saratoga [0] (123e11e0@gateway/web/freenode/ip.
19:16:52 Quit jlbiasini_ (Ping timeout: 248 seconds)
19:28:12 Join jlbiasini_ [0] (
19:29:45 Quit crose (Ping timeout: 246 seconds)
19:39:01***Saving seen data "./dancer.seen"
19:40:27 Quit Substanzlos1 (Read error: No route to host)
19:41:29 Join Substanzlos [0] (
19:44:59 Join webguest44 [0] (
19:47:03webguest44Can Rockbox be used on the Samsung Z5, I know it's not usable on YP-Z5, but is it unusable, meaning I can use it sort of.
19:52:22 Quit webguest44 (Quit: CGI:IRC (Ping timeout))
20:04:08 Join rdn [0] (
20:05:29saratogaARM9E has 32x16 shifted into 32 MACs, but only limited saturation IIRC
20:05:47saratogai think they added it in v5 or v6, but its not very useful because its slow and crappy
20:06:16saratogaits not super useful anyway since with 32 bit samples you shouldn't really need it for most things
20:06:46 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
20:12:18 Nick DormantBrain is now known as SuperBrainAK (
20:15:22 Join crose [0] (
20:18:43fs-bluebotBuild Server message: New build round started. Revision c323381, 217 builds, 21 clients.
20:25:51fs-bluebotBuild Server message: Build round completed after 429 seconds.
20:30:52fs-bluebotBuild Server message: New build round started. Revision 1357e0a, 217 builds, 21 clients.
20:33:34 Join wodz [0] (
20:36:27fs-bluebotBuild Server message: Build round completed after 336 seconds.
20:38:14wodzTorne: What is the usage of locking cache way (not individual cache lines)?
20:38:44Tornemaking the logic simpler, i expect
20:40:20wodzTorne: So how do you use such feature from software pov?
20:40:34 Join wowaname [0] (~wow@
20:41:12Tornearrange your code so the bits you want to lock end up in the same ways, potentially :)
20:42:47wodzTorne: You mean lock way0, prefetch, unlock way0, lock way1 ?
20:43:03Tornedepends, you'd ahve to read chip docs
20:43:08Tornethe ways locking works vary wildly
20:43:16Tornereally rather annoying :)
20:44:21kugelwhat's a cache way?
20:44:22wodzOk assuming I can figure out how to put something in cache way and then lock this way, how this can be used to speed up things in rb?
20:44:53Tornewodz: if you can find hot loops in codecs and cachealign them and lock them you may find decoding is better
20:45:03Tornesimilar to the improvements we get by putting stuff in iram, but more so :)
20:45:17Torneyou may need to lock different things at different times, though
20:45:25Tornesince you probably can't lock muc
20:45:33Tornemany cachesyou can only lock, say, 50%
20:45:48wodzhmm, sounds complicated in practice
20:45:52Torneyes, it is.
20:45:59Torneit's very hard to use to any real benefit
20:46:23Torneit has occasinoally been used for unusual things like hardware security
20:46:46saratogafunfact: on older PP CPUs, the IRAM is actually faster than the L1 cache
20:47:07Tornealso that, yeha. if you actually have iram in the specific sense then it may be just as good as the cache, or even better
20:47:21Tornethe cache in the PP isn't really the same architecture as a real L1 cache ;)
20:47:32Torneit's external, whcih is not a thing people have typically done with L1
20:47:48Torneso the IRAM is actually more like TCM and the cache is more like L2 cache :)
20:48:27wodzon rk27xx I have only 4kB of Iram but 16kB of cache where I can lock way so you know
20:48:48wodzbut this seems fairly complicated to gain anything
20:48:49Torneyeah, so you may be able to treat the cache like another 4kb or 8kb of iram
20:48:59Tornebut you would have to lay the code out in ram correctly
20:49:19Torneand lock/unlock the right bits at the right times
20:49:34Torneand it may well be that the cache and prefetch logic are already doing a good enoguh job
20:49:48Torneand your manual cache management may even make it slower since it takes time to execute the prefetch/lock code :)
20:50:22Tornesomething we could potentially do better on chips with cache is just using explicit prefetch/preload instructions on chips that have those
20:50:31Tornewithout needing to lock anything
20:51:48 Quit Rower (Quit: Hmmm...)
21:13:14 Join rela [0] (~x@pdpc/supporter/active/rela)
21:18:47 Join Borg^Queen [0] (~arklinux@arklinux/tester/borgqueen)
21:18:55Borg^Queenhello people
21:20:54amayerBorg^Queen: hello
21:21:08Borg^Queenhello wowaname and amayer
21:21:36Borg^QueenGoing to try to build rb from git, here in case I hit a wall.
21:22:03Borg^QueenI can't find a prebuilt that supports nano g2
21:23:18Borg^Queenyes, I didn't get to the problem. I'm running an older system that has served me so well, I'm not willing to part with it.
21:23:32Borg^QueenSo I would need older versions capable of supporting nano g2
21:25:05gevaertsIn that case I'd recommend running some older release
21:25:09gevaertsThose are all available
21:25:23Borg^QueenAye, looking for one.
21:26:13Borg^QueenI have release ... m1.0.9
21:26:26Borg^QueenI had forgotten that I had it, so I fired it up.
21:26:28gevaertsThat's not a rockbox release
21:26:38Borg^Queenah ok
21:26:56Borg^Queenthank you for the link , checking now
21:28:09 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
21:28:28 Part LinusN
21:30:09Borg^Queendoes anyone know off hand where version started to support ipod gen2 and archos 5
21:30:53saratogathe archos 5 was never supported
21:31:05saratogathe nano2G was a couple years ago
21:31:36Borg^QueenThose are the devices I have
21:35:26saratogano i mean why are you asking
21:35:33saratogawhy do you want to know when support was added?
21:36:36Borg^Queenso I can get the oldest version which is likely to work on older distro rel
21:36:57saratogaI don't understand?
21:37:11Borg^QueenIt's ok
21:39:05***Saving seen data "./dancer.seen"
21:39:09amayerBorg^Queen: the newest version of the firmware still works on ipod nano 2g
21:39:37Borg^Queenyes, I understand that, I wanted to build the utility.
21:39:55Borg^QueenSorry just realized I misspoke meself.
21:40:17Borg^QueenRaised a Gaelic speaker
21:40:23gevaertsBuilding an older rockbox utility is unlikely to achieve much
21:40:26saratogaah so you're asking about building rockbox utility
21:40:27amayeryou can guild the new utility and it should work
21:41:04Borg^Queenamayer: currently trying with latest, I got an error last time I wasn't able to figure out, which is why I'm here.
21:41:05gevaertsrockbox utility gets its information about versions from the server, so using an older version to try to get older rockbox versions won't work
21:41:07Borg^Queenusing git
21:41:23Borg^Queengevaerts: yes
21:41:24amayerBorg^Queen: what was your error?
21:41:37Borg^Queenstill in compile mode
21:42:01Borg^Queenit will be about 20 mins, which is why I was looking for an older version whilst waiting.
21:42:46saratogagevaerts: he mispoke above, hes not trying to get an old version of rockbox, just a version of the utility that will build on whatever hes running
21:43:19 Nick Borg^Queen is now known as BigHairGuy (~arklinux@arklinux/tester/borgqueen)
21:43:37gevaertsStill, older versions of rockbox utility will have trouble with some newer targets...
21:43:44 Nick BigHairGuy is now known as Borg^Queen (~arklinux@arklinux/tester/borgqueen)
21:43:49Borg^QueenI thought as much
21:48:54Borg^Queenit seems easy to install it manually
21:54:13 Quit y4n (Quit: MOTHER EUROPA CALLING ME!)
21:55:13 Quit rdn (Remote host closed the connection)
21:59:04Borg^QueenI love being able to use ogg, it was such a bloody bother to convert to mp3 all the time.
22:03:07wowanamedoes anyone know how to update the bootloader
22:03:12wowanamei have the image
22:03:33wowanameipod classic 160g
22:06:52wowanamealso, i seem to fail at %pv as a progress bar
22:07:59Borg^Queenwowaname: same here, lots of traffic, but it's moving
22:09:06wowanameBorg^Queen» what are you replying to
22:09:44Borg^Queenyou progress, are you trying to down using git?
22:10:01wowanamei am talking about themes
22:10:03wowanamethe %pv tag
22:10:09Borg^Queenah sorry misunderstood
22:11:07wowanamelol i guess
22:11:44Borg^Queensorry multitasking, working on 5 computers and translating in as many languages.
22:11:47wowanameoh the font converter is one way
22:11:55wowanameBorg^Queen» what's your native language
22:12:15wowanamedamn you're doing well trying to keep up
22:12:29Borg^QueenI also speak english, spanish, some italian and a little french, mostly the curses.
22:12:39wowanamei want to learn japanese
22:12:43funmanplease move social chat to #rockbox-community
22:12:54wodzthis isnt social chat chaps
22:13:08Mode"#rockbox +o funman" by ChanServ (ChanServ@services.)
22:13:10Mode"#rockbox +b *!*wow@184.242.79.*" by funman (~fun@rockbox/developer/funman)
22:13:10Kick(#rockbox wowaname :wowaname) by funman!~fun@rockbox/developer/funman
22:13:10 Quit fs-bluebot (Remote host closed the connection)
22:13:11Borg^QueenProblem is these languages use the same words, but with different meanings so I have to be careful not to mix them up.
22:13:28 Join fs-bluebot [0] (
22:14:11 Join thomasjfox_ [0] (~thomasjfo@rockbox/developer/thomasjfox)
22:14:46 Quit thomasjfox (Ping timeout: 256 seconds)
22:21:29 Join Transmogrifox [0] (
22:23:42 Quit Transmogrifox (Client Quit)
22:28:33 Quit rela (Read error: Connection reset by peer)
22:29:42 Nick thomasjfox_ is now known as thomasjfox (~thomasjfo@rockbox/developer/thomasjfox)
22:31:57 Join Insider[away] [0] (
22:31:57 Nick Insider[away] is now known as Insider (
22:32:18Insiderhi all
22:34:59InsiderI'm a disappointed Creative X-Fi3 user.
22:35:22InsiderDisappointed due to the very low volume output of the European versions.
22:37:33amayerInsider: is this rockbox related or just a limitation on the device itself?
22:37:54Insidernope, an imposed limitation by Creative.
22:38:37Insiderso I'm trying to figure out how to overcome this limitation. Someone say to me to upload a non original firmware.
22:39:02n1sthere's some EU regulation on max power or something in portables
22:39:16Insideryes, this regulation is just ridiculous.
22:40:00InsiderAre European ears more sensitive than American one? What the reason of this normative?
22:40:14n1sfor many players you can just flash a firmware for a different region but i'm not sure that works on that player. Also this isn't strictly rockbox related i think
22:40:14saratogathe eu has laws and such
22:40:17Insiderthe volume is definitely lower than non european version.
22:40:23saratogatry flashing a different firmware, or run rockbox
22:40:33InsiderI have the american firmware (
22:40:52Insiderbut actually I cannot flash it.
22:41:05InsiderActually I have the same version, but european
22:41:17Insiderand looks like that the installation abort.
22:41:54InsiderI don't know how to do it manually.
22:42:09Insiderjust copying in the root of the device?
22:42:22pamauryInsider: no that won't work, you need some custom updater
22:42:36pamauryare you using windows or linux ?
22:42:45Insiderwindows, but I have virtual box
22:42:52Insiderto virtualize linux
22:43:06pamaurywith linux that can be done pretty easily
22:43:17Insideroh good
22:43:45InsiderI have already downloaded the rockbox dev image
22:44:43jlbiasini_pamaury: I have a release candidate for you: g#569
22:44:46fs-bluebotGerrit review #569 at : touchpad: Disable touchpad on softlock. by Jean-Louis Biasini (changes/69/569/18)
22:44:53pamauryyou will need a special file, I'll build it for you, you also need a recent version of our source code
22:45:40pamauryjlbiasini_: thanks, I'll have a look at that in a minute
22:46:12jlbiasini_it is separated from touchscreen so that we can wait to have a test on g#532
22:46:14fs-bluebotGerrit review #532 at : touchscreen: touchdev_enable implementation by Jean-Louis Biasini (changes/32/532/8)
22:47:08Insiderpamaury: I have to do a gitcheck out?
22:47:19jlbiasini_before to implement touscreen with it. I have to change this last pacth into one that will make it usuable after g#532 is pushed
22:47:21fs-bluebotGerrit review #532 at : touchscreen: touchdev_enable implementation by Jean-Louis Biasini (changes/32/532/8)
22:47:36jlbiasini_I meant g#569
22:47:38fs-bluebotGerrit review #569 at : touchpad: Disable touchpad on softlock. by Jean-Louis Biasini (changes/69/569/18)
22:47:55pamauryInsider: that's the simplest option yes
22:48:15Insiderok, I'm going to do it.
22:48:21jlbiasini_first we set this new setting touch menu > then we can add touchscreen to it
22:48:37Borg^Queenok I've completed dling rockbox from git
22:49:00 Join ikeboy [0] (
22:49:06jlbiasini_then I have to check a few touchpad target that are not defined as touchpad to see if they are worth/possible to add it
22:49:14Borg^QueenShould I mkdir build and then run from it?
22:49:46 Quit fs-bluebot (Remote host closed the connection)
22:50:02 Join fs-bluebot [0] (
22:50:02pamauryInsider: download this:
22:50:03jlbiasini_then I will also change touchpad_sensitivity to touchdev_sensitivity for the case some touchscreen might have a use of it
22:50:15gevaertsBorg^Queen: no. You can run from anywhere. I don't think you need that for rockbox utility though
22:51:18Borg^Queenwhat do you recommend I do from here. I have several compile recommendations, don't know which to follow.
22:52:30Insiderpamaury: done.
22:52:36jlbiasini_finally we can also add the lock only touchdev on softlock. It's really easy with touchdev_enable. We can set a switch in action linked to a setting that does either lock or touchdev_enable
22:53:11pamauryInsider: when you have a full git checkout, using the command line, go to the root of the source tree, then go into utils/imxtools/sbtools and run "make"
22:55:33 Quit petur (Quit: Nettalk6 -
22:56:26gevaertsBorg^Queen: for rockbox utility, is what you need I suspect
22:56:44Borg^QueenThanks already there and reading
22:56:59Insiderpamaury: I have the following error: ubuntu@ubuntu-VirtualBox:~/Desktop/rockbox/utils/imxtools/sbtools$ make
22:56:59Insidergcc -O3 -g -std=c99 -W -Wall `pkg-config −−cflags libusb-1.0` -DCRYPTO_LIBUSB -c -o elftosb.o elftosb.c
22:56:59InsiderPackage libusb-1.0 was not found in the pkg-config search path.
22:56:59DBUGEnqueued KICK Insider
22:56:59InsiderPerhaps you should add the directory containing `libusb-1.0.pc'
22:57:54InsiderI haven't do the 'configure' and make
22:57:55pamauryInsider: you need to install the libusb-1.0.0-dev package
22:58:04Insiderdescribed in rockbox.txt
22:58:04pamaurythere is no configure for this tool
22:59:11Insiderok, installed.
22:59:15jlbiasini_kugel: where can I set this list acceleration value?
22:59:22Insiderpamaury: still errors: ubuntu@ubuntu-VirtualBox:~/Desktop/rockbox/utils/imxtools/sbtools$ make
22:59:22Insidergcc -O3 -g -std=c99 -W -Wall `pkg-config −−cflags libusb-1.0` -DCRYPTO_LIBUSB -c -o elftosb.o elftosb.c
22:59:22InsiderPackage libusb-1.0 was not found in the pkg-config search path.
22:59:22InsiderPerhaps you should add the directory containing `libusb-1.0.pc'
22:59:40Insideroops, sorry wrong cut & paste.
22:59:47Insiderelftosb.c:433:5: error: too few arguments to function sb_write_file
22:59:53InsiderI have this error.
23:00:09pamauryah right, wait a second
23:00:33Borg^Queenah ok I need qt 4.5, I have 4.4
23:00:36fs-bluebotBuild Server message: New build round started. Revision 99f3ca1, 217 builds, 21 clients.
23:00:58pamauryInsider: I've just committed the fix, update the code by running "git pull −−rebase"
23:01:01jlbiasini_kugel, pamaury: g#571
23:01:03fs-bluebotGerrit review #571 at : fuze+/keymaps: Implement the ACTION_STD_MENU by Jean-Louis Biasini (changes/71/571/1)
23:01:11funmansbtoelf.c:80:40: erreur: ‘elf_std_printf’ undeclared (first use in this function)
23:01:56funmanelftosb built fine though
23:02:32Borg^Queenwhere can I find the rbutility source?
23:02:46Insiderpamaury: other error now: sbtoelf.c:80:40: error: elf_std_printf undeclared (first use in this function)
23:02:57Borg^Queenone old enough to run on qt 4.4
23:03:13pamauryInsider: update again, i committed another fix
23:04:09Insiderpamaury: done
23:04:14 Quit amayer (Quit: Leaving)
23:04:19funmanpamaury: inlined from ‘sb1_write_file’ at sb1.c:162:11:
23:04:19funman/usr/include/x86_64-linux-gnu/bits/string3.h:51:3: attention : call to __builtin___memcpy_chk will always overflow destination buffer [enabled by default]
23:04:33funmanit builds though but it's a bit scary ^_^
23:04:39Borg^QueenAll I see it rb itself
23:05:27Borg^Queenfound it
23:06:56InsiderI have 5 executable: sbloader, sbtoelf, rsrctool, elftosb and elftosb1
23:07:01fs-bluebotBuild Server message: Build round completed after 385 seconds.
23:07:02fs-bluebotBuild Server message: New build round started. Revision 8dc7970, 217 builds, 21 clients.
23:07:08pamauryInsider: now plug your xfi3, while plugged hold the power button and while holding the power button, press the reset the button (and keep power for a few seconds)
23:07:42Insiderlinux is virtualized, I think that windows will recognize it
23:07:46Insidernot linux
23:08:10gevaertsBorg^Queen: the RockboxUtilityDevelopment wiki page was updated to mention 4.5.0 (from 4.4.0) in March 2010. Git commit 17f56379 suggests that the requirement for 4.5.0 is still older. I suspect getting a usable rockbox utility to build with 4.4.0 will be a challenge...
23:08:21pamauryInsider: you need to give the usb device to linux, if you can
23:08:29pamauryotherwise that will get tricky
23:08:33pamauryfunman: interesting my compile doesn' catch that
23:08:33Borg^Queengevaerts: aye, thanks
23:08:41Borg^QueenI'll give it a go anyway
23:09:02InsiderIf I plug the xfi3, I can then let virtual box recognize it
23:09:09InsiderI will try
23:09:21funmanpamaury: i am running ubuntu 13.04 : gcc 4.7.2 / glibc 2.17
23:09:26Mode"#rockbox -o funman" by funman (~fun@rockbox/developer/funman)
23:09:52Insiderpamaury: Can I restore my original firmware in future?
23:09:59pamauryInsider: yes
23:10:51Borg^Queenok I'm going to try manual installs, it can't be that hard.
23:10:54pamaurycan you give the usb device to linux ? if you run "lsusb", you should get something like "066f:3780 STMP3780/i.MX23 recovery device"
23:12:24pamauryfunman: gcc 4.8.1 here, did you tweak the compile options ?
23:13:12fs-bluebotBuild Server message: Build round completed after 370 seconds.
23:13:20funmanpamaury: no
23:14:06pamauryInsider: ^
23:15:08funmanpamaury: try adding -D_FORTIFY_SOURCE=2 ?
23:16:23pamauryfunman: it does the trick
23:17:03funmanten points for ubuntu
23:20:46Borg^QueenGod love the cli.
23:21:20funmanpamaury: i wonder what would happen if we use that flag to build rockbox itself
23:21:25pamauryInsider: did you manage to give the usb device to linux ?
23:21:43pamauryfunman: we would need a recent version of gcc, and probably too many errors to cound them
23:21:51Insideryes, I cannot enter the recovery mode
23:22:04funmanpamaury: i doubt we need a recent version of gcc, that's been in ubuntu for years
23:22:23funmanalthough support for it is (entirely?) in glibc so we'd need to adapt our headers
23:22:44Borg^Queenshould I use a daily build for the nanog2?
23:23:01funmanBorg^Queen: current build
23:23:18Borg^Queen2013-08-21 Rockbox 3656 KB?
23:23:31pamauryInsider: did you follow my procedure exactly ? First plug, then slide the button to power and while doing that, use a pen to press reset, continue to slide the power for a few seconds
23:23:32Insiderlsusb say: Device 004: ID 041e:4169 Creative Technology, Ltd
23:23:50Borg^Queenfunman: thanks
23:23:53Insiderlsusb say: Device 004: ID 041e:4169 Creative Technology, Ltd
23:24:10pamauryyou need to get 066f:3780 and the screen will stay black
23:24:30 Quit Hadaka (Ping timeout: 241 seconds)
23:24:46funmanpamaury: hmm it seems it's all done by glibc, although i guess stack smashing detection is in gcc
23:25:47Borg^Queencan I load my own fonts into it?
23:25:55Insiderthe screen is black, virtual box say: unknown device 064e:D213
23:26:09Insiderand I cannot connect it to linux
23:28:12pamaurythis is not the right id, it should be 066f:3780
23:28:25pamauryand report as a HID device
23:29:47wodzInsider: this is different device I guess. Google suggest this vid:pid is for Suiyn Camera
23:30:32Insideroh wodz you say right
23:30:46Insiderit report that device also with xfi3 disconnected
23:30:58Insiderhow to check if I enter the recovery mode correctly?
23:31:06Insiderany way from windows?
23:31:20 Join Hadaka [0] (
23:32:05pamauryInsider: maybe in device manager, it should appear
23:32:22pamauryin theory it should report has a human interface device
23:33:09Insiderah ok
23:33:13Insiderfound it
23:33:24Insider066f 3780
23:33:29Insiderfrom hardware device
23:33:32pamaurygood, so now can you give it to the virtual box ?
23:33:49Insidersigmatel inc rom recovery?
23:34:33Insider I have it
23:34:39Insider066f:3780 SigmeTel, Inc.
23:35:17pamaurynow in your console, still in sbtools, you need to run: sudo sbloader
23:35:33pamaurywhere is the path to the file you downloaded from dropbox
23:36:12Insidersome text appears
23:36:15Insiderin the display
23:36:17Insiderof xfi3
23:38:11pamaurygood, now your device should appear with as a mass storage device
23:38:14Insidernow the lsusb reports: Bus 001 Device 006: ID 041e:2020 Creative Technology, Ltd Zen X-Fi 2
23:38:20Borg^Queenok it's not allowing me to copy .rockbox into the ipod
23:38:29pamauryextactly, oops wrong id by the way, I need to fix that, anyway
23:38:29Borg^QueenI checked all the perms too
23:38:44pamauryyour device should be something like /dev/sdb probably
23:38:54pamauryrun fdisk /dev/sdb, then "p" and pastebin the output
23:39:06***Saving seen data "./dancer.seen"
23:39:21Borg^Queenok hang on
23:39:46Insiderit show nothing
23:39:53Insider"Command (m for help):
23:39:59Insiderprobabily isn't sdb?
23:40:06InsiderHow I can check it?
23:40:08gevaertsBorg^Queen: does dmesg say anything about filesystem corruption?
23:41:04InsiderI have an sdhc card mounted
23:41:07pamauryInsider: check in dmesg | tail
23:41:11Borg^Queenno, nothing about fs cor
23:41:19pamaury(make sure the mass storage device belongs to the linux box)
23:41:56Borg^Queenisn't it fdisk -l to see the attached drives?
23:42:12pamauryah yes
23:42:17pamauryvery relevant, thanks
23:42:19Insider[ 3231.991700] sd 6:0:0:1: [sdc] No Caching mode page present
23:42:19Insider[ 3231.991708] sd 6:0:0:1: [sdc] Assuming drive cache: write through
23:42:19Insider[ 3232.001258] sdb: sdb1 sdb2 sdb3 sdb4
23:42:19Insider[ 3232.009848] sdc: sdc1
23:42:19Insider[ 3232.114789] sd 6:0:0:0: [sdb] No Caching mode page present
23:42:19***Alert Mode level 1
23:42:19Insider[ 3232.114797] sd 6:0:0:0: [sdb] Assuming drive cache: write through
23:42:20***Alert Mode level 2
23:42:20Insider[ 3232.114805] sd 6:0:0:0: [sdb] Attached SCSI removable disk
23:42:20***Alert Mode level 3
23:42:20Insider[ 3232.123737] sd 6:0:0:1: [sdc] No Caching mode page present
23:42:21***Alert Mode level 4
23:42:21Insider[ 3232.123744] sd 6:0:0:1: [sdc] Assuming drive cache: write through
23:42:21***Alert Mode level 5
23:42:21Insider[ 3232.123751] sd 6:0:0:1: [sdc] Attached SCSI removable disk
23:42:38Borg^Queensdc2 is likely it
23:42:52pamauryInsider: avoid pasting many lines here, use
23:42:59Insiderah ok, sorry
23:43:15 Join ender1 [0] (
23:43:16InsiderI have 2 drive because I have inserted an sdhc of 16 gb
23:43:36pamauryit should be sdb
23:43:44 Quit ender` (Ping timeout: 256 seconds)
23:43:44 Quit thomasjfox (Quit: Konversation terminated!)
23:43:46Borg^Queengevaerts: This happened after I did a cli install of the ipodpatcher
23:44:05pamaurycan you check again with fdisk /dev/sdb and paste the output when pressing "p" (use
23:44:20gevaertsBorg^Queen: what exactly is the error you get?
23:44:26Borg^Queenhang on
23:44:29Insidersame as before
23:44:38Insiderit print nothing, only a prompt: "Command (m for help)."
23:44:40Borg^QueenCould not make folder /home/mnt/sdb2/.rockbox
23:44:47Borg^Queenthis is from with konqueror
23:44:59Insideridem for sdc
23:45:10Insidermaybe the fdisk syntax is different?
23:45:34Insideractually I have the 16 gb mounted
23:45:39InsiderI can see the files inside it
23:45:50pamauryInsider: did you press "p" as I said
23:46:07InsiderI have it
23:46:45Insiderhere the result:
23:46:59Borg^Queenthe ipodpatcher seems to have erased .rockbox
23:47:09gevaertsIt doesn't do that
23:47:19Borg^Queenit's gone now
23:47:33gevaertsIs the device still mounted?
23:47:34Borg^Queenand I can recopy it over
23:47:43Borg^QueenI can see all the usual dir
23:47:47pamauryInsider: good, now just to check, run "sudo hexdump -C /dev/sdb2 | head"
23:48:22CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:48:22*gevaerts doesn't really know
23:48:30gevaertsUnless unplugging without unmounting is involved
23:48:45Borg^Queenbut, I can try running a fsck
23:49:04Borg^Queenbut then wouldn't it fail to mount?
23:49:13pamauryInsider: ok good
23:49:35pamaurynow you will need your american version of the creative software
23:49:52InsiderI have it
23:50:02pamauryactually this is the step of the procedure where you can take any firmware file and flash it, disregarding version downgrading and so on
23:50:27 Quit n1s (Quit: Ex-Chat)
23:50:45pamaurynow run: sudo dd bs=512 seek=4 of=/dev/sdb2
23:50:52pamaurythen then "sudo sync"
23:52:22***Alert Mode OFF
23:52:41jlbiasini_pamaury: there still a few stuff to be corrected in g#569
23:52:43fs-bluebotGerrit review #569 at : touchpad: Disable touchpad on softlock. by Jean-Louis Biasini (changes/69/569/18)
23:52:49jlbiasini_shouldn't be long
23:54:05Insiderhow much time it require on average?
23:54:57Insiderok done
23:54:59pamauryInsider: don't remember for the xfi3, probably a minute or so
23:55:32Insiderand now?
23:57:07pamaurydisconnect your device and pray ;)
23:57:44pamauryand power it again, if everything went well, it should boot your newly flashed firmware
23:58:38pamaurycheck the version in the OF settings
23:58:56Insiderit's changed

Previous day | Next day