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 2010-03-18

00:02:18 Part froggyman
00:04:59 Quit Schmogel (Quit: Miranda IM! Smaller, Faster, Easier.
00:07:23 Quit saratoga (Ping timeout: 252 seconds)
00:08:26 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
00:10:16 Quit ender` (Quit: Consultation, n. Medical term meaning "to share the wealth.")
00:19:24 Join webguest41roko [0] (
00:19:47 Join RadicalR [0] (
00:23:40 Join Strife1989 [0] (
00:24:18 Join CaptainKewl [0] (
00:26:22 Quit Strife89 (Ping timeout: 265 seconds)
00:28:34***Saving seen data "./dancer.seen"
00:28:35 Quit Strife1989 (Read error: Connection reset by peer)
00:28:59 Join Strife1989 [0] (
00:30:48 Join robin0800 [0] (
00:32:34 Quit CaptainKewl (Read error: Connection reset by peer)
00:33:02 Part toffe82
00:34:02 Nick Strife1989 is now known as Strife89 (
00:36:29 Quit webguest41roko (Quit: CGI:IRC)
00:48:48 Quit r0b- (Ping timeout: 260 seconds)
00:50:17 Quit xavieran (Ping timeout: 246 seconds)
00:56:14 Quit bertrik (Ping timeout: 246 seconds)
01:10:54 Quit komputes (Read error: Connection reset by peer)
01:11:13 Quit Adubb (Read error: Connection reset by peer)
01:18:20 Quit perfectdrug_ (Quit: perfectdrug_)
01:23:49 Join r0b- [0] (
01:28:01 Quit hebz0rl (Quit: Ex-Chat)
01:30:15 Quit scorche|sh (Ping timeout: 276 seconds)
01:31:44 Quit DerPapst (Quit: Leaving.)
01:39:17 Join komputes [0] (~komputes@ubuntu/member/komputes)
01:39:19 Join saratoga [0] (~9803c6dd@gateway/web/freenode/x-zsqusexkeghgjdzg)
01:40:25 Join Strife1989 [0] (
01:40:35 Nick Strife1989 is now known as Strife89|Desktop (
01:49:03 Quit robin0800 (Remote host closed the connection)
02:14:37 Quit MethoS- (Remote host closed the connection)
02:27:49 Join CaptainKewl [0] (
02:28:11 Join scorche|sh [0] (~scorche@rockbox/administrator/scorche)
02:28:36***Saving seen data "./dancer.seen"
02:36:37 Join Tomis2 [0] (~Tomis@
02:38:47 Quit Tomis (Ping timeout: 265 seconds)
02:38:48 Nick Tomis2 is now known as Tomis (~Tomis@
02:48:24 Join MrShlee [0] (
02:49:11MrShleeMy iPod classic blew up (until the linux4nano team work their magic for a bootloader) I'm looking for a new mp3 player running rockbox. any recommendations?
02:49:17MrShlee80GB+ wanted.
02:52:59S_a_i_n_tMrShlee: It's all subjective my "good" or "best" (or thers) will differ from yours. Have a look at the supported players on the RB mainpage, and pick one :D
02:53:12S_a_i_n_t*others rather.
02:56:11 Quit jordan` (Ping timeout: 265 seconds)
02:57:40 Quit adnyxo (Quit: Leaving)
03:04:22 Quit komputes (Ping timeout: 246 seconds)
03:22:41 Join I3uckwheat [0] (~chatzilla@
03:23:37I3uckwheathas rock box come out for the sansa fuze v2 yet
03:25:00*S_a_i_n_t directs I3uckwheat to the Rockbox main page.
03:27:48I3uckwheathas rock box come out for sansa fuze v2 yet
03:28:40S_a_i_n_tI3uckwheat: check the Rockbox main page, it lists all supported devices.
03:31:40 Quit I3uckwheat (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
03:39:59 Quit Barahir (Ping timeout: 260 seconds)
03:40:24 Quit saratoga (Ping timeout: 252 seconds)
03:41:22 Join Barahir [0] (
03:41:24Connor1Music is I:\Connor's Complete Music Collection
03:41:59 Quit MrShlee (Ping timeout: 240 seconds)
03:48:27 Quit Tomis (Read error: Connection reset by peer)
03:48:40 Join Tomis [0] (~Tomis@
03:56:08 Quit kugel (Remote host closed the connection)
04:05:22 Join jordan` [0] (~jordan@
04:15:05 Quit mc2739 (Remote host closed the connection)
04:21:15 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
04:28:40***Saving seen data "./dancer.seen"
04:31:08 Quit TheSeven (Disconnected by services)
04:31:22 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:31:32 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:42:49 Quit Barahir (Read error: Operation timed out)
04:55:07 Join Rob2222 [0] (
04:57:36 Quit Rob2223 (Ping timeout: 260 seconds)
04:58:12 Join phanboy4 [0] (
05:15:19 Quit DV_ (Read error: Connection reset by peer)
05:54:24 Quit Strife89 (Read error: Connection reset by peer)
05:54:52 Join Strife89 [0] (
05:58:04 Quit Strife89|Desktop (Ping timeout: 246 seconds)
06:06:28 Quit Strife89 (Quit: Rebooting.)
06:08:12 Join Strife89 [0] (
06:19:00 Join FOAD_ [0] (
06:19:20 Quit soap (*.net *.split)
06:19:20 Quit preglow (*.net *.split)
06:19:20 Quit killan (*.net *.split)
06:19:20 Quit mt (*.net *.split)
06:19:20 Quit scorche (*.net *.split)
06:19:20 Quit yosafbridge (*.net *.split)
06:19:20 Quit jordan` (*.net *.split)
06:19:20 Quit scorche|sh (*.net *.split)
06:19:20 Quit RadicalR (*.net *.split)
06:19:20 Quit chaos (*.net *.split)
06:19:20 Quit GodEater (*.net *.split)
06:19:20 Quit leavittx (*.net *.split)
06:19:20 Quit Battousai (*.net *.split)
06:19:20 Quit stavrob (*.net *.split)
06:19:20 Quit tchan (*.net *.split)
06:19:20 Quit krazykit (*.net *.split)
06:19:20 Quit dionoea (*.net *.split)
06:19:21 Quit avacore (*.net *.split)
06:19:21 Quit ChanServ (*.net *.split)
06:19:21 Quit FOAD (*.net *.split)
06:19:21 Quit FlynDice (*.net *.split)
06:19:21 Quit Llorean (*.net *.split)
06:19:21 Quit lostlogic (*.net *.split)
06:19:21 Quit Tuplis (*.net *.split)
06:19:21 Quit doomcup (*.net *.split)
06:19:21 Quit tipi^ (*.net *.split)
06:19:25 Nick FOAD_ is now known as FOAD (
06:23:56 Quit JdGordon (Read error: Connection timed out)
06:24:35 Join JdGordon [0] (
06:27:46 Join shai [0] (
06:28:44***Saving seen data "./dancer.seen"
06:31:10 Join CGL [0] (~CGL@
06:52:45 Join elcan [0] (
07:02:14 Join n1s [0] (~n1s@
07:05:55 Quit fyrestorm (Read error: Connection reset by peer)
07:06:14 Quit CaptainKewl (Remote host closed the connection)
07:09:31 Join DV_ [0] (~DV@
07:17:47 Quit DV_ (Ping timeout: 264 seconds)
07:20:45 Join xavieran [0] (
07:21:02*n1s wonders what the point of screenshots of the text viewer is
07:23:32 Quit anewuser ()
07:32:57pixelmais the forum down or is this just a problem in my end?
07:34:13 Quit n1s (Ping timeout: 258 seconds)
07:36:56JdGordonfind here
07:41:20 Quit CGL (Quit: Saliendo)
07:42:14JdGordonany suggestions for the tag letters for "track changing?" (so you can know if its about to change or has just changed +/- user timeout)
07:42:21JdGordon%pC was suggested in the forum
07:46:41 Join DV_ [0] (~DV@
07:56:46 Join Horscht [0] (
07:58:55 Quit Horschti (Ping timeout: 245 seconds)
08:08:09 Join Zagor [0] (
08:08:16 Join phanboy_iv [0] (
08:08:56CIA-5New commit by 03jdgordon (r25239): 2 new tags: ...
08:11:26 Quit phanboy4 (Ping timeout: 265 seconds)
08:14:17 Join scorche|sh [0] (~scorche@rockbox/administrator/scorche)
08:14:17 Join scorche [0] (~scorche@rockbox/administrator/scorche)
08:14:17 Join fyrestorm [0] (
08:14:17 Join Barahir [0] (
08:14:17 Join jordan` [0] (~jordan@
08:14:17 Join RadicalR [0] (
08:14:17 Join soap [0] (~soap@rockbox/staff/soap)
08:14:17 Join preglow [0] (
08:14:17 Join chaos [0] (~chaos@gentoo/user/ch4os)
08:14:17 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
08:14:17 Join killan [0] (
08:14:17 Join leavittx [0] (~leavittx@
08:14:17 Join Battousai [0] (~bryan@gentoo/developer/battousai)
08:14:17 Join mt [0] (~chatzilla@rockbox/developer/mt)
08:14:17 Join stavrob [0] (
08:14:17 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
08:14:17 Join yosafbridge [0] (
08:14:17 Join FlynDice [0] (
08:14:17 Join krazykit [0] (
08:14:17 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
08:14:17 Join dionoea [0] (
08:14:17 Join lostlogic [0] (~lostlogic@rockbox/developer/lostlogic)
08:14:17 Join Tuplis [0] (
08:14:17 Join doomcup [0] (
08:14:17 Join avacore [0] (
08:14:17 Join tipi^ [0] (
08:14:17 Join ChanServ [0] (ChanServ@services.)
08:14:17Mode"#rockbox +o ChanServ " by
08:14:43 Quit JdGordon (Changing host)
08:14:43 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
08:15:12 Nick shai is now known as Guest43559 (
08:15:12 Nick Zagor is now known as Guest13876 (
08:16:19 Nick Guest43559 is now known as shai (
08:16:36 Quit shai (Quit: Leaving)
08:16:55 Join shai [0] (
08:20:18CIA-5New commit by 03jdgordon (r25240): fix yelllow and add those tags to the debug output
08:21:17 Quit mt (Ping timeout: 245 seconds)
08:21:53 Join ender` [0] (
08:24:14 Join DerPapst [0] (
08:25:06 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:28:07JdGordonwhat the heck hapened with the 1/2g ipod delta?
08:28:15JdGordon-2k?! with no code change
08:28:48***Saving seen data "./dancer.seen"
08:30:24 Join pondlife [0] (~Steve@rockbox/developer/pondlife)
08:30:34 Part pondlife
08:31:03 Join mt [0] (~chatzilla@
08:31:14 Quit mt (Changing host)
08:31:14 Join mt [0] (~chatzilla@rockbox/developer/mt)
08:32:16 Quit Guest13876 (Quit: Clint excited)
08:32:35 Join Zagor [0] (
08:32:35 Quit Zagor (Changing host)
08:32:35 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:36:59 Join S_a_i_n_t_ [0] (S_a_i_n_t@
08:38:26 Quit S_a_i_n_t (Ping timeout: 260 seconds)
08:40:50 Nick S_a_i_n_t_ is now known as S_a_i_n_t (S_a_i_n_t@
08:53:39linuxstbJdGordon: What is the default time for %pS and %pE ? Also, isn't "elapsed" in ms, not HZ?
08:54:40 Quit DV_ (Ping timeout: 248 seconds)
08:54:57linuxstbJdGordon: Also, what's with the "+ +" in;r2=25239;pathrev=25239 ?
08:55:44 Join liar [0] (
08:56:00JdGordonwtf? keyboard wierdness, not sure how tha happened
08:56:06JdGordondefault time works out to 10s for both
08:56:08JdGordonor should be
08:56:35linuxstbLooking at the code, the default seems to be 0, unless I'm misreading.
08:56:55JdGordonhow is that compiling even? ...elapsed++state->... ?
08:58:06JdGordondefault is 10, parser.c line 1113
09:00:22linuxstbAh OK, I hadn't updated that file... ;) Can that be documented?
09:00:27CIA-5New commit by 03jdgordon (r25241): woops, elapsed is ms not HZ and how did that extra + get in there?
09:01:18 Join petur [0] (~petur@rockbox/developer/petur)
09:01:33BagderZagor: can you understand how some of the 150-190 speed machines in the build farm never complete builds, while some < 100 completes several?
09:01:59Bagderthere seems to be a discrimination that could need some attention
09:03:07Zagorthere looks to be a bug in the average speed calculation, which causes rbmaster to give a bunch of bootloaders to gevaerts for example
09:03:19Bagder <= this is good example of this
09:03:29Bagderw1ll14m-w1ll14m did 6 builds
09:03:32Zagorhis average is listed as 128 when in fact it is one of the fastest machines in the cluster
09:03:35Bagderdeepthought-ender did none
09:11:29Bagderwell, in my case I have my core2 duo laptop that isn't a very slow machine that often ends up not completing a single build
09:11:39Bagderand it seems wrong to me
09:12:19Bagderespecially the last few days when I've had it hooked up to a 10mbit uplink
09:16:40pixelmaI had the problem at the low end too, with my build client just being slightly faster than the slowest ones. The latter always got the bootloaders and mine a sim which it seldom finished
09:16:52Bagderyeah, something like that
09:17:30Bagderit seems the slowest machines get more builds than the ones slightly faster somehow
09:19:17Zagorthe primary problem is the wrong average calculation. it corrupts the whole planning. last time, gevaerts was assigned one very easy build which it completed in ~23 seconds and then started chasing all the slower clients
09:19:40Zagoruntil that is fixed, it's very difficult to get a clear view of the other problem
09:19:49Bagderright, and at least that's an obvious problem that will skew most things
09:20:52pixelmagevaerts in person? ;)
09:21:36Bagderyes, I think I saw him come running down the hall here!
09:23:17 Join wodz [0] (
09:24:22wodzcan buildclient be run on machine behind NAT?
09:24:32pixelmalooks like there was a build messup in the previous round again with the 1st gen Nano (looking at the deltas)
09:24:46wodzwill be P4 1.5GHz helpfull?
09:25:14Zagorwodz: absolutely. all machines are helpful.
09:25:16Bagderwodz: it basically cannot hurt, so it's fairly easy to just fire it up and see what it can do
09:25:36wodzok I will setup buildclient then
09:25:46Zagormy Atom N270 contributes to every build round!
09:27:31wodzwhat about 64bit system? Does it makes any problems as a buildclient? I have some 64bit debians rather unused
09:28:51Zagorwodz: no problems
09:29:15wodzok I will explore how I can contribute to build farm than
09:29:51wodzIs the build client aware of multicore?
09:30:31Zagorby default it uses all cores/hyperthreads, but you can limit to a specific number if you want
09:32:16pixelmaAlexP: doesn't matter which backdrop
09:32:51pixelmaif there was no fundamental change recently I missed
09:33:51CIA-5New commit by 03zagor (r25242): Limit number of rows returned.
09:37:34 Join DV_ [0] (~DV@
09:52:16LloreanJdGordon: Do you plan to write a manual entry for the new tags any time soon, or are you just going to follow the trend of ignoring such things?
09:53:21Zagorfound the problem. some boots and especially the wpscheck builds are so small that if you take a full second to complete them you get a very low score
09:53:39Zagorand the system does not support fractions of a second
09:57:19Zagorit's exacerbated by the fact that I use a conservative 33% median speed value for planning, rather than the average speed
09:59:32LloreanJdGordon: The general agreement was that the minimum was an FS entry of a full text description like you'd put in the manual, yours is just a rough description of it.
10:03:48 Join robin0800 [0] (
10:04:53 Join anewuser [0] (anewuser@unaffiliated/anewuser)
10:04:58 Join einhirn [0] (
10:12:27 Quit DV_ (Ping timeout: 252 seconds)
10:13:20 Join DV_ [0] (~DV@
10:14:21 Quit DV_ (Read error: Connection reset by peer)
10:14:42JdGordonare there any major objections to upping the skin buffer size radsio skin can go in before the resisable buffer patch get finished?
10:14:57 Join DV_ [0] (~DV@
10:17:17 Join merbzt [0] (~benlar@
10:18:03 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
10:20:47linuxstbJdGordon: Shouldn't things be done in the correct order? i.e. implement resizable buffer before things which require a resizable buffer?
10:21:19JdGordonS_a_i_n_t: ping?
10:21:50JdGordonlinuxstb: well, the fm skin doesnt depend on the resizing, it does need a slightly larger buffe though
10:23:10JdGordonpamaury: you're doing dircache fiddling yeah? any idea why the init order for dircache in the sim and target are different?
10:23:29JdGordonon sim its setting_load(); init_dircache(true); init_dircache(false);
10:23:39JdGordonon target the settings apply is between them
10:24:35pamauryJdGordon: yes, I'm always tweaking dircache :) I can't answer you now but iirc, there is difference because one tries to load the cache from the disk and the other not, or perhaps it's about transparent build or not. If you're ready to wait for a few minutes I can answer your question :)
10:24:52JdGordonim not going anywhere :)
10:25:04JdGordonit looks like dircache shouldnt even work on target with that orering!
10:26:32pamauryWhen I was fiddling with usb in simulator, I had a deep look at the init functions and there are big difference between targets and sim :)
10:27:50pamauryOn the other hand, in the sim, much of the initialization part is useless I think
10:28:24 Quit TheSeven (Ping timeout: 252 seconds)
10:28:52***Saving seen data "./dancer.seen"
10:29:10pamauryok, so init_dircache takes a parameter which tells whether dircache_init should be called. so the first time it's true, then it's false. That's sensible
10:29:40JdGordonglobal_settings.dircache is I think undefined at that point though
10:29:44JdGordonon target anyway
10:29:48pamauryIt also seems that init_dircache checks whether settings are loaded or not
10:29:51JdGordonon the sim it will be set to the config
10:30:19JdGordonah no, it will always be false thre
10:30:37 Quit einhirn (Ping timeout: 265 seconds)
10:30:54pamauryHum, there is this fucking EEPROM thing that doesn't make any sense to me so if we ignore it, then the first call will be equivalent to dircache_init
10:31:27pamauryThe second one will display a nice screen during rebuild
10:32:20pamauryI don't understand what the settings have to do with that
10:32:49JdGordonto not init dircache if the user dosnt want it?
10:33:41linuxstbIsn't that EEPROM thing the h120, when Rockbox is flashed to NOR? IIRC, that will cache the dircache, as it can tell whether the device may have been modified between boots.
10:34:41pamauryyes and no, iirc it was doing strange thing with EEPROM settings that don't seem to have any link with EEPROM
10:36:21pamauryJdGordon: I think you're right, during the first call, if the settings have never been assigned, then either it's undefined or more probably lobal_settings.dircache is false
10:36:52JdGordondoes that actualy change anything though?
10:37:10JdGordonI'm tyring to find out why boot splashes apparently arent using the sysfont, and dircache is one of those
10:37:54pamaurywell, it's just that the first call is equivalent to dircache_init apparently, so it's weird to this function two times, with the first call being trivial
10:38:03 Quit DV_ (Read error: Connection reset by peer)
10:38:16 Join einhirn [0] (
10:38:21 Join DV_ [0] (~DV@
10:38:35pamauryI can't help you with this, I know nothing about that :) But the first call will not trigger a splash, I'm nearly sure
10:39:07pamauryI you have doubts, replace the init_dircache(true) by dircache_init
10:39:23JdGordonno, looks like it wont
10:40:09pamauryBut perhaps this difference makes sense in the sim ? Sounds strange but we never know
10:45:42wodzI have question about iriver h300 bootloader code. charger_inserted() is defined in firmware/powermgmt.c and returns power_thread_inputs & POWER_INPUT_CHARGER; power_thread_inputs is updated in power_thread(). Now examing I can't see power_thread() symbol so how does it work than?
10:46:11JdGordonbe very careful with the h300 bootloader.. svn will brick it
10:47:08wodzI am not messing with that bootloader I am only using it as a template for my own (MPIO HD200). I do have BDM so I am on the safe side
10:49:36wodzon my target charger_inserted() doesn't work and I can't figure out what is the code workflow. Under debugger variable power_thread_inputs seems to be empty so I am wondering how it is updated in bootloader context
10:49:54wodzon normal build I assume it is updated by power_thread()
10:54:46S_a_i_n_tJdGordon: belated pong.
10:55:15JdGordonare you sure its the sysfont and not just hta your userfont is very similar?
10:56:39S_a_i_n_twell, my user font is Helvetica 12, and I'm pretty sure I can tell the difference between the two.
10:57:09S_a_i_n_t*maybe* I'm messing it up, but it looks like 8pt sysfont on the splashes to me.
10:59:53JdGordonvery wierd.. I just had another look on the h300 sim and it was dfinitly using the correct fonts
11:00 development. It's only when I'm using that one font.
11:00:38S_a_i_n_tNot "new" I guess, I just hadn't noticed it until now.
11:02:23JdGordonis that font file corrupt maybe?
11:02:58S_a_i_n_tI'm replacing it with a newer version now.
11:05:35S_a_i_n_tHmmm. OK, that task cab closed I guess. I feel slightly sheepish now. Helvetica Bold from the "extras" page works fine. Not sure how the one I compiled is any different, but it is apparently :/
11:05:55S_a_i_n_tshit...sorry about that.
11:09:43JdGordonno worries
11:13:38 Quit phanboy_iv (Ping timeout: 265 seconds)
11:20:28JdGordondoes anyone have any good reasons why we dont immediatly draw images in skins when they show up? instead of (What we do now) drawing all images in a viewport at he end of the vewport loop?
11:23:15gevaertsYou basically mean, why don't we have a predictable draw ordering?
11:23:23gevaerts(and controllable)
11:23:52 Join flydutch [0] (
11:24:42gevaertsI guess that this just dates back to times when an advanced WPS meant that you showed both metadata *and* a progress bar...
11:25:23JdGordonso its maybe something to think about fixing?
11:26:21gevaertsIf nobody can give a good reason, I'd consider it a bug, yes
11:26:33 Join einhirn_ [0] (
11:29:07linuxstbIs it perhaps to enforce drawing images after any text? So if text clears a "line", the image isn't wiped?
11:29:15CIA-5New commit by 03funman (r25243): sd-as3525v2: sd_get_info() is already in common SD code
11:29:18CIA-5New commit by 03funman (r25244): Clip+: correct usb product id
11:29:50 Quit scorche|sh (Ping timeout: 246 seconds)
11:29:52 Quit einhirn (Ping timeout: 246 seconds)
11:30:09 Quit anewuser ()
11:37:27 Quit m3dlg (Ping timeout: 276 seconds)
11:38:44wodzis there central page with statistics how buildclients perform?
11:51:37 Join evilnick__ [0] (
11:52:45 Join ender [0] (
11:53:52 Join funman [0] (
11:53:55 Quit funman (Changing host)
11:53:55 Join funman [0] (~fun@rockbox/developer/funman)
11:55:00 Quit evilnick_ (Ping timeout: 256 seconds)
11:55:02 Quit ender` (Ping timeout: 246 seconds)
11:59:33 Join adnyxo [0] (
12:17:17 Nick evilnick__ is now known as evilnick (
12:17:39 Quit evilnick (Changing host)
12:17:39 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
12:21:07 Quit RadicalR (Quit: Nettalk6 -
12:23:20 Join MethoS- [0] (~clemens@
12:28:55***Saving seen data "./dancer.seen"
12:30:03 Join webguest47 [0] (
12:31:06 Quit webguest47 (Client Quit)
12:31:27 Join m3dlg [0] (~m3dlg@
12:37:11 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
12:37:23funmani get USB interrupts on my fuze but both device & endpoint irq status registers are 0x0
12:48:06 Quit m3dlg (Ping timeout: 276 seconds)
12:53:57 Join watto [0] (~watto@
12:59:42 Join DV__ [0] (~DV@
13:00:29 Join PaulJam [0] (
13:03:51 Quit DV_ (Ping timeout: 268 seconds)
13:05:36pamauryNobody interested in FS #11118 about the FAT driver ?
13:09:53TorneI am, but haven't had time to test it yet
13:18:25ranmachanfunman: Maybe it's the IRQ_ENRD0 usb status change interrupt?
13:19:25ranmachanBTW, if I enable DEBUG build, I get a compile error in apps/gui/skin_engine/skin_parser.c
13:19:36ranmachan(debug_wps undefined)
13:20:38funmanranmachan: should it call the i2c isr?
13:20:53funmanranmachan: slap JdGordon for that :)
13:21:54*FlynDice has found clip+ uSD and manages to get the card to STBY state but not TRAN yet. The card makes it through init just fine though.
13:22:05ranmachanfunman: Good question, that's one of the things I'd like to try, but haven't tried yet.
13:22:26ranmachanMeanwhile I had fun rewriting ascodec to use interrupts :)
13:22:42funmanFlynDice: good! did you look at the FIXME for CGU_BASE+0x3c clock register?
13:23:17FlynDicefunman: No have'nt seen that yet
13:24:45funmanranmachan: nice
13:26:29 Join JohannesSM64 [0] (
13:28:05ranmachanIs there an easy way to postpone the ascodec_write in system_init()?
13:28:15ranmachanOr is it maybe safe to enable interrupts at that point?
13:28:39ranmachanATM I have a special polling version of write just for that...
13:29:57funmanranmachan: enabling interrutps should be safe after setup_vic()
13:30:00 Join froggymana [0] (
13:30:22funman(i was wondering why polling was needed and why system_init() used it)
13:30:27pamauryI just saw "FS #10112 - Rockbox abort: search of title when the database is created by illegal codepage. " which is bug related to the fact that mp3 tags have no preferred code page. There has been a discussion recently about that no ?
13:30:36funmani believe rockbox enable interrupts in kernel_init()
13:32:14ranmachanHmm, or rather than just interrupts, since I use the wakeup system I think the kernel also has be initialized sufficiently at that poit.
13:35:06ranmachanWould it be acceptible to introduce system_init_late() in main.c and do the ascodec_write from there?
13:37:40funmanor move it into power_init() for example since it's realted to power
13:38:43funmanand call enable_irq() in system_init() because power_init() is called just before that in main()
13:41:15FlynDicefunman: you're a genious! or have supernatural powers... I added the shift for CGU_BASE+0x3c and am now browsing the uSD!
13:42:36funmanjust wild guessing, i hadn't tried to touch again this register fearing it wouldn't work ..
13:43:29FlynDiceNeeds a little cleanup first but I'll commit shortly
13:43:34*funman waits eagerly
13:43:44topikprint business cards: funman, wild guessing genius
13:44:35funmantopik: i had guessed you'd do that!
13:44:51topiksure thing. pick them up later
13:45:26topikgreat job on the newer samsas guys
13:45:28funmanFlynDice: hm i wonder if the bootloader still works after this change
13:46:05funmani guess (!) that everything will be clear once we understand how CGU_PERI divider works
13:46:19FlynDiceI'll test that before I commit
13:47:21funmani can test if you want, if it doesn't work let's just put a #ifdef BOOTLOADER
13:49:22ranmachanHmm, the i2c controller sucks. For safety I still have the 'while (i2c_busy())' in there, but I was hoping that since it raised the interrupts it wouldn't be busy anymore when we get around to submit the next command (when another one is queued)
13:50:13ranmachanBut: read_ctr=1267 write_ctr=200 busywaits_after_read=1779 busywaits_after_write=13794
13:50:47funmanranmachan: is there a noticeable speed difference with current method?
13:51:08funmani guess if wakeup_wait() schedules another thread we can't compare decently
13:52:16ranmachanWell, in theory the cpu can spend time in another thread while the i2c controller is busy instead of waiting.
13:52:21S_a_i_n_tw00t! Symmetry is now multilingual!
13:52:45ranmachanI haven't cross-checked the above busywait measurement with the unpatched version yet.
13:55:50rasherHow well does charging work on the e200 these days? As fast as in OF?
14:03:02 Join Connor_ [0] (
14:03:08 Quit einhirn_ (Quit: Miranda IM! Smaller, Faster, Easier.
14:05:09 Join Connor__ [0] (
14:06:21 Quit Connor1 (Ping timeout: 260 seconds)
14:08:06 Quit Connor_ (Ping timeout: 264 seconds)
14:09:04ranmachanHmm, if I increase the i2c clock from 400KHz to 4MHz I have less busywaits :)
14:13:33 Quit JdGordon (Quit: Leaving.)
14:22:43wodzI am at the point that bootloader loads rockbox image to the memory and complains about bad checksum
14:23:43 Join Schmogel [0] (
14:24:32 Join hebz0rl [0] (
14:25:17CIA-5New commit by 03FlynDice (r25245): SansaAMSv2: Add support for uSD cards to sd-as3525v2.c and activate hotswap and multidrive.
14:25:20 Join TheSphinX^ [0] (
14:28:56***Saving seen data "./dancer.seen"
14:33:13FlynDicegreen, back later!
14:35:14 Join ichthys [0] (~jaaron@
14:39:42 Nick fxb__ is now known as fxb (
14:40:04funmanwodz: rockbox firmware field has 2 fields to check if it loaded correctly from storage: length & a checksum
14:40:26funmanlength is at position 0 + FIRMWARE_OFFSET_FILE_LENGTH of the file, and data at position 0 + FIRMWARE_OFFSET_FILE_DATA
14:40:36funmanranmachan: isn't 400kHz the maximum authorized for i2c ?
14:41:33wodzfunman: and where is CRC stored?
14:42:11funmanhave a look at bootloader/common.c: load_firmware()
14:42:53wodzfunman: from load_firmware() it seems that file length is not stored in firmware
14:43:37funmanah sorry, it stores the model number instead
14:44:11funmanlength is used in firmware/rolo.c on archoses
14:44:27wodzfunman: so order is CRC MODEL DATA right?
14:44:43 Join fred_99 [0] (~fred@
14:45:34funmanwodz: looks like it (also have a look at tools/scramble.c which add this header)
14:45:46 Join stoffel [0] (
14:45:49 Join n1s [0] (~n1s@rockbox/developer/n1s)
14:48:43ranmachanfunman: Yeah, but I thought 'hey, it's both on die, let's try something higher :)'
14:48:43 Join evilnick_B [0] (~0c140464@rockbox/staff/evilnick)
14:49:02ranmachan4MHz seems to be the upper bound though, with 8MHz it doesn't boot.
14:49:26funmanFlynDice: nice, works fine even in bootloader! thanks :)
14:50:12fred_99battery time left on an X5 : are the CURRENT_something variable the only constant to setup to have a battery time left ?
14:50:29 Join m3dlg [0] (~m3dlg@
14:51:00wodzfunman: does it mean I have to hack scramble.c to add new target also?
14:51:03funmanfred_99: should be, try changing them to some value and see if the time changes
14:51:31funmanwodz: yep, look at svn log to see when new targets were added, you have to add the same model number than in configure i think
14:52:14wodzthis make clear why it doesn't boot than
14:52:31 Quit froggymana (Quit: CGI:IRC)
14:52:45fred_99funman: I set up CURRENT_NORMAL and CURRENT_BACKLIGHT with value like 1100mAh divided by the battery_bench time, and it seems to work.
14:54:11fred_99funman: I can put a patch if it can help.
14:54:13funmanfred_99: that's correct for CURRENT_NORMAL, CURRENT_BACKLIGHT is added to CURRENT_NORMAL to calculate current use when backlight is on
14:55:28fred_99funman: ah .... I didn't do this exactly .....
14:58:07funmanI can't write to µSD either
14:59:08 Quit Hillshum (Ping timeout: 264 seconds)
14:59:30funmanfred_99: CURRENT_NORMAL = 1100/hours_lcd_off, CURRENT_BACKLIGHT = 1100/(hours_lcd_off-hours_lcd_on)
15:01:29fred_99funman: for exemple (in my case), CURRENT_NORMAL = 1100mAh / 7h15 ~ 150 and CURRENT_BACKLIGHT = (1100mAh / 4h26) - CURRENT_NORMAL ~ 100
15:01:59fred_99funman: do you agree ?
15:02:30 Quit Rob2222 (Read error: Connection reset by peer)
15:02:33funmani have no opinion, post your runtimes with the patch on flyspray so we can check
15:04:00fred_99funman: ok .... thanks
15:11:12 Quit DV__ (Ping timeout: 260 seconds)
15:11:14 Quit DerPapst (Quit: Leaving.)
15:12:38 Quit m3dlg (Ping timeout: 258 seconds)
15:13:08 Join Rob2222 [0] (
15:13:25 Join DerPapst [0] (
15:13:42wodzfunman: what do You mean by 'the same model number than in configure'?
15:13:46 Quit DerPapst (Client Quit)
15:15:36 Join Farthen [0] (
15:16:12funmanclip+ has target_id=66 in tools/configure , and modelnum=66 in tools/scramble.c
15:21:00wodzfunman: look for irivers iriver h100 has target_id=11 in configure, and modelnum=1 in scramble.c
15:21:16wodzthe same for iaudios
15:23:23funmanah it's not in configure, it has to be the same MODEL_NUMBER defined in firmware/export/config/xxmodel.h
15:24:10funmanideally target_id should be MODEL_NUMBER
15:24:53wodzYUPI!!!! I booted rockbox on mpio hd200 :-)
15:26:39wodzand how should I choose MODEL_NUMBER?
15:27:32funmanunique number (grep for it in config/*.h to see if no other defines it)
15:28:37wodzAll I can say is that adding new target to rockbox is a real pain.
15:29:17 Quit robin0800 (Ping timeout: 265 seconds)
15:29:37wodzI am not talking about hardware and so on but on source structure and documentation
15:30:22funmanwodz: did you see ?
15:30:57wodzof course
15:32:29funmani agree there is room for improvements but once you've finished adding a target you probably want to work on it instead of on the target tree
15:32:35 Join saratoga [0] (~9803c6dd@gateway/web/freenode/x-kyfhztesikadsffz)
15:32:41 Join robin0800 [0] (
15:32:49saratogadoes anyone have an actual arm-linux system they can test something on?
15:33:22funmani have arm926-ejs not too far, just need 5 minutes to plug it again
15:34:14saratogafunman: would you mind trying my latest libtremor patch when you get a chance
15:34:29saratogasee if it gets correct output and maybe bench mark it if thats not too much trouble
15:34:36 Part LinusN
15:34:47 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:35:50saratogait works on QEMU but trying real hardware would be nice, plus it takes 30 minutes to compile in QEMU :)
15:37:46n1ssaratoga: i'm impressed you do the work to push your improvements upstream
15:37:51 Quit hebz0rl (Read error: Connection reset by peer)
15:39:30kugelFlynDice: good work! :) I just noticed the fuzev2 OF sets up GPIOA isr too but didn't think further about it, looks like it was for microsd
15:39:59gevaertssaratoga: I have some. I can try later today
15:41:29funmansaratoga: is there a sample program to test output/benchmark in tremor svn ?
15:42:14gevaertsOne arm5 and one arm7
15:44:32funmansaratoga: ogg_int32_t isn't defined in asm_arm.h (needed for MULT32 proto)
15:46:08funmankugel: you could skip lcd code and enable buttonlight only in rockbox.sansa to see if storage works
15:46:38kugeland how does that help?
15:47:05funmanpower on fuzev2: if buttonlight is on, storage works
15:48:45kugelI assume it works, whether it does or not doesn't get me further though
15:49:28 Quit funman (Quit: free(random());)
15:50:08 Join toffe82 [0] (~chatzilla@
15:51:14kugelI have the strong feeling that the fuzev2 also has slighly different versions, it does something strange with GPIOA 0 and 5
15:51:16 Quit ranmachan (Ping timeout: 256 seconds)
15:51:21 Quit simabeis_ (Ping timeout: 258 seconds)
15:51:26 Quit wodz (Quit: CGI:IRC)
15:51:31kugeland I think backlight is also done through sw pwm, but I don't know for sure
15:51:41 Quit bzed (Ping timeout: 248 seconds)
15:52:12 Quit knittl (Ping timeout: 245 seconds)
15:55:03 Join knittl [0] (
15:56:41 Join bzed [0] (
15:57:22 Join CGL [0] (~CGL@
15:58:22 Join ranmachan [0] (
15:58:57 Join simabeis [0] (
16:04:16 Join Strife1989 [0] (
16:09:21 Nick fxb is now known as fxb__ (
16:09:32 Join hebz0rl [0] (
16:13:34saratogafunman: sorry, "make example" then "./ivorbisfile_example < file.ogg > out.raw"
16:14:10 Quit TheSphinX^ (Quit: XChat)
16:15:55saratogawhich linux distro was this tested on?
16:18:43saratogafunman: could you try commenting out the include to asm_arm.h in fft-ffmpeg.c?
16:18:50saratogai think that was added by mistake at some point
16:19:40saratogai put the instructions to test here
16:19:47saratogaif anyone else wants to try
16:24:39kugelGPIOA_4 seems to indicate some fuzev2 revision
16:25:03merbztsaratoga: how much faster is it compared to the tremor mdct ?
16:26:59saratogamerbzt: about 10% faster verses are well optimized version of the tremor IMDCT
16:27:15 Join DerPapst [0] (
16:27:19saratogasorry I mean vorbis gets 10% faster
16:28:02saratogaoverall its probably something like 20-25% faster for just the mdct, verses stock tremor's mdct (no rockbox optimizations) probably something like 50% faster
16:28:07merbztno change in memory/code footprint ?
16:28:17saratogait uses 2KB more RAM, though that can probably be fixed
16:28:24saratogaor at least reduced
16:28:57***Saving seen data "./dancer.seen"
16:29:01saratogai don't remember the exact numbers, but I think we saved something like 3MHz over stock mdct, then when we rewrote it we got another 3 or so MHz faster
16:29:10 Quit mikroflops (Remote host closed the connection)
16:29:31saratogabut i'd love to see some numbers for the patch above, not just extrapolations from rockbox's hacked up version
16:30:07merbztcool work anyway
16:31:11merbztwas the half_mdct way of doing stuff to hard to implement?
16:35:04 Join m3dlg [0] (~m3dlg@
16:35:18 Quit Strife1989 (Remote host closed the connection)
16:39:17saratogamerbzt: not too hard I think, but its not a huge improvement so we haven't gotten around to it
16:39:31saratogatheres still a lot more places this code could be optimized
16:39:38saratogai expect to squeeze another 2-3 MHz out of it over time
16:39:58saratogaeventually folding in the windowing code to the last part of the IMDCT would probably help a lot
16:40:09saratogaliba52 in rockbox already does this
16:40:31saratogawould save a loop and probably some load/stores
16:57:12 Quit mt (Ping timeout: 240 seconds)
16:57:33Unhelpfulsaratoga: didn't i say that and you said that it's generally accepted the best way is to window separately? :P
16:57:47saratogaUnhelpful: I don't think so
16:58:04saratogafolding it in obviously saves you at least some loop overhead, its just much harder to do in a general way
16:58:22saratogasince different codecs have different blocks
16:58:35Unhelpfuland some load/store overhead given that at *some* point we may have to deal w/ data that is not in iram?
17:00:02merbztyou might trigger register starvation also
17:00:19saratogaregister starvation is a forgone conclusion on arm
17:02:06saratogarockbox optimization would be a lot more fun if MIPS had won the embedded space
17:02:16Unhelpfuland i imagine the code in question would be asm, so we can decide carefully how much of what to load, and when
17:02:41 Join funman [0] (~fun@rockbox/developer/funman)
17:03:45funmansaratoga: running the tests, example didn't build with the patch because of missing MB declaration (any program using the lib wouldn't link). I moved it to asm_arm.h and added back the memory barrier
17:04:22 Quit m3dlg (Ping timeout: 256 seconds)
17:05:10saratogafunman: did you have to manually define ARM_ASM?
17:05:24saratogait sounds like the make script didn't handle it for me in QEMU because it compiled and ran fine for me
17:05:29 Nick ender is now known as ender` (
17:05:59saratogahuh probably some QEMU weirdness then
17:06:01saratogadoes it build now?
17:06:22funmanyep, i'll paste the patch i used
17:07:28FlynDicefunman: sd writes are no longer disabled for sd-as3525v2.c is that correct?
17:07:37saratogaby the way, did it decode correctly?
17:08:02funmanmd5 differ
17:08:23 Quit robin0800 (Remote host closed the connection)
17:09:14saratogafunman: yeah it probably will since the transform uses a different algorithm, mostly i'm curious if it sounds like audio
17:09:17 Join komputes [0] (~komputes@ubuntu/member/komputes)
17:09:43saratogathe algorithm is correct so mostly i want to make sure I haven't made some horrible mistake porting it
17:10:17funmanif i use ffplay -f u32le i got a float exception
17:10:46 Join webguest04 [0] (
17:11:04saratogafunman: does "./ivorbisfile_example < file.ogg > out.raw" work?
17:11:15 Quit webguest04 (Client Quit)
17:11:43funmanyes, same output size for both
17:12:21saratogasox −−rate 44100 -c 2 −−bits 16 -s out.raw out.wav
17:12:25saratogasee if its audio or just crap
17:12:45saratogahopefully your distro has sox
17:13:28saratogasorry have to run
17:14:37funmanworks with -t raw
17:15:03 Join mt [0] (~chatzilla@
17:17:15funmansaratoga: results with my diff to your diff
17:17:49funmanthe patched version spends more time in user time it seems
17:19:53funmanlistened to a whole file: excellent audio
17:20:52funmanFlynDice: writes are still disabled : #if 1 return -1 # else sd_transfer_sectors(...) #endif
17:21:54FlynDiceyes, I just worked through that, enabling writes seems to not work at the moment...
17:31:34funmanoops i spotted an error
17:32:13funmanremoving the timeout check (max=0x40000) in send_cmd locks the clip, timeout failing is not detected
17:35:02 Join mikroflops [0] (
17:35:09 Quit n1s (Ping timeout: 252 seconds)
17:39:28saratogafunman: which arm cpu is this anyway?
17:40:23CIA-5New commit by 03funman (r25246): sd-as3525v2: correctly check send_cmd return value (which is boolean)
17:42:15funmanhm write is still locking up
17:43:32saratogafunman: do i need to do something special to apply that patch?
17:43:51saratogaoh its a diff against the patch not the code
17:44:43funmanbtw why did you remove memory barrier?
17:45:06saratogai didn't realize i did
17:45:15saratogaprobabaly just got lost when syncing the files
17:45:28saratogafunman: could you diff against the source tree, I can't figure out how to apply this patch
17:45:59saratogaplus its mostly just changing the patch headers to french :)
17:50:10 Join B4gder [0] (~daniel@rockbox/developer/bagder)
17:50:24funmanwrite still locks up if i remove CMD_CHECK_CRC_BIT, other threads are still running
17:50:34 Quit pamaury (Quit: Quitte)
17:53:22saratogafunman: maximum size of that pastebin was exceeded? was there anything changed beside misc.h and asm_arm.h? i can fix those myself
17:53:37 Part knittl
17:53:46 Join phanboy_iv [0] (
17:54:10funmani don't think so
17:55:48funmani moved MB from misc.h to asm_arm.h, commented out include asm_arm.h in fft-ffmpeg.c
17:56:02saratogasounds good
17:56:14saratogaany chance you could try this patch before I send it to the mailing list?
17:58:07saratogahuh did the patch actually get slower on your system then stock? 37.51s vs. 42.61s
17:59:21funmanlooks like it but i made only 1 run though
17:59:54saratogathats odd, it was faster on x86
18:01:53 Quit petur (Quit: work->home)
18:08:07 Join jfc^2 [0] (
18:08:20 Join __arbingordon [0] (
18:09:38funmansaratoga: only user time is significative, right?
18:10:17funmanTorne: any chance you could try mkamsboot on last released clipv2 OF ?
18:11:27Tornenot until the weekend; at work atm and won't be home tonight
18:11:34saratogafunman: yeah but the patched one took more user time :)
18:11:43funmanoriginal, 3 runs: 39.84, 40.82, 40.77
18:11:43 Quit _arbingordon (Ping timeout: 264 seconds)
18:11:43 Quit jfc (Ping timeout: 264 seconds)
18:11:53funmanyep :/
18:12:08funmanpatched one: 1st run 40.65
18:13:14 Join bertrik [0] (
18:14:34 Quit bertrik (Client Quit)
18:14:36funmanso it's only marginally faster, not even clearly faster than original
18:15:16 Join bertrik [0] (
18:15:43saratogafunman: I don't understand why, in rockbox is it substantially faster on the CPU
18:15:52funmanhm i didn't specify -mcpu flags
18:16:03saratoga"on that cpu"
18:16:10bertrikfunman, does the as3525v2 really have 1 MB IRAM?
18:16:25funmanbertrik: the Clip+ at least, yes
18:16:35funmanbut it looked a bit slower than DRAM
18:17:17saratogafunman: did you compile the stock tremor, or use one that came with the OS?
18:18:26bertrikok, there was some joke going around about it having double the clipv1 size (640k) and I thought that maybe an mistake was made thinking that 0x100000 was twice as much as 0x50000
18:19:25funmantrue i made the mistake, but i also made it when testing so i really tested 0x100000
18:19:52funmani really deserve slapping for that :p
18:20:06kugelI need assistance ... I can't get this lcd to work
18:20:10 Quit bertrik (Client Quit)
18:20:38funmansaratoga: testing with CFLAGS=-O3 -mcpu=arm926ej-s
18:20:47kugelI don't even know whether SSP is involved or not. lcd_init_device does messes with ssp, but the other lcd functions look similar to the fuzev1 ones
18:23:40funmankugel: could it be that there are 2 different screens ?
18:23:42kugelif someone wants to help me, I can offer my disassembly, my local code and/or a spare fuzev2
18:24:45kugelit looks like the OF looks gpio4 for some fuze revision; but I can only see references to the corresponding global variable in the init functions
18:25:01kugellcd_window_*, clear_display() all deal with dbop
18:25:33 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
18:25:42funmansaratoga: with -O3 -mcpu=arm926ej-s: original medium time (3 runs) : 37.59s, patched medium time (3 runs) : 34.72s
18:26:28saratogafunman: ok so its a little faster
18:26:32saratogawhats the clock speed btw?
18:27:02funmannot sure, 250 or 300
18:27:11saratogaand the track length?
18:27:41saratogaoh 246 seconds
18:28:21funman says 256.6-297MHz so I suppose 297MHz
18:28:27kugelI should check if the dbop is activated properly
18:28:33saratogaso you're doing 42Mhz realtime decoding
18:28:50saratogai suppose the percentage difference is small since the rest of the decoder is much slower then in rockbox
18:29:00***Saving seen data "./dancer.seen"
18:29:14funmankugel: you have 2 fuzev2 ?
18:29:41funmansaratoga: you want another test, or can i shut it down?
18:29:47saratogafunman: thats fine
18:29:54saratogai'd really rather you work on the clip anyway :)
18:30:20funmani am looking at usb on the fuze and ranmachan is looking at recordign on c200v2
18:30:43kugelfunman: yes
18:30:48funmanranmachan: btw i just though perhaps we could use i2c interrupts for headphone plug/unplug detection?
18:31:41 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:32:08funmankugel: i'd be glad if i can help you somehow, i'll give you my address
18:35:50 Join m3dlg [0] (~m3dlg@
18:37:11ranmachanI could test which interrupt the headphone detection generates I suppose, it's not necessarily going to be the i2c interrupt.
18:37:43ranmachanNote that "The detection is only working as long as the headphone stage is in power down mode"
18:38:07funmanhm so we can detect plug but not unplug
18:38:10ranmachanSo I suppose you can't detect unplugging the headphones while music is playing over the headphones
18:38:45AlexPpixelma: So where is the pink transparent then?
18:40:04ranmachanIRQ 9 is marked "AUDIO IRQ" in the vic interrupt sources, so I'd assume that's the one that'll get triggered
18:40:39funmanranmachan: i see you dumped the CGU_* regs while playing, powered over USB, how hard would it be to dump the regs with battery power, in the case the OF changes its clocking while powered by USB ?
18:41:43ranmachanHmm, I thought I also did that...
18:42:04funmani believe you had unplugged the battery at that time
18:42:40funman < i wanted to compare with what we do
18:42:55fred_99funman: when you say "CURRENT_BACKLIGHT = 1100/(hours_lcd_off-hours_lcd_on)" are you sure ? looking in the config files it seems that the constant CURRENT_BACKLIGHT is almost all the time smaller than CURRENT_NORMAL, which is not the case in your calcul. Can you confirm ?
18:43:48funmanfred_99: you're right, it's (1100/hours_lcd_off) - (1100/hours_lcd_on), i suck at maths :)
18:44:22fred_99funman: ok thanks ..... so do I ;)
18:45:31fred_99funman: the opposite should work perfect ;)
18:45:40funmanranmachan: also could you look at MCI_CLOCK : 0xC8000004 and 0xC8020004 ?
18:47:57pixelmaAlexP: in all the other bitmaps (besides backdrops) and not on greyscale
18:48:30AlexPOK, I've been under a misapprehension for years then :)
18:48:40funmanranmachan: i'll first look at those registers, if there is some significant difference we can assume it'll be the same with battery power
18:48:53pixelmaAlexP: you could have tried ;)
18:49:19AlexPyeah, could have
18:50:40 Join jgarvey [0] (
18:50:55funmanranmachan: ah 0x0-0x40 are mirrored at 0x40-0x80
18:51:20ranmachanYeah, I know :)
18:51:27ranmachanIIRC the Datasheet was claiming otherwise
18:51:55funmancould be sandisk diffs from the original design, they put a pl180 instead of the nand controller at least
18:52:46ranmachanHmm, or maybe not. Looks like I was just misreading it?
18:52:48gevaertssaratoga: I'll have some more results for you in a few minutes
18:53:02saratogagevaerts: thanks
18:53:12saratogagevaerts: does your system have IRAM?
18:53:39gevaertsany easy way to find out?
18:53:47gevaertsit's my phone
18:53:50funmanthey run plla at 216MHz and fclk to half of that: 108MHz
18:56:12 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:56:26ranmachanfunman: [edit3] added
18:56:39ranmachanMCI_CLOCK is 0x301
18:59:32funmanonly CGU_IRQ & CGU_INTCTRL change
19:00:20gevaertssaratoga: has my results. Numbers are user time in seconds, averaged over 5 runs
19:00:26funman+ bit 15 of CGU_PERI = SDMCI (i assume you had no card and it's on with USB mode)
19:00:30gevaertsThe output sounds as expected
19:00:47gevaertsI used the rockbox testfiles
19:01:18gevaertsI did similar things as funman to get it to compile, otherwise nothing special
19:01:34saratogabtw, do you have IRAM on that system?
19:01:40ranmachanfunman: I just put a write watchpoint on 0xC80F0014 and it's toggling between 0ef3c38d and 0ef3c3d1 (on battery, playering mp3)
19:02:35funmanmciclk means NO widebus, powersaving enabled, pclk/4 == 13.5MHz
19:02:41funmanFlynDice: ^
19:03:25 Quit m3dlg (Ping timeout: 246 seconds)
19:04:09funmanyes very interesting: it changes extmem_clk
19:05:04gevaertssaratoga: I'm trying to find out. It's an OMAP3430, which has a cortex-a8
19:05:40*gevaerts doesn't really know where to find this information
19:05:40funmanhm no, extmem_clk stays at 54MHz but pclk changes between 27MHz & 54MHz
19:05:48funmanranmachan: how fast does it toggle?
19:05:57ranmachanfunman: sub_1660 is doing the toggling
19:06:08saratogawell if you don't know its ok, was just curious if enabling it would make any difference
19:06:13ranmachanfunman: Good question
19:06:23gevaertsI also wouldn't know how to use it :)
19:06:40ranmachanAt 0x1712 and 0x171c are the writes
19:07:42funmanthere's also calls to code changing pll freqs in there
19:08:15funmanit asks for 65MHz but got the closer possible freq (54)
19:10:08funmanUSB phy is still enabled with usb unplugged
19:11:19funmanide clk at 108MHz
19:12:03funmanmemstick at 72MHz
19:12:49gevaertssaratoga: has the numbers converted to MHz as well
19:12:51 Join perfectdrug [0] (
19:12:59 Quit B4gder (Ping timeout: 256 seconds)
19:13:02ranmachanThe toggles look semirandom, if I can trust TIMER2...
19:13:17ranmachanAt least not at regular intervals.
19:13:37 Join captainewkl [0] (~2669ecc2@gateway/web/freenode/x-lzdaohivferjdoaj)
19:14:08funmanlike several time per seconds, or each 10 seconds?
19:14:08 Quit captainewkl (Client Quit)
19:15:26ranmachanSeveral times per second
19:15:46ranmachanFor each resume I'd hear a very short bit of music
19:16:02 Quit kugel (Remote host closed the connection)
19:16:39ranmachanTIMER2_CONTROL: 000000e0
19:17:27ranmachanTIMER2_VALUE between resumes: 00008867 000056f9 00001e8d 0000b5f5 00009993 00000bd2
19:19:07ranmachanHmm, at 16bit it might just be wrapping around multiple times though...
19:19:16ranmachanSince prescaler is 1
19:19:52 Quit CGL (Ping timeout: 246 seconds)
19:20:21 Join FOAD_ [0] (
19:21:28funmanwell anyway it's fast
19:21:34 Join panni_ [0] (
19:22:17 Quit tipi^ (Ping timeout: 246 seconds)
19:24:09 Quit FOAD (Ping timeout: 265 seconds)
19:24:09 Nick FOAD_ is now known as FOAD (
19:25:41 Quit Horscht (Quit: Verlassend)
19:30:51saratogagevaerts: wow your system is really fast
19:31:52 Quit DerPapst (Quit: Leaving.)
19:32:23fred_99sorry ... .I'm really stupid but I don't see how to submit a patch in FlySpray. Can anybody help me
19:32:39 Quit funman (Quit: free(random());)
19:34:11saratogafred_99: add new task button on the flyspray page
19:36:41fred_99saratoga: I don't see this button on my page
19:36:55 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:36:57mtfred_99: Are you registered and logged in ?
19:37:27fred_99saratoga: I'm not sure, I received the mail and confirm, but log or not my page is the same
19:37:58saratogafigure out if you're logged in or not, then come back here if it still doesn't work
19:38:29fred_99how do you see if you are logged in ?
19:39:00fred_99I will try with another browser .... and see what appens
19:41:04saratogawhat are you submitting on flyspray?
19:42:02mtfred_99: Also, look right under the rockbox logo on flyspray, if you see empty text fields you're not logged in.
19:42:27 Quit hebz0rl (Quit: Ex-Chat)
19:43:13fred_99mt: thanks .....It's ok now
19:44:13fred_99saratoga: I don't know how to read ..... thats it ...... confirmation code pre-filled with my email ..... but now it's OK
19:44:28fred_99I have the famous button
19:44:35 Join gill0r [0] (
19:45:10mtGood news everyone
19:45:16mtRockbox accepted for gsoc
19:45:18fred_99the wiki says it's better to ask here before submitting a patch, so I will
19:46:14Bagdermt: cool!
19:46:24 Join captainewkl [0] (~2669ecc2@gateway/web/freenode/x-yzkdbowswcwgvhwr)
19:47:00fred_99my patch only add the CURRENT_NORMAL and CURRENT_BACKLIGHT values to the iaudiox5.h file
19:47:22fred_99is it OK to submit ?
19:47:49fred_99saratoga: OK .... so I try
19:48:31AlexPmt: Cool - where is that said?
19:49:05 Quit __arbingordon (Quit: `)
19:49:07mtAlexP: I just asked danderson on #gsoc
19:49:25BagderI got a notification about it too it seems
19:49:26AlexPah right, good stuff. I just noticed that the list hasn't been published yet :)
19:49:38AlexPexcellent :)
19:49:51Bagder"Your Organization Application for "Rockbox" in Google Summer of Code 2010 has been accepted."
19:50:01fred_99saratoga: do I have to choose 3.4 when I did it with the SVN ?
19:50:36Bagderlet me rephrase that: I got 7(!) notifications about this
19:50:37saratogafred_99: if this is so difficult maybe you could just tell me what the current normal value is and i will type it myself
19:50:39Bagderall identical
19:50:41mtBagder: Beware .. you might receive like 6+ more acceptance mails. There's some bug with their automailer. :)
19:52:00 Quit stoffel (Remote host closed the connection)
19:52:31AlexPmt: So are you thinking of applying by any chance? :)
19:53:11fred_99saratoga: sorry to bother ..... it was suppose to be the first, and certainly would have been the last ..... as you like CURRENT_NORMAL = 152 and CURRENT_BACKLIGHT = 96
19:53:27saratogafred_99: i don't think those values are reasonable
19:53:46mtAlexP: Probably yes :). I'll just have to check some stuff by the end of the month to be able to determine whether I'd have enough time.
19:53:56Bagdermt: I don't get any mails, I get them in their weirdo gsoc web app
19:54:12Bagderah no
19:54:17Bagderwrong, I get mails too
19:56:02fred_99saratoga: reading a bin powermgmt.c .... and listening to funman , and testing it .... OK for not so long to trust it ;) ..... it was my calcul
19:56:44saratogai think you're off by at least a factor of 2
19:56:45fred_99saratoga: I keep it and in fue days when I will know more I will submit the patch
19:57:04saratogahow did you calculate that value?
19:58:02fred_99 CURRENT_NORMAL = 1100mAh / 7h15 ~ 150 and CURRENT_BACKLIGHT = (1100mAh / 4h26) - CURRENT_NORMAL ~ 100
19:58:50 Join petur [0] (
19:58:50 Quit petur (Changing host)
19:58:50 Join petur [0] (~peter@rockbox/developer/petur)
19:59:01saratogafred_99: the wiki lists 17+ hours for the X5, why did you use 7 hours?
20:01:00fred_99fred_99: because it should be for a x5L but mine is x5 , and I have rockbox for several years now and changed the battery once, and never be abble to lissten to music more than 6 or 7 hours
20:01:54 Join DerPapst [0] (
20:01:57saratogathe battery type doesn't change the amount of power the player uses
20:01:57fred_99saratoga: so the read on the battery_bench file for me looks good
20:02:09fred_99I agree
20:02:11saratogaand if your player is only getting 6 or 7 hours theres obviously something wrong with your battery
20:02:29saratogaso assuming your broken battery works as well as a new battery is not a very good idea
20:02:50fred_99I only use rockbox .....
20:03:05fred_99I already changed my battery
20:03:15fred_99soldering and everything
20:03:36saratogais your new battery half the capacity of the old one?
20:03:39fred_99and with the new ol one
20:03:55AlexPGSOC list is out btw:
20:04:00gevaertsfred_99: maybe you've already said this, but which codecs do you use?
20:04:01fred_99and the new new one ..... the working time was the same
20:04:05saratogalooking at the wiki, it looks like maybe 65mA and 25mA for the backlight
20:04:21gevaerts says 8 to 12 hours for the X5
20:04:22fred_99let me check
20:04:54fred_99gevaerts: you mean for the test ?
20:05:04gevaertsor in general
20:06:32pixelmathe X5Ls should have 35 hours (according to the manufacturer and considering that my M5L got almost 52 hours with fresh batteries...)
20:06:40fred_99gevaerts: in general it depends a lot
20:07:56 Quit Strife89 (Read error: Connection timed out)
20:08:06fred_99pixelma: I guess I have a wrong config from the beginning
20:08:30 Join Strife89 [0] (
20:08:31fred_99gevaerts: let me 2 minutes I check wich codec I used
20:08:32saratogaK5L is a 2200mah battery?
20:08:38gevaertsfred_99: also, do you use EQ?
20:08:41pixelmanah, the Iaudio batteries seem to degrade relatively quickly
20:08:49fred_99gevaerts: no
20:09:26fred_99gevaerts: standard wps a bit modified
20:09:45saratogaboth benches I've seen look to use about 65 mA so i'm going to use that
20:11:34fred_99saratoga: next time I will change my battery, I will do a bench
20:11:34 Quit Strife89 (Client Quit)
20:11:34saratogaok just put it on the wiki or something, no need to tell me about it
20:11:45CIA-5New commit by 03saratoga (r25247): Add runtime estimation for the iaudio X5.
20:12:57fred_99gevaerts: ogg is OK or you need the codec
20:12:58fred_99gevaerts: sorry I'm always confusing between container and codec
20:13:20gevaertsin this case it's clear enough :)
20:15:29ThomasAHFlynDice: I can confirm that after r25245 reading from an 8GB SDHC card works fine in the Clip+ 8GB
20:16:09gevaertsfred_99: unless you use really high bitrates, vorbis should give reasonably good runtimes. mp3 might give you a bit more though
20:16:14 Quit flydutch (Quit: /* empty */)
20:17:14Bagdersome happy fellow could update /topic
20:17:19fred_99gevaerts: ffvorbis
20:17:46fred_99gevaerts: ~240 kbps for the bench
20:18:34fred_99gevaerts: I suppose it depends of the frequency ..... but MP3 128 it's no way for me
20:18:54gevaertsfred_99: have a look at for the full details
20:18:59fred_99gevaerts: I don't have dog ears but not less than 200
20:19:06gevaertssure :)
20:19:32fred_99gevaerts: it was VBR of course
20:20:03gevaertsBasically, with the current rockbox code, on coldfire mp3 needs less CPU than vorbis, for similar quality
20:20:17Topic"Please read before speaking: | Please direct offtopic/social chat to #rockbox-community | Rockbox has been accepted for GSOC 2010! Potential students see" by ChanServ (ChanServ@services.)
20:20:41fred_99gevaerts: thanks I will have a look, but I won't reencode my CD's which are in ogg ..... because I don't think at all I cann double the playing time
20:20:41gevaertse.g. 320kbps mp3 needs about as much (or even slightly less) as 96kbps vorbis
20:20:54 Quit phanboy_iv (Read error: Connection reset by peer)
20:21:11gevaertsoh, it definitely won't double. Basically anything that's less than 50MHz on that table is reasonable
20:21:30fred_99gevaerts: not an integrist but mp3 is not free, and quality in lower
20:21:40gevaertstry musepack then :)
20:21:47fred_99gevaerts: but not the subject
20:22:06domonokywavpack or any other lossless codec is even better :-)
20:22:24gevaertsdomonoky: not on a hard disk player I'd say
20:22:38gevaertsAlso not according to that table :)
20:22:46pixelmaespecially on those that only have 16MB RAM
20:22:58gevaertswavpack needs more CPU than mp3 or mpc
20:23:04domonokygevaerts: its better for quality, not for runtime.. you have to make some tradeoff :-)
20:23:19 Join stooo [0] (
20:24:06fred_99I will have a look ..... thanks
20:24:22pixelmaflac is very efficient to decode but I don't know how much it makes up for the disk spinnimg
20:25:56 Part stooo
20:28:31fred_99last thing , on the iaudioruntime page there is a battery bench file from PhilipBarton showing something like 7 hours too
20:29:01***Saving seen data "./dancer.seen"
20:30:32fred_99and another showing 17h is with a CF instead of the HDD, and another stops at something like 7 hours and the guy write 17 hours in the table
20:31:06fred_99I will try my patch for some time and will come back
20:31:17 Quit liar (Quit: partey)
20:31:24 Quit fred_99 (Quit: Ex-Chat)
20:36:49 Quit PaulJam (Ping timeout: 240 seconds)
20:41:05 Join froggyman [0] (
20:49:51 Quit chaos (Read error: Operation timed out)
20:49:53 Quit Llorean (Quit: Leaving.)
20:50:18 Join chaos [0] (~chaos@gentoo/user/ch4os)
20:50:19 Join Llorean [0] (
20:50:19 Quit Llorean (Changing host)
20:50:19 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
20:53:42 Quit parafin (Ping timeout: 260 seconds)
20:56:32 Join parafin [0] (
20:56:37 Join MaadMan [0] (
21:11:25 Part watto
21:12:54 Quit pixelma (Disconnected by services)
21:12:55 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
21:13:09 Quit amiconn (Disconnected by services)
21:13:11 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
21:13:14 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
21:13:18 Join einhirn [0] (
21:13:37 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
21:20:57 Join raket2 [0] (
21:21:53 Join dottedmag [0] (~dottedmag@altlinux/developer/dottedmag)
21:22:06saratogawhats the file size limit for fat32 in rockbox?
21:22:40raket2hey peoples. strange problem. my battery dies when i run chkdsk on my ipod so I can never complete the scan. What say you? I have an error on the 120gb drive, so I need to fix it. (aside of replacing the battery of course)
21:23:08 Join m3dlg [0] (~m3dlg@
21:23:18evilnick_Braket2: Can you charge it in the OF first and then try?
21:23:26raket2OF ?
21:23:34evilnick_BOriginal Firmware
21:24:11raket2sure i could try. so charge it in the OF, but should i revert it to rockbox once done?
21:24:17saratogaare you running the scan in the OF?
21:24:24raket2i ran it under rockbox
21:24:27saratogai don't think doing it in rockbox is a good idea
21:24:44raket2oh crap. ok
21:24:45gevaertsdo the scan in the OF too, or you might get battery problems again
21:25:12saratogai thought fat32 had a 4GB limit?
21:25:26raket2ok, I shall try soon, and hopefully that will work
21:25:28evilnick_Bsaratoga: Sorry, it is 4GB. My bad
21:25:41AlexPraket2: Rockbox on your ipod doesn't charge at full speed over USB, so if you are doing disk intensive things the battery can run down while in use
21:26:20raket2AlexP: but it does charge at full speed under the OF ?
21:26:23AlexPAnd scanning for errors is pretty disk intensive :)
21:26:25AlexPraket2: yes
21:26:35raket2cool! bbl
21:29:22 Quit raket2 (Quit: Leaving.)
21:30:09 Join phanboy4 [0] (
21:32:39 Join GHF [0] (~meow@unaffiliated/ghf)
21:39:21 Join p3tur [0] (~petur@rockbox/developer/petur)
21:41:29 Quit JohannesSM64 (Quit: WeeChat 0.3.2-dev)
21:50:42 Quit bluebrother (Disconnected by services)
21:50:43 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
21:57:46 Join anewuser [0] (anewuser@unaffiliated/anewuser)
22:02:55domonokywe could use a new "project news" entry about the successfull gsoc news :-)
22:05:14 Join Bug2000 [0] (~bug@unaffiliated/bug2000)
22:05:21Bug2000I just found out rockbox lies.
22:05:59Bug2000I run in on my Sansa Clip v1 and it works perfectly fine.
22:06:26Bug2000Except for the fact that I reduced the volume up to -74db. At which point rockbox change the sound icon to no sound.
22:06:30 Part dottedmag
22:06:46Bug2000Yet, I can still hear the playback of the music. Which in other words, means it's not silent even though rockbox claims it to be.
22:07:09 Quit einhirn (Ping timeout: 246 seconds)
22:07:49LloreanSo mute isn't yet implemented on a port that isn't entirely complete yet?
22:07:54LloreanHave you filed a proper bug report on this?
22:09:33gevaertsMaybe if you turn it to maximum for a while, your ears will be rockbox compatible and you won't hear anything at -74
22:09:46Bug2000Llorean: I wouldn't care if it's not implanted, heck, even if it wouldn't have mute. I just have problems with it claiming it's mute :P
22:10:07LloreanBug2000: Then file a proper bug report.
22:10:11Bug2000gevaerts: lol. That's an interesting idea.
22:10:14Bug2000Llorean: K, sec.
22:11:00Bug2000Uha. >.< I forgot, I'm not using the latest version. Hopefuly I'll remember to upgrade in the morning and see if I can still hear at lowest sound level, if so I'll hopefully report it.
22:11:08Bug2000Right now I'm too tierd to do so. Sorry for bothering.
22:11:49Bug2000gevaerts: Still, it's pretty funny to hear at -74 as it's quieter then other sounds around. Whenever it's the clicks made by clicking the sansa clip buttons or the clock.
22:13:08 Join fred_99 [0] (
22:14:46 Quit fred_99 (Client Quit)
22:15:02 Join fred_99 [0] (
22:19:07 Join einhirn [0] (
22:26:13 Quit rvvs89 (Ping timeout: 260 seconds)
22:27:26 Join rvvs89 [0] (
22:29:02***Saving seen data "./dancer.seen"
22:30:59leavittxHmm, my third plugin is ready! A game this time: It's first battleship for rockbox, isn't it?
22:31:57domonokyleavittx: nice graphics !
22:33:01mtNice indeed. :)
22:33:05leavittxdomonoky: thanks! It's hand-made in gimp :)
22:34:05 Join [foo [0] (~Nadia@
22:34:50[fooHi there
22:36:28 Quit jordan` (Ping timeout: 240 seconds)
22:36:49leavittxBy the way: is there any chance to include it (or/and my other stuff) in official rb?
22:37:40gevaertsAssuming it has no licensing issues, it works, and it has documentation, sure
22:38:27[fooCould I ask gsoc questions?
22:39:11[fooI am interested in making rockbox an outstanding mobile app
22:39:32[foocross-platform (hopefully :)
22:39:42 Part froggyman
22:39:56domonokyleavittx: to get plugins commited: make sure they work on all platforms where possible, make a manual entry (pure text is enough) and bug us here to commit it :-)
22:40:10 Join jordan` [0] (~jordan@
22:40:51domonoky[foo: on which plattforms are you interessted ?
22:41:05 Quit m3dlg (Ping timeout: 258 seconds)
22:41:23[fooI already have some iphone knowledge
22:41:39[foobut if thats an issue I could take android
22:42:10domonokyiphone is not ideal, because we surely cant get into the appstore, so it would only work on jailbroken devices.
22:42:14gevaertsif the goal is to make things portable, that shouldn't make too much difference
22:42:27LloreanRemember that Rockbox is GPL and C, meaning that wherever you bring it, it must be able to accommodate the license and language.
22:42:38[foothats for sure
22:42:43[fooidea is that
22:42:43leavittxdomonoky: is it ok to have lots of bitmaps for different screen resolutions?
22:42:47domonokyso android is better, but there are also other possible targets. (WiMo, Maemo, other linux based phones) :-)
22:43:03domonokyleavittx: yes. take a look at how other plugins do it.
22:43:08[foowhy not to make a Model layer crossplatform
22:43:10gevaertsOr even just host-based PC apps
22:43:23[foowhile Controller and View - platform specific
22:43:29leavittxdomonoky: #if defined(BLAH) :)
22:43:40 Join Adubb [0] (~aldubuc@
22:43:41domonokyleavittx: there is a system in place to provide images of different resolutions.
22:43:54 Quit fred_99 (Quit: Ex-Chat)
22:44:24domonokyleavittx: i think the ifdef is only needed in the correct SOURCES file :-)
22:44:44 Quit evilnick_B (Quit: Page closed)
22:44:48gevaerts[foo: a lot will depend on how exactly you see this project. Everyone here probably has his or her own ideas about how this should work. Some think that something like the current sim is fine (i.e. sdl and our GUI code), while others want to integrate it more
22:45:08 Join fred_99 [0] (~fred@
22:45:16Llorean[foo: Rockbox is already fairly split into the apps and firmware layer, where apps is mostly more cross-platform style code while firmware is more device/hardware specific.
22:46:24domonoky[foo: do you have a rockboxed mp3player ?
22:46:46[foounfortunately, no
22:47:04[foojust iTouch and some not suitable transcend
22:47:37[foodo I need to have one?
22:47:44*domonoky recommends [foo to play a bit with some rockbox UI Simulator, to get a feel how rockbox is..
22:47:56leavittxdomonoky: ok, though defining different keymaps and bitmaps is quite difficult, I'll try to do that. Thanks (:
22:48:08 Join CGL [0] (~CGL@
22:48:23gevaertsno, certainly not for the app project
22:48:41[fooI`ll play with it anyways
22:48:46domonokythe rockbox as app project should probably mostly use touch input.
22:48:59LloreanDepends on the device.
22:49:22LloreanI know that if it were on my phone, I'd generally prefer to use the hardbuttons so I don't need to constantly take it out of the pocket to adjust things.
22:49:23domonokysure, if it has many buttons, we should use it. :-)
22:49:39*gevaerts tends to think that all this is not important :)
22:49:48LloreanThat being said, the D2 simulator (or other touchscreen device) would be a good place to start for general functionality, since *most* likely app devices will have a touchscreen and/or mouse input
22:50:03gevaertsGet the lower level infrastructure working right, the UI is a detail! ;)
22:51:01domonokyyes, the lower level is the important part for this project. But the UI sim is still good to get a feel of rockbox for people whithout rockbox experience and rockboxed dap :-)
22:51:20[fooso I thing the best path would be like split current APP layer into managable pieces which could be later reused with any external UI framework
22:53:21AlexPfred_99: Please don't test things here
22:53:52 Join robin0800 [0] (
22:54:12 Join Soap_Hotel [0] (
22:54:23Soap_HotelWoo Wee! Congrats on GSOC 2010!
22:54:30fred_99AlexP: sorry it's what i was thinking about
22:54:39*gevaerts still isn't convinced that redoing the GUI layer using whatever framework the target platform uses is a good idea
22:55:46Lloreangevaerts: I think in situations where Rockbox is likely to be full screen (many/most mobile devices) our current UI is pretty good / acceptable
22:55:55gevaerts[foo: as a general warning, mentors are by no means decided yet, so even if my name is currently next to this idea, that doesn't mean you have to listen to me
22:56:11AlexPNobody else does :)
22:56:12LloreanOn a PC, I'd much rather have something the ability to have something slimmer with a lot of options accessible through standard menus / widgets.
22:56:56gevaertsLlorean: that's part of my thinking. The other part is that we're slowly getting everything to be themable, using native widgets will throw that away
22:57:54LloreanI don't think they're at all necessary anywhere that Rockbox "takes over" the device as something that's either permanently fullscreen or minimized.
22:58:41gevaertsIf I were to do this project, I'd start with the threading model
22:58:46LloreanLike XBMC on PC has what I'd consider somewhat "Rockbox like" menus in the sense that it's a series of nested lists that can be navigated entirely with keyboard or entirely with mouse, and no native widgets at all.
22:59:21gevaertsyes, it's not as if there's no precedent for media player apps to ignore all common sense and do their own thing :)
23:00:02gevaertsAnyway, I'm also not opposed to make using native widgets possible
23:00:43gevaertsI just think that having an app that plays back audio without dropouts and with minimal CPU usage has slightly higher priority :)
23:00:58LloreanMemory usage on such devices is also important.
23:01:21gevaertsah, yes. The buffering code also could use work
23:01:51LloreanWell, I mean most devices can manage giving up 4-8MB of RAM that Rockbox can just monopolize, I'd imagine.
23:01:56LloreanBut it's probably not ideal.
23:02:11gevaertsAlthough, to be honest, there I'd wait for the buflib rework first. That should make it much easier to change things
23:02:23domonokythe buffering code works fine with atleast ~600kb buffer, so that should be doable on mobile phones :-)
23:02:32*gevaerts 's phone could afford to give more than enough to rockbox
23:03:06LloreanOne big thing would probable be the ability to support multiple screen sizes with a single build.
23:03:26 Quit fred_99 (Remote host closed the connection)
23:03:58LloreanAnd, I guess, configurable controls for devices with varying buttons or button IDs with a single build.
23:04:09 Join GeekShadow [0] (Antoine@reactos/tester/GeekShadow)
23:04:11LloreanI guess it's significantly just removing all the dependencies on one build = one hardware.
23:04:23gevaertsIs that really necessary? A full rockbox build is a few megabytes; having one build per common screen size, packaged in one executable, would work just as well I think
23:04:44LloreanWhat about devices that can rotate the screen, for example?
23:04:51*domonoky thinks that isnt really needed. we already manage 100+ builds, a few more dont hurt :-)
23:05:12domonokyLlorean: at least for the beginning, we can just not rotate :-)
23:05:22 Join Kitr88 [0] (
23:06:11gevaertsyes, while rotation is nice, I happen to own a device where the media player app doesn't support rotation while the webbrowser does. People seem to survive
23:06:31 Quit Kitar|st (Ping timeout: 240 seconds)
23:06:55LloreanWould it really be that difficult to support multiple screen sizes now?
23:07:11gevaertsit depends
23:07:26 Quit einhirn (Ping timeout: 264 seconds)
23:07:37LloreanWe've got viewports for most screens already, which basically restricts those screens to an imaginary screen size anyway, right?
23:07:57gevaertsFor plugins it wouldn't be straighforward at all I think, but for those we don't have to care much since most of them don't make that much sense in an app anyway
23:08:22gevaertsthe list screens should work just fine
23:08:23AlexPI wouldn't envisage most of the plugins being included here
23:08:42gevaertsIs there anything in rockbox that's not a list or a wps?
23:08:48AlexPMaybe the odd music related one, and they don't tend to have lots of bitmaps to deal with
23:09:19Lloreangevaerts: There's a few screens like the quickscreen or equalizer, but in effect they're still lines of text (or text-like bars)
23:09:25 Quit robin0800 (Quit: No Ping reply in 180 seconds.)
23:09:49 Join robin0800 [0] (
23:10:10gevaertsIf you use native widgets the question goes away obviously
23:10:16 Quit MaadMan (Quit: Verlassend)
23:14:13 Join fred_99 [0] (~fred@
23:14:59 Quit fred_99 (Client Quit)
23:22:09saratogaLlorean: a lot of things use defines for the screen size, doing that at runtime rather then compile time will be somewhat less efficient
23:22:28saratogaIIRC this was discussed in more detail some years ago for another player but i can't remember the details
23:22:30 Join fyre^OS [0] (
23:23:00 Join jfc^3 [0] (
23:23:34 Join ender [0] (
23:23:56 Nick ender is now known as ender| (
23:25:09Lloreansaratoga: Won't most of these things need to be adapted to runtime sizing anyway to work with viewports?
23:25:18LloreanEspecially with the themeing trends.
23:25:34 Join GeekShado_ [0] (Antoine@
23:25:35saratogai don't know the details
23:26:21 Join GeekShad__ [0] (~Antoine@
23:26:38 Quit Kitr88 (Ping timeout: 246 seconds)
23:26:38 Quit FlynDice (Read error: Connection reset by peer)
23:27:01 Quit jfc^2 (Read error: Operation timed out)
23:27:02 Quit fyrestorm (Ping timeout: 240 seconds)
23:27:04 Join FlynDice [0] (
23:27:08 Quit ender` (Ping timeout: 240 seconds)
23:28:00 Quit DerPapst (Quit: Leaving.)
23:28:41 Quit GeekShadow (Ping timeout: 245 seconds)
23:29:01 Quit moos (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
23:29:23 Quit amiconn (Disconnected by services)
23:29:25 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
23:29:51 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
23:30:08 Quit GeekShado_ (Ping timeout: 260 seconds)
23:30:31 Join bzed_ [0] (
23:31:05 Join MethoS-- [0] (~clemens@
23:31:09 Quit Soap_Hotel (Quit: CGI:IRC (EOF))
23:31:24 Quit MethoS- (Write error: Broken pipe)
23:32:20 Quit jgarvey (Quit: Leaving)
23:32:46 Quit rvvs89 (*.net *.split)
23:32:46 Quit p3tur (*.net *.split)
23:32:46 Quit FOAD (*.net *.split)
23:32:46 Quit komputes (*.net *.split)
23:32:46 Quit Schmogel (*.net *.split)
23:32:46 Quit advcomp2019_ (*.net *.split)
23:32:46 Quit Galois (*.net *.split)
23:32:46 Quit Utchybann (*.net *.split)
23:32:46 Quit Hadaka (*.net *.split)
23:32:47 Quit jvd (*.net *.split)
23:32:47 Quit beta2k (*.net *.split)
23:32:54 Join Kitar|st [0] (
23:34:00 Quit bmbl (Quit: Bye!)
23:35:57 Join einhirn [0] (
23:36:10 Quit einhirn (Client Quit)
23:36:11 Join phanboy_iv [0] (
23:36:11 Quit bzed (Ping timeout: 260 seconds)
23:36:17 Join rvvs89 [0] (
23:36:17 Join p3tur [0] (~petur@rockbox/developer/petur)
23:36:17 Join FOAD [0] (
23:36:17 Join komputes [0] (~komputes@ubuntu/member/komputes)
23:36:17 Join Schmogel [0] (
23:36:17 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
23:36:17 Join Galois [0] (
23:36:17 Join Utchybann [0] (
23:36:17 Join Hadaka [0] (
23:36:17 Join jvd [0] (~syscrash@poipu/developer/syscrash)
23:36:17 Join beta2k [0] (
23:36:23 Nick bzed_ is now known as bzed (
23:36:43[fooso the main idea of rockbox app is to make it native on a) x86 b) some portable device
23:36:55[fooam I right?
23:37:33Bagderthe main idea is to run rockbox as an app, as a guest in another OS
23:38:06Bagderthere are several such OSes
23:38:25 Quit GeekShad__ (Ping timeout: 258 seconds)
23:38:43 Quit phanboy4 (Ping timeout: 256 seconds)
23:38:45[fooi.e. current UI SIM -> native *nix/Win -> some embedded OS (IphoneOS, Android..(
23:39:08 Nick ender| is now known as ender` (
23:40:02BagderI don't think so, no
23:40:13Bagderif thouse "->" implies an order of work
23:40:40[foobut this is stated on current gsoc page
23:40:41linuxstbBagder: That order was my suggestion (and others agreed, or at least didn't object...).
23:40:54*Bagder doesn't keep up
23:41:00Bagderwell, that's one way to do it
23:41:13[fooSuggested goals
23:41:13[fooA mid-term goal could be to undertake the refactoring of the existing Rockbox code and produce a Rockbox application capable of running in a dexktop environment using SDL
23:41:13[fooThe remaining part of the summer could be spent porting this Rockbox application to a portable device.
23:41:28 Quit domonoky (Read error: Connection reset by peer)
23:41:50Bagderso if you read that already, what is the question again?
23:42:19gevaertsI think it doesn't make much difference. The first steps are not about the target OS but about the code structure and kernel work anyway I think
23:42:32Bagderyes I agree
23:42:58linuxstbYes, that was my intention. i.e. don't get bogged down in the details of a particular target - do the important work (restructuring Rockbox) first.
23:43:21[fooso, what's the best starting path?
23:43:31gevaertsAs soon as the target tree work is done, and it doesn't use the preemptive-multithreading-with-only-one-running-thread style of cooperative multithreading anymore, you can start thinking about the target you like
23:44:25[foook. I think I got your point
23:46:15[foothanks for the info. I'll be back after checking the SIM and the code
23:46:24gevaertsgood plan :)
23:46:26 Quit rvvs89 (*.net *.split)
23:46:26 Quit p3tur (*.net *.split)
23:46:26 Quit FOAD (*.net *.split)
23:46:26 Quit komputes (*.net *.split)
23:46:26 Quit Schmogel (*.net *.split)
23:46:26 Quit advcomp2019_ (*.net *.split)
23:46:26 Quit Galois (*.net *.split)
23:46:27 Quit Utchybann (*.net *.split)
23:46:27 Quit Hadaka (*.net *.split)
23:46:27 Quit jvd (*.net *.split)
23:46:27 Quit beta2k (*.net *.split)
23:47:01[foowhat time are u here regularily? (gevaerts?)
23:48:27gevaertsMost people are here in euro evening times. I tend to be online during daytime as well, but I'm at work then so I don't always reply quickly
23:50:13 Join rvvs89 [0] (
23:50:13 Join p3tur [0] (~petur@rockbox/developer/petur)
23:50:13 Join FOAD [0] (
23:50:13 Join komputes [0] (~komputes@ubuntu/member/komputes)
23:50:13 Join Schmogel [0] (
23:50:13 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
23:50:13 Join Galois [0] (
23:50:13 Join Utchybann [0] (
23:50:13 Join Hadaka [0] (
23:50:13 Join jvd [0] (~syscrash@poipu/developer/syscrash)
23:50:13 Join beta2k [0] (
23:50:20[fooc u
23:50:21gevaertsbut again, don't concentrate on me particularly. I volunteered as a mentor for this project, but other people might still do that as well, or I might end up mentoring another project...
23:51:00[foowhom should I then ask stupid questions?
23:51:15linuxstbNo-one in particular - just ask in this channel.
23:51:16gevaertsjust ask them here, anyone can answer them :)
23:52:12 Join stripwax [0] (
23:52:20 Quit robin0800 (Quit: No Ping reply in 180 seconds.)
23:52:27 Join hd [0] (
23:52:27 Quit hd (Changing host)
23:52:28 Join hd [0] (~jd@Wikipedia/HellDragon)
23:52:45 Join robin0800 [0] (
23:53:18 Join mc2739_ [0] (~mc2739@rockbox/developer/mc2739)
23:53:23 Join AlexP_ [0] (~ap@rockbox/staff/AlexP)
23:53:43 Quit elcan (*.net *.split)
23:53:43 Quit jd (*.net *.split)
23:53:43 Quit SirFunk (*.net *.split)
23:53:43 Quit AlexP (*.net *.split)
23:53:43 Quit rasher (*.net *.split)
23:53:53 Quit mc2739_ (Client Quit)
23:54:04 Quit [foo (Quit: Leaving)
23:54:35 Quit ender` (Quit: Pandas are the least racist animals: they're black, white and asian!)
23:55:22 Join Darkknight512 [0] (
23:56:42 Quit rvvs89 (*.net *.split)
23:56:42 Quit p3tur (*.net *.split)
23:56:42 Quit FOAD (*.net *.split)
23:56:42 Quit komputes (*.net *.split)
23:56:42 Quit Schmogel (*.net *.split)
23:56:42 Quit advcomp2019_ (*.net *.split)
23:56:42 Quit Galois (*.net *.split)
23:56:42 Quit Utchybann (*.net *.split)
23:56:42 Quit Hadaka (*.net *.split)
23:56:42 Quit jvd (*.net *.split)
23:56:42 Quit beta2k (*.net *.split)
23:56:58 Quit pamaury (Quit: Page closed)

Previous day | Next day