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 2010-08-03

00:00:28stripwaxkugel - do you have a 'current build' (or equivalent :-) for RaaA Android host?
00:00:55kugelstripwax: nope :(
00:01:08kugelyou can build one yourself though!
00:01:15gevaertskugel: briefly. I hope I'll get back to following things properly tomorrow.
00:01:19gevaertsNice work though
00:01:19stripwaxcan now! :)
00:01:31kugelthanks :)
00:01:38 Quit piggz (Ping timeout: 265 seconds)
00:06:10 Join krazykit [0] (~kkit@
00:11:56kugelstripwax: if you have any questions feel free to ask
00:13:18 Quit petur (Quit: Leaving)
00:13:24stripwaxkugel - none yet (although I don't have eclipse on this machine so won't be able to try out the latest version for a little while), but looking great so far! I especially liked the nugget that rockbox seems less cpu intensive than the default media player.
00:15:23 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
00:23:41freddyb0.22 seconds? Am I reading that right?
00:23:42CIA-6New commit by 03bieber (r27671): Theme Editor: Fixed rendering of %ca tag
00:24:11Tornefreddyb: yes
00:25:31CIA-6r27671 build result: All green
00:25:48bieberAlexP: Fixed your %ca bug :)
00:26:17AlexPbieber: ta :)
00:26:57CIA-6New commit by 03bieber (r27672): Theme Editor: Added a default label to switch block in devicestate.cpp to be safe
00:27:08 Quit bmbl (Quit: Bye!)
00:27:09 Quit ender` (Quit: Whenever you cut your fingernails, you will find a need for them an hour later.)
00:28:13jthundleyS_a_i_n_t: thank you so much for the tip earlier, I just finished translating my old custom theme to the new format :)
00:28:49CIA-6r27672 build result: All green
00:32:39CIA-6New commit by 03bieber (r27673): Theme Editor: SkinDocument class now checks file extension when loading a document and sets the cs device setting accordingly
00:32:59 Quit stripwax (Quit:
00:34:30 Quit domonoky (Ping timeout: 265 seconds)
00:34:47CIA-6r27673 build result: All green
00:35:46 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
00:37:51 Quit Jerom (Read error: Connection reset by peer)
00:38:43 Join Jerom [0] (~heidi@
00:49:03***Saving seen data "./dancer.seen"
00:57:13 Part jthundley ("Konversation terminated!")
01:02:08JdGordonkugel: \o/ nice commit
01:02:20 Quit n1s (Quit: Lämnar)
01:03:21 Quit evilnick_B (Quit: Page closed)
01:09:37 Quit Jerom (Read error: Connection reset by peer)
01:11:47kugelJdGordon: thanks :)
01:14:08 Quit JdGordon (Ping timeout: 265 seconds)
01:14:35kugelphew, that was tricky
01:14:46kugelI need to change over to a service for proper background music
01:18:08bieberkugel: How does the interface work in RaaA? Does it just behave like a normal RB touchscreen target?
01:18:27bieberI'm totally wishing I had a functioning Android phone to try it out on
01:18:57kugelbieber: yes
01:24:00 Join Strife89 [0] (
01:29:32 Quit bug2000 (Ping timeout: 240 seconds)
01:29:43 Quit MethoS- (Read error: Connection reset by peer)
01:30:30TheSevenfyi, the ipod nano 3g nor flash was just dumped and decrypted. we're facing yet another efi hell.
01:31:05TheSevenall classic generations will be very similar to this and can very likely be dumped using the same code
01:31:10 Join saratoga [0] (9803c57f@gateway/web/freenode/ip.
01:34:31S_a_i_n_tTheSeven: Linky, please?
01:35:11TheSeveni don't have the dump yet
01:35:23TheSevenserpilliere just reported success some minutes ago
01:35:29saratogais the linux4nano channel still logged?
01:35:44TheSeventheoretically yes
01:35:58TheSevenhowever our logbots are quite shaky sometimes
01:36:08saratogacould you PM me the link
01:36:18saratogagmail seems to have lost the email
01:37:01TheSevenhowever, there wasn't much being said regarding this on irc
01:38:12S_a_i_n_tbieber: Is there a way I can load an existing theme in its entirity into the themeeditor?
01:38:24bieberDo you mean from a ZIP file?
01:38:26S_a_i_n_tI assume it's possible, but I can't figure it out.
01:38:40S_a_i_n_tbieber: Not necessarily.
01:38:48S_a_i_n_tjust, "load the whole theme"
01:39:11bieberThere's no unzipping feature (I should probably add one), but if you just extract the theme into a folder, "Open project," and open the .cfg file it should all be there for you
01:40:19kugelI think I get hit by
01:40:45S_a_i_n_tbieber: Hmmmm...that's what I thought.
01:40:56S_a_i_n_tFor some reason none of the images are displaying.
01:41:07 Join bug2000 [0] (~bug@unaffiliated/bug2000)
01:41:28S_a_i_n_tI'll grab a newer version of the editor.
01:41:43bieberDo you mean in the theme previews?
01:42:02bieberThey don't show up in the "Project Files" list, if that's what you mean
01:43:26S_a_i_n_tbieber: Theme Prievew, yes.
01:43:51S_a_i_n_tAnother thought I had was the option to switch that possible?
01:43:59S_a_i_n_t(for the editor itself)
01:44:05bieberHmm, could you show me the theme? I haven't had any problems getting images to display before
01:44:18bieberFile -> Preferences lets you change fonts/colors in the editor :)
01:44:31S_a_i_n_tSeeing my code in a non-monospaced font looks *weird* ;)
01:44:35S_a_i_n_toh, cool.
01:45:08bieberIt _should_ try to select a monospace font by default, but Qt doesn't seem to be the best at finding an ideal default font
01:45:24S_a_i_n_tbieber: I may be using a more than slightly dated version...I'll grab the newest build I can and try it there.
01:45:54bieberOkay, let me know how it goes
01:48:32S_a_i_n_tdo all the files need to be in the same directory, or can it have the subdir structure as intended for use on device?
01:48:46bieberIt has to be in the same dir structure as the device
01:49:02bieberwps images in /wps/themename/ and so on
01:49:33bieberThe only real difference is that there's no obligation to make your root directory .rockbox, it'll automatically change that for you when you export
01:50:16S_a_i_n_tSweet, one last thing *very minor) is I find it a little overwhelming that the editor "fullscreens" on first run.
01:50:33S_a_i_n_tIf you have a rather large monitor its a little overwhelming ;)
01:50:56bieberI could have sworn I set it to 1/4 the screen size on first run, lemme check on that
01:51:04bieberYou don't have multiple screens perchance, do you?
01:51:06S_a_i_n_tAs I said though, in the scheme of things that's *very* minor.
01:51:33S_a_i_n_tRe: multiple monitors, no.
01:51:40*kugel tries the emulator
01:53:33bieberHmm, it starts out small for me, it's got to be something wrong with getting the screen geometry on Windows
01:54:03kugelfuck fuck fuck fuck
01:54:27kugelRaaA won't be usable on a wide range of devices
01:54:46S_a_i_n_tbieber: What do the colours mean?
01:54:51S_a_i_n_t(text colour)
01:55:11S_a_i_n_tIn general, I see red (perhaps it is marone in this case) as "bad"
01:55:20 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
01:55:21bieberDo you mean in the preferences dialog?
01:55:38S_a_i_n_tI see a lot of red, yet I also see "The document parsed successfully"
01:55:57bieberOh, the default highlighting for tags is red
01:56:07S_a_i_n_tHmmm, no, I mean in the large central code viewing panel.
01:56:07bieberGood point there
01:56:42bieberIf you want to change the highlighting colors, it's under the "Highlighting" tab in the preferences
01:56:47S_a_i_n_tIt made me thing I was doing something wrong initially, until I noticed the "blah parsed correctly"
01:56:54bieberI probably should come up with a better default color
01:56:59bieberI'm just not sure which one to use
01:57:13JdGordon|keep red for errors :)
01:57:19S_a_i_n_tvarious shades of green?
01:57:28bieberBut green says comment
01:57:32kugelrasher: ping
01:57:45bieberHmm, I just stumbled on a dark shade of gold I rather like
01:57:52S_a_i_n_tgrren (roughly) universally means "good", and red (again, roughly) universally means "bad"
01:59:07bieberHehe, I think my perception is warped by my current IDE's default highlighting :P
01:59:29S_a_i_n_theh ;)
02:00:18S_a_i_n_tbieber: One last, last there a way to cycle through animations/'simulated' track playback progress?
02:00:47bieberYeah, bring up the timer panel from the "View" menu
02:00:52 Quit Rob2222 (Quit: Rob2222)
02:00:59S_a_i_n_tExcellent, thanks again.
02:01:12bieberThat lets you step through simulation time. The animations aren't perfect because of the intervals it updates at, but it gives you a good rough idea
02:01:35bieberThe play/etc. buttons at the bottom control the player's play mode, not the timer, btw, in case that's ambiguous
02:02:03kugelsaratoga: saratoga:
02:02:25bieberS_a_i_n_t: If you come up with a set of highlighting colors you like, feel free to send them along to me and I'll switch the defaults. In general, I'll defer to just about anyone on design decisions ;)
02:02:46bieberGraphic design, that is
02:02:56kugelinstall it, run it, then press home/start the terminal emulator. run top at look if org.rockbox has "fg" under PCY
02:03:13S_a_i_n_tbieber: Last, last, last thing: Is it possible to specify if the "track" being "played" has AA or not?
02:03:42bieberYep, it's under Playlist/Song info in the device options panel
02:03:57 Join Rob2222 [0] (
02:04:16bieberTowards the very bottom of the tab
02:04:33saratogakugel: installing
02:04:35S_a_i_n_tthanks, quite easy to get lost in the tabs ;)
02:04:42JdGordon|kugel: haha we really need a new sysfont for big screen targets :)
02:05:08kugelthat apk comes with a bigger font at least
02:05:15bieberYep, those are a little bit of a mess. If you have any suggestions for rearranging them, by the way, let me know
02:05:16saratogakugel: just gives me a blank screen
02:05:38saratogaoh wait
02:05:40saratoganow its going
02:05:44saratogatook 30+ seconds
02:05:51JdGordon|kugel: no PCY column in top?
02:06:01kugelsaratoga: it installs the themes etc on the first run
02:06:34kugelI have " PID CPU% S #THR VSS RSS PCY UID Name" in both my legend's and the emulators top
02:07:21saratogahow do the controls work?
02:07:32kugelgrid mode :)
02:07:38JdGordon|what is PCY?
02:07:51kugelI'm assuming policy
02:08:12kugelit denotes whether that process is a fore- or background process
02:09:10saratogaok figured out the controls, it plays music
02:09:29saratogait seems to skip once and a while
02:09:31JdGordon|woo \0/
02:10:21kugelsaratoga: it's possible that the dream is too slow to handle the UI at the same time
02:10:21saratogaget test codec working so i can benchmark
02:10:33kugelanyway, can you run top for me? :)
02:10:50saratogaits very similar to the gigabeast CPU
02:10:57saratoga450MHz ARM11 I think
02:11:06saratogahmm just crashed without me touching it :(
02:11:26saratogaha and it doesn't like it when the screen rotates :)
02:11:36JdGordon|kugel: I assume you've seen the force closes trying to open it from the notifications area?
02:11:38saratogabut i guess thats our wonderful graphics drivers
02:11:40kugelah yes, I forgot to mention that
02:11:45JdGordon|and sound works fine on the N1 :)
02:12:02 Quit efyx (Remote host closed the connection)
02:12:05kugelJdGordon|: yes, I've seen that
02:12:38kugelguys, can you finally tell me what top says under PCY on your devices? :)
02:13:01JdGordon|tell me another way to get to it... no PCY column
02:13:42JdGordon|28% cpu playing music :/
02:13:42kugelI don't know another way
02:13:56JdGordon|and 40% ram
02:14:02kugelthat's the ui. go to the main menu at it will be around 6-7%
02:14:13JdGordon|just not good enough for a first attempt! :D
02:14:22saratogakugel: about 20-23% CPU with the app minimized
02:14:53saratogaseems pretty stable as long as I don't have teh GUI up
02:15:10kugelquite the opposite here :)
02:15:13JdGordon|yeah, music in the background, the display shouldnt be updating at all.... this is what we were talkking about last night. I can make the themes not update at all if the display isnt visible
02:16:22S_a_i_n_tbieber: One thing I have just noticed in the editor: "%s%al%?Sr<%pe %Sx(of) %pp|%pp %Sx(of) %pe>" displays as " 10 20 "
02:16:55kugelsaratoga: also no PCY under top?
02:17:08S_a_i_n_tI can hand you the theme if you like, still (aaaargh, been a LONG time) a WIP
02:17:36S_a_i_n_tby the way, images display fine in the new(er) version I am using.
02:17:41S_a_i_n_tbieber: ^
02:18:43saratogahow do i make top only show the most busy processes, it doesn't fit all on my phones screen
02:19:17kugelcan you scroll?
02:19:50bieberS_a_i_n_t: Looking at that line now
02:19:57JdGordon|kugel: can rockbox be closed?
02:20:08saratogayeah but i get a tenth of a second before it scrolls me back :)
02:20:09kugelnot yet
02:20:47bieberS_a_i_n_t: What _is_ %Sx?
02:20:48S_a_i_n_tbieber: Do you want the theme?
02:20:54bieberI can't find it in CustomWPS
02:21:08S_a_i_n_tbieber: That's because it isn't there ;)
02:21:16bieberOkay, found it
02:21:20bieberI had no idea that tag existed :P
02:21:26S_a_i_n_tIt's "Translate <string> to user selected lang"
02:21:33bieberWhich is why it's not rendering for you ;)
02:21:43S_a_i_n_tAha ...makes sense.
02:21:46bieberI'll go ahead and implement that tonight, I've gotta run for now
02:21:59S_a_i_n_tI guess you also need all the lang files in the editor then.
02:22:18S_a_i_n_t(for %Sx to work)
02:22:27saratogaleaving the GUI up causes more skipping and about 60% cpu load with the app minimized
02:22:36bieberFor now I'm just going to display the text
02:23:15kugelsaratoga: try ctrl+c, you could scroll back then
02:23:16S_a_i_n_tbieber: Aha...good compromise.
02:23:25bieberI don't know if translation needs to be testable in the editor, but I at least want to get _something_ displaying there
02:23:33saratogayeah thats what i ended up doing, just took me a while to figure out how to do a control c on a keypad :)
02:23:54S_a_i_n_thowever, it'd be nice to have the lang files in though, as the line *wil;l* fail if the string isn't present in the lang file.
02:23:59S_a_i_n_tbieber: ^
02:24:08kugel"ball+c" in the terminal I just installed
02:24:23S_a_i_n_tSo, I'd say translation needs to be implemented in the editor.
02:24:35bieberI honestly know just about nothing about translation at the moment, I'll have to read up on it
02:24:54S_a_i_n_tJdGordon|: ...really?
02:25:02kugelsaratoga: did you go to the main menu in rockbox before leaving the app?
02:25:20kugelI have not implemented stopping lcd updates so it might help
02:25:29S_a_i_n_tThe idea is for the editor to make a file that *guarantees* working-ness on device no?
02:25:49JdGordon|the most I would expect would be to have the english strings know, no need to actully translate them
02:26:00S_a_i_n_tA wrong string being translated (or attempted to) is a pretty good way to make a theme not load.
02:26:13S_a_i_n_tAh, good point on that.
02:26:30JdGordon|how is %St handled?
02:26:31S_a_i_n_tyeah, it definitely needs a current english.lang
02:26:48S_a_i_n_tre: %St ...pass.
02:27:05S_a_i_n_tThat's black magic as far as I'm concerned. Never had to use it.
02:27:36bieberS_a_i_n_t: It's really more of a suggestion than a guarantee ;)
02:27:56S_a_i_n_tI know only *roughly* how it works on device, no idea about it in the editor.
02:28:17bieberAre translation files customarily distributed with themes? If so, it should be easy enough to implement, otherwise it's really beyond the scope of what I want the editor to cover
02:28:27JdGordon|they are not
02:28:54S_a_i_n_t"translation files" are the .lang files themselves.
02:29:17JdGordon|for completeness it would be nice if it could download the latest english.lang from svn and parse it, but not a high priority imo
02:29:23S_a_i_n_t%Sx9<string>) just points to a string present in 'english.lang'
02:29:44S_a_i_n_tbah, %Sx(<string>)
02:30:16saratogakugel: if I go to the main menu its about 20%, otherwise more like 65% or so, and music skips
02:30:33 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
02:30:45S_a_i_n_tie. You can't just translate %Sx(foo) without "foo" being a valid string in englisg.lang
02:30:52S_a_i_n_tbieber: ^
02:31:00 Join fdinel [0] (
02:32:33bieberI see
02:32:55bieberI'll definitely put it on my time-permitting to-do list, it just won't be an urgent priority at the moment
02:33:11S_a_i_n_tOh, sure.
02:33:40kugelsaratoga: it's about 25% vs 6-8% on my legend, your dream is quite slow it seems
02:33:59kugelsaratoga: but you at least get "fb" under PCY, that's broken on my legend
02:34:45kugelon various other htc phones as well
02:35:10saratogamaybe because I have 2.2?
02:35:45kugeleither that, or because you don't have an official htc image
02:36:26kugelI need rasher to test it on his froyo'd desire
02:36:43kugelif the issue is there too I see no hope for RaaA on HTC phones
02:37:22kugelit makes running raaa in the background practically impossible
02:37:54kugel"bg" means a hell lot of stuttering audio when you only swipe the fingers over the display
02:38:02saratogawould this include your phone?
02:38:28kugelI hope htc fixed it in 2.2
02:38:43JdGordon|you cant bump the thread priority or something?
02:39:02kugelthat doesn't help
02:39:03S_a_i_n_tbieber: Another observation is that if backspace is held, the code on screen is not updated until backspace is released. So, if you hold backspace you get the impression it does nothing, but when releasing it you realise it deleted a large chunk ;)
02:39:05JdGordon|is this still not using the asm code?
02:39:22S_a_i_n_tbieber: Same thing goes for Delete
02:40:14kugelevery bg process is put into some low prio corner with all others bg process. they all together get very little cpu time from the kernel
02:40:24saratogamost of the CPU time here probably isn't the codecs themselves, which even in plain c will be pretty fast on such CPUs
02:40:57saratogahmm actually it might be
02:41:02saratogai need to play with this more
02:41:12 Quit DerPapst (Quit: Leaving.)
02:41:36JdGordon|disableing the peak meter should help a bit also.. forces 25FPS refresh rate on the wps
02:42:44saratogaouch my phone just rebooted :(
02:43:03saratogaapparently i need to get out of rockbox before plugging in USB, not just pause it
02:45:17FlynDicekugel: Cool themes now! froyo on my EVO if I can test anything for you?
02:47:51kugelFlynDice: I only need to know if rockbox runs fg process if you leave it to the home screen
02:49:04***Saving seen data "./dancer.seen"
02:51:12 Join Strife1989 [0] (
02:51:14FlynDiceIt continues to play and there is an icon up in the notification area if that's what you're getting at...
02:51:48saratogaflac is about 9%, so maybe the codecs are just really slow without ASM
02:52:08saratogai would expect flac to be maybe 3x faster then MP3
02:52:52 Quit bieber (Ping timeout: 260 seconds)
02:53:07 Join bieber [0] (
02:53:18kugelFlynDice: you can see the scheduling group if you run top in a terminal or adb shell
02:53:23 Quit Strife89 (Ping timeout: 240 seconds)
02:54:28FlynDiceI'm not rooted, can I do that with USB debugging?
02:59:53kugelsaratoga: for the asm optimizations to work you somehow need apps/ know about the arm cpu, but not firmware/
03:00:34 Quit freddyb (Quit: ChatZilla 0.9.86 [Firefox 3.5.9/20100401213457])
03:01:06JdGordon|just ignore it in firmware/ ifdef ANDROID?
03:03:51 Quit bieber (Ping timeout: 265 seconds)
03:04:06 Join bieber [0] (
03:06:11kugelJdGordon|: did it really use 40% ram?
03:06:21JdGordon|thats what is says
03:06:48kugeldoes it mean 40% of the entire ram, or of the 10MB heap reserved for each dalvik vm(i.e. app)
03:06:50saratogakugel: do you have a config file for RAAA?
03:07:16kugelsaratoga: firmware/export/config/application.h if you mean that
03:08:32JdGordon|kugel: I'm giong to guess the later because the numbers are all around the 40% mark
03:08:38kugelJdGordon|: I have plenty free ram with rb running
03:08:44saratogakugel: i guess there should probably be an application-android.h somewhere that defines CPU_ARM
03:10:25kugelJdGordon|: it should use about 10MB (a bit more probably) in total, 8 of which is the audio buffer
03:11:03JdGordon|yeah, thats what I thought... if its 40% of the heap then its fine
03:11:40saratogaeventually we'll probably want to do hardware specific optimizations on RAAA targets
03:12:13kugeldo we?
03:12:43 Join Strife89|Laptop [0] (
03:13:07kugelI think speed is not important enough to make it dependant on certain phones
03:13:20kugelI would rather have the same binary to run on as many phones as possible
03:13:55saratogai suppose android phones are close enough that it doesn't matter
03:14:11saratogabut at very least the system should be aware that its on ARM
03:14:34saratogausing DSP cores might be a nice option too if the NDK exposes them
03:16:03 Quit Strife1989 (Ping timeout: 240 seconds)
03:23:05JdGordon|kugel: how is the java<->c bridge done? i mean can the c code send messages or something to java?
03:23:28kugelJdGordon|: look at the source code :)
03:23:49JdGordon|simple example, can rockbox ask for the android keyboard?
03:23:49kugelyou can call c functions from java and java methods from c
03:24:27JdGordon|so.. yes?
03:25:10kugelI would think so
03:25:24 Quit komputes (Quit: I haven't slept for ten days, because that would be too long.)
03:31:48*JdGordon| wonders if we can use the notification area like a remote lcd :)
03:51:36 Quit kugel (Remote host closed the connection)
03:52:47 Join stooo [0] (
03:58:21FlynDicekugel: (logs) EVO with froyo, cpu 21%, S, #THR 10, VSS 169804K, RSS 26404K, PCY fg
04:02:52bieberS_a_i_n_t: That's for performance purposes. On a good sized skin, the editor was hanging pretty badly when the delete key was held because it was rendering too frequently, so now there's a brief delay between editing and rendering
04:14:26 Quit anewuser (Quit: FAMICOMPO 7 - August 15 Through August 31 JST -
04:19:18 Quit Barahir (Ping timeout: 260 seconds)
04:20:46 Join Barahir [0] (
04:22:20 Quit bieber (Ping timeout: 252 seconds)
04:22:32 Join bieber [0] (
04:27:54 Quit amiconn (Disconnected by services)
04:27:57 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:28:03 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:28:45 Quit pixelma (Disconnected by services)
04:28:47 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:29:02 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:31:51 Quit RemDeprived ()
04:39:12 Quit TheSeven (Ping timeout: 276 seconds)
04:40:36 Quit elcan (Remote host closed the connection)
04:40:39 Join elcan [0] (
04:43:24 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:49:06***Saving seen data "./dancer.seen"
04:58:18 Part toffe82
05:02:29 Join larrythechamp [0] (
05:06:16 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
05:20:01 Join Tarvis [0] (
05:21:07 Quit larrythechamp (Quit: Leaving)
05:23:55 Join BHSPitLappy [0] (~BHSPitLap@unaffiliated/bhspitmonkey)
05:34:38 Quit Horscht (Quit: Verlassend)
05:40:22 Quit Tarvis (Quit: /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit /quit)
05:51:04 Quit rob (Ping timeout: 260 seconds)
06:02:46 Nick Strife89|Laptop is now known as Strife89 (
06:04:56 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:07:05 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
06:16:35 Quit bug2000 (Ping timeout: 246 seconds)
06:17:04 Join rob [0] (
06:25:47 Join r0b- [0] (
06:28:27 Join bunnyboi [0] (
06:29:37 Quit rob (Ping timeout: 276 seconds)
06:31:36 Quit BHSPitLappy (Read error: Operation timed out)
06:34:09 Join captainkewllllll [0] (
06:34:27 Quit captainkewllllll (Client Quit)
06:49:07***Saving seen data "./dancer.seen"
06:49:34 Quit kramer3d (Quit: Leaving)
06:54:10 Join rob [0] (
06:54:19 Quit r0b- (Read error: Connection reset by peer)
06:55:53 Join r0b- [0] (
06:58:52 Quit rob (Ping timeout: 276 seconds)
07:01:39 Quit lestatar (Ping timeout: 260 seconds)
07:19:21 Join bieber_ [0] (
07:29:50 Join LinusN [0] (linus@rockbox/developer/LinusN)
07:44:51 Join Buschel [0] (
07:48:02CIA-6New commit by 03bieber (r27674): Theme Editor: Implemented Sx tag in renderer
07:49:58CIA-6r27674 build result: All green
07:52:18 Join fyre^OS [0] (
07:55:09 Quit fyrestorm (Ping timeout: 240 seconds)
07:56:46 Join milz [0] (
08:05:38 Join lestatar [0] (
08:09:06FlynDicekugel: (logs) <−− cut & paste from top command on HTC EVO with 2.2 FROYO update, with rockbox playing music very nicely, after going back to home screen.
08:13:44 Quit S_a_i_n_t ()
08:18:06 Quit saratoga (Ping timeout: 252 seconds)
08:21:47 Join Zagor [0] (bjst@rockbox/developer/Zagor)
08:22:22 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
08:22:39 Quit lestatar (Ping timeout: 240 seconds)
08:25:30 Join lestatar [0] (
08:26:00 Quit bieber_ (Quit: No Ping reply in 180 seconds.)
08:26:16 Join bieber_ [0] (
08:26:21 Quit fyre^OS (Read error: Connection reset by peer)
08:33:34 Quit bieber_ (Remote host closed the connection)
08:35:30 Join BHSPitLappy [0] (~BHSPitLap@unaffiliated/bhspitmonkey)
08:47:22 Join perfectdrug [0] (
08:49:09***Saving seen data "./dancer.seen"
08:50:39 Join MethoS- [0] (~clemens@
08:59:56 Join funman [0] (~fun@rockbox/developer/funman)
09:00:18 Join hebz0rl [0] (
09:01:02 Join S_a_i_n_t [0] (S_a_i_n_t@
09:07:14S_a_i_n_tJdGordon|: Instead of "%xd(playmode, %mp)", howabout "%xd(playmode, %mp(5,3,2,4,1,1))" I see that as still being pretty readable, and it has the benifit of not needing an offset value if you want the first subimage to be an arbitrary number instead of "1" and also allows for the same image to be called more than once.
09:07:20S_a_i_n_tIf I need to explain more, shoot.
09:07:59JdGordon|not a chance :)
09:08:07S_a_i_n_tI'd probably still use the "%xd(A,1)" + subline method myself, but it more closely represents the intended change for %xd.
09:08:08JdGordon|that would mean adding more to the language
09:08:36JdGordon|the parser wouldnt know that %mp(5,4,3,2,1) isnt a parameter list
09:08:45solrizeanyone know the size of the lua plugin and how much ram it uses? tx
09:09:12S_a_i_n_twasn't the deal with the new parser (or one of them) its apparent ability to be expanded with ease?
09:09:25funmansolrize: check the file
09:09:29JdGordon|I guess if you were crazy you could do %?mp<5|4|3|2|1> but that defeats the purpose of the feature
09:09:33solrizefunman, thanks
09:09:59JdGordon|S_a_i_n_t: sure, but %mp() looks like a param list to the parser
09:10:26S_a_i_n_tJdGordon|: Could you do "%?mp<5|4|3|2|1|1|2|3>" etc.?
09:10:41S_a_i_n_tAnd, I guess the brackets were a poor choice
09:10:53JdGordon|probably, but that seems just as silly as the long form
09:10:57solrizefunman, where do i find that file? it's not in
09:11:28JdGordon|I guess I could support that with a tiny bit of work
09:11:39S_a_i_n_tIt seems more sensible to me personally, and the ability to call a subimage more than once, or to call subimages out of "order" would be a bg plus IMO
09:12:38*S_a_i_n_t is trying to compromise ;)
09:12:55JdGordon|you mean make it so complicated that noone uses it :D
09:13:20funmansolrize: it's in the build folder (apps/plugins/lua)
09:13:28S_a_i_n_tI think its too restricted presently, people don't have much reason to use it.
09:13:35S_a_i_n_tIt only covers very specific cases.
09:15:17JdGordon|the first step would be chaning it to the %xd(name, number) format.. alloiwing number to be skin code is a bigger change which would mean this whole thing would probably happen after 3.7
09:15:26JdGordon|but we want 3.7 to be finisehd breaking things
09:15:43S_a_i_n_tI think the current way of doing it is fine personally, readability would be increased simply by the "%xd(A,1)" change.
09:16:21 Quit r0b- (Read error: Connection reset by peer)
09:16:22S_a_i_n_tBut I'm trying to find a way to (sensibly) expand the proposed addition, if it simply must go in.
09:17:53S_a_i_n_t"%?mp<%xd(A,1)|%xd(A,2)|%xd(A,3)|%xd(A,4)|%xd(A,5)>" seems pretty damn readable to me, personally.
09:18:01JdGordon|the code thing is more of a reason to want to change the format
09:18:08 Join petur [0] (d408b802@rockbox/developer/petur)
09:18:23 Part perfectdrug ("Leaving")
09:18:26JdGordon|like I said (probably not early enough), the code thing would be a later addition
09:19:13 Join Szpila [0] (~sszpila@
09:19:56S_a_i_n_tLet me get one thing straight's not *adding* any functionality is it? Just readability (which seems terribly subjective)...or am I wrong?
09:20:09S_a_i_n_tIs there something not currently possible this would add?
09:20:51 Join perfectdrug [0] (
09:21:37S_a_i_n_tThen I guess the poll should really be not "shpuld X or Y be added or dropped", but "is X really so terribly hard to read it warrants changing?"
09:22:03 Join ender` [0] (
09:22:04JdGordon|yes and no
09:22:13pixelma%xd(A,number) will probably look worse compared to %xd(Ae) once you need two digits which could be covered by one letter in the old way, so everything between 10 and 52
09:22:36JdGordon|more "does this code feature look useful enough to break themes now with the possibility of the addition at some later time"
09:22:58S_a_i_n_tpixelma: I had thougt of that as well
09:23:09JdGordon|pixelma: without thinking... r is which image number?
09:23:19JdGordon|W is which?
09:23:41JdGordon|sure it might be more readable, but no way could it be considered more understandable
09:24:34S_a_i_n_tRegardles if its digits or'll still have to count out the subimages one by one if you want them assigned in a "non ordered" way.
09:25:06S_a_i_n_tcounting out A B C D is just the same as 1 2 3 4 to me.
09:25:13pixelmafor me it doesn't matter much, in which "place" which subimage is as you usually only count the one time you write it
09:25:19JdGordon|IMO if your strips are not in the same arer as the tag you are doing it wrong
09:25:47S_a_i_n_tit doesn't *need* to be ordered.
09:26:01S_a_i_n_tIt sensible, sure...but hardly necessary.
09:26:33JdGordon|what about the %Li (list icon) tag... that has 31(?) options...
09:27:17*S_a_i_n_t doesn't see the point there...sorry.
09:27:22pixelmaI have a WPS which has one big strip combining all the same sized icons. The first are in order of the corresponding condfitional, then adds the next ones to it - still in order but not starting at 1 or a
09:27:51S_a_i_n_tpixelma: I do the same in some themes, I expect others do also.
09:28:20S_a_i_n_tIf at all possible I see loading one image a lot better than loading 5 smaller ones.
09:28:48JdGordon|I think you are both crazy :)
09:29:04S_a_i_n_tNo, we're themers ;)
09:29:15JdGordon|you say tomato
09:29:19S_a_i_n_tYou write the code, we use/abuse it :P
09:30:02pixelmaS_a_i_n_t: same here... and as the idea of the WPS already made it necessary to put the icons in own viewports this came "for free"
09:30:20S_a_i_n_tpixelma: Indeed.
09:30:28pixelmathe solution to positioning was already there
09:31:25 Quit stooo (Ping timeout: 265 seconds)
09:33:39JdGordon|no, I stand by what I said. There is no logical reason to *want* to put all images in a single strip. sure you dont need as many %xl lines, but if you want to have the strip ordered there is N times as many tags to fix, and it doesnt let you/others swap out the strip completly
09:33:51JdGordon|and counting with letters isnt bloody natural!
09:34:55*S_a_i_n_t considers 1 load, as opposed to say...5 a WIN.
09:34:57pixelmaI really don't see how %xd(image,%mp) is more readable because you need to know how many conditional states %mp has (can be worse with other conditionak) and you lose "control" if you want to use less
09:35:36pixelmaas in "not fill all conditional branches"
09:35:58S_a_i_n_t"%xd(image,%mp)" also looks completely foreign to all the other skin code.
09:36:43S_a_i_n_tNot being able to load individual images conditionally is a bit crap too IMO.
09:36:59S_a_i_n_tIt does seem that this is such a *terribly* limited use case to me.
09:37:20JdGordon|pixelma: lets pretend I didnt suggest that form, do you not agree that %xd(Fh) is less understandable than %Xd(playmode, 5) ?
09:38:20pixelmaI don't agree as soon as I start imagining a long conditional line with the suggested syntax
09:39:12*S_a_i_n_t does not, if I want to see what image 'F" is, I look at the %Xl line, naming the images sensibly is just as good IMO
09:39:34pixelma%xl would be more suitable
09:39:45JdGordon|%?mp<%xd(a,0)|%xd(a,1)|%xd(a,2)|...> is no worse than %?mp<%xd(aa)|%xd(ab)|%xd(ac)|...>
09:40:18S_a_i_n_texcept when it comes to could digits
09:40:20 Quit BHSPitLappy (Ping timeout: 260 seconds)
09:40:30pixelma%?mp<%xd(playmode,0)|%xd(playmode,1)|%xd(playmode,2)|...> is
09:40:45S_a_i_n_tAz would be A26...which seems odd to me.
09:40:47JdGordon|pixelma: nothing is stopping you from naming them with one letter
09:41:12pixelmaok, then I load "a.bmp"?
09:41:28JdGordon|what? no... the %xl line doesnt change
09:41:57S_a_i_n_tAn additional way to do something that can be done (in a better, more flexible way) isn't really necessary...that's the way I look at it.
09:42:07JdGordon|if you change the way you think about strips then the benefit is huge when the number param can be a tag
09:42:42S_a_i_n_tI only see it as limiting, for reasons expressed above.
09:43:23JdGordon|limiting suggests that it removes features which it doesnt
09:43:28S_a_i_n_tif you want to use the codec tag, but only want to display a few just...can't.
09:43:32JdGordon|you would be free to use it or not
09:43:43S_a_i_n_tlimiting doesn't suggest it removes anything.
09:43:48JdGordon|my only aim is to make things more flexible and possibly nicer for new people
09:44:07S_a_i_n_tit's just why would I use X when Y can do it better?
09:44:25JdGordon|because Y is only better because your method of doing Z is wrong
09:44:38JdGordon|fix Z and X is much better than Y
09:44:39S_a_i_n_tIf more flexibility is the aim, then I honestly don;t agree with using a tag as a subline param
09:45:07S_a_i_n_tfor the fact that all subimages *must* be used.
09:45:13JdGordon|ok, so think about simplicity, readability
09:45:28S_a_i_n_tthere's no way to say "right, I only want to use subimages 3,4,5 and 9 here"
09:45:30pixelmawho decides how doing Z is right or wrong?
09:45:54JdGordon|S_a_i_n_t: untill anything is actually coded it is silly saying "X cant be done"
09:46:53 Quit mc2739 (Ping timeout: 265 seconds)
09:47:04S_a_i_n_tif simplicity/readability is the case...I don;t agree either. A user would probably still have to/want to use the "subline" way of doing this...and two ways of doing a similar thing in the same set of code is not something I consider to be more readable
09:47:22 Join Rob2223 [0] (
09:47:39 Quit perfectdrug (Quit: Bye)
09:47:46pixelmathe "number lets you more easily understand which position the subimage is" argument is not very interesting to me. It's true, but if I want to know which part of the subimage is really shown, I'd still start counting with the whole image in front of me to see which is "number 14", then there is no difference if it is "n" or "14"
09:48:27 Join einhirn [0] (
09:48:34 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
09:49:25S_a_i_n_twell, there's a difference in the fact that larger numbers can be represented in a smaller fashion, but that's about it.
09:49:54S_a_i_n_tpersonally, I'd rather use "j" than "11"
09:50:06pixelmame too
09:50:22JdGordon|I will concede that without being able to use the tag for a number there is little point, but with that ability then all your arguments are flawed. absolutly everything can be done in less text
09:50:58 Quit Rob2222 (Ping timeout: 240 seconds)
09:50:59pixelmaI don't think so
09:51:00S_a_i_n_tindividual conditional elements?
09:51:19S_a_i_n_thow would that look, as an example.
09:53:09JdGordon|give me a concrete example how it would actually be needed
09:53:10S_a_i_n_tfor me, it is *really* easy t see what is happening in "%?<mp<%xd(Aa)|%?C<%xd(Ab|%xd(Ac>|%xd(Ad)|etc.
09:53:42pixelmayou are missing some )s there ;)
09:53:43JdGordon|that line would NOT be a good candiadte for the change anyway
09:54:04S_a_i_n_tso...original format wins then?
09:54:14pixelmayou said "everything" can be done in less text−−-
09:54:24pixelmaor ...
09:54:43JdGordon|ok, %xd(blaa, %xx)
09:54:54JdGordon|where %aa evaluates to "%?<mp<%xd(Aa)|%?C<%xd(Ab|%xd(Ac>|%xd(Ad)|etc.
09:55:37S_a_i_n_twhere are the conditional elements defined in "%xd(blaa, %xx)"?
09:56:07S_a_i_n_tI don't see *how* %xd(blaa, %xx) can == %?<mp<%xd(Aa)|%?C<%xd(Ab|%xd(Ac>|%xd(Ad)|etc.
09:56:11JdGordon|you gave me an arbitrary example and I was just showing you that untill anything is coded then i can do anything
09:56:30S_a_i_n_tif it can, I definitely don;t see it as more readable, as it is impossible to see what is happening.
09:56:48JdGordon|because that is not a real world example
09:56:58JdGordon|look at the cabbie image lines.. they are the ideal candidate for this
09:59:51JdGordon|show me a real line in a skin where adding 50% more chars to the line would make it >50 chars and I'll prove to you this is better
09:59:59JdGordon|imaginary examples are pointless
10:00:31*S_a_i_n_t suggests it should ultimately be considered by "what is the point, really, in having two ways to do the same thing"
10:00:32JdGordon|and 50chars isnt even that long
10:00:44S_a_i_n_tI really don;t think that "less is better" in all exaples.
10:01:45JdGordon|I dont see it as comeplete duplication
10:01:58S_a_i_n_tAs far as readability goes, your proposed way is...smaller. The current way it is incredibly easy to see how the line flows, and exactly what it is doing and where...even if it is twice as long.
10:02:13S_a_i_n_tIt's not like the length of the skin file is an issue.
10:02:20JdGordon|I disagree
10:02:29 Join mitk [0] (~mitk@
10:02:57JdGordon|the old/current way requires you to know the tag value order is 2 places. with my proposal you need to know the order in the bmp ONLY and the number of images
10:03:06 Quit Judas_PhD (Remote host closed the connection)
10:03:27JdGordon|%xd(playmode, %mp) means "I dont care what the order is, just display the right image"
10:04:02S_a_i_n_tIt also prevents you from selectivly including/discarding an image from that subimage set.
10:04:03 Join Judas_PhD [0] (
10:04:13JdGordon|%?mp<%xd(Aa)|.....|> requires ou to a) know the order for the image, b) know the order for the code and NOT get it wrong, c) not get lost or miscount the |'s
10:04:42S_a_i_n_tAll things people manage to do fine now.
10:04:56JdGordon|only because there isnt a cleaner option
10:05:26JdGordon|selective in/exclusion could be fixed with another tag also
10:05:31S_a_i_n_tIf your proposed option indeed was cleaner, I may agree.
10:05:45S_a_i_n_texactly...adding *another* tag.
10:05:55S_a_i_n_twhat's the point? it can all already be done.
10:06:25JdGordon|%xd(playmode, %zz(%mp, +, 0,2,4,6)) is STILL easier to understand after the initial wtf
10:06:49S_a_i_n_tif you say so...
10:06:56S_a_i_n_ttwo themers disagree though.
10:07:18JdGordon|I'm still waiting for a counter example
10:07:31S_a_i_n_tthe code already in place?
10:08:16S_a_i_n_tyou're confusing "concise" with readable.
10:08:27JdGordon|%?mp<%xd(Fa)|%xd(Fb)|%xd(Fc)|%xd(Fd)|%xd(Fe)> is straight out of cabbie... => %xd(playmode, %mp)
10:08:35S_a_i_n_tbeing able to walk the line through in ones head is a real plus IMO
10:08:49JdGordon|you're head is screwed on backwards then!
10:09:02S_a_i_n_t"%?mp<%xd(Fa)|%xd(Fb)|%xd(Fc)|%xd(Fd)|%xd(Fe)>" is perfectly readable to me.
10:09:14S_a_i_n_tI'd have to check to see how many tags %mp has, etc.
10:09:22S_a_i_n_t(in your example)
10:09:26JdGordon|you have to do that anyway
10:09:32JdGordon|but only when writing
10:09:41JdGordon|what about when debugging or modifiying someone elses skin?
10:09:50S_a_i_n_tsure, but simply by looking at "e", "sweet, there's 5"
10:09:59JdGordon|scorche|sh: is there anyway to see who voted in the forum polls?
10:10:27S_a_i_n_tI'm the "no" if it helps.
10:10:47JdGordon|S_a_i_n_t: and *why* would you ever do that? if you want to change the image then you dont care about the skin code
10:11:11JdGordon|if you want to add something to that one option then yes this doesnt help you
10:11:59pixelmawhat would "%xd(playmode, %zz(%mp, +, 0,2,4,6))" mean? And it doesn't have your advertised advantage of "you only need to know in one place how many branches a conditional has" anymore
10:12:09*S_a_i_n_t fails to see how "if you want to change the image then you dont care about the skin code" works at all.
10:12:19 Quit Strife89 (Quit: Bed.)
10:13:08JdGordon|pixelma: that was a stupid example to prove a point about an arbitrary example... that would mean "from the %mp tag only use the value if it is 0,2,4,6"
10:13:19 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:13:39JdGordon|S_a_i_n_t: if you want to change the play icon... why would you need to see where it is displayed?
10:14:37S_a_i_n_tJdGordon|: As the subimages may not be ordered exactly in the way that the tag uses them.
10:14:47 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
10:14:57 Quit bunnyboi (Quit: Ex-Chat)
10:15:16pixelmahe deiced that to be the "wrong way" anyway
10:15:41pixelmaI can't see how S_a_i_n_t is a bitmap strip but whatever
10:15:58pixelmaS_a_i_n_t: JdGordon| decided that
10:16:04JdGordon|if your strips are not sane then sure you need to look in the code, but if they in fact are sane then you only need to know which bmp it is in
10:17:00S_a_i_n_tim my "battery" bitmapstrip for example, my first 4 images are "charger plugged" "charged" and two for an animation to display during "charging" is that not "sane"?
10:17:14S_a_i_n_tin your way, I'd need to split all those...which is crap.
10:17:19JdGordon|no you wouldnt
10:17:36JdGordon|stop adding extra requirements and saying it cant be done
10:17:41pixelmamy bitmap strip is completely sane, thanks
10:17:45JdGordon|nothing is in code so anything can be done
10:18:05JdGordon|pixelma: oh for crying out loud, dont take it personally
10:18:15S_a_i_n_twell, ok, but its certainly an example of "un-ordered" subimages != "wrong"
10:18:57JdGordon|you are yet to show me an example of a REAL WORLD usage where my answer couldnt be "this is obviously not suited to this" or "we can fix that"
10:19:20S_a_i_n_tI thought I said that.
10:19:33S_a_i_n_tIt can already be done, in a completely sane way IMO.
10:19:44pixelmaJdGordon|: I don't. But I think the discussion becomes a bit silly. Because you claim things themers do currently and which are perfectly possibly being the "wrong way" - while stating before that you don't write WPSs
10:20:01JdGordon|dupleication also says that about half a dozen tags should be removed, namely "replay gain?"
10:20:54JdGordon|pixelma: think of it this way. as someone who doesnt code skins because they are painful to write/understand, I'm looking for ways to make it easier for new people
10:20:57S_a_i_n_tSo, just because we can we should implement 2 or 3 ways to do everything?
10:21:11JdGordon|convincing you two is obviously impossible because you are set in the old way
10:21:45S_a_i_n_tNo, that's totally untrue...I just fail to see the "new" way as being any "better" on "right"
10:22:19S_a_i_n_tI'm totally willing to try new things, if they make sense.
10:22:46 Join Kitar|st [0] (
10:22:59S_a_i_n_tI already have a way which is already coded to do everything you have discussed that I (and other I assume) find perfectly readable.
10:23:14S_a_i_n_tThe sheer volume of themes submitted agrees with that I think.
10:23:18 Join et_ [0] (4d7eee3c@gateway/web/freenode/ip.
10:23:35JdGordon|I disagree
10:23:46JdGordon|the volume of themes just means that people managed to work out the current system
10:23:53n1sJdGordon|: pong
10:23:58JdGordon|it doesnt say anything about how easy/hard it is
10:24:11JdGordon|n1s: dont worry, was hoping to get a h300 test but commited now :)
10:24:18n1sah, ok
10:24:28S_a_i_n_tAnd who's to say people will really find your way any easier?
10:24:48JdGordon|which is why I was always happy to keep the old system around
10:25:12pixelmaas I said earlier, I don't think your way is easier to understand - *especially* when only reading. You say that now we would only need to know how many branches a conditional has when creating the image then not anymore when writing. With the change, you would need to look up how many subimages your playmode.bmp has - so what do you gain?
10:25:29S_a_i_n_tAnd, that's fine...but I only see the point of adding the "new" way if i actually adds something.
10:26:47pixelmaand the example line with "only use state 2,4,x is completely incomprehensible to me
10:27:22JdGordon|pixelma: I dont understand that comment, and that example was made up anyway and would probably not be needed
10:28:50JdGordon|S_a_i_n_t hasnt shown me an example where that request would even be needed
10:29:35pixelmaevery other theme lets the stop image blank
10:29:48S_a_i_n_theh, was just typing that.
10:30:10S_a_i_n_ta stop image is only needed in the .sbs, and will probably be a different resolution.
10:30:26S_a_i_n_thaving "stop" in the wps is pointless.
10:30:36S_a_i_n_tthere's a fine example.
10:30:46JdGordon|a fine example of what?
10:31:15pixelmanot using all conditional branches
10:31:18S_a_i_n_t"not wanting to use the full set of states the tag offers"
10:31:26S_a_i_n_tI thought that was the current dispute.
10:31:48JdGordon|the wps will never go into the STOP state for more than a few milliseconds so no, that isnt an example
10:33:01S_a_i_n_tIt's really hard to argue this when your proposed exaple is theoretically infinite.
10:33:53JdGordon|no it isnt... the worst tag (except the ones which scale) has 31 options... I am yet to see a theme with it probably because it is so hard to use
10:33:58S_a_i_n_tyou're saying "but it *could* do this, and that, and that", eventually it *would* be a complete duplication of the current way of doing this.
10:34:21S_a_i_n_tJdGordon| You haven't seen Symmetry then?
10:34:37S_a_i_n_tthat uses %Li
10:34:41JdGordon|that sets them all to one image doesnt it?
10:35:00S_a_i_n_tyep, but, that's besides the point.
10:35:14JdGordon|not at all
10:35:38JdGordon|obviously in that case you would use %?Li<%xd(image)>
10:36:12S_a_i_n_tassuming you wanted all the states filled though.
10:36:16S_a_i_n_twhat if you didn't?
10:36:45JdGordon|it depends how many you want to actually use
10:36:58S_a_i_n_tthen you either need to use the current way of doing it, or tag on some other functionality to the "new" method.
10:37:27S_a_i_n_twith the current way, it is *really* easy to see where a state doesn't have anything defined.
10:38:03JdGordon|no, you are wrong
10:38:35S_a_i_n_tOh...right, that's because you keep saying "well, it isn't coded it can do anything"
10:38:43pixelmaworse would be my "home stereo" theme where I want to show an image when e.g. bass/treble is *not* 0 so I know I need to load the correct cfg. On the M5 bass/treble can be set in a range of -24 to +24 IIRC - even while the complete conditional is loooong and I can miscount, I don't see how a looong bitmap strip that has the same picture in 48 places and nothing in the other would help me here
10:38:47S_a_i_n_tSo, how can we even argue for or against it?
10:39:02JdGordon|if you dont want to show all 30, (maybe you want ot reuse 1 5 times) the logical way is to just copy that image in the bmp strip so later you can change it much more simply
10:39:37S_a_i_n_tJdGordon|: doesn't that needlessly waste skin buffer.
10:39:46S_a_i_n_ts/doesn't that/that does/
10:40:01JdGordon|you would use FAR more allocating the extra conditionals and options
10:41:10JdGordon|each skin element is 16bytes (I think), each token adds another 24 + strings from the skin are added ontop of that
10:41:25S_a_i_n_tWhen you actually have some code, I'll be happy to revisit this conversation. For now, it is nothing but unproductive. Anything we say you can counter with "well, it isn't written yet, so it can do anything"
10:41:34JdGordon|whereas a 8x8x16 bmp is barely noticable
10:42:06JdGordon|how will starting coding change that? untill i see actual real world examples I dont know what actually needs adding
10:42:33CIA-6New commit by 03bieber (r27675): Theme Editor: Removed markup comments from CodeEditor files (it was originally a Nokia example, so it was marked up with comments for their ...
10:42:40S_a_i_n_tWell, based on the current code and its possibilities *nothing* needs adding...that's my point.
10:42:56 Quit mc2739 (Ping timeout: 240 seconds)
10:43:42S_a_i_n_tAnd having actual code to show would stop this "well, it isn't written yet, so it can do anything" business.
10:43:49 Join DerPapst [0] (
10:44:00S_a_i_n_tWe would also have a specific example of what you believe "cleaner" to be.
10:44:28 Quit ender` (Read error: Connection reset by peer)
10:44:31CIA-6r27675 build result: All green
10:44:32JdGordon|%?mp<%xd(Fa)|%xd(Fb)|%xd(Fc)|%xd(Fd)|%xd(Fe)> is straight out of cabbie... => %xd(playmode, %mp)
10:44:41JdGordon|I'm waiting for a counter example
10:45:00 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
10:45:01 Join ender` [0] (
10:45:04JdGordon|im heading home... take a 45min break to find your worst image line example
10:45:07S_a_i_n_tI gave one "%?mp<%xd(Fa)|%xd(Fb)|%xd(Fc)|%xd(Fd)|%xd(Fe)>" is *much* easier to walk through.
10:45:21 Quit JdGordon| (Quit: leaving)
10:47:53pixelmathis example just proves that things can be written differently and that JdGordon thinks it's more readable while S_a_i_n_t and me don't think so... so I don't know what another example should "counter"
10:48:31S_a_i_n_tMe either.
10:49:13***Saving seen data "./dancer.seen"
10:51:30S_a_i_n_tMy ffw/rew animation is a pretty good example of "how the *fuck* would this work...and would it be any more readable than now":
10:52:15S_a_i_n_tbasically, ffw loads the images in the bitmapstrip in forward order, rew loads them in reverse.
10:52:57S_a_i_n_tOnce you get your hard around its sheer length, it's realy easy to see what is happening IMO
10:56:16pixelmaexample from cabbiev2: %?mm<|%xd(Ea)|%xd(Eb)|%xd(Ec)|%xd(Ed)> -> leaving out the conditional branch for "Repeat off"
10:56:55S_a_i_n_tAll my themes do that.
10:57:13pixelmaI already know his answer "put a blank image into the strip"
10:57:50pixelmaand I seem to have completely baffled everyone with the bass/treble example :P
10:57:51S_a_i_n_twhich needlessly wastes buffer IMO, "Oh, but we have more buffer it's OK"
10:59:10S_a_i_n_tpixelma: It didn't baffle me thankyou! ;)
10:59:47S_a_i_n_tI'd like to see my seek animation in the "new" format.
11:00:30 Quit bmbl (Ping timeout: 258 seconds)
11:01:03S_a_i_n_ts/like/possiblyvery much dislike/ ...though, I'm still open.
11:02:18S_a_i_n_tThe point I think is being missed is that I'm not (nor do I think you are pixelma)saying "No, don't add this as additional functionality" but rather "is it really needed in the first place, if so...why?"
11:02:48pixelmathe bass/treble example is real btw., I just didn't finish the theme yet but I want to one day (and the line is already there as a proof of concept though)
11:03:29S_a_i_n_tmy seek animation is real also...I have a battery animation that I'd need to pastebin.
11:03:34S_a_i_n_tit is *looooooong*
11:05:41S_a_i_n_tpixelma: A real line, in a current theme of mine:
11:05:58S_a_i_n_tit just *seems* daunting, but couldn't be simpler IMO
11:07:14pixelmawow... iCatcher does similar things (also seek, pause and battery charging animations etc.), not to the same extreme but close
11:09:00 Quit et_ (Quit: Page closed)
11:09:48 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
11:10:09S_a_i_n_tThe new parser will let me cut that down *dramatically* though...
11:10:14 Quit MethoS- (Remote host closed the connection)
11:10:33S_a_i_n_tIt's just an example of a really intensely long line, that is still dead easy to read (IMO)
11:15:35 Quit bmbl (Quit: Verlassend)
11:15:44 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
11:16:32S_a_i_n_tpixelma: The new parser lets me convert that line to:
11:16:46S_a_i_n_tSo, I think that is a bit of a win ;)
11:17:17S_a_i_n_tI'm all for the new parser, but "readability" is so terribly subjetive I think.
11:17:50pixelmamy standard "not home" theme doesn't use different image definitions in conditionals much as the choice is made by using conditional viewports mostly. Idea was to show things in different colour depending on state (e.g. battery low, volume muted or above 0dB). I achieve this by using different viewports (with set %Vb/f) and monochrome bitmaps which take foreground and background colour
11:18:24pixelmaon c200 that is - and in a bit cut down version with grey shades on M5
11:19:02S_a_i_n_tHmmm, have you gone through your themes looking for code the new parser lets you do better?
11:19:30pixelmano, I haven't even updated everything yet
11:19:40S_a_i_n_t;) took me a while also.
11:21:09pixelmabut I don't have a lot of animations so I doubt it ;) . I'm trying to think if %Vb/f seperately would help me here but can't think of how currently
11:22:21pixelmahmm... can you pit %Vb/f in a conditional?
11:22:30pixelmaor put
11:22:41S_a_i_n_tsublines nested in conditionals is the only thing I have really found like "OMG, that is *so* much better" so far.
11:23:15S_a_i_n_tpixelma: I'm not sure...I doubt it.
11:23:21S_a_i_n_tpossibly though.
11:25:44pixelmamaybe then I'd only need one viewport and loading the three foreground colours conditionally, instead of 3 differnt viewports loaded conditionally. Would help readability
11:26:16S_a_i_n_tWell, not really...just make smaller code (IMO)
11:28:00pixelmawell one viewport instead of three... taking care of when which is loaded. It looks like this in the old syntax (snippet) . One thing is nice though, I don't have to use conditionals in the viewport though
11:29:18n1soooh, very nasty, our flac deoder doesn't fail gracefully for files with too large block sizes
11:30:29 Quit Judas_PhD (Quit: Aaarrrggghhh!!!)
11:31:05n1sit does give some very nice rythmic bursts of whitenoise though
11:31:44 Join bimbel [0] (~Miranda@unaffiliated/bmbl)
11:31:44 Quit bimbel (Client Quit)
11:33:26pixelmaS_a_i_n_t: but that doesn't have to do with the proposed image and conditional change - in fact, I currently only show one battery icon as "explanation" of the value that follows
11:34:39 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
11:36:59 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:40:56 Part Szpila
11:45:55 Quit S_a_i_n_t (Ping timeout: 245 seconds)
11:48:59CIA-6New commit by 03nls (r27676): flac: error out if max blocksize is larger than we can handle.
11:50:52CIA-6r27676 build result: All green
11:51:32pixelmaZagor: it's about the FM and V2 Recorder manuals. You changed the platform file name and some* options to archosfmrecorder - it was recorderv2fm or similar before because the manual "treats" both the same, you can't see this from the platform file and option names anymore though and may wonder what the V2 one does (you'll only find a few lines in the code and manual stating this)
11:52:38 Quit sasquatch (Read error: Operation timed out)
11:54:32pixelma*) not all options that would have needed renaming have been. Leading to things wrongly included or excluded but I go through this now slowly step by step, because some things can also be written more correctly while at it
11:56:57Zagorpixelma: is the question "was this intentional or a mistake"?
11:57:33 Join efyx [0] (
11:59:39pixelmathe question is more or less how to deal with it - keep pretending in the naming that there is no V2 manual? It's confusing me a bit - the combined name was clear, now it's not that obvious that this is for both manuals, separating them completely would mean unnecessary code duplication
12:00:08pixelmathe combined name wasn't "clear" but gave a hint at least
12:07:10 Quit Dreamxtreme__ (Ping timeout: 245 seconds)
12:10:49 Join sasquatch [0] (
12:11:45 Join S_a_i_n_t [0] (S_a_i_n_t@
12:12:06 Quit linuxguy3 (Read error: Operation timed out)
12:13:08JdGordonS_a_i_n_t: (nice timing :) ) lets try a different approach... the line you pasted really just says "I want the *correct* image for the %mp tag drawn" right?
12:14:15 Join linuxguy3 [0] (
12:15:22JdGordonactually, forget it... I'm going to do the whole patch and see how it goes, if it goes in after 3.7 and people complain that the themes are broken again, well, I wont accept the blame
12:16:31funmanno hurry for 3.7
12:16:40funmanwe can take the time to finish themes work
12:17:02JdGordonbut in reposone to I'd say %?if(%mp, eq, 4)<%t(0.1)%xd(Fa);%t(0.1)%xd(Fb)>
12:17:34JdGordonfunman: well, I dont tihnk there is any support for postponing releases except for bug fixes which this couldnt be considered as
12:18:11JdGordonthat line would be read as "if %mp equals 4 then do the block in the true branch"
12:18:45JdGordonpixelma's supposedly scary example would be done the same way
12:18:54 Join anewuser [0] (anewuser@unaffiliated/anewuser)
12:19:50pixelmawhere, what?
12:20:20pixelmaalso, what's %?if
12:20:43funmanJdGordon: on the other side, is there support for pushing releases?
12:21:22JdGordonfunman: imo releases are stupid and a waste of time :D I'm in a very small minority though
12:22:01JdGordonpixelma: which I assume is what is referring to
12:22:08funmanJdGordon: we'll be two then
12:22:26JdGordon%if is a new tag which lets you do basic logic
12:23:00funmanno %switch or %for ? :/
12:23:24JdGordon%? is bassically a swicth anyway
12:24:19JdGordon"for" could be done easily as a counter
12:24:47funmanhow can i exit rockbox on android?
12:25:00JdGordonforce close it in the settings :)
12:25:40 Quit anewuser (Ping timeout: 240 seconds)
12:26:22pixelmaso, how is that %?if line more understandable, despite being shorter
12:26:55 Join anewuser [0] (anewuser@unaffiliated/anewuser)
12:27:54JdGordonyou honestly dont tinhk that is easier to read than "%t(0.1)%?mp<|||%xd(Aa)|%xd(Ae)>;%t(0.1)%?mp<|||%xd(Ab)|%xd(Ad)>;%t(0.1)%?mp<|||%xd(Ac)|%xd(Ac)>;%t(0.1)%?mp<|||%xd(Ad)|%xd(Ab)>;%t(0.1)%?mp<|||%xd(Ae)|%xd(Aa)>;%t(0.1)%?mp<|||%xd(Af)|%xd(Ah)>;%t(0.1)%?mp<|||%xd(Ag)|%xd(Ag)>;%t(0.1)%?mp<|||%xd(Ah)|%xd(Af)>"
12:28:06JdGordonactually I just noticed my line is wrong
12:29:49JdGordon%if( %mp, eq, 3)<%xd(animation, %loop(0.1, 1, 5, 1)>%if( %mp, eq, 4)<%xd(animation, %loop(0.1, 1, 5, -1)>
12:30:20JdGordonno need to try to figure out which subline and which branch and which conditional you are in
12:32:48 Quit AndyI (Ping timeout: 248 seconds)
12:33:10JdGordonwe could go RPN and do %?eq(%mp, 3)<|> even :)
12:39:35 Join AndyI [0] (~pasha_int@
12:42:12*S_a_i_n_t personallt thinks %if( %mp, eq, 3)<%xd(animation, %loop(0.1, 1, 5, 1)>%if( %mp, eq, 4)<%xd(animation, %loop(0.1, 1, 5, -1)> looks *terrible*
12:42:21S_a_i_n_tAnd prefers his massive subline.
12:42:53S_a_i_n_tits implementing a buttload of things for stuff that can already be done.
12:43:21JdGordonthat doesnt mean the current way is good
12:43:34S_a_i_n_tNor does it mean yours is any better.
12:44:06JdGordonbetter is obviously subjective, so at least we can agree it is different
12:44:32 Quit AndyI (Ping timeout: 248 seconds)
12:44:57JdGordonand no, %if(%mp, eq, 3) is FAR more readable and FAR less likely to cause problems than %?mp<||||...>
12:44:58S_a_i_n_tafter you implement all this though, it *will* ne an almost complete duplicate of current what motivation is there to keep the current code in place?
12:45:05*S_a_i_n_t thinks that answer is easy, none.
12:45:10S_a_i_n_tAnd, it sucks balls.
12:45:32S_a_i_n_tand no, %if(%mp, eq, 3) is FAR more readable and FAR less likely to cause problems than %?mp<||||...> <−−says who?
12:46:02S_a_i_n_tI thought the idea was to simplify it?
12:46:03JdGordonsays me... and I'm sure I can get others to agree... counting bars sucks ass
12:46:10S_a_i_n_tA;;; I see there is overcomplication.
12:46:12JdGordonand your line wouldnt be possible to be removed...
12:47:15S_a_i_n_t"I bet I can get others to agree" just sounds like a majority vote threat....not looking at the "if it 'aint broke, don;t "fix" it" principle.
12:47:33S_a_i_n_tIf I thought what you were doing was actually "fixing" anything, I wouldn;t care.
12:48:11S_a_i_n_tQuite simply, it ain't broke. All I have seen you say is really "I (you, personally) don;t understand it"
12:48:26S_a_i_n_t...learn to understand it, why implement a brand new way?
12:49:03JdGordonbecause IMO the current way is shit, and the benefit of being able to code means I can do something about it
12:49:15***Saving seen data "./dancer.seen"
12:49:24JdGordonI tinhk its pretty stupid to think *I* dont understand it
12:49:51S_a_i_n_tYou understand the code, but not how a WPS should be written optimally IMO
12:50:15S_a_i_n_tthe fact two themers outright dissagree has to say
12:50:51JdGordonI disagree
12:50:59JdGordonit just says they are close minded :)
12:51:02JdGordonstuck in the old way of doing things
12:51:05S_a_i_n_tBut I know enought by now to know that whatever the conclusion, you'll do it, fuck it.
12:51:16 Join AndyI [0] (~pasha_int@
12:51:42JdGordonI probably wouldnt have got the extra ideas if you didnt ocme up with nonesense examples :)
12:51:59S_a_i_n_tthat's bullshit and you know it.
12:53:53JdGordonbugger... does the theme editor not ocmpile in svn?
12:54:29JdGordonbieber: ping?
12:55:19S_a_i_n_tI wonder, seriously why you even bothered to ask anyones opinion about this.
12:55:48S_a_i_n_tYou were given it, but, you didn't like it so it doesn't matter to you...that's crap.
12:56:06JdGordonobviously to get the discussion for the ideas to be finished before starting codeing... obviously I didnt care what you were going to say.... :/
12:58:24S_a_i_n_tWell, yeah. I don't see how that can have any air of sarcasm.
12:59:59 Join stooo [0] (
13:00:53 Quit milz (Read error: Connection reset by peer)
13:03:27pixelmawell, it boils down to "a different way" and after reading the examples 5 times I can see that it's easier to read in *some* cases but *only* if you implement all the other stuff. IMO that's much too much at once and I'd like to also be convinced that this is simpler and at least not too bin- or RAM-size heavy too. If you want so much at once, you should rather think about creating a whole new theme language, well thought through from the beginning
13:03:27pixelma :\
13:04:14 Quit stooo (Client Quit)
13:04:38S_a_i_n_tOr pick something to "imporve" that is flawed to begin with.
13:04:45pixelmathat "eq" thing was only understandable btw. after realising that it stands for "equal"
13:04:51S_a_i_n_t*improve even
13:05:49JdGordonpixelma: heck, I could make it "equals" instead of "eq" to make it more readable
13:06:42S_a_i_n_tI can't see what happens (or would happen) if you want tag states to be independently conditional.
13:08:30 Join mt_ [0] (~mtee@
13:08:43 Join milz [0] (
13:08:56S_a_i_n_t%?mp<%xd(Aa)|%?C<%xd(Ab)|%?xd(Ac)>|%xd(Ad)|etc. does that work in your scenario?
13:09:16JdGordonit doesnt
13:09:30JdGordonI'm not replacing anything... im giving you another option
13:09:35*S_a_i_n_t has no use for this new implementation.
13:09:36 Quit mt (Ping timeout: 240 seconds)
13:09:40 Nick mt_ is now known as mt (~mtee@
13:09:44JdGordonok, fine, good for you then
13:10:31pixelmaI'd find another option a waste then too
13:11:43S_a_i_n_tIMO if you're going to duplicate features, the newer implementation should at least be able to do what the old one does first, before adding to it.
13:11:59pixelmayou can already get the results with the current implementation. If the new one doesn't add much wrt possible results - having to things doing the same is a waste
13:12:46S_a_i_n_tseeing different code doing the same thing in a theme defeats the "readability" argument also.
13:12:51JdGordongevaerts: how do I get libskin_parser to compile with -g for checkwps?
13:14:05funmanJdGordon: add -g in CFLAGS
13:14:10JdGordonGCCOPTS in the checkwps Makefile does have -g
13:14:29funmanlib/skin_parser/Makefile ?
13:15:34funmanbtw perhaps we should always force -g
13:15:59funmanand strip the binaries in specific cases (e.g. when creating android package)
13:16:20JdGordonah got it
13:16:26JdGordonI think I just had to make clean && make :p
13:16:43pixelmamaybe JdGordon's new syntax can be a translation layer in the theme editor > he can read it all he wants and it won't cost anything on target. No functionality is lost or duplicated on the target and one can read it the old way too </actually semi-serious>
13:16:47 Join Topy [0] (
13:17:09*JdGordon mashes head on keyboard
13:17:33JdGordonyou understand that the current code is less optimised for target than my suggestion right?
13:17:38JdGordonin absolutly every way
13:17:50S_a_i_n_tyet, it does more.
13:18:58funmanat least, it is a good flaming material
13:19:04pixelmayes, that's why it's not optimised for target. Having two implementations just can't be
13:19:35pixelmathis time I really mean "size" of the parser too
13:20:17S_a_i_n_tWhere did the specific need for this pop up? Did anyone (other than yourself) actually say "Gee, the current way %xd works regarding sublines is really confusing and impossible to read"?
13:20:30 Quit funman (Quit: leaving)
13:20:31 Quit T44 (Ping timeout: 265 seconds)
13:24:52AlexPJdGordon: Don't bother arguing - he's a "themer" and therefore all knowing about themes
13:25:31S_a_i_n_tYeah, that's construstive...thanks for that.
13:25:42AlexPAs was much of your argument
13:25:58AlexPFor what it is worth, I quite like the idea, but I'm not bothered enough to insist
13:27:19S_a_i_n_tThink of it this way, would another tetris plugin, that actually does less than the current one, have any shot of being committed?
13:27:39AlexPI just love lines like "[12:49:50] <S_a_i_n_t> You understand the code, but not how a WPS should be written optimally IMO"
13:27:40gevaertsSee rockblox1d
13:27:43n1swe did include such a plugin
13:27:44AlexPSo condescending
13:27:53AlexPI am the themer, only I can have an opinion!
13:29:34AlexPYou can of course take any opinion you like on something, but to discount someone's opinion as they aren't a "themer" and you are the king of all themers (apparently) is just ridiculous
13:30:05S_a_i_n_tThat's your perception of it, and you can have that if you want...but you're incorrect.
13:30:17AlexPIt is how you come across
13:30:24JdGordon*ahem* time to shut the fuck up :) this isnt a social channel
13:30:52S_a_i_n_tit's a pretty fucking antisocial channel for the past few hours...
13:30:52 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
13:31:17AlexP"[12:50:14] <S_a_i_n_t> the fact two themers outright dissagree has to say" implies very heavily that people on the other side of the argument aren't "themers" and therefore their opinion doesn't count
13:31:40S_a_i_n_tNo, it doesn't.
13:31:47S_a_i_n_tThat's just the way you want to read it.
13:33:18S_a_i_n_tIf I wanted to say, or even thought anyones opinion didn;t count...I'd damn well say so.
13:34:06JdGordonNOOOOOO!!!!!!!!!! CRAP! I think my parser optimisations last week may have screwed me here
13:34:29 Join TehSpawn [0] (
13:35:18 Quit TehSpawn (Client Quit)
13:35:43*gevaerts doesn't really like the idea of multiple ways to achieve the same things (even if they do differ in some features), but he hasn't followed the discussion enough to state a firm opinion
13:36:02 Join SpawnHyuuga [0] (
13:36:16 Join kugel [0] (
13:37:00*kugel also dislikes redundancy
13:37:21 Quit SpawnHyuuga (Client Quit)
13:39:50 Join piggz_ [0] (~piggz@
13:51:29*JdGordon needs to thank bieber again for making it so easy to add a new tag :D
13:55:06 Quit kugel (Ping timeout: 240 seconds)
13:57:16 Join kugel [0] (
13:58:50 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
13:59:19 Quit piggz_ (Ping timeout: 246 seconds)
14:02:39 Nick fxb__ is now known as fxb (
14:04:52 Quit anewuser (Quit: FAMICOMPO 7 - August 15 Through August 31 JST -
14:06:24JdGordonETOOMUCHINDIRECTION element->params[0].data.code->data
14:14:52CIA-6New commit by 03jdgordon (r27677): make -vvv display the parse tree in checkwps. Fix a potential bug in the parser where recursive tags (the playlist viewier) would share params with ...
14:16:51CIA-6r27677 build result: All green
14:22:30 Quit feisar_ (Ping timeout: 260 seconds)
14:24:16 Quit bluebrother (Ping timeout: 240 seconds)
14:24:35 Join fdinel [0] (
14:25:07 Quit TheSeven (Read error: Connection reset by peer)
14:26:11 Join bluebrother [0] (
14:26:11 Quit bluebrother (Changing host)
14:26:11 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
14:28:32kugelFlynDice: ping
14:40:30JdGordonbieber: can you let me know what I need to change to make FS #11516 work in the editor please?
14:41:55JdGordonif anyone wants to play I would welcome comments (other than "this adds nothing")
14:49:17***Saving seen data "./dancer.seen"
14:55:35 Join AndChat| [0] (
14:55:55AndChat|"this adds nothing" is not a valid reason anymore?
14:55:55 Quit kugel (Ping timeout: 245 seconds)
14:56:22 Nick AndChat| is now known as kugel (
14:57:02 Quit mitk (Quit: Leaving)
14:57:29 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
14:57:43 Join robin0800 [0] (~robin0800@
14:58:02 Quit efyx (Remote host closed the connection)
14:59:40 Join fdinel [0] (
15:00:40 Quit fdinel (Client Quit)
15:02:15 Join fdinel [0] (
15:02:58JdGordonkugel: it is, I'm just making it known I would ignore that particular one
15:03:26 Quit antil33t (Read error: Connection reset by peer)
15:03:32 Join antil33t [0] (
15:04:34 Join efyx [0] (
15:06:50 Join detrick [0] (
15:07:42 Join feisar [0] (
15:08:10 Nick feisar is now known as Guest69382 (
15:09:05 Join piggz_ [0] (~piggz@
15:14:37 Join r0b- [0] (
15:17:37 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:26:51 Quit r0b- (Ping timeout: 276 seconds)
15:30:23 Join r0b- [0] (
15:30:50FlynDicekugel: pong
15:33:01kugelFlynDice: does it play music well even if if it's in the background, I.e. if you use other apps?
15:34:52FlynDicekugel: I think so, lemmee test it some. Played fine while I downloaded a terminal app and went through all the settings pages. I'll test it harder right now.
15:35:57 Part LinusN
15:41:07 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
15:41:42 Join jgarvey [0] (
15:41:59preglowkugel: you have an android build hanging around for people to test?
15:42:53kugelpreglow: I linked one for 320x480 tonight
15:43:50preglowwee, that's my reso
15:44:01kugelit seems to run also on other resolutions but it'll be a bit awkward
15:46:00preglowbout when?
15:46:16FlynDicekugel: It plays fine even when using other apps. I played several games with rockbox playing and there were no hiccups at all.
15:48:01kugelpreglow: 2:02, you need 2.x for it
15:48:10preglowgot 2.x
15:49:08kugelFlynDice: cool, at least on your phone it works as it's supposed to
15:49:36preglowi can't find the url for the life of me
15:49:45kugelvarious HTC phones ship a broken Android
15:51:59kugelsaratoga: saratoga:
15:52:38kugelpreglow: ^
15:54:48 Quit slck (Ping timeout: 276 seconds)
15:55:19tmztkugel: borken how?
15:55:26 Quit kugel (Quit: Bye)
15:55:53 Join kugel [0] (
15:57:38kugeltmzt: startforeground() doesn't work as advertised
15:58:18tmztkugel: you might ask romainguy about that in #android during the afternoon pacific time
15:58:44tmztare you using the media scanner or doing that internally to rockbox?
15:58:59tmztand will rockbox have access to the sd card directly (similar to simulator I guess)
16:00:32tmztto which? sorry for so many questions so quickly
16:01:14tmztare you using the media scanner api?
16:01:21kugelit has direct access to the sdcard
16:02:02tmztmedia scanner isn't very good anyway, thats seems to be the consensus and a source of many questions in the channel
16:02:46kugelwe use whatever api rockbox has for this
16:03:27 Join slck [0] (Venci@No.Real-Love-For.Me)
16:04:42 Quit r0b- (Ping timeout: 240 seconds)
16:10:15 Quit mt (Ping timeout: 258 seconds)
16:12:14 Join mt [0] (~mtee@rockbox/developer/mt)
16:12:29 Join r0b- [0] (
16:17:14kugeltmzt: does romainguy have a clue
16:17:43tmztkugel: he developed the main android apis, I would think so
16:17:57tmzt(he's a google employee)
16:18:04tmztoh, that should be #android-dev not #android
16:18:59kugelI was in that channel a few times now. most of the people there are no help at all
16:19:03JdGordonbieber: please have a look at FS #11517 which adds optional tag param groupings to the parser (so it guesses which type of param has been given). let me know if you have comments... I have a few ideas which would need this but maybe there is a better way to add this?
16:19:40 Quit JdGordon (Quit: Leaving.)
16:19:41 Quit piggz_ (Ping timeout: 245 seconds)
16:20:18tmztkugel: if I see him I'll mention it, they are usually good if you know what to ask
16:30:44 Join panni_ [0] (
16:37:56 Quit kugel (Remote host closed the connection)
16:45:39 Join saratoga [0] (9803c20d@gateway/web/freenode/ip.
16:45:50preglowhow to change font in this android business? :P
16:48:03*evilnick_B points to the topic and stares at preglow :)
16:48:58 Quit mt (Ping timeout: 258 seconds)
16:49:21***Saving seen data "./dancer.seen"
16:49:49pixelmaevilnick_B: he's testing RaaA...
16:49:57preglowwhat, i should read the android docs which dont exist?
16:51:24evilnick_Bpreglow, pixelma: Apologies for that
16:54:14 Join mt [0] (~mtee@rockbox/developer/mt)
16:56:04 Join newnicj [0] (
16:57:07 Quit newnicj (Client Quit)
16:57:27 Part Zagor
16:58:55 Join captainkewll [0] (2669ecc2@gateway/web/freenode/ip.
17:00:13preglowheh, shit crashes if i tilt the phone
17:01:00 Join Judas_PhD [0] (
17:03:30saratogai noticed that
17:03:38saratogai think its because we hardcode the resolution in rockbox
17:03:50saratogaprobably doesn't like the LCD resolution changing on the fly :)
17:03:53gevaertsRaaA is only stable if the phone is kept stable :)
17:03:58preglowah, so that's how it's handled
17:04:09tmztyou basically have to recreate the activity when the orientation changes
17:04:09preglowit's pretty sluggish
17:04:14preglowbut madly impressive
17:04:43preglowthis has got to be one of the most complex android apps ever
17:04:46preglowif not the most complex
17:05:12saratogais the long term goal to reimplement the lists as (hardware accelerated) Android widgets, or to leave everything using the NDK rockbox code on the CPU?
17:05:57preglowandroid lists would definitely be the way to go, if you ask me
17:06:17tmztdoes this work on pre 2.2 versions?
17:06:28tmztare you using the newest NDK and opengl directly?
17:06:38preglowworks on 2.1
17:06:39tmztor should that be 2.1
17:07:25 Nick YPSY is now known as Ypsy (
17:08:42 Quit DarkJuju (Remote host closed the connection)
17:10:00 Quit milz (Remote host closed the connection)
17:13:12saratogai wonder how hard it would be to redo the list code for Android
17:14:37tmztthen what is the goal here?
17:15:50 Quit antil33t ()
17:16:06 Join milz [0] (
17:26:49 Join funman [0] (~fun@rockbox/developer/funman)
17:34:17 Join stoffel [0] (
17:37:53 Join DarkJuju [0] (juju2143@gateway/shell/
17:40:26 Quit efyx (Remote host closed the connection)
17:54:34 Quit petur (Quit: *plop*)
17:55:37funmankugel: ping
17:57:10 Join komputes [0] (~komputes@ubuntu/member/komputes)
17:57:28 Join clone4crw [0] (
17:59:00 Quit clone4crw (Client Quit)
17:59:05 Quit detrick (Quit: Leaving)
17:59:13 Quit Utchybann (Ping timeout: 260 seconds)
18:03:58 Quit robin0800 (Quit: Leaving)
18:07:36 Join Utchybann [0] (
18:13:25 Quit Stummi (Excess Flood)
18:13:33 Join Stummi [0] (
18:21:00 Join piggz_ [0] (~piggz@
18:26:13 Quit DarkJuju (Remote host closed the connection)
18:26:59 Quit piggz_ (Ping timeout: 265 seconds)
18:27:51 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:28:05 Join piggz_ [0] (~piggz@
18:32:18 Quit bieber (Ping timeout: 265 seconds)
18:32:33 Join bieber [0] (
18:39:33 Quit piggz_ (Ping timeout: 265 seconds)
18:44:21 Join piggz_ [0] (~piggz@
18:49:24***Saving seen data "./dancer.seen"
18:52:36 Quit bieber (Ping timeout: 276 seconds)
18:52:52 Join bieber [0] (
18:53:57 Quit jfc (Read error: No route to host)
18:54:39 Join jfc [0] (
18:55:13*funman can't build rockbox for android
18:56:49 Quit AndyI (Ping timeout: 248 seconds)
19:00:29 Join clone4crw [0] (
19:01:36 Quit clone4crw (Client Quit)
19:02:21 Join root [0] (
19:02:35 Nick root is now known as Guest3476 (
19:02:36 Quit bmbl (Quit: Verlassend)
19:02:46Guest3476Hi, everyone
19:02:53 Join AndyI [0] (~pasha_int@
19:03:08Guest3476At the moment I'm trying to unbrick a Sansa Fuze V2 that I got off ebay
19:03:33funmanhow do you know it's bricked?
19:04:01 Join bmbl [0] (
19:04:03Guest3476Doesn't power up, nothing on screen
19:04:06 Quit bmbl (Changing host)
19:04:07 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
19:04:20tmztfully charged?
19:04:21Guest3476But it seems to be recognised by Linux
19:04:24funmancheck SansaAMS for specific problems, this problem is known, explained, and there is a solution.
19:04:28Guest3476I wouldn't know, I assume so
19:04:39Guest3476Well, when I run lsusb, it shows up
19:04:52 Join DarkJuju [0] (juju2143@gateway/shell/
19:04:58 Join bimbel [0] (~bmbl@unaffiliated/bmbl)
19:05:00Guest3476and when I run fdisk -l, it also is recognised, but there is no partition table
19:05:04 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
19:05:05 Quit bmbl (Client Quit)
19:05:13 Nick bimbel is now known as bmbl (~bmbl@unaffiliated/bmbl)
19:05:35Guest3476so I tried to dd its OF, but it seems to only write about 4 MB out of the 15MB
19:06:08funmanGuest3476: the unbrick page on the wiki says the device must be ~ 1GB
19:06:29funmanyou shouldn't have opened it anyway
19:06:49Guest3476It's a 4gb model, and reporting about that much
19:07:28funmanhow much sectors? (look in dmesg)
19:08:58simonrvn'how many'
19:09:46Guest3476unless I just don't see it, dmesg isn't saying
19:10:09Guest3476unless you mean blocks
19:10:25Guest34768265 512-byte logical blocks
19:10:47funman8265 * 512 is far away from "about 4gb"
19:13:24 Join robin0800 [0] (~robin0800@
19:14:16Guest3476yes, my bad on that.
19:14:55 Join einhirn [0] (
19:15:06funmanstop messing with unbricking, pray for not having broken your fuzev2, and follow the advice on
19:15:13gevaertsDoes that bit need the OF (as downloaded) written to it, or something slightly different?
19:15:26funmangevaerts: SansaAMSUnbrick should be clear about it
19:15:37 Quit bieber (Ping timeout: 252 seconds)
19:15:54 Join bieber [0] (
19:16:11 Quit einhirn (Client Quit)
19:18:47Guest3476what precisely am I looking for in the Specific Problems page?
19:19:28funmanlast line before 'Port status'
19:20:13 Quit bmbl (Quit: Verlassend)
19:21:00Guest3476ah. no hardware power off.
19:22:04Guest3476I don't suppose desoldering the battery would be a quicker alternative than waiting a couple days
19:22:21funmanwhy? it's much quicker if you can solder it back
19:22:25 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
19:22:43 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:23:00funmanFWIW, there is a hardware power-off, but somehow it doesn't work in this case. No idea why :/
19:25:30 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:26:02 Quit stoffel (Ping timeout: 260 seconds)
19:26:50kugelfunman: pong
19:27:30Guest3476i really wish i knew what state the battery was in
19:27:41funmankugel: i could build an apk after some fixes to wma and 1 hour messing with eclipse
19:27:58kugelwma doesn't built right
19:28:07kugelbut that shouldn't be a problem :)
19:28:10funmanhowever i can't install it - not sure why: my phone won't tell me why
19:28:30 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:28:36kugelfunman: is it signed?
19:29:12kugelI noticed it (silently) rejects unsigned apks, but eclipse should sign it for you if you run it
19:29:35funmanah, it works after desinstalling the previously installed apk (upgrade didn't work)
19:30:35funmani just removed 'debuggable' field
19:31:01 Quit Guest3476 (Quit: leaving)
19:33:59 Quit milz (Remote host closed the connection)
19:34:09funmankugel: can i send you patches?
19:34:32funmani'm tracking gsoc-android-2 branch
19:35:24kugelyou could also commit to svn
19:35:56funmanah nice
19:36:21 Quit bertrik (Ping timeout: 245 seconds)
19:36:43kugelrasher: can you try the apk I linked to tonight (at 2:02)?
19:37:03kugelI'm wondering if the problem is fixed in htc's froyo
19:38:40 Join liar [0] (
19:41:06 Join stripwax [0] (
19:41:31CIA-6New commit by 03funman (r27678): configure: fix LCD width/height dialog for android
19:41:38CIA-6New commit by 03funman (r27679): wma: move inline functions into .h file ...
19:42:23funmanhelp message didn't correspond
19:42:38Bagder"Please enter" would sound a lot better in my ears
19:42:39funmanso it should have been 'fix help message'
19:43:23CIA-6r27678 build result: All green
19:43:25funmanor just 'Enter the LCD xxx' ?
19:43:57funmankugel: + echo "Enter the LCD width (default: 320)"
19:44:00funmanworks for you?
19:44:29kugelwell, it was probably confusing becuase it says 320x480 in the parenthesis, but it said "separated with enter" :)
19:44:43CIA-6New commit by 03funman (r27680): Typo
19:45:10funmanah, i skipped this, "320x480" was a long enough description ;)
19:45:31CIA-6r27679 build result: All green
19:45:44kugelfunman: the problem with wma is possibly that I added -std=gnu99 and inline of c99/gnu99 behaves differently as the inline in gnu89 (the gcc default)
19:46:18pixelmadoes this commit message mean you introduced a typo? ;)
19:46:18kugelthe headers that ship with the android toolchain complain a lot, including about _STDC_ not defined
19:46:37funmani googled and found this (on gcc 4.3 release notes) but didn't check the exact difference between the two. I don't think we specifically need extern inline, do we ?
19:47:16CIA-6r27680 build result: All green
19:47:40kugelit doesn't make sense to me anyway
19:48:00 Quit DerPapst (Quit: Leaving.)
19:48:26funmanwe need a new header to put memset16()
19:49:01kugelIIRC we have a memory.h, but it would also fit into string-extra.h I guess
19:49:28funmanwhen the host memory.h is used we don't have it
19:50:13kugeloh, I didn't know that exists on hosts
19:50:20kugelit just includes string.h on my system
19:51:27 Join milz [0] (
19:52:42funman/media/bordel/rockbox/apps/main.c:189: warning: function declared 'noreturn' has a 'return' statement
19:53:03 Join root [0] (
19:53:09 Nick root is now known as Guest8399 (
19:53:12 Quit bieber (Ping timeout: 240 seconds)
19:53:26 Join bieber [0] (
19:53:40CIA-6New commit by 03funman (r27681): android: declare audiohw_set_volume()
19:53:45CIA-6New commit by 03funman (r27682): system-android: cast void* to uintptr_t*
19:55:20funmanbtw the code sourcery people said they 'added my bug report into the database' and it should be fixed at next upgrade ("fall")
19:55:26CIA-6r27681 build result: All green
19:55:31kugel&stack wasn't void*, it was volatime uintptr_t*
19:55:36 Quit pamaury (Remote host closed the connection)
19:55:49Guest8399fuman: earlier I was the one dealing with a supposedly bricked fuzev2. I had to reboot my machine. I've disconnected the black (-) wire for the battery, and it the sansa is being reliably recognised. at this point, I am not sure what to do
19:56:33kugelfunman: I would prefer CONFIG_PLATFORM & PLATFORM_ANDROID over defined(ANDROID)
19:56:41 Quit milz (Remote host closed the connection)
19:56:45funmani copied the def from sound.c ^^
19:57:04funmankugel: but stackbegin/end are uintptr_t* (above)
19:57:07CIA-6r27682 build result: All green
19:57:14funmanGuest8399: does it work?
19:57:45kugelfunman: no
19:57:57kugel"uintptr_t *stackend;"
19:58:17kugelerr, that's what you said :)
19:58:54rasherkugel: it seems to work. What am I looking for?
19:59:09kugelrasher: does it play music well when you are in other apps?
19:59:19funmankugel: why uintptr_t* btw?
19:59:32funman(why the '*' also)
19:59:36rasherkugel: ah.. let me see.. where do I put music?
19:59:53Guest8399it is recognised, but nothing else seems to work.
19:59:58 Join DerPapst [0] (
20:00:00CIA-6New commit by 03funman (r27683): faad: move inline function in header
20:00:04funmanGuest8399: what's "recognized" ?
20:00:25kugelrasher: it might be news for you, but historically you can put music whereever you want for rockbox :P
20:00:51*funman misses something
20:00:52kugelfunman: stackend/begin are declared uintptr_t[] on native builds
20:01:16 Join stoffel [0] (
20:01:18kugelit comes from the linker script
20:01:25rasherkugel: I get an "error accessing playlist control file", I assume that's known?
20:01:26Guest8399shows up in lsusb as SanDisk corp, and as a 4mb device without a partition table in fdisk
20:01:35funmankugel: why volatile?
20:01:47CIA-6r27683 build result: All green
20:01:48kugelfunman: to make gcc put it on stack
20:01:56rasherkugel: seems to work beautifully
20:02:04funmanshouldn't it be static instead?
20:02:16kugelrasher: that's not expected, no. didn't it create a /sdcard/rockbox directory for you?
20:02:18funmanhmm i see
20:02:55rasherkugel: hm, now I got a bit of stutter during a screen transition. Managed to open google earth and move around a bit without hiccup though
20:03:12kugelthat's only to get a semi-valid stack address. but effectively it's not used anywhere anyway (only for stack usage statistics which is disabled for the main thread)
20:03:13rasheralso, trying to re-enter Rockbox crashes it
20:03:43kugelrasher: alright. you'd do me a favor if you could check top when it's running
20:03:56rasherkugel: I don't see a /sdcard/rockbox, no
20:03:57kugelwhat it says under PCY for the org.rockbox process
20:04:08 Join BRi7X [0] (
20:04:48kugelit's a bug if it didn't, but you can create it yourself (it's where rockbox stores config.cfg, etc)
20:05:34BRi7Xhowdy, I just bought a used Gigabeat S30, the hard drive appears to be corrupted as I can't quite copy any files to it. I was wondering if it'd be possible to replace the hard drive inside of it with a 40GB hdd from a Gigabeat F?
20:06:01evilnick_BBRi7X: It's the wrong connector inside, unfortunately
20:06:28evilnick_BThe S uses a ZIF connector whereas the F uses IDE
20:07:07gevaertsBRi7X: what exactly do you mean by "the hard drive appears to be corrupted"?
20:07:08BRi7Xthat's the same connector as an iPod 5G, the ZIF, isn't it?
20:07:09rasherkugel: 9633 35% S 10 172696K 26288K fg app_70 org.rockbox
20:07:33kugelso I can expect my phone to work properly once I get a froyo update
20:07:50BRi7Xwell, when i first opened the box, I've tried to sync music with WMP using the original firmware, it seemed to have gone on fine, but the interface was incredibly laggy on the device and nothing would work, that was my first clue
20:08:01funmanand again: Problems: "Unable to resolve target 'android-4'"
20:08:04kugelrasher: it says "bg" on my phone, due to a bug in the OS. that bg makes rockbox barely usable if it's in the background
20:08:12rasherkugel: surely there must be a way?
20:08:19rasherI mean, there are other media players
20:08:20funmancan you help me decipher this?
20:08:31rasherAlso, oh my, 35% cpu usage
20:08:34funmanoh done..
20:08:58BRi7Xgevaerts: secondly, after beast patching it and revealing the two partitions, copying the rockbox software onto it stayed stuck at 1% and 0kb/s.
20:09:48kugelI guess the builtin media player / mediaplayer java class doesn't have this problem for some reason
20:10:11 Join milz [0] (
20:10:17gevaertsBRi7X: right. Yes, it could be the disk of course
20:11:36BRi7Xgevaerts, might you have any suggestions as to what i should do? or is there any chance whatsoever that i'd be able to adapt the connector from my gigabeat F hard drive over onto the S?
20:12:32gevaertsBRi7X: I think the first thing to do would be to make sure this is indeed the problem, possibly by having the OF reformat the disk and see if the problems remanin
20:12:39 Join watto [0] (~watto@
20:13:45gevaertsAs for adapting the F drive to the S, I'm not sure. Adapters do exist, but I don't know if the case has enough room. I think the S firmware should handle the disk itself file
20:13:56*gevaerts would *never* move a disk from the F to the S. He much prefers his F (although he would get an X if he found one)
20:14:15BRi7Xthe F was nice, it just kept falling apart, i have a few pictures...
20:14:25 Quit milz (Remote host closed the connection)
20:14:30 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
20:14:33gevaertsYes, it tends to do that :)
20:14:54BRi7Xgevaerts: i remember a while ago my friend wanted to put his 80GB ipod HDD into his gigabeat, we ran into the ZIF connector problem, and researched and researched then just sort of gave up
20:15:06BRi7Xbut i mean, it seemed possible, we just didn't put forth the effort
20:15:33gevaertsYou'll have to decide if it's worth trying, or if getting a new ZIF disk is a better idea...
20:16:50BRi7Xit seems to be working now, i'm not sure what i did haha
20:16:53bertrikslashdot has an article about apple chargers, with a link to an article revealing the magic resistor values in the charger
20:17:23BRi7Xgonna test it with the new blind guardian flac really quick.
20:18:27kugelrasher: it still plays fine?
20:18:43rasherkugel: oh I stopped it
20:18:48kugelthe cpu usage can be reduced
20:19:17kugeldriving our UI is quite expensive, but that can of course be stopped when it's in the background or when the backlight is off
20:19:30kugelalso, the peakmeter on the wps makes it even worse
20:19:47rasherI sure need a larger theme
20:19:57kugelI expect the desire can do rockbox sans UI with under 5% cpu usage
20:20:51funmankugel: do you use eclipse or CLI tools?
20:20:55rasherIt doesn't help that I've never used a touchscreen target (and the screen size is invisible to me)
20:21:09rasherokay, playing music now, in the background
20:21:10kugelfunman: eclipse. but I made the apk once via CLI
20:21:31funmaneclipse throws me a 'Conversion to Dalvik format failed with error 1'
20:21:39rasheror not.. messed up
20:21:43kugelrasher: on my phone it already drops out when I visit the app menu and scroll a bit
20:22:06rasherhow do I leave the wps?
20:22:23funmani also needed this diff:
20:22:42rasherI guess I could read the manual..
20:24:41rasherTop-left, it turns out. Trying to make it skip now.. any ideas?
20:25:09kugel"make it skip"?
20:26:25 Quit hebz0rl (Quit: Ex-Chat)
20:26:54 Quit bmbl (Quit: Verlassend)
20:27:10 Quit mt (Ping timeout: 258 seconds)
20:27:49 Join milz [0] (
20:28:12rasherSeems to hover around 10-15% cpu usage now that I left the WPS
20:29:08kugelIIRC it still does two lcd updates per second even in the main menu
20:30:33 Join kyle__ [0] (
20:31:30kugelrasher: 10-15% already beats the built in media player, so it's not that bad. but it can go down more, I had my phone at 6-8%
20:31:53 Quit stoffel (Remote host closed the connection)
20:31:55funmankugel: what's missing for using asm code?
20:32:07kugelpreprocessor/makefile magic
20:32:27n1si think performance is not the thing to focus on at this early stage
20:32:43rasherYeah, I wasn't seriously suggesting that
20:33:00rasherAlso, if we're not even using asm code yet, it should go way way way down
20:33:33n1srasher: maybe, our asm code isn't really written for the kind of advanced cores in these phones though
20:33:37funmankugel: can i work on that without stepping on your toes?
20:33:40 Quit milz (Ping timeout: 276 seconds)
20:33:46rashern1s: well, I imagine it'd still go down?
20:33:53rasherJust not as much as it could
20:34:09funmann1s: you can only target armv5te or armv7-a, nothing in between afaiu
20:34:45n1sprobably, most of it is just written for arm7tdmi so stalls etc are not taken into account
20:34:55kugelIIRC virtually all android phones have armv6 or armv7
20:35:00 Join Dreamxtreme [0] (~Dreamxtre@
20:35:21n1son armv7 neon could be very interesting i think
20:35:27kugeln1s: most of our asm doesn't give the speed ups it gives for PP on other arm targets
20:35:56kugelI guess we could safely built armv6 binaries, even the g1/dream is armv6
20:36:12n1skugel: true, and that is mostly because those other targets are pipelined so need careful instruction sceduling to get good performance
20:36:22 Quit Dreamxtreme (Client Quit)
20:37:40 Quit DerPapst (Quit: Leaving.)
20:38:21saratogaturning on ASM will probably make most of the codecs faster just because the fixmul type functions will be written correctly
20:38:41saratogatelling gcc to do a 32x32=64 bit mul in c code tends to result in some very ugly ASM on 32 bit cores
20:41:35kugeldoesn't armv6 have some nice instructions for that?
20:42:31n1sit has 32*32=32hi
20:42:50funmankugel: are you sure we can build for armv6? NDK mentions only armv5 and armv7
20:43:18kugelfunman: -march=armv6 ?
20:43:47 Join Dreamxtreme [0] (~Dreamxtre@
20:44:05kugelwhat the NDK mentions probably refers to the build script it provices
20:44:07n1sarmv5 or armv6 will make very little difference i think, both in regard to our asm (very little armv6 asm written) and gcc
20:45:09funman <- define CPU_ARM
20:46:02kugelthat's all?
20:46:26bertrikfunman, I could reproduce the hang on alarm powerdown on my clip+ :|
20:46:34funmankugel: it builds, but codecs don't start
20:46:50kugeldid you rebuild them?
20:46:54funmans/don't start/crash rockbox/
20:47:18kugelthe __div0 breaks codec api
20:47:50funmanhm i forgot something
20:48:05*gevaerts tries to remember...
20:48:38funmankugel: i forgot to remove __div0 from the .c
20:49:28***Saving seen data "./dancer.seen"
20:50:06kugelIMO we should wait until test_codec works before switching on the asm
20:50:18kugelit might turn out that many of it is not faster or even slower
20:51:13funmanwhat's missing for plugins?
20:51:29kugeldunno, I haven't even tried to build them
20:51:59 Nick Ypsy is now known as YPSY (
20:52:03kugelI guess most will work, but I have not implemented the get_user_file_path() thing for plugins (some write to disk)
20:52:33funmanmp3 works fine
20:52:48funmanperhaps top can give a clue?
20:53:27*funman requests a 'shutdown' menu
20:54:17 Quit Dreamxtreme (Quit: Don't follow me)
20:54:47funmankugel: can I commit the diff without config.h ? __div0 / system-arm.c only apply to native builds anyway
20:55:18kugelso it only needs a single config.h change for enabling?
20:55:23kugelsounds good
20:55:43 Nick fxb is now known as fxb__ (
20:57:14*funman wonders where __div0 went in sim builds
20:57:32kugelsim builds don't define CPU_ARM
20:57:42kugelbecause sim.h #undefs CONFIG_CPU
21:00:17 Join Dreamxtreme [0] (~Dreamxtre@
21:00:31CIA-6New commit by 03funman (r27684): system-arm.c/__div0 are for ARM native builds only ...
21:02:11CIA-6r27684 build result: All green
21:03:23 Quit kyle__ (Remote host closed the connection)
21:07:29 Join einhirn [0] (
21:07:35 Join milz [0] (
21:08:19funmanGuest8399: any progress?
21:10:21 Join DerPapst [0] (
21:11:44funman"there's no way to reset the fms"
21:12:44funmansaratoga: what makes you think there is a high chance of bricking the clipv2?
21:12:57Guest8399Nope :(
21:14:26funmando you have the battery unsoldered? (1 wire is enough)
21:14:37Guest8399Yeah, the black wire
21:15:13funmanwhen i tried recovery mode, i had either no device at all, either the 1GB one. I don't remember if i saw a 4MB device
21:15:52Guest8399i was looking around on the SanDisk forums earlier. This seems to be a kind-of common problem with the v2s
21:16:54bertrikooh, that FMS looks nice
21:17:55funmanGuest8399: i'm not aware of bricked fuzev2 except mine
21:18:25Guest8399I did try shorting out those two recovery pins with a small screwdriver, and it didn't seem to make a differnce
21:18:47funmantry again and again and again, it's hard to get it at first try
21:19:03Guest8399I do not know how this got bricked, i got it off ebay for about $20. It would be real nice to know
21:19:26Guest8399I have to get going, I will be back in an hour or so.
21:20:34 Quit stripwax (Quit:
21:26:03 Join darkenvy [0] (60fe6f59@gateway/web/freenode/ip.
21:26:21darkenvyhey can someone help me out?
21:26:57darkenvyI didn't want to post on the forums because I can see this as something someone will say 'do a search fisrt' (which I did btw)
21:28:22 Part watto
21:28:23funmanGuest8399: did it happen when you plugged USB cable?
21:28:49funmandarkenvy: if you have a question just ask
21:29:17darkenvyremap keys on an ipod 5th generation? the play button doesnt work nor does the hard reset. The only way to reset is by removing battery cable
21:29:49evilnick_Bdarkenvy: You'd need to download the source, change what the buttons do then recompile
21:30:11darkenvyand always boot into rockbox. if it goes into original ipod firmware then it must die which takes forever or I must open it up and remove the battery cable
21:31:19darkenvyah. thats the only way? no ini file or nuttin? :P
21:36:04n1sdarkenvy: you can install in a special way with ipodpatcher so that the original firmware is removed and dualbooting into it is impossible
21:36:17pixelmaalso changing the keymap in Rockbox won't change the hard reset button combo
21:36:35 Part piggz_ ("Konversation terminated!")
21:37:24n1sdo this and just ignore the backup step if you want to be unable to dualboot
21:38:04 Join Jerom [0] (~heidi@
21:38:21darkenvyoh sweet
21:38:57darkenvyokay so two problems down
21:39:13darkenvyhas anyone ever heard of an unresetable ipod?
21:39:34darkenvymenu and center does absolutely nothing. both buttons work and both click too
21:42:51evilnick_BIt not working would indicate a hardware issue, but if they work separately then... /me is stumped
21:43:20pixelmadid you try to switch hold to on and off before and did you hold both buttons for a long time?
21:44:10 Quit simonrvn (Quit: brb, silly client)
21:45:23darkenvyyes I did that too
21:45:58darkenvyonce I basicly lock my ipod into rockbox ill have options to shutdown and reboot so its kinda fine..... untill it freezes :\
21:46:36 Join simonrvn [0] (
21:46:45darkenvyyou know what?! whats a great rockbox supported player with a large HDD? (I wanted to upgrade my Ipod hdd to 240gb soooo...... I mean LARGE)
21:47:14AlexPIf you want more than 80 GB you need to upgrade the disk
21:51:23amiconnn1s, kugel: You will definitely notice the effect of bulding for armv6 whem trying ape :)
21:51:29darkenvywhats a good rockbox player that I can upgrade? (same kind of HDD right?)
21:51:43BRi7Xdarkenvy: gigabeat F, i'd say
21:52:07AlexPtrouble with the gigabeat F is the 50 pin hd connector
21:52:07BRi7Xthough gigabeats can be hokey
21:52:17BRi7Xi just got a gigabeat S and the hard drive is HOKEY
21:52:26AlexPSo you are limited at 80 gb (and that drive is very hard to find), or you need an adaptor
21:53:02AlexPThe gigabeat S OF is braindead and doesn't support drives bigger that 137 GB (so 120 GB max)
21:53:13AlexPto put a big drive in your only real option is an ipod video
21:54:02darkenvyhey anyone selling an ipod here? lol
21:54:15BRi7Xwell i bought a used one (because they don't exist anymore, so i had to), and the 30GB hard drive in it is fairly corrupted, or so it seems. file transfers fail most of the time. I wanted to put one of the SEVERAL 40GB hard drives from my Gigabeat F laying around, but apparently it's a different connector
21:54:17darkenvyI have a 6th gen 120gb
21:54:27AlexPBRi7X: Yes, it is
21:54:45BRi7Xare they adaptable though?
21:54:46darkenvyill take it off your hands Bri7x
21:54:49AlexPI have a 120 GB drive in my S and it works great, but sadly I can't go bigger (stupid toshiba bootloader)
21:55:09AlexPBRi7X: If you can find an adaptor that'll fit in the case, then yeah
21:55:22AlexPPeople have put massive drives in e.g. iriver H100 with an adaptor
21:55:48AlexPThe 50 pin to ZIP adaptors do exist, so many other players become upgradeable if you can fit an adaptor in
21:55:48darkenvyoh the gigabeat! I always thought it was an ugly player :\ no offense
21:55:55AlexPdepends which gigabeat
21:56:04AlexPThe F is fairly, yes
21:56:11AlexPThe X and S are much better IMO
21:56:34darkenvyselling ipod anyone? lol
21:57:12BRi7Xthe black F is cool
21:57:47BRi7Xwill this fit in the S?: or is that even the right one?
21:58:20AlexPYou don't want a 2.5" HD
21:59:00AlexPIf you want to put a drive from the F in the S, then you want a 50 pin IDE 1.8" HDD to ZIF adaptor, if such a thing even exists
21:59:27AlexPThe other way round certainly does
21:59:35gevaertsBoth ways exist
21:59:49AlexPOK :)
21:59:52gevaertsThat doesn't mean the case has enough room of course
22:00:06AlexPThat'd be my primary concern
22:00:34BRi7Xyeah when i got it i was surprised about how tiny it was
22:02:39BRi7Xi haven't been inside the gigabeat S yet, so i don't know what fits and what won't, i need a tiny screwdriver.
22:02:49saratogafunman: not a high chance, but AFAIK the only sansa anyone bricked was a V2 clip
22:03:45 Quit milz (Remote host closed the connection)
22:03:56AlexPBRi7X: The beast is a bit of a nightmare to get apart (opening the case is OK, and changing battery/disk, but to remove everything is hard)
22:04:34BRi7Xwell at the moment i'm concerned with the hard drive... the F was so easy to screw around with, i was fairly surgeonic (word?) with it
22:04:54funmansaratoga: was it you who proposed the AMSv2 to be stable for next release?
22:07:23 Join mt [0] (~mtee@rockbox/developer/mt)
22:07:39 Quit darkenvy (Quit: Page closed)
22:11:12 Join milz [0] (
22:12:01funmanbertrik: if make fullzip doesn't work, try blaming kugel :)
22:12:15saratogafunman: no, i have no opinion on that
22:12:47funmanIMO these crashes are pretty annoying even if rare, it could cause even more users opening their players and breaking them
22:12:47saratogabut i would like to disable rebooting into the OF on USB insert, since that seems to crash some players badly enough that they can't reboot until the battery dies
22:13:00funmanyeah i think the problem is reboot
22:13:14saratogai think otherwise crashes are rare
22:13:25funmani'm ok with disabling reboot until we harass pamaury enough to fix USB :)
22:13:27saratogai don't think i've seen one on my players in months
22:14:03funmansame but i use more clipv1 than the amsv2
22:14:47BRi7X :'(
22:14:49saratogai use a mix of all 3 clips, they seem pretty similar aside from battery life
22:18:00funmanclipv2 noise annoys me a lot :/
22:21:22preglowkugel: what's broken about some phone's android install?
22:22:07 Join stripwax [0] (
22:22:30 Quit stripwax (Client Quit)
22:23:21BRi7Xhrmm, putting my gigabeat S in the freezer... bad idea or awful idea?
22:24:41Guest8399funman: I'm back, no progress here :/
22:24:53bertrikBRi7X, probably bad for the battery
22:25:06xvellosaratoga (or others) : just got a clip+ and got familiar with the rockbox code. Are there some "junior jobs" to start contributing ?
22:27:05 Join darkenvy [0] (60fe6f59@gateway/web/freenode/ip.
22:27:23darkenvyI recieved a 'can't load linux.bin' on this ipod
22:27:39darkenvyon the stable build, and current
22:27:54pixelmadid you start it with the Play button (I believe it was)?
22:28:44darkenvythe ipod wont wake up at all witht that button
22:29:10darkenvyits completely broke. oddly enough though no matter which state (boot loader or official ipod firm) the menu+center wont hard reset either
22:29:26darkenvyhmmm I coulda sworn that was a firmware issue
22:29:46funmanxvello: you could start by looking on the tracker
22:31:49darkenvyhow to fix linux.bin error?
22:31:50bertrikxvello, all of the basic stuff has already been implemented and all of the easy bugs have probably been fixed ... :)
22:32:11 Join Strife89 [0] (
22:32:12BRi7Xchkdsk /R taking an ETERNITY for a 30GB hard drive, and "queefing" sounds coming from the gigabeat, tell tale sign it's partially physically corrupted?
22:32:24bertrikthere's still this odd bug when powering down with an alarm set though
22:32:54funmanperhaps it would be easy for someone who looks at it with a fresh and innocent mind
22:34:16AlexPxvello: Just find something you want to work on, and get going :)
22:37:54mtxvello: There's also this : - They're not really "junior jobs", but you could maybe find something there that might inspire you.
22:38:26funmanstart reading, it's full of hidden bugs ;)
22:39:00xvellofunman: reading what ?
22:39:04xvellomt: thanks for the pointer
22:39:12funmanthe source
22:39:42mtxvello: you're welcome.
22:39:47xvellookay, will see what I can catch
22:47:55 Quit bieber (Ping timeout: 276 seconds)
22:48:11 Join bieber [0] (
22:49:29***Saving seen data "./dancer.seen"
22:49:57darkenvyany fix to cant find linux.bin?
22:50:06 Quit funman (Quit: free(random());)
22:53:24 Quit Guest8399 (Quit: leaving)
22:54:53 Join clone4crw [0] (
22:54:59darkenvyI cant find anything on this error surprisingly
22:57:25gevaertsdarkenvy: the bootloader tries to load ipodlinux if "play" is held during boot. Maybe your play button is stuck?
22:57:52darkenvyoh wow that could be a possibility
22:58:19darkenvyhow to I rebuild rockbox to delete this function?
22:58:32saratogagrep the source for that message, comment out the block its in
22:58:35darkenvyor the bootloader
22:58:39saratogaprobably somewhere in the bootloaders folder
22:58:46saratogathen compile a new bootloader
22:59:03darkenvyokay ill be back to tell ya how it went
23:01:22darkenvywoah where the eff is the source outlet? all of these are just the folders you throw on the ipod
23:03:05 Quit milz (Remote host closed the connection)
23:04:44 Quit slck (Remote host closed the connection)
23:04:47 Join slck [0] (Venci@No.Real-Love-For.Me)
23:09:52 Join milz [0] (
23:18:05 Quit slck (Quit: Requested.)
23:18:19 Join slck [0] (
23:24:38darkenvyshould I use the ipodpatcher to use my compiled bootloader?
23:29:47 Quit clone4crw (Quit: leaving)
23:29:48 Join Buschel [0] (
23:30:48Buschelmt: did you see my comments and screenshots?
23:31:01 Quit ender` (Quit: The propeller is just a big fan in front of the plane used to keep the pilot cool. When it stops, you can actually watch the pilot start sweating.)
23:36:01 Join anewuser [0] (anewuser@unaffiliated/anewuser)
23:36:20mtBuschel: Yeah seen them. I'm trying to compile r27402 to be able to compare svn and your patches against the ref decoder.
23:36:40mtbut I'm having troubles with my tmp dir now that I'm trying to resolve.
23:38:07mtBuschel: Also, somewhere along the optimisations, our decoder started outputting more samples than ffmpeg's, how is that ?! (I guess those could easily be the spikes we're seeing)
23:38:32Buschelmt: r27406 can be switched back to float for verification?
23:40:06Buschelmt: no, if the spikes would be caused by additional samples the diff would show massive difference from the first spike on (as the time domain would then be shifted between both output wav's)
23:40:34 Quit AndyI (Ping timeout: 248 seconds)
23:40:57mtBuschel: I think yes. (also, do you mean 402 ? :) )
23:41:23mtBuschel: The difference on matlab between 2 wav files (rockbox - ffmpeg) was massive
23:41:39Buschelmt: regarding outputting more samples -> I don't know. I assume this was introduced with the very first rockbox port's version.
23:42:21Buschelmt: our implementation needs to cut the first 1024 samples (for the 173kbps-file). it starts with 1024 zeroes.
23:43:16Buschelmt: I did this in an experimental code change. afterwards I could directly diff the wav-output from our flac-decoder with the wmapro-output
23:43:43 Quit bertrik (Ping timeout: 258 seconds)
23:44:43Buschelmt: btw, the diff between svn/flac is (minor) larger than patched/flac. which could be a sign that svn might not be optimal... a direct comparison between float and fixedpoint decoder would be better, of course
23:44:43 Join AndyI [0] (~pasha_int@
23:45:30mtBuschel: before the optimisations I could subtract the 2 wav files on matlab, (both matrices were of the same dimensions) ... now I have to truncate our matrix to be the same size (number of samples) as ffmpeg's. So those extra samples haven't been introduced early with the port.
23:45:49mtBuschel: ^ (re: more samples)
23:46:42 Quit evilnick_B (Quit: Page closed)
23:46:44mtI'm trying to compile a non-optimised fixed point version of the code to compare it against the original ffmpeg decoder.
23:47:25Buschelmt: well, a float-version on pcsim could also be helpful. this was the starting point, right?
23:47:54 Quit jgarvey (Quit: Leaving)
23:47:56xvelloyay, found a bug ;)
23:48:27mtBuschel: Yes. I started with floating point decoder from ffmpeg's libavcodec.
23:49:00mtxvello: Great, get fixing then. ;)
23:49:28 Quit domonoky (Read error: Connection reset by peer)
23:49:39xvellomt: yeah, tomorrow, it's almost midnight here
23:49:43xvello'night people
23:49:51mtgood night
23:49:56Buschelgood night
23:50:43gevaertsxvello: what sort of bug?
23:52:27 Quit milz (Remote host closed the connection)
23:54:03xvellogevaerts: in the FM tuner, when switching stations with left-right on preset mode, the station name isn't updated, while it is if one selects a station directly via the list
23:54:16xvelloshouldn't be too hard to fix
23:54:24 Join Xerion_ [0] (~xerion@
23:54:29gevaertsfamous last words :)
23:54:52AlexPI think update issues in the fms is quite long standing
23:55:00AlexPSpeak to JdGordon about it :)
23:55:05 Quit Xerion (Read error: Connection reset by peer)
23:55:05 Nick Xerion_ is now known as Xerion (~xerion@
23:55:10xvellogevaerts: ^^
23:56:54 Join milz [0] (
23:56:58gevaertsxvello: I don't actually know much about the UI code. It's just that how hard a bug is tends to be unpredictable

Previous day | Next day