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

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

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

#rockbox log for 2010-08-02

00:02:41 Quit ender` (Quit: Do you remember when you only had to pay for windows when you broke them?)
00:03:00CIA-6New commit by Buschel (r27662): Use higher precision for libwmapro's quant_tab[].
00:05:00CIA-6r27662 build result: All green
00:07:36kugelif someone wants to have a look, I get this when compiling wma for RaaA
00:07:47kugelonly on android though, it seems
00:12:00Buschelkugel: try to declare those two in wmafixed.h instead of wmadeci.c −− the implementation is done in wmafixed.c
00:12:23kugelBuschel: I think the problem is that they're declared inline?
00:12:41kugelor part of
00:13:57Buschelor you may compare it to libfaad (static inline, but implemented in the same file)
00:15:18Buschelmt: how did you generate the fixed point tabs in wmapro? seems like they do not use full precision but were generated out of floats.
00:16:26Buschelmt: it might be a good idea to re-create them from doubles −− just to keep the maximum precision. the sin-tabs were not converted by myself yet.
00:16:53Buschelgood night!
00:16:55 Quit bieber (Ping timeout: 276 seconds)
00:21:07 Quit Buschel (Ping timeout: 240 seconds)
00:21:52 Quit BHSPitMonkey (Quit: Ex-Chat)
00:23:24*kugel found the breakage
00:23:41kugeltoo late though, I started again :/
00:34:38 Join bieber [0] (
00:46:54 Quit bieber (Ping timeout: 264 seconds)
00:48:34***Saving seen data "./dancer.seen"
00:50:08 Join bieber [0] (
00:54:14 Quit DerPapst (Quit: Leaving.)
00:54:19 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
01:03:31 Quit n1s (Quit: Lämnar)
01:05:36 Join anewuser [0] (anewuser@unaffiliated/anewuser)
01:07:42 Quit Jerom (Read error: Connection reset by peer)
01:11:05 Join phantomofopera [0] (
01:11:53phantomofoperahi, what themes are the best, cleanest, focusing on the text rather than album art
01:12:05phantomofoperabut are ALSO pretty
01:12:22phantomofoperaand simple
01:14:50phantomofoperai went through the whole list in "theme installation" dialog window...
01:14:57phantomofoperathere seems to be nothing that fits the bill
01:15:37phantomofoperamaybe except for classical_dark
01:16:58 Quit kugel (Remote host closed the connection)
01:18:25 Quit phantomofopera (Quit: CGI:IRC (EOF))
01:20:39 Join bunnyboi [0] (
01:20:47 Quit JdGordon (Ping timeout: 240 seconds)
01:37:51 Join Leif [0] (
01:48:54 Quit bmbl (Quit: Bye!)
01:52:06 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
02:20:48 Quit joxx (Disconnected by services)
02:48:38***Saving seen data "./dancer.seen"
03:00:57 Nick Ypsy is now known as YPSY (
03:14:51 Quit MethoS- (Remote host closed the connection)
03:20:16 Quit mikroflops (Ping timeout: 260 seconds)
03:20:39 Join mikroflops [0] (
03:23:29 Join Mr_Sensitive [0] (Dreamxtrem@
03:24:43 Quit Dreamxtreme (Ping timeout: 240 seconds)
03:30:53 Join Strife89 [0] (
03:32:10 Quit BlakeJohnson86 (Read error: Connection reset by peer)
03:35:44 Join Strife1989 [0] (
03:38:27 Quit Strife89 (Ping timeout: 240 seconds)
03:41:02 Quit bunnyboi (Quit: Ex-Chat)
03:52:55 Join Judas_PhD [0] (
04:00:20 Nick DarkJuju is now known as [J] (juju2143@gateway/shell/
04:11:04 Join BlakeJohnson86 [0] (
04:15:33 Nick [J] is now known as juju2143 (juju2143@gateway/shell/
04:15:58 Nick juju2143 is now known as DarkJuju (juju2143@gateway/shell/
04:16:12 Join Barahir_ [0] (
04:19:08 Quit Barahir (Ping timeout: 240 seconds)
04:23:09 Quit mc2739 (Ping timeout: 258 seconds)
04:23:37 Quit Judas_PhD (Quit: Aaarrrggghhh!!!)
04:25:04 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
04:27:06 Join Judas_PhD [0] (
04:27:27 Quit Judas_PhD (Client Quit)
04:27:33 Quit pixelma (Disconnected by services)
04:27:35 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:27:40 Quit amiconn (Disconnected by services)
04:27:42 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:27:48 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:27:51 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:30:43 Join Judas_PhD [0] (
04:35:28 Join bob_brain [0] (
04:35:58 Quit mc2739 (Ping timeout: 248 seconds)
04:37:47 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
04:38:35bob_brainI have installed all cross compilers and have verified that they are in my path but still recieve the error "the compiler you must use (/build/prebuilt/linux-x86/arm-eabi-4.4.0/bin/arm-eabi-gcc) is not in your path!" I know that I have the arm-eabi-gcc folder in my /usr/local/bin but am confused about the error
04:38:43 Quit TheSeven (Ping timeout: 276 seconds)
04:39:03bob_brainI am using the git repository from the RaaA git
04:39:44bob_brainI woulnd't mention it normally but since I am able to compile other builds, it seems strange that this one won't
04:43:43 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:48:40***Saving seen data "./dancer.seen"
04:52:19 Quit slck (Quit: Reconnecting.)
04:52:25 Join slck [0] (Venci@No.Real-Love-For.Me)
04:53:33JdGordon|bob_brain: the compiler isnt in your path then :)
04:53:36JdGordon|echo $PATH
04:54:05JdGordon|how did you install the compiler?
04:54:17bob_brainJdGordon|: The standard rockbox compiler is in my path as I am able to compile other arm targets
04:54:49 Quit skim32 ()
04:54:55bob_brainJdGordon|: I think that the custom source code has been modified to use a different location of the comiler
04:55:17JdGordon|that path seems odd
04:55:21bob_brainI am looking at the modified configure script right now
04:55:48JdGordon|if this is for android, maybe the NDK supplies a prebuilt gcc?
04:56:28bob_brainIt is the same version as the rockbox compiler
04:56:36bob_brainbut looking for it in a strange place
04:57:28JdGordon|untill that tree is merged into svn you really need to talk to kugel for help...
04:57:44bob_brainthat is the part of the configure script which defines the location if I am correct
04:58:31JdGordon|you can try chanign that to the usualy arm-eabi-gcc
04:58:31bob_brainthe $Android_NDK_Path must be the variable I'm looking for
04:58:42bob_brainhmm will do
04:58:46bob_brainmaybe that will work
04:58:59JdGordon|although using the $ADNROID_NDK_PATH is probaly better
04:59:07bob_brainor else I need to get the compiler which that variable is pointing
04:59:08JdGordon|have you donwloaded the android ndk?
04:59:12bob_brainnope :)
04:59:58bob_braindownloading now, thanks for the help
05:03:10 Join Tyr0 [0] (
05:04:33 Quit bob_brain (Ping timeout: 258 seconds)
05:07:07 Quit captainkewllllll (Quit: Colloquy for iPod touch -
05:26:17 Quit [Saint] (Ping timeout: 246 seconds)
05:30:50 Join S_a_i_n_t [0] (S_a_i_n_t@
05:39:15 Quit Horschti (Quit: Verlassend)
05:45:08 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:02:56 Join bunnyboi [0] (
06:07:59 Join hebz0rl [0] (
06:08:54 Nick bunnyboi is now known as Jennifur (
06:17:24 Quit Tyr0 (Quit: Tyr0)
06:23:36 Quit Leif (Ping timeout: 260 seconds)
06:29:56 Quit Jennifur (Quit: IRC is just multiplayer notepad)
06:33:57lestatarhi folks...quick question re: AArt on Fuze v2 please... :-)
06:34:41lestatarnot a complaint at all, but just noticed that my aart is kinda blurry, especially where there is legible text on the image
06:35:13lestatarused AlbumArtExtractor, jpg, 125x125 average pixel size
06:35:37lestatarthe original embedded art is good quality 500x500 or a bit better.
06:36:17lestatarIs the blurry-ness due to combination of smallish images and not so great display hardware on teh Fuze?
06:37:25 Join Leif [0] (
06:38:09 Quit anewuser (Quit: FAMICOMPO 7 - August 15 Through August 31 JST -
06:38:45 Join Buschel [0] (
06:39:43 Quit AndyI (Ping timeout: 248 seconds)
06:47:11 Join AndyI [0] (~pasha_int@
06:48:44***Saving seen data "./dancer.seen"
06:58:02 Quit maraz (Ping timeout: 258 seconds)
07:07:19 Join maraz [0] (
07:10:18 Join fyrestorm [0] (
07:10:45 Quit panni_ (Read error: Connection reset by peer)
07:16:15 Quit Leif (Ping timeout: 264 seconds)
07:21:30S_a_i_n_tlestatar: Just a heads up, but there is absolutely NO point in having AA that exceded the maximum height/width of the DAPs display.
07:21:39S_a_i_n_tYou *may* get better results with smaller AA.
07:21:59S_a_i_n_tAll having 500x500 AA is doing on your DAP is taking up room.
07:23:01S_a_i_n_tOh...sorry, I read that incorrectly.
07:23:07S_a_i_n_tThe AA is 125x125?
07:24:56S_a_i_n_tIt could also be that the fuze has a poor LCD, but, I seem to remember it being quite decent at 50x50px.
07:29:29S_a_i_n_tlestatar: Maybe you might get better results by resizing the AA to the size that the theme uses, using a known decent image editor, so it doesn't have to resize it on device.
07:50:16 Join stoffel [0] (
07:53:11 Join JanDo [0] (~JanDo@
07:54:26 Quit JanDo (Client Quit)
07:58:01 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
07:59:41JdGordon|Unhelpful: the skins dont redraw AA every update anymore, would it be possible to only save the origional image on the buffer and resize on draw?
07:59:57JdGordon|save a bit of buffer and potentially allow next aa to work more easily?
08:04:31 Join Buschel [0] (
08:05:45Buscheln1s: Can you take a look at the latest patch in FS #11498 ? A change to the Coldfire asm implementation for vector_fixmul_scalar() is needed... It would be great if you could do this change and submit an updated patch.
08:06:02 Quit Buschel (Client Quit)
08:21:39 Quit n17ikh (Ping timeout: 240 seconds)
08:23:56 Join stooo [0] (
08:24:16 Quit stooo (Client Quit)
08:29:34 Join mitk [0] (~mitk@
08:30:37 Join n17ikh [0] (
08:32:21 Join powell14ski_ [0] (
08:42:25pixelmaJdGordon|: wouldn't the savings on the buffer be dependent on the size of the original image then?
08:42:45JdGordon|yes, but isnt that image there anyway?
08:43:11JdGordon|or even resize it on load to the largest requested size and then resize again
08:43:50pixelmano idea
08:44:22 Join LinusN [0] (linus@rockbox/developer/LinusN)
08:46:53pixelmaI wanted to ask you if you knew how checkwps now deals with %Vb (%Vf respectively) on greyscale. The question arose because the theme under uses %Vb(333333) and IMO should be rejected by checkwps then and not appear on greyscale target themes (the theme seems to be rejected by the parser on target it seems, haven't tried myself though)
08:48:18JdGordon|%Vf/b are ignored on target now
08:48:26JdGordon|I can change it to reject if you really want
08:48:38pixelmaand how do I set grey shades?
08:48:46***Saving seen data "./dancer.seen"
08:48:50 Join ender` [0] (
08:48:56JdGordon|oh ah yes, ok.. no
08:48:58JdGordon|hehe :D
08:49:11JdGordon|that might be a bug happening then
08:49:18JdGordon|you use them for shades also, they arnt ignored
08:49:32*pixelma needs to update her greyscale WPS to see :\\
08:52:03pixelmacheckwps should reject WPSs that won't work on target, so %Vb/f with an RGB number. There's another issue with themes on greyscale which set bright on dark globally and not in the WPS, that's a theme site problem though
08:54:05gevaertsCan some forum admin ban 'zlzqq'?
08:57:36simonrvngeneric spam bot *sigh* (seeing from the 'name')
08:59:34JdGordon|pixelma: remind me in 3 hours or so :) should be a tiny fix
08:59:56pixelmaok, I'll try to keep it in mind
09:00:00JdGordon|or apps/gui/skin_engine/skin_parser.c somehwere in parse_viewportcolours or something
09:03:09pixelmaare you saying that it would currently not be rejected on target? I just *think* it is because someone in the forums said his album art wouldn't work with this WPS on the H120 while it works with other themes, so my assumption was that he would only see the built-in WPS because it would be rejected by the parser
09:04:19*S_a_i_n_t has seem checkwps let in other garbage before that didn't load on device.
09:05:14 Quit mc2739 (Ping timeout: 276 seconds)
09:06:43 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
09:08:16JdGordon|yeah, pretty sure it doesnt make sure the colour is in the right format
09:09:02JdGordon|how much effort do we want to put into checkwps? is pass/fail enough now that we have the theme editor?
09:15:22 Join Zagor [0] (
09:15:23 Quit Zagor (Changing host)
09:15:23 Join Zagor [0] (bjst@rockbox/developer/Zagor)
09:17:55amiconnJdGordon: Afaik the original image is not buffered, only the resized one. Buffering the original would be a potentially huge waste
09:18:08 Nick Strife1989 is now known as Strife89 (
09:18:17*S_a_i_n_t isn't sure forcing use of the themeeditor is such a good idea.
09:19:02JdGordon|amiconn: OK
09:19:35JdGordon|lets say the wps wanted 100x100 and the sbs wanted 25x25, if we only stored 100x100 would the sbs image look absolutely crap?
09:19:48JdGordon|although how good could it look at the size anyway
09:20:39amiconnI'm not sure, but afaik the image is loaded and resized twice in this case
09:21:49JdGordon|yes, I'm asking waht if we only stored it as 100x100 and resized it to 25x25 on demand?
09:21:52*amiconn never tried this sbs stuff
09:22:30JdGordon|I'm pretty sure we only draw the images on a full redraw now so the reclaimed buffer might be better than wasting the cpu
09:22:47JdGordon|(Also I'm asking about being able to do next aa with this)
09:23:10amiconnResizing from a bitmap in rockbox format is kinda difficult. Not that difficult for colour, quite difficult for greyscale
09:23:37 Join einhirn [0] (
09:23:39amiconnBut even for colour this would interfere with dithering
09:24:00JdGordon|so the quality would be bad? or the code just isnt there?
09:24:54amiconnThe code isn't there, and the quality would be bad...very bad (depending on pixel format and scale factor)
09:25:34JdGordon|ok, there goes that idea then
09:30:34 Quit Mr_Sensitive (Read error: Connection reset by peer)
09:30:45 Join Mr_Sensitive [0] (Dreamxtrem@
09:31:44 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
09:35:03 Quit Strife89 (Ping timeout: 240 seconds)
09:36:39 Join honkytonkwillie [0] (
09:44:51lestatarguys, thanks a lot, appreciate it as always
09:45:13lestatarno need for me to batch resize all the images, album art extractor is fine
09:45:25lestatarthanks for the tips tho :-)
09:46:47lestatarthe question actually came up on anythingbutipod forum from someone else - just decided to pick your brains here on it
09:47:47S_a_i_n_tYeah, I'm not absolutely 100% sure about it, but you may get a better result if the AA doesn't need to be resized by the player for the theme to use it.
09:48:25 Join Rob2222 [0] (
09:51:14 Quit stoffel (Remote host closed the connection)
09:51:51 Quit bieber (Ping timeout: 260 seconds)
09:52:11 Join bieber [0] (
09:52:16 Quit Rob2223 (Ping timeout: 264 seconds)
09:52:55lestataryeah, figured that saint, esp with theme designers specifying certain sizes. like i said, not a prob for me, just wondering.
09:53:22lestatarbtw, any word on usb v2 fuze support? i'd dearly love to kiss the OF goodbye for good :-)
09:53:48lestatari understand you folks have made some great progress on other platforms
10:00:16 Nick delt_away is now known as delt (
10:13:53 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:16:15 Part Zagor
10:16:34 Join Zagor [0] (bjst@rockbox/developer/Zagor)
10:22:18 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
10:27:50AlexPgevaerts: Done
10:30:37 Join jgarvey [0] (
10:34:28 Join DerPapst [0] (
10:41:35 Join bimbel [0] (~Miranda@unaffiliated/bmbl)
10:41:44 Quit bmbl (Disconnected by services)
10:41:48 Nick bimbel is now known as bmbl (~Miranda@unaffiliated/bmbl)
10:45:37 Nick delt is now known as delt_zZzZ (
10:48:49***Saving seen data "./dancer.seen"
10:49:26 Quit JdGordon| (Quit: leaving)
10:50:45 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:54:21 Quit Judas_PhD (Quit: Aaarrrggghhh!!!)
11:00:53 Join mischasworld [0] (
11:01:36 Join petur [0] (~petur@rockbox/developer/petur)
11:01:38 Join Dreamxtreme__ [0] (Dreamxtrem@
11:03:03 Quit Mr_Sensitive (Ping timeout: 240 seconds)
11:10:38CIA-6New commit by pamaury (r27663): Fix a logf bug which would trigger undefined behaviour(and probably crash) when both DEBUG, CONFIG_PLATFORM & PLATFORM_HOSTED are enable
11:12:29CIA-6r27663 build result: All green
11:27:37 Quit mischasworld (Ping timeout: 240 seconds)
11:41:09 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
11:46:05 Quit petur (Quit: reboot)
11:48:17 Quit sasquatch (Quit: WeeChat 0.3.2)
11:51:30JdGordonpixelma: whats the issue with the colours? grey accepting colour? or colour accepting grey?
11:52:02 Join petur [0] (~petur@rockbox/developer/petur)
11:52:25 Join sasquatch [0] (
11:57:00 Join dfkt [0] (dfkt@unaffiliated/dfkt)
12:00:02 Join mischasworld [0] (
12:17:05 Quit mischasworld (Ping timeout: 276 seconds)
12:17:52 Quit bieber (Ping timeout: 265 seconds)
12:18:08 Join bieber [0] (
12:48:50***Saving seen data "./dancer.seen"
12:59:41 Join watto [0] (~watto@
13:05:49 Quit Rob2222 (Quit: Rob2222)
13:10:44 Join Rob2222 [0] (
13:10:45JdGordonIs anyone against the ability to use more than one letter as the image identifier in skins?
13:11:33JdGordonactually ignore that :)
13:11:37JdGordoncbb with the discussion/patch now
13:14:30*S_a_i_n_t is against it, but doesn't really see the point...are there any themes even close to 52 images?
13:14:42S_a_i_n_tI can't particularly be bothered either though ;)
13:15:12S_a_i_n_tman, bad typo... s/is/isn't/
13:16:16 Quit TheSeven (Ping timeout: 264 seconds)
13:16:53 Join T44 [0] (
13:20:42 Quit Topy44 (Ping timeout: 265 seconds)
13:22:55S_a_i_n_tHmmmm, /me spots another Nano1G user in the forums complaining of unclean shutdown (without a low battery warining) causing legacy items left drawn on the screen.
13:23:11S_a_i_n_tYay....kinda, at least it isn't just me.
13:30:49 Quit jgarvey (Quit: Leaving)
13:36:47 Quit bug2000 (Ping timeout: 260 seconds)
13:38:08 Join Rob2223 [0] (
13:38:51 Quit Rob2222 (*.net *.split)
13:38:52 Quit Barahir_ (*.net *.split)
13:38:53 Quit AlexP (*.net *.split)
13:38:53 Quit scorche (*.net *.split)
13:38:55 Quit parafin (*.net *.split)
13:38:55 Quit jae (*.net *.split)
13:40:14 Join jae [0] (
13:40:16 Join Barahir [0] (
13:40:43 Join AlexP [0] (~alex@rockbox/staff/AlexP)
13:43:17 Join parafin [0] (
13:44:14 Join scorche [0] (~scorche@rockbox/administrator/scorche)
13:49:09 Join bug2000 [0] (~bug@unaffiliated/bug2000)
13:49:28pixelmaJdGordon: grey accepting RGB values, seeing as the theme using %Vb(333333) is on the greyscale theme pages
13:49:47JdGordonyeah ok, checked the logs because I was impatient :)
13:49:58JdGordonfixed locally... I'll commit tonight some time
13:51:05 Join MethoS- [0] (~clemens@
13:51:08pixelmathanks. I wondered a bit if it's because a single "3" would be a valid grey shade
13:51:59JdGordonexactly the reason
13:52:20JdGordona tiny bit of extra error checking and its fixed
13:52:29pixelmaah, ok
14:04:37 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:13:24JdGordonbieber: small request of the editor (which will make sense once I commit this....) renaming a viewport should ask if you want to rename all references (i.e in %Vd() tags)
14:16:52JdGordonwho can I annoy about the manual?
14:17:22S_a_i_n_tDepends who wanted what fixed ;)
14:18:21S_a_i_n_tThe "propper" suggestion would probably be the tracker though
14:18:50JdGordonI'm fixing viewports to allow proper names to give an idea how useful (or not) it would be for images.. looking into if the manul needs fixing
14:19:04kugelS_a_i_n_t: I need a 320x480 cabbie quite soon
14:19:11JdGordonviewports is the easier change to make :) no silly %xd(Aa) tag to worry about
14:19:38S_a_i_n_t"%xl(foo)" and "%Vd(foo)"?
14:19:43JdGordonkugel: use 320x240 and be happy with 1/4 of the lcd being blank
14:20:09JdGordonS_a_i_n_t: yeah, %v though
14:20:22*S_a_i_n_t shudders
14:20:28JdGordonyou dont have to use it
14:21:09S_a_i_n_tI just imagine it looking weird, that's's functionally identical.
14:21:46 Join efyx [0] ( it though, is there still a limit on identifiers?
14:22:16 Quit bieber (Ping timeout: 264 seconds)
14:22:29 Join bieber [0] (
14:22:31JdGordonbut %?C<%Vd(aa)|%Vd(no aa)> is certainly nicer than one letters and 4 lines of comments
14:23:04JdGordonthe only limit is you cant put a comma in the name
14:23:34S_a_i_n_tto some, possibly. My comments only ever explain optional features, I let the code explain the rest.
14:23:56 Quit bluebroth3r (Ping timeout: 240 seconds)
14:25:48 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
14:36:24 Quit S_a_i_n_t (Ping timeout: 276 seconds)
14:37:45UnhelpfulJdGordon: i'm not sure what that would actually save? and do we mean an original image in 8-bit color-or-greyscale, or at native depth? the latter is fairly worthless for further scaling.
14:39:59Unhelpfulif you're going to scale it again, i suppose you'd want to scale it to the maximum size that will fit the screen on load, and store it at 8bits/channel. then scale that *down* on demand. i don't really think this would save us much, though, since the bitmap data would be 3/2 as large on color targets and 4x as large on greyscale
14:40:38JdGordonok fair enough
14:40:42Unhelpfulwe also lack at present a means to scale an in-memory image, although the glue code to feed one to the scaler should be fairly trivial.
14:41:25JdGordonI was thinking for next AA, if the image is already on the buffer (but in the wrong size) that could be scaled down so it isnt in ram 4 times
14:41:27kugeldidn't uchida post a patch which does in-memory scaling?
14:42:28kugelthe embedded albumart patch
14:42:51Unhelpfulthat's part of the embedded AA patch?
14:42:56JdGordonanyone got a sbs which sets the ui viewport conditionally?
14:44:19dfktdoes disk tidy not support wildcards for "other" files? it works well with the eq presets i added in the "other" section, but it doesn't clean *.jpg, and so on:
14:44:50kugelUnhelpful: it looks like
14:45:05Unhelpfulkugel: are you sure you don't mean scaling from jpeg data that is in memory?
14:45:24 Join S_a_i_n_t [0] (S_a_i_n_t@
14:46:01kugelUnhelpful: I do mean that, you said we lack that
14:47:26Unhelpfuli don't think decoding jpeg on-demand is a good idea. sloooooow.
14:48:53***Saving seen data "./dancer.seen"
14:50:17CIA-6New commit by jdgordon (r27664): make grey targets reject viewport colours if theyare more than one letter (i.e %Vb(333333) )
14:50:25CIA-6New commit by jdgordon (r27665): Allow viewport labels to be mostly free text instead of only one character. If this is useful the same will be done for images
14:52:02CIA-6r27664 build result: 1 errors, 0 warnings (jdgordon committed)
14:52:31JdGordonodd error
14:53:56CIA-6r27665 build result: All green
14:54:13 Join oxman [0] (
14:54:19oxmanhello :)
14:55:30 Quit mitk (Quit: Leaving)
14:56:22oxmanI've a problem with my Sansa clip+, gather runtime doesn't work, and if i do a song rating i get an "no info", gather runtme is on, i've try to initialize again and again, and update too, without success, gather runtime still doesn't work
14:56:24oxmanany idea ?
14:56:57oxman(i using r27663)
14:57:01S_a_i_n_trating is set via the DAP, not shared between databases IIRC
14:57:02JdGordonhave you got the database set to auto update and load to ram?
14:57:11S_a_i_n_tie. I don't think you can export a rating.
14:57:30oxmani've set to auto update yes, and trying with enable to ram and without
14:57:33oxmanthe both doesn't work
14:58:24S_a_i_n_tAs for the rating, try setting it via the context menu in the WPS
14:58:32S_a_i_n_tI think you'll find that wil work.
14:58:37oxmani get "no info" when i doing that :)
14:58:55S_a_i_n_tAh...where do you get the "no info"?
14:59:01S_a_i_n_twhat displays that, and where?
14:59:08oxmanafter push "song rating" in the context menu
14:59:50S_a_i_n_tand, you have set a rating first? (via the context menu)?
14:59:59oxmanmy god... now its work, 2 hours on the problem !
15:00:03oxmanand now it working :/
15:00:47JdGordonrasher: I want to deprecate a skin tag instead of outright breaking it. would it be possible to make the theme site be able to handle it if I changed checkwps to print a error message if the tag is found?
15:01:17JdGordonof course, now owuld be the time to break it anyway...
15:01:19oxmanwell "thanks" :)
15:01:19 Quit antil33t (Read error: Connection reset by peer)
15:01:24 Join antil33t [0] (
15:01:40 Join Leif [0] (
15:09:51*JdGordon points people to
15:11:32S_a_i_n_tis it an actual poll, or do you need to "reply" your opinion?
15:11:52*S_a_i_n_t sees nothing clickable.
15:12:06 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
15:12:40JdGordonyou might have to log in to vote?
15:12:56S_a_i_n_tAh... ;)
15:13:36JdGordonkugel: you did the backlight stuff... is there an event when the backlight comes back on?
15:13:47JdGordonadd a ? before those ...
15:15:55kugelJdGordon: no
15:16:13JdGordoncan I add one?
15:16:23S_a_i_n_tyou naming of "palymode" confused me, as "playmode" is likely to be the image name comprising the bitmap strips, not a name of the subimage.
15:16:43JdGordontypo :) I'lll fix it
15:17:06kugelJdGordon: yea, why not
15:17:27JdGordonare you using the lcd/backlight in android?
15:17:37JdGordonoff = gui hidden
15:17:58*S_a_i_n_t also wonders why " (This would eventually be removed, sometime in the future, more than 1 release away)", but other than that..."meh"
15:17:59 Quit krazykit (Quit: bbl)
15:18:01JdGordonS_a_i_n_t: hang on! thats your type
15:18:15S_a_i_n_tdo you not get it?
15:18:24kugelJdGordon: you mean you want an event when the display because readable again (!= backlight on) like there is one for when it becomes unreadable (LCD_ACTIVATION_EVENT)?
15:18:29S_a_i_n_t"playmode" comprises of "play" stop" etc.
15:18:42S_a_i_n_tSo it isn;t likely to be a subimage of %mp
15:18:58JdGordonkugel: yeah
15:19:06JdGordonto force a skin redraw
15:19:30kugelJdGordon: I think it would be better to simply move the send_event() call out of the if (enable) in each driver (and pass enable as param)
15:20:02kugelit's send from the driver's lcd_enable()
15:21:00AlexPJdGordon: If you are going to change it, I'd do it now rather than let it keep working then break in the future
15:21:12kugelJdGordon: oh wait, there's an event for when it becomes readable again. there's no event for when it becomes unreadable
15:21:38kugelthe wps already forces skin redraw (add_event(LCD_EVENT_ACTIVATION, false, wps_lcd_activation_hook);)
15:21:56JdGordonAlexP: well yeah, me too, but if really noone wants it then I guess I wont bother
15:22:02*S_a_i_n_t was reasonable happy knowing old behaviour would be reserved...but, now knowing it won't be, doesn't see the point.
15:22:17 Quit Kitar|st (Ping timeout: 264 seconds)
15:22:23JdGordonkugel: yeah, but that doesnt work :)
15:23:18kugelbut it should (and did before)
15:25:57JdGordonnot since 3.4
15:26:47JdGordonmid january
15:26:48JdGordonif we knew when the lcd wasnt visible anymore we could do a lot less updates also
15:29:37 Quit kramer3d (Ping timeout: 240 seconds)
15:30:40kugelsomeone needs to explain the rationale behind sim_codec_load_ram()
15:31:00kugelJdGordon: that's what's being done
15:31:36JdGordononly in wps.c
15:32:23kugelI had that implemented for the sbs too
15:32:52kugelit's in sb_skin_update()
15:34:00kugelwhat do you mean with "it doesn't work"? there's a minimal delay which is inevitable but you shouldn't notice it most of the time (not if backlight fading is enabled anyway)
15:35:23kugelJdGordon: all reporters use a h300 it seems? maybe it's more related to the remote. because I haven't noticed that bug yet
15:35:48JdGordonhmm, good point
15:37:45 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:38:05JdGordonn1s: ping?
15:39:07S_a_i_n_tHmmm, so "%xd(playmode, %mp)" means, each subimage from the image is a respective segment in the %mp tag?
15:39:14*S_a_i_n_t is trying to understand this.
15:39:49S_a_i_n_tIf so, doesn't that bork the current ability of being able to use subimages out of sequence as animations commonly do?
15:40:04JdGordonyes and yes
15:40:17S_a_i_n_tthen, it's crap. sorry.
15:40:17JdGordonyou would only use that if you wanted that behaviour though
15:40:41S_a_i_n_tI read that the "old" (current" behaviour is doomed though.
15:41:00JdGordon%xd(Aa) is doomed
15:41:10JdGordon%xd(a,1) is not
15:41:37 Join Kitar|st [0] (~Kitar_st@
15:41:48 Join Kitr88 [0] (
15:42:10JdGordon%?mm<|%xd(Ea)|%xd(Eb)|%xd(Ec)|%xd(Ed)> would either be %xd(E, %mm) or %?mm<|%xd(E,1)|%xd(E,2)|%xd(E,3)|%xd(E,4)>
15:43:41S_a_i_n_tthe problem I see with "%xd(E, %mm)" is the image *has* to have the same ammount of subimages as the tag has values...unless it had an offset value, which overcomplicates things.
15:44:11S_a_i_n_tA lot of themes don;t use a "stop" subimage for example, as its pointless if there's no .sbs
15:44:14JdGordonthe %xl tag wouldnt change so it would still be told how many images are there
15:44:32 Join jgarvey [0] (
15:44:35JdGordonthe qeustion is which image to use if %mm is too big, probaly the last or none
15:45:02 Quit Leif (Ping timeout: 260 seconds)
15:45:35S_a_i_n_tIt seems nice in principle but there seems to be too many things it depends on to be "smooth".
15:45:40JdGordonthis of course would only be useful if your images are set up to handle them
15:45:41 Quit Kitar|st (Ping timeout: 240 seconds)
15:45:57S_a_i_n_tI'm in the "if it aint broke, don't fix it" camp.
15:46:23S_a_i_n_tI just really don't want to see another big theme-breaker any time soon.
15:47:04JdGordonthe counter argument is now is the perfect time to do another break
15:47:48kugelJdGordon: sb_skin_update() sets do_full_update to false, a subsequent call for the other screen would not have a full update
15:47:56S_a_i_n_tIt is? Why people are still reeling from the last one and the documentation still doesn't accurately reflect the last?
15:48:37S_a_i_n_tI see the point, but it doesn't seem as simple as it does in casual conversation.
15:48:41JdGordonkugel: yeah, saw that in wps.c also... I never did like that mostly global do_full_update :/
15:50:17JdGordonwhat is really needed is more cleanup :/ a single skin manager which keeps track of all the load skins and makes sure the corect update is done for each
15:51:05JdGordonthat might even allow on demand skin un/loading!
15:59:38JdGordonah, the wps_sync_data isnt shared between the sbs and wps, only between sbs and rsbs, and wps and rwps
16:01:29 Part LinusN
16:03:18 Join fdinel [0] (
16:04:34 Join Strife89 [0] (~Strife89@
16:04:49 Join stoffel [0] (
16:12:46CIA-6New commit by jdgordon (r27666): Fix FS #10926 - h300 wps wouldnt be updated correctly when the backlight turns off because the remote is being naughty
16:14:44CIA-6r27666 build result: All green
16:14:51kugelJdGordon: the wps.c code looked correct to me
16:15:11JdGordonthats where the bug was actually reported
16:15:16kugelunless the skin engine modifies do_full_update?
16:15:51 Join piggz [0] (~piggz@
16:16:09JdGordonthe if in wps.c was always true so do_full_update would get to false straight away
16:16:26piggzhi, rockbox on my ipod video 5.5g no longer works on my sony dock...used to work fine, im using the current build, anything i can try?
16:16:56S_a_i_n_tpiggz: Yep, finding the build that broke it.
16:16:58JdGordoneither way, that removed line should have been inside the if() block
16:17:36piggzS_a_i_n_t: first, is it a known issue?
16:18:01S_a_i_n_tpiggz: Did you check the tracker?
16:18:55 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
16:24:34 Quit JdGordon (Quit: Leaving.)
16:25:19piggzS_a_i_n_t: i went back to 3.6 and it still didnt work, so i reset the settings, rebooted, turned off my hi-fi, and it started working
16:25:23piggzgo figure
16:25:31piggzi'll put curent back on iy now
16:26:13S_a_i_n_tYou probably turned off accessory power supply.
16:26:43pixelmais power supply important on a dock? No idea how Ipod accessories work though...
16:26:54S_a_i_n_tIIRC it is on by default for the releases, not sure about the current builds though.
16:27:21pixelmaI don't think it's on for the releases
16:27:27S_a_i_n_tpixelma: I'm under the impression ic can be important for some makes of dock, yes.
16:27:33S_a_i_n_tHappy to be corrected though.
16:27:52piggzfor info, its a sony dock
16:28:07Torneaccessory power supply and/or lineout may need to be enabled
16:28:20Tornesome docks/accessories don't actually draw power from the ipod but use the accessory power pin to detect its presence
16:28:35S_a_i_n_tYay, conformation from an iMan ;)
16:29:03 Join domonoky [0] (
16:29:05 Quit domonoky (Changing host)
16:29:06 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
16:41:30 Quit S_a_i_n_t (Ping timeout: 240 seconds)
16:42:36 Join S_a_i_n_t [0] (S_a_i_n_t@
16:42:53 Join _jhMikeS_ [0] (~jethead71@rockbox/developer/jhMikeS)
16:42:53 Quit jhMikeS (Disconnected by services)
16:47:05 Join panni_ [0] (
16:48:55***Saving seen data "./dancer.seen"
16:50:47 Nick YPSY is now known as Ypsy (
16:51:51 Quit antil33t (Read error: Connection reset by peer)
16:51:57 Join antil33t [0] (
16:58:17 Quit bieber (Ping timeout: 240 seconds)
16:58:33 Join bieber [0] (
16:59:46 Quit domonoky (Read error: Connection reset by peer)
17:00:58 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
17:02:45 Part Zagor
17:05:56 Join Judas_PhD [0] (
17:06:27kugelif anyone is interested, kugel-rb.git/shortlog/refs/heads/gsoc-android-2"> now has the java bits so if anyone wants to build RaaA for android go ahead
17:06:29 Quit Strife89 (Quit: Putting away my lappy for now.)
17:07:38 Quit bieber (Ping timeout: 260 seconds)
17:07:53 Join bieber [0] (
17:10:14 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
17:11:51 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
17:22:43*kugel needs a cabbie, quickly... :)
17:24:06S_a_i_n_tStretch the backdrops from the 320x240 Cabbie ;)
17:25:52S_a_i_n_tHmmm, it's beyond me...but the math to work out the y values for each viewport/image wouldn't be *impossible to calculate I guess.
17:34:00pixelmaisn't there a cabbiev2 graphics package still? I don't know what picture size the original is though
17:35:21kugelI'm too lazy to code the wps though
17:36:02*S_a_i_n_t finds that comical considering the coding kugel does, and the fact that only y values and AA size really need be changed.
17:37:32S_a_i_n_t(drop everything down from mid-screen, and put a big fucking AA image with a "fallback image" in the centre...done ;))
17:37:40kugelit's not that easy though. the 320x240 is landscape, therefore a completely different layout
17:37:54S_a_i_n_tAh, right...gotcha.
17:38:29S_a_i_n_toh, you mean "the 320x240 (with touch regions)" :P
17:38:58 Join {phoenix} [0] (
17:39:30S_a_i_n_tas far as I remember the video is 320x240...and touch regions would need changing anyway.
17:39:38S_a_i_n_tI *can* do it, just not now.
17:40:25*kugel hugely dislikes scaling
17:40:53S_a_i_n_tthe original GFX pack is in the forums.
17:41:02pixelmacouldn't you base the layout off the gigabeat or e200 version?
17:41:03kugelapparently the biggest backdrop in the cabbiev2 graphics package is too small :(
17:42:01S_a_i_n_tkugel; Aren't there .svgs in there?
17:42:08S_a_i_n_tthey scale *way* better.
17:42:23kugelI grabbed the package from, is there another one?
17:42:47 Join LambdaCalculus37 [0] (803b0d0c@rockbox/staff/LambdaCalculus37)
17:43:12S_a_i_n_tHmmm, that does seem to be it. I seem to recall there being SVGs on hand.
17:43:22kugelI too :/
17:45:39 Quit Judas_PhD (Quit: Aaarrrggghhh!!!)
17:45:43*S_a_i_n_t decides he hugely dislikes the variations present in the colours of the varying Cabbie backdrops.
17:45:58S_a_i_n_tSome are black & yellow...some decidedly orange.
17:47:27 Quit bmbl (Quit: Bye!)
17:48:05 Nick Ypsy is now known as YPSY (
17:48:10pixelmait's also fun that the 128x128 version used bigger icons in the cabbiev2 "status bar" than the 176x132 for quite a while. I think I changed that with the 128x128 greyscale versions for the MPIO but you see that they were done by different people
17:48:57S_a_i_n_tOne day, One Day...if noone beats me to it.
17:49:13S_a_i_n_t+(I'll make a set of SVGs)
17:50:13kugelS_a_i_n_t: icons.psd contains everything in huge
17:50:30S_a_i_n_tfor greyscale too?
17:50:45S_a_i_n_tAh...well, it's a start.
17:51:01 Join freddyb [0] (
17:51:01pixelmamaybe ZincAlloy has SVGs
17:51:23S_a_i_n_tI think the .PSGs are all his workings.
17:52:15 Quit Ramsey[LC] (Quit: (Killed (NickServ (GHOST command used by dissident))))
17:52:53 Quit TheSeven (Ping timeout: 264 seconds)
17:53:33 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
17:53:34 Quit _jhMikeS_ (Disconnected by services)
17:59:24pixelmaS_a_i_n_t: the greyscale icons were all "hand made" by me. The first versions were just scaled down to the 4 greys from the colour icons and looked... well... weird
17:59:29 Quit petur (Quit: later)
17:59:33pixelmasame with monochrome
17:59:47S_a_i_n_tyeah, I've tested that ,yself ;)
18:00:04S_a_i_n_tthough...anything in greyscale/monochrome looks weird :P
18:00:18S_a_i_n_tbah... *myself
18:00:48pixelmapff... Nano junky
18:03:53 Join RemDeprived [0] (
18:04:00 Join Strife89 [0] (~Strife89@
18:05:17 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:09:07freddybWhat happened to anti-aliased fonts? Was there an objection to it or just lack of interest? I made a few small fixes and I think I have it working now.
18:09:36 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
18:10:10*Torne doesn't personally see the point; fonts are already too big :)
18:10:34Torneif you're going to make fonts bigger by adding an alpha mask, then fix FS #11168 first :)
18:13:49 Join Hillshum [0] (4ba5f419@gateway/web/freenode/ip.
18:16:22 Quit LambdaCalculus37 ()
18:17:26freddybHaha. It seems to me that FS #11168 targets the userfont (60000) but it's much more likely to over run the skin fonts (10240)
18:18:47TorneThat's people's own fault, though :)
18:19:09Tornethe problem is not just that fontcache loading is slow on some targets..
18:19:19Tornethe problem is that this applies to the default userfont used by cabbiev2 on those targets
18:19:29Torneobviously fixing fontcache loading being slow would be ideal :)
18:19:41Tornethe patch i wrote is just a workaround to let hte memory usage be a little flexible
18:21:34freddybYou mean it's slow when it boots or slow all around?
18:21:55Torneboot time
18:22:04Tornei guess it would also be slow if there were a lot of fontcache misses
18:22:11RemDeprivedkugel: if you get something testable with Android, I would like to give it a try. I have a HTC HD2 hijacked to run android 2.2
18:22:17Tornebut most people don't have any noticable number of misses as most people's tracknames are in fairly basic latin chars
18:22:24 Quit Strife89 (Quit: VBS Day 1 - Done. Finally.)
18:22:25Tornewith maybe a minimal helping of accents
18:22:38Tornethe problem is that the boot time on ipodvideo goes up by 6 *seconds* if you select a font that needs fontcache
18:22:46Tornethat's just absolutely insane
18:23:01kugelRemDeprived: clone my git repo
18:23:05Torneand cabbiev2's font is too big, so yeah
18:23:16Tornedefault install of rockbox on ipodvideo boots 6 seconds too slow
18:23:37Tornethis is unimpressive :)
18:25:45S_a_i_n_tThough, said user doesn't *know* it's too slow until changing themes ;)
18:25:55S_a_i_n_tJust say the other themes loading fast is the bug :P
18:26:40Tornewell, it might still put people off, especially since we hav eto boot up every time
18:27:07Torneit's some kind of horrible perfect storm of failure between the fontcache being stupid and the sector emulatio nbuffer being stupid as well
18:27:12Torne15-Helvetica is >60000 bytes
18:27:16Torneso it provokes the bad behaviour
18:27:22Torne14 or 16 are fine :)
18:27:26S_a_i_n_twould RB be able to "sleep" the iPods...and just doesn't, or can it just not be done?
18:27:47TorneS_a_i_n_t: it can be done, but it's nontriviail
18:27:54Torneit's not just a matter of finding some "sleep" register
18:27:58Tornethe CPU turns off
18:28:06Torneso, when it powers back on it executes Apple's flash bootloader
18:28:10Torneit just leaves power to RAM
18:28:20Torneit's equivalent to suspend-to-ram on PCs
18:28:36Torneso what we'd hasve to do is work out wha tapple's flash bootloader expects to find in ram in order to resume from sleep
18:28:42Torneand set up the ram to look enough like that that it worked.
18:29:20S_a_i_n_tthere've been problems before with the magic the OF expects in RAM, no?
18:29:27TorneYes, there are *ongoing* problems with that magic
18:29:36Torneit's what causes the bootup-needs-reset issue, i think
18:29:45Tornei've done what i think should solve it but it doesn't work for everyone
18:30:03S_a_i_n_tie. me. ;)
18:30:05Tornethe OF basically never expects to be actually powered off fully other than when the battery has totally died
18:30:19S_a_i_n_tcoinceidentally, has there been anyone else or is it *just* me?
18:30:19Torneit always expects at minimum the contents of IRAM to be retained and to be something it wrote itself
18:30:23Tornethere are others, yes
18:30:30freddybHmmm..Some of my AA-Fonts load much faster than 6 seconds on my iPod Video and they're BIG. Is it because of the number of glyphs?
18:30:30Torneyou don't count though, we decided, i thought
18:30:47Tornefreddyb: it's whether the filesize is over MAX_FONT_SIZE, nothing else
18:30:55Torneit's whether it uses the fontcache or not.
18:31:04Tornefontcache loading is stupendously slow, regular loading is fine
18:31:24Torneit probably only affects the models with 1024-byte sector hard disks, also
18:31:32Tornei.e. the 5.5G
18:31:48TorneS_a_i_n_t: the other fix didn't fix it for you either, or am i thinking of osmeone else? :)
18:32:11kugelhm, somehow gimp lacks a filter for the rockbox logo
18:32:15S_a_i_n_tTorne: Yeah, that's me. Neither "fix" worked 100%
18:32:23Torneright, i had no explanation for that one
18:32:28Torneso i was pretending you didn't exist ;)
18:32:40S_a_i_n_tkeep the averages clean ;)
18:32:54Torneas far as I know nobody with an ipodvideo has problems with the new fix
18:32:58 Quit DerPapst (Quit: Leaving.)
18:32:58 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:33:09Tornethe people who have issues, other than you, are 1) other models and 2) worked fine with the original fix
18:33:26Tornethis reduces my motivation to look into it more, i'm afraid, since i'm an ipodvideo owner
18:33:42S_a_i_n_tHey, totally fair.
18:33:47Tornealso, the loader i've been disassembling to look for more info is the ipodvideo one, so it may well be that i won't find anything that way
18:34:01S_a_i_n_tIf it did it more than once every couple of days it'd be a bit more annoying.
18:34:03Torneit's enough effort to disassemble any substantial proportion of *one* loader ,let alone the, er, 30-odd there are
18:34:09Torne(different versoins as well as models)
18:34:34TorneSince *either* fix was included, a lot less people have asked for help with the problem
18:34:47Torneso I think we can reasonably conclude that for most people, *both* fixes worked fine
18:34:59Torneand thus the latter one is probably better since it doesn'thave any side effects
18:34:59S_a_i_n_tconsiderably less since the second.
18:35:15Tornewell, lots of people asked about the side effects of the first one
18:35:25Tornei don't think anyone complained about still having to reset it
18:35:27Torneother than you :)
18:35:36*S_a_i_n_t never even saw it, but suspects it may be a Flash thing.
18:37:08Tornewell, the obvious side effect (the icon on shutdown) didn't show up on all models, depended on what kind of lcd they had i think
18:37:16Torneand the waking-up-randomly one never happened for me :)
18:37:20*Torne shrugs
18:37:38TorneIt's just too much of a pain to account for all the possible stupid things the apple OF could be doing without knowing exactly what it's up to
18:37:53Torneand knowing that for sure would need someone to reverse *every* version, or at least the latest version for every model :)
18:38:09*S_a_i_n_t saw neither, but still had "need to reset for power on".
18:38:17S_a_i_n_tNext years GSoC? :P
18:38:53kugelhm, damn
18:38:54Tornethat would be slightly sadistic :)
18:39:01S_a_i_n_t"GSoC Submission: Pwn iPod Bootloaders/OF"
18:39:16TorneWell, an alternative would be to just suck it up and write a flash based bootloader for rockbox
18:39:21Torneand chuck apple's code out entirely
18:39:26Tornethat would make supporting suspend much easier :)
18:39:32Tornebut.. yeah. risk of bricking
18:40:04S_a_i_n_tAh...right, Flash only.
18:40:22Tornethe flash-based versoin of iloader is only for nano2g
18:40:32Tornethe iloader for PP ipods is just a disk based loader like ours
18:40:48S_a_i_n_tAnd, its NOR based.
18:40:48TheSevenadapting it to newer ipods should be rather easy, but to older pp-based ones won't be
18:40:58Torneyou'd basically have to write the whole thing again
18:41:09TheSevenwell, the driver side of things :)
18:41:10Torneother than the easy/boring bits
18:41:12TorneWell yeah
18:41:17Tornebut those are the 90% that takes most of the work :)
18:41:37Torneand the PP ipods have no DFU so if you mess it up, brick
18:42:15TheSevenso you'll need a 100% safe flashing process + dfu stub
18:42:17Tornethere's a way to amke it a one-time bricking risk so you can update the bootloader safely later, but you can't avoid that one time ;)
18:42:37TorneTheSeven: well, not DFU but just loading the regular Apple loader, i suspect
18:42:48Torneinstall our loader as a seperate flash image like the diag/disk modes
18:42:57Torneand just replace the first page of the flash with a dualboot stub
18:43:09Tornethere's enough room, i think
18:43:16Tornethe flash isn't particularly full
18:43:23TheSeventhey aren't executing directly from flash, are they?
18:43:33TheSevenon the nano2g, everything is copied to IRAM before being executed
18:43:47Tornenano2g is NAND, though, no?
18:43:59Tornepp ipods use NOR flash, the initial part is certainly executed directly from flash
18:44:02Torneit copies itself to ram later
18:44:22TheSevennano2g is mask rom + nor + nand data flash
18:44:34Torneok, so the PPs are the same but without ht emask rom :)
18:44:41Torne(and with data over ATA)
18:44:50Tornethey just start at the first word of the NOR flash
18:44:57Torneso there is absolutely no recovery if you trash it
18:44:59 Join Strife89 [0] (
18:45:54TheSeventhis sounds like writing a self-contained recovery loader to the first page is a good idea
18:46:02Tornewell yes
18:46:13Tornethough it would be easier to just branch back to apple's
18:46:14TheSeventhe thing is just that you have no way to test it before flashing it
18:46:28Tornesomeone with a pogo pin flashing device :)
18:47:05Torneit's a pretty boring nor part, you cna flash it externally ifyou're willing to solder or buy a flashy pogopin thingy
18:47:06TheSevenon the nano2g, i can just corrupt the nor to enter dfu, which is very similar to the context the nor loader is usually executed from, so i can verify the loader before flashing it
18:47:19Tornethis is the reason nobody has looked at this, anywya, is the point
18:47:27TheSeven(but a broken loader with a valid checksum could still brick it)
18:47:41 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
18:47:42Torneit might be useful to solve a few issues, but whether the effort is worth it is debatable
18:47:58Tornealso you'd pretty uch have to nick a whole load of the OF's init code anyway
18:48:10Tornesince it does a whole bunch of setup that i don't think we have any idea about
18:48:18TheSevenif you have the possibility to test it easily, it might be worth the effort, but otherwise it isn't
18:48:49 Quit pamaury (Remote host closed the connection)
18:48:56***Saving seen data "./dancer.seen"
18:49:30*TheSeven continues to hunt that devious bug
18:54:41 Quit AndyI (Ping timeout: 248 seconds)
19:00:59 Join AndyI [0] (~pasha_int@
19:01:03 Join mt [0] (~mtee@rockbox/developer/mt)
19:08:55 Join DerPapst [0] (
19:11:07freddybTorne: I switched the glyph-cache loading to numerical order instead of reversed and the font loading is nearly instant now.
19:11:31 Quit delt_zZzZ (Quit: gotta fix some hardware shit, bbl)
19:11:43 Join Judas_PhD [0] (
19:12:10oxmanDo you have any solution to disable "Refreshing your media" on Sansa clip+ each time you plug off the usb ?
19:12:44 Join toffe82 [0] (
19:13:13bertrikoxman, no, not that I know
19:13:21 Quit jhMikeS (Ping timeout: 260 seconds)
19:14:03 Join jhMikeS [0] (
19:14:04oxmanI put MTABLE.SYS read-only, it don't work, try to corrupt MTABLE.SYS, useless too
19:14:06 Quit jhMikeS (Changing host)
19:14:06 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
19:16:36Tornefreddyb: what?
19:16:44Tornefreddyb: what did you actually do, I don't understand
19:16:51Torneit doesn't load them in numerical order, forward *or* reversed
19:17:01Torneit loads them in the order they are in the glyphcache file
19:17:05Tornewhich is the most-recently-used order
19:17:58*Torne has to leave for now, but if you link a patch i'll have a look :)
19:21:51 Join bunnyboi [0] (
19:22:22logiclostw0000 audio on android
19:22:26 Nick logiclost is now known as lostlogic (
19:22:37lostlogichmmm, rockbox on my palm pre... want.
19:26:00kugellostlogic: :)
19:27:41 Quit Rob2223 (Quit: Rob2223)
19:31:04 Join Rob2222 [0] (
19:31:48 Join rob [0] (
19:32:01 Quit rob (Client Quit)
19:35:09 Quit r0b- (Ping timeout: 258 seconds)
19:36:52 Join captainkewlll [0] (2669ecc2@gateway/web/freenode/ip.
19:37:03 Quit Kitr88 (Ping timeout: 240 seconds)
19:40:22 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:48:18 Join Ramsey[LC] [0] (~RamseyLC]@
19:52:21 Join Buschel [0] (
19:53:37 Quit stoffel (Remote host closed the connection)
20:01:37 Join stripwax [0] (
20:01:50mtBuschel: The patch I posted wasn't final (regarding comments), I just wanted to see what people thought of the introduced error. (also, I don't really know how to correctly interpret tiny_psnr results, I'll still compare the plots to see the avg error and the max error)
20:06:12 Quit Hillshum (Quit: Page closed)
20:12:33mtBuschel: The max error (between rockbox's mdct and wma pro's) is 0.02 (a pulse) - avg is 0.00022 on a 320kbps sample.
20:13:14mthmm ... but I still hear noise in the background
20:19:55 Join Wookie [0] (
20:20:38 Join devilhat [0] (
20:20:38 Quit Wookie (Client Quit)
20:21:31Buschelmt: 0.02 = 0.02%
20:21:41 Join rob [0] (
20:22:12mtBuschel: no .. 2% :(
20:22:34devilhatHi all. Anybody had any success running Rockbox on an ipod 120gb classic? Seems to be sixthgen, but RB only supports up to fifthgen,,,?
20:22:58S_a_i_n_tyou just answered your own question ;)
20:23:09S_a_i_n_tIf its on the main page, its supported.
20:25:05devilhat:) What's "5.5g" then? Fifth-and-a-half generation?
20:25:37S_a_i_n_t5.5 != 6 ;)
20:26:11 Join Buschel_ [0] (
20:26:43 Join ggg [0] (
20:27:10Buschel_mt: hmm, that's a lot... is the diff the same between your patch and mine?
20:27:13devilhatFound CNet review of 5.5g :)
20:27:15 Quit bunnyboi (Quit: Ex-Chat)
20:27:41 Quit Buschel (Ping timeout: 258 seconds)
20:27:53 Nick Buschel_ is now known as Buschel (
20:27:55mtBuschel_: That's your v04.. I'll test mine now
20:28:47devilhatThe reason I'm asking is that my Photo today reported it couldn't find rockbox.ipod and when I rebooted, said it couldn't find primary partitiion. :'( So is it new player time?
20:29:14S_a_i_n_tdevilhat: Have you tried the obvious things?
20:29:28devilhatAnd I won't buy another player I can't run Rockbox on :)
20:29:31S_a_i_n_t(checking the filesystem, restore using iTunes)
20:30:51devilhatWindows sees it as an external drive, but with 0B usedspace and 0B freespace, Windows disk checks do nothing
20:31:02 Join petur [0] (~petur@rockbox/developer/petur)
20:31:13S_a_i_n_tAnd, restore using iTunes...?
20:31:23S_a_i_n_t(its pretty hard to break an iPod)
20:31:34mtBuschel_: same figures
20:31:34devilhatnot got itunes installed - worth a go?
20:31:49AlexPCheaper than a new player
20:31:50 Join antgel [0] (
20:31:59S_a_i_n_tDo you *want* to throw out a potentially perfectly fine player?
20:32:10S_a_i_n_tah, yes...what AlexP said.
20:33:08antgellooks like the trusty H340 is on its way out. it's survived a battery and hard drive replacement back in the day, but now it's starting to sound thin at times, not sure why, it's not the jacks
20:33:23antgelit has been through a lot though *sigh*
20:33:39devilhat<sigh> and I didn't want to have to use itunes again, ever :) anything freeware/foss equivalent?
20:34:05devilhatyamipod, eg?
20:34:22antgelanyway looking for recommendations for a modern player that runs rockbox well, that will do all the H340 does. obviously i would expect a second-hand H340 to be at a similar time in its life, so looking for a newer model
20:34:35antgelall my music is in vorbis. any recommendations?
20:34:49S_a_i_n_tdevilhat: It can be done, but, put simply isn't worth the hassle.
20:35:12S_a_i_n_tInstalling iTunes is the only guaranteed way of restoring an iPod to factory settings as I understand it.
20:35:20AlexPI sympathise with the dislike of itunes, but it is much easier for this
20:35:26devilhatok then, itunes it is <sob> :)
20:35:44S_a_i_n_tA few minutes, then uninstall it, no biggie ;)
20:36:10 Join anewuser [0] (anewuser@unaffiliated/anewuser)
20:36:51 Part Buschel
20:37:42 Join Buschel [0] (
20:38:03Buschelmt: any change if you set WMAPRO_FRACT to <17 ?
20:39:45 Quit ggg (Quit: CGI:IRC (EOF))
20:40:42 Quit anewuser (Ping timeout: 240 seconds)
20:42:09 Join anewuser [0] (anewuser@unaffiliated/anewuser)
20:42:41devilhatok, installing itunes now...
20:44:52mtBuschel: the noise is more audible with FRACT < 17
20:45:48Buschelmt: and the other way (>17, e.g. 20)?
20:45:56mtyeah just tried with 19
20:46:01mtand sounds much better
20:46:10*mt plots
20:47:22 Quit anewuser (Ping timeout: 276 seconds)
20:47:47Buschelmt: sounds like the fract part is too low for the introduced error
20:48:00mtavg error is much better .. max is still the same
20:48:57***Saving seen data "./dancer.seen"
20:50:56mtBuschel : avg = 0.0001 (0.01%)
20:51:01mtmax still 2%
20:53:31Buschelavg should not be more than 0.003% (1/32768). the 2% spike is always at the same position?
20:54:18devilhatitunes installed, running now...
20:54:38Buschelmt: or is the avg that high because you have several spikes?
20:55:04AlexPdevilhat: You don't need to keep us up-to-date with every step :)
20:56:27mtBuschel: Several spikes
20:56:41devilhatok :) What am I looking for, at what point does it restore to factory settings? in the initial setup, or do I have to tell it to?
20:57:15Buschelmt: ok. did you try to go back to WMAPRO_FRACT (17) and apply the latest patch from FS #11498?
20:57:28Buschelmt: this should give +8 bits of fract part
20:57:33devilhatCos it;s already asked about turning CD autorun on on a system with no CD drive, so I don't rate its intelligence much... :)
20:58:00mtBuschel: No. I'll check..
21:00:50S_a_i_n_tdevilhat: I'm really not trying to be an ass, but your best bet is to google "how to restore an iPod <my iPod model>"
21:02:10devilhatS_a_i_n_t: no ass taken, will do - thanks for help so far all :)
21:02:50mtBuschel: avg = 0.008% .. max still the same, with several spikes too.
21:03:01 Join anewuser [0] (anewuser@unaffiliated/anewuser)
21:03:59Buschelmt: ok, at least the larger fract part leads to similar results...
21:04:20Buschelmt: you are testing on sim or one target?
21:04:51mtBuschel: on sim .. I'm comparing the output wav files of both mdcts
21:09:32antgelso, er, nobody like to recommed a modern device to replace my battered H340?
21:09:39antgelexplained ^^^^^
21:09:58S_a_i_n_tthe Rockbox Wiki buyers guide might help.
21:10:03mtAlso for some reason, after the recent optimisations, our decoder outputs a different number of samples than ffmpeg's
21:10:20antgelS_a_i_n_t: thanks. i find some of these things tend to be out of date but will have a look
21:11:27*mt will bisect for that tomorrow ..
21:11:34 Quit S_a_i_n_t (Read error: Connection reset by peer)
21:11:36mthave to get some sleep now
21:12:02 Join bunnyboi [0] (
21:12:40Buschelmt: just download a tool to check wav-difss. will keep you updated
21:12:46Buschelhave a good night!
21:13:21mtThanks :)
21:13:27CIA-6New commit by kugel (r27667): A few post-fixes to the get_user_file_path() commit. ...
21:13:36 Join S_a_i_n_t [0] (S_a_i_n_t@
21:15:16CIA-6r27667 build result: All green
21:21:29 Join Jerom [0] (~heidi@
21:21:57 Part watto
21:22:01 Part antgel
21:24:47 Quit stripwax (Quit:
21:25:20Buschelmt: the spikes occure each 2k samples. sounds like the frame length
21:29:22 Quit bunnyboi (Ping timeout: 265 seconds)
21:30:29 Quit FOAD (Quit: I'll be back)
21:35:24Buschelmt: overall analysis -> max diff = +/- 600 (of +/- 32768). between the spikes the noise floor is +/-30. <= v04-patch
21:36:42 Join FOAD [0] (~dok@
21:39:52 Join jfc [0] (
21:40:34 Quit bieber (Ping timeout: 248 seconds)
21:40:48 Join bieber [0] (
21:42:29 Quit hebz0rl (Quit: Ex-Chat)
21:43:49Buschelmt: with patch for larger FRACT_PART applied -> max error still the same. avg. error comparable...
21:47:33Buschelmt: but in this version the 2k-spikes disappear. there are other spikes left, that somehow like filter responses...
21:47:52Buschelmt: need to correct myself -> with larger fract part the error is 1/3 of v04.
21:51:36 Quit bieber (Ping timeout: 265 seconds)
21:51:51 Join bieber [0] (
21:52:29 Join stooo [0] (
21:53:49 Part stooo
21:57:07 Quit S_a_i_n_t (Read error: Connection reset by peer)
21:57:49 Join elcan [0] (
21:58:12 Join S_a_i_n_t [0] (S_a_i_n_t@
22:11:01devilhatok, have done a restore, hasn't helped, but am connecting to pc in disk mode. can I use a disk partitioner to create a new primary partition? or will that turn my music player into a portable hard drive? :)
22:11:33Tornedid itunes think it succeeded?
22:12:16devilhatno, i disconnected and reconnected, and it wanted to restore it again :(
22:12:25devilhatand then again
22:13:25devilhatbut if I hold down play+select as it boots, it drops into disk mode and windows sees it, no problem
22:13:51devilhatit's just that the properties show it to be a zero-capacity drive :(
22:14:40Tornewhat does the disk manager say about the partitions?
22:14:48Tornether should be a gap of 100ish MB
22:14:53Tornethen a partition on the rest of the disk
22:15:11S_a_i_n_twhat happened exactly when you restored the device?
22:15:11S_a_i_n_twhat did iTunes ssay/do?
22:15:12S_a_i_n_tI assume it didn't automatically detect an iPod needing restore on connect like it should.
22:17:04devilhatTorne: seems to be 1 55.89Gb "Healthy" partition
22:17:34Tornei assume this worked before? or did you get it in this state?
22:18:22S_a_i_n_tThe ideal thing to perform the task of formatting (if needed) and restoring the iPod is iTunes (however unfortunate that may be)...still interested in how the restore ended.
22:19:00devilhatyeah, was working over the weekend, plugged it into pc to put new music on, no apparent problems: turned on this evening and it's grumbling about no primary partitions
22:19:20Torneok, do you know how to get to diagnostic mode?
22:19:37Tornereboot it by holding select+play, then hold select and rewind as soon as the apple logo comes up
22:20:03Tornehit manual, then pick io -> disk -> smart info
22:20:06Torneor soimething like that
22:20:11Tornei forget what it's called exactly
22:20:17Tornedon't have my ipod handy
22:20:32devilhatSaint: plug it in, itunes pops up,click "Restore", "Restoring ipod..."
22:20:47devilhatTorne, thanks much didn't know about that mode :)
22:21:22S_a_i_n_tdevilhat: Sounds like it never ended, you may have disconnected it while the (sometime very hard to see) progress bar on the device was still doing its thing.
22:21:33devilhatSaint: "Your ipod restored to factory settings"
22:21:45S_a_i_n_tAh, Hmmm.
22:22:14devilhatbeep of disconnection, ipod reboots, rockbox boot loader chokes with "No partition found"
22:22:34S_a_i_n_tdevilhat: Does the OF work?
22:22:46S_a_i_n_t(maybe a silly question, didn;t see you mention that)
22:23:33devilhatcan't boot into it, but the disk mode screen looks like the original firmware, if that's what OF means :)
22:23:58S_a_i_n_tthere's two "disk modes" really.
22:24:52S_a_i_n_tdeliberately entering diskmode, and the OF dropping to diskmode is different as I understand it.
22:25:02devilhatyeah, there would be :grr: :)
22:25:30devilhatok, going to try torne's diagnostic mode
22:25:46*kugel commits his android progress
22:26:58 Quit Buschel (Ping timeout: 248 seconds)
22:27:43 Join Buschel [0] (
22:28:24 Join stripwax [0] (
22:28:49devilhatTorne: ok, i have SMART data up - is that what you mean?
22:31:28Torneyes, what does it say?
22:32:20 Join bba1973 [0] (
22:33:11devilhat"Retracts: 57, reallocs: 0, Pending sectors: 0,Poweron Hours: 233, Start/Stops:113
22:33:32Tornehm, that's ok then
22:34:53CIA-6New commit by kugel (r27668): Rockbox as an application: Commit current Android port progress. ...
22:36:57CIA-6r27668 build result: All green
22:37:54kugeldo we have some sort of .gitignore in svn?
22:37:55bba1973I keep getting data aborts with my iPod Photo while trying to commit the database. It used to work fine. Any ideas on how to fix it? Restoring it with iTunes several times didn't help, and I've tried reinstalling Rockbox with rbutil and manually. Could my logic board be dying?
22:38:13kugelor svn ignore
22:38:24devilhatSaint: first time I restored, it downloaded a new version of firmware, never did that again...
22:38:25Tornedevilhat: sorry, dunno what to suggest..
22:38:35Tornebba1973: probably not, you probably have a file with a bad tag
22:39:08bba1973You mean a music file with a messed up ID tag?
22:39:28bba1973I've got most of my music on my iPod 5.5G with Rockbox, and it works fine.
22:39:43devilhatTorne: Consult my local Qualified (unofficial, natch ;) ) Apple Repairer? ;)
22:40:27freddyb Torne: If I read the glyphs in disk order and then re-read them in .glyphcache order that will put the order right again?
22:41:21Tornebba1973: i'd start with whatever isn't on there then :)
22:41:24S_a_i_n_tdevilhat: The reason it only downloaded the firmware once is that it stores it on the PC and does not remove it during an install.
22:41:28 Quit {phoenix} (Remote host closed the connection)
22:41:35S_a_i_n_terrr...uninstall rather
22:41:40Tornefreddyb: should do, barring the data being too big for the cache
22:42:24devilhatSaint: OK, gotcha. Wondered if it was detecting the most uptodate firmware on the ipod and hence not installing
22:42:26freddybThe .glyphcache file is font independent, right?
22:42:38Tornefreddyb: the ther idea that was suggested was just dumping the entire contents of the cache to disk instead of just the list of glyphs
22:42:47Torneyes, it only has one
22:42:51bba1973I'm going to play all my tracks while going through the file browser instead of the database. Maybe I'll find the offending one.
22:43:11S_a_i_n_tdevilhat: You can force it to check for the current version, and it will grab one if there is one.
22:43:13Torneso the cache may well refer to more glyphs than will fit in the cache using the current font
22:43:26Tornesince it might've been generated from a smaller font
22:43:29S_a_i_n_tBut there hasn't been a newer revision in iPod FWs in *ages* now.
22:43:44Tornerestore always reinstalls the firmware
22:43:49Torneonly upgrade does nothing if it's up to date
22:44:07 Join jthundley [0] (
22:44:32S_a_i_n_tAh, correct indeed.
22:44:33jthundleyhi guys
22:45:36bba1973It's cutting off the first letter of some albums and songs. It's also messing up the screens. Some pixels aren't the color they're supposed to be, but the Apple firmware seems to be working fine.
22:45:49jthundleyI just updated rockbox on my sansa e200v2 and my custom theme isn't working. I have everything commented out and it's failing on the second line (loading the background) for some reason. The UI simulator with −−debugwps isn't telling me anything
22:46:13jthundleyI'm re-reading the customwps wiki page and it looks like I'm doing everything correctly
22:46:24devilhathmmm. ok, looks like it might be a drive issue rather than rockbox then
22:46:30S_a_i_n_tjthundley: post the line it is failing on.
22:47:15*jthundley tries
22:47:18S_a_i_n_tCustomWPS is out of date, use the manual.
22:47:37piggzis there any hope for rockbox supporting the video chip on an ipod video?
22:47:53 Join bunnyboi [0] (
22:48:40 Quit stripwax (Quit:
22:48:41S_a_i_n_tjthundley: Use the manual for your device, it in itself is not a *complete* list of theme tags, but moreso tham CustomWPS. And CustomWPS has blatantly incorrect syntax.
22:48:42jthundleythanks for the tip S_a_i_n_t, didn't know that. It's actually still reverting to the default theme for some reason
22:48:59jthundleyI'll check out the manual, thanks :)
22:49:01***Saving seen data "./dancer.seen"
22:49:43 Quit Strife89 (Quit: Departure.)
22:50:21bba1973Doing drive scans in diagnostic mode.
22:50:22 Quit S_a_i_n_t (Quit: "back soonish")
22:53:00 Quit bertrik (Quit: :tiuQ)
22:53:32jthundleyis the UI simulator picky about end of line in wps's?
22:54:05jthundleyI'm running the UI simulator on a windows machine but editing my theme on my linux machine
22:57:00 Quit captainkewlll (Quit: Page closed)
23:07:39 Quit bba1973 (Quit: CGI:IRC)
23:08:54devilhatOK, many thanks to all and to S_a_i_n_t when they're back :) be good. :)
23:09:38 Quit devilhat (Quit: CGI:IRC (EOF))
23:09:55 Quit Buschel (Ping timeout: 276 seconds)
23:10:44 Quit bunnyboi (Ping timeout: 260 seconds)
23:21:35CIA-6New commit by alex (r27669): Manual: %Vl tag was in the viewports section but missing from the tag table in the tags appendix.
23:23:36CIA-6r27669 build result: All green
23:37:00 Join fml [0] (
23:37:29fmlkugel: hello. What do you say to this: (amiconn doesn't like the binsize change from r27656, given that it should be a no-op on real targets :\\)
23:38:39CIA-6New commit by alex (r27670): Manual: Fix typo
23:38:50kugelfml: maybe it's not a complete no-op in every case, but for the most part
23:39:38fmlkugel: but it was advertized as such
23:39:47 Quit freddyb (Read error: Connection reset by peer)
23:40:13kugele.g. there were some cases where the compile time constant was part of the format string in snprintf (ROCKBOX_DIR "/backdrops"). in those cases it need to be changed to "%s/backdrops, get_user_file_path()..."
23:40:16 Join freddyb [0] (
23:40:21kugelthe call itself is a no-op still
23:40:43CIA-6r27670 build result: All green
23:43:25 Quit fml (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
23:46:39 Quit jgarvey (Quit: Leaving)
23:49:40 Join stripwax [0] (
23:53:02 Join S_a_i_n_t [0] (S_a_i_n_t@

Previous day | Next day