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 2009-08-18

00:00:00rasherZagor: that's strange
00:00:32 Join truthtaco_ [0] (
00:00:55gevaertsJdGordon|: we go through (parts of) the OF on most targets
00:01:18JdGordon|not like the mr500
00:01:27JdGordon|stop your bloody pedantiness
00:01:30 Quit TruthTaco (Read error: 60 (Operation timed out))
00:01:33ZagorJdGordon|: generic builds?
00:01:57JdGordon|Zagor: like checkwps, or tools... just running a .sh or a make
00:02:40gevaertsJdGordon|: why does that difference matter? What's so wrong about the mr500 way that it would make it a no-go for support?
00:02:42 Nick truthtaco_ is now known as TruthTaco (
00:02:58JdGordon|I didnt say it would... I just said it sucks
00:03:13JdGordon|no other target has to be booted as an extra step
00:03:24 Join robin0800 [0] (
00:03:59 Part domonoky
00:04:03rasherZagor: I can't figure out why it wouldn't work :\
00:05:28gevaertsJdGordon|: I agree there. It would be a lot nicer not to have that feature
00:06:05JdGordon|I'm bringing my jtag-ed mr500 to new york.. maybe we can have a play
00:06:11Zagorrasher: "SELECT * FROM `log` WHERE client = 'lillebror1-zagor' AND TYPE = 'completed' AND revision =22381" shows different data than the graph, at least
00:08:35Zagoryay, first 1000+ speed reading
00:09:38kugelJdGordon|: what is the problem about the rec statusbar?
00:09:49JdGordon|its there!
00:10:03JdGordon|its actually completly different than the regular one
00:10:16JdGordon|and not possible to skin jsut yet
00:10:34 Join Thundercloud [0] (
00:10:35 Quit dfkt (Read error: 104 (Connection reset by peer))
00:11:19 Join robin0800_ [0] (
00:11:42 Quit bluebrother ("leaving")
00:12:35Zagorrasher: now it worked! weird.
00:13:01rasherZagor: I poked it.. but nothing that shouldn't have been working previously I think..
00:14:22 Quit robin0800 (Read error: 60 (Operation timed out))
00:17:00kugelJdGordon|: eh, making the old statusbar and new one play nicely together is nasty
00:17:24JdGordon|play together how?
00:17:56kugeluser either via the statusbar setting
00:18:21JdGordon|yeah, nup forget it... complelty remove the old one except in the rec screen
00:18:30JdGordon|(longer term goal)
00:19:10kugellonger term goal. How about this can get quicker in than custom vp?
00:19:13JdGordon|hmm.... this might not work so nicely.. each display will need its own statusbar skin line
00:19:33JdGordon|I dont tinhk we can reasonably move on this untill the skin struct is much smaller
00:19:39kugelit already has I think?
00:19:53JdGordon|no, its resized in code
00:20:14JdGordon|"it has already" to what?
00:20:21kugel"each display will need its own statusbar skin line"
00:20:32JdGordon|screen size I meant
00:21:14kugelyou confuse me now
00:21:23JdGordon|the skin code
00:21:30JdGordon|to fit the screen properly
00:21:42JdGordon|also, crap... the images neeed to come from disk which wont be nice
00:22:55JdGordon|wwe shuold be able to get around that with some hacks, but wont be nice
00:23:04JdGordon|unless we drop the images?
00:23:39kugelI think we might be able to use little hacks
00:23:52JdGordon|or statusbar only tags
00:24:19kugelon the other hand, images from disk arent that bad, loaded once only anyway
00:24:49kugelor statusbar only tags. although I see now real reason to not offer them to the wps also.
00:24:54JdGordon|the viewport tag would be really nice if it could do -X for the placement bits.. so it starts X pixels from the bottom/right of the screen
00:26:47kugelshould be possible to implement
00:27:34JdGordon|I wonder how many people would be upset if the default statusbar was redically simplified? only showing text and have it centered
00:27:40JdGordon|that would greatly simplify things
00:27:44JdGordon|or left aligned
00:28:18JdGordon|pixelma: any thoughts on that?
00:28:51kugelare you serious?
00:28:54pixelmacan't follow at the moment
00:28:58amiconnSince the statusbar is also part of the usb screen, and early usb must not access the disk, there must be a built-in default statusbar
00:29:22kugelthe disk won't be accessed
00:29:29JdGordon|amiconn: it would all be loaded well before usb is started
00:29:36kugelit's working like a wps, loaded once.
00:29:56JdGordon|kugel: my thinking is that I would imagine most people either disable the statusbar, or would use a pretty one once all this goes in
00:30:07amiconnEarly usb must not access the disk *at all*. The rockbox binary comes from flash rom in this case
00:30:10JdGordon|just like the wos
00:30:23pixelmaisn't early USB when you start Rockbox in USB mode?
00:30:31kugelJdGordon: let's just add some %SB tags for those tiny built-in icons..
00:30:38saratogaanyone around who is familar with filterbanks
00:30:44saratogagoogling is just making me more confused
00:30:56JdGordon|kugel: I tihnk we can scrap them completly
00:31:31JdGordon|we could also ship a file along with the needed .bmp's for those that want it
00:31:56JdGordon|seen as cabbie has the sb disabled on almost all(all even?) targets
00:32:05kugelyea, or we make it like the hardcoded wps and use simple tags
00:32:22pixelmaa fallback WPS is there too, why not a similar approach there?
00:32:26JdGordon|its not possible to keep the look consistant without having a #ifdef hell
00:32:56JdGordon|pixelma: we would, but I'm proposing a single skin setup for all targets, not a target specific one for each
00:33:02kugelwhy? we can insert the .sb file into the .c file before compiling
00:33:04JdGordon|the wps has 2 screen sizes
00:33:22JdGordon|kugel: and who is going to set them all up?
00:33:39*JdGordon| thinks getting rid of those built in icons is a useful change
00:33:53CIA-61New commit by zagor (r22386): Added support for release tags. Patch by Dominik Wenger (FS #10525).
00:35:07kugelJdGordon|: the ifdef hell could also be in external .h files
00:35:25JdGordon|I tihnk avoiding it completly is better
00:35:34*kugel is undecided
00:35:40JdGordon|think about it.. to do it properly you are talking at least 3 viewports
00:35:51JdGordon|+ adding tags to handle the inbuilt icons
00:36:05kugel1 vp should be enough
00:36:14JdGordon|no way
00:36:25JdGordon|target specific spaces to pad it?
00:36:38pixelmaJdGordon|: for the complete skin? And do you only mean the fallback WPS/skin or also stuff like cabbiev2 (I doubt the latter would work with all the different screen sizes and aspect ratio etc.)
00:37:02JdGordon|I only mean the fallback skin
00:37:08kugeljust a premade .sb in wps/ for each target
00:37:20JdGordon|kugel: the fallback cant touch the disk
00:37:38kugelthat's why it should be automagically turned into a .h or so
00:37:45JdGordon|I already said I think it would be good to have a target specific skin for each
00:38:08kugelisn't that what I said?
00:38:08 Quit saratoga ("CGI:IRC (EOF)")
00:38:16mcuelenaere Anyone opposed to commiting FS #10534 (make scrollbar width variable)? I could make binsize hit smaller by only enabling it for touchscreen targets (don't see much use for non-touchscreen targets..)
00:38:27JdGordon|mcuelenaere: go for it!
00:38:28kugelgo for it
00:38:47Zagornew round speed record: 182 seconds
00:38:55kugelccache ftw!
00:38:59JdGordon|kugel: the difference is you want to hard code that back in.. I'm saying thats not needed, the hardcoded fallback should be dead simple
00:39:15JdGordon|Zagor: you can do better than that!! :)
00:39:18kugelbut that's going to hit the disk, isn't it?
00:39:40pixelmaok, couldn't the distribution of the fallback icons be calculated from the screenwidth and then put into the fallback .sb file (like LCD_WIDTH minus all icons width divided by number of all icons -1 -> gets you the spacing
00:39:41kugelloading icons?
00:40:15JdGordon|kugel: put it this way.. cabbie is the default theme, that has to come from disk, and it doesnt have the sb enabled... if cabbie fails to load then the bar comes from a very boring hardcoded with no images, and the hardcoded wps
00:40:26pixelma(number of all icons - 1) ;)
00:40:28JdGordon|ideally users would never see the hardcoded sb (like the wps)
00:41:03kugeli think it shouldn't be too hard to have the fallback including icons built in
00:41:17JdGordon|they are not needed... and waste space
00:41:19kugelgenerated from .sb files in wps/
00:41:25CIA-61New commit by mcuelenaere (r22387): FS #10534: Make scrollbar width variable through a setting (useful for touchscreen targets)
00:41:33JdGordon|you have to add special handling for them which isnt needed
00:41:42JdGordon|noone is going to use the current bar anyway
00:41:53kugelI don't think so
00:41:57JdGordon|and hitting the disk to load a prettier one isnt going to cause any issues
00:41:58*mcuelenaere was also thinking of adding an option to position the scrollbar on the left
00:42:01pixelmaI might
00:42:02kugelit's a valuable thing on smaller screens
00:42:13JdGordon|mcuelenaere: go for it!
00:42:20JdGordon|although real RTL would be better
00:42:44JdGordon|OK, I'll try explaining once more....
00:42:55JdGordon|for hebrew/arabic
00:43:18mcuelenaereah you mean the text, I'm only talking about placing the scrollbar on the left
00:43:24mcuelenaereI think it's easier to handle on-target
00:43:41pixelmawith my suggestion I don't think you need different sb files (and I think it's currently done similarly in the fixed statusbar now - if not fixed positions, some calculated from the right side)
00:44:04JdGordon|pixelma: kugel: think about the wps, there is a very boring wps which is bassically the "error" wps, cabbie is the default. if cabbie cant load then the error one is shown... same for the statusbar, we would ship a "" which is the same as todays, and the error one would be similar but much more boring
00:44:59kugelthe error wps is build-in though
00:45:19kugelyou say it shouldn't be
00:45:36pixelmacabbiev2 uses the small icons in list on small screens - and on small screens I think having the old statusbar is nicer as it doesn't waste so much space
00:45:37JdGordon|no... I'm saying there will be ONE built in, just like the wps, which is boring
00:46:44 Join fdinel [0] (
00:46:51pixelmain lists, in the WPS it's ok as you don't have so many lines anyways
00:47:09CIA-61New commit by mcuelenaere (r22388): Don't redefine SCROLLBAR_WIDTH (aka fix yellow)
00:47:25JdGordon|pixelma: yes, the current bar would be there, but it would come from disk instead of hardcoded
00:47:40JdGordon|there would be an error fallback on hardcoded incase it cant load
00:47:47 Join aaron424 [0] (
00:48:28pixelmaI don't know
00:49:01JdGordon|kugel: did that make it clearer?
00:49:08kugelyea, I don't like it
00:49:51JdGordon|arg, why?
00:50:06amiconnJdGordon|: As already mentioned, the default statusbar must be available without hitting the disk at all, otherwise early usb will break
00:50:23kugelI like %SB tag idea better
00:50:32kugelmaking the icons free for use in any skin
00:50:37JdGordon|amiconn: it wont be a problem, slightly less pretty but not a problem
00:50:43mcuelenaerehmm should I bring LANG_LEFT & LANG_RIGHT back from the dead or use LANG_SCROLLBAR_LEFT & LANG_SCROLLBAR_RIGHT?
00:51:02amiconnThat means the statusbar code must not even try to read the disk unless told explicitly (i.e. by loading the settings)
00:51:05JdGordon|kugel: the STATUSBAR icons are not skinnable... they SUCK.... they are a waste
00:51:15kugelmcuelenaere: I guess the latter one, it might still be translated in some files
00:51:40kugelwaste? how can 8x8 icons be a waste?
00:51:48JdGordon|because they arnt useful
00:51:57mcuelenaerekugel: but doesn't that duplicate language strings? I can imagine a generic LANG_LEFT & _RIGHT being useful in other places..
00:52:14amiconnI think this bar can be even simpler than the current one though, because in that mode there will never be playback (obviously)
00:52:24kugelmcuelenaere: ugh, I meant the former*
00:53:08JdGordon|kugel: I would be very surprised if many people used the hardcoded bar, so anything added to code for it is a waste
00:53:20 Join Strife1989 [0] (
00:53:25kugelI can imagine that
00:53:38JdGordon|what? that people would or wouldnt use it?
00:54:06kugelI would use the current statusbar on the clip, possibly on my yh920 too
00:54:25JdGordon|ok, but why would you care if it comes from disk or is in the code?
00:54:41*amiconn will probably use the default bar on most of his targets
00:54:57kugelfor a) consistent fallback sb, and b) as for the reasons amiconn mentioned
00:55:01gevaertsamiconn: the early-usb statusbar issue could also be "solved" by having the usbscreen provide more (i.e. some) information itself
00:55:05JdGordon|if you want to then change it a bit to suit you, you then have to have 2 copies of it in ram
00:55:14 Quit jgarvey ("Leaving")
00:55:38JdGordon|kugel: how can you say that if the whole thing is identical on every target is not consistant?
00:55:41pixelmamcuelenaere: where's LANG_LEFT/RIGHT used currently? Just trying to find out if there could be context dependent translations that did not just simply translate to left/right (and where the translater could at least need some hint that it is more generic now, if it wasn't already)
00:55:56amiconnWhy use anything that wastes more space than that, or even fiddle with it? It works, and has all the necessary information...
00:56:00kugelit's not the same, but it's still consistent over targets, like cabbiev2
00:56:10mcuelenaerepixelma: it's deprecated, so I guess nowhere
00:56:29JdGordon|you cant undeperatce a string without breaking voice...
00:56:33JdGordon|or, I assume not
00:56:36amiconnThe only exception will probably be the targets with very high pixel density... i.e 3 out of 15 for me
00:56:41pixelmaJdGordon|: ok, then it'll show up as different anyway
00:56:51pixelmaerr mcuelenaere
00:57:08mcuelenaeregrep only finds references in apps/lang/*
00:57:26mcuelenaereJdGordon|: what do you mean?
00:57:40kugelmcuelenaere: I don't really understand capping of the scrollbar to 6 and 20 (20 is OK, but 6 could even be lowered)
00:57:59mcuelenaereah yes, those are just arbitrary
00:58:00JdGordon|mcuelenaere: LANG_LEFT is already deprecated right? I tinhk if you undeprecate it you will break current voice and lang files?
00:58:20kugelyou better ask rasher for that
00:58:21mcuelenaereI don't know, I'm not really familiar with voice files
00:58:37kugelI guess undeprecating + moving it at the end would work
00:58:41mcuelenaerebut I guess I could mess a bit with (or what's it called?)
00:59:07JdGordon|kugel: that will definetly break it
00:59:34kugeldeprecated stuff is ignored, it's not in the lng or voices after genlang
00:59:47kugelisn't it?
00:59:50mcuelenaerekugel: what do you suggest as a lower limit? 1?
00:59:53JdGordon|its not ignored.. iirc its kept in the generated .lng as a placeholder
00:59:59JdGordon|to keep the enum correct
01:00:21mcuelenaereand perhaps the upper limit could be bigger (for 640x480 targets)
01:00:26JdGordon|mcuelenaere: and an upper limit of 30 even...
01:00:37kugelmcuelenaere: 3 probably, I'm not sure if the scrolling part of the bar is still visible at 2
01:00:45JdGordon|the range is free... so you could go as far as LCD_WIDTH even!
01:00:46kugelthe scrollbar has a 1px border too
01:01:29*gevaerts proposes an upper limit of LCD_WIDTH/10
01:01:35 Quit bertrik ("Leaving")
01:01:45*mcuelenaere uses gevaerts' proposal
01:01:58JdGordon|I know pronounce you man and scrollbar!
01:02:00kugelthen also update the manual accordingly :)
01:02:09kugelI'd like to see how you add this to the manual
01:02:30JdGordon|if the bar has a border of 2x1 pixels then make the minimum 3 and subtract 2 from it when drawing...
01:02:44pixelmakugel: and your changes?
01:02:45mcuelenaerehmm I don't even have the manual toolchain installed..
01:02:51 Quit efyx_ (Remote closed the connection)
01:03:14 Nick robin0800_ is now known as robin0800 (
01:03:33 Join J-23_ [0] (
01:03:35kugelpixelma: I'll do it, BdN3504 wanted to look at the ui viewport for me
01:03:43 Nick J-23_ is now known as Guest23768 (
01:03:57 Quit robin0800 (Remote closed the connection)
01:04:38 Quit J-23 (Read error: 104 (Connection reset by peer))
01:05:48CIA-61New commit by mcuelenaere (r22389): * Fix red ...
01:07:09 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.0.13/2009080315]")
01:08:13 Join robin0800 [0] (
01:10:00 Quit Zagor ("Clint excited")
01:12:17 Quit GeekShado_ ("The cake is a lie !")
01:15:51 Nick Strife1989 is now known as Strife80 (
01:16:00 Quit stripwax (Read error: 104 (Connection reset by peer))
01:16:15kugelJdGordon|: the other nice thing about the skin-statusbar is that it will bring the touch regions into the ui (hopefully)
01:16:17***Saving seen data "./dancer.seen"
01:17:17kugeland, since .sb and .wps should be compatible, possibly staying in the wps on the remote while browsing menus in the main unit (and vice versa)
01:17:20 Quit Thundercloud (Read error: 101 (Network is unreachable))
01:17:54JdGordon|I'm not agreeing with anything untill you agree that a minimal fallback .sb is the way to go :D
01:19:56kugelstroppy cow ;)
01:24:17 Quit robin0800 (Remote closed the connection)
01:27:58 Join aaron424 [0] (
01:29:21 Quit trisiak (Read error: 110 (Connection timed out))
01:35:44 Part toffe82
01:37:26CIA-61New commit by mcuelenaere (r22390): Add setting to position the scrollbar on the left or right
01:38:45 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.0.13/2009080315]")
01:42:58 Join TheCoolGman [0] (n=446e5e7d@gateway/web/cgi-irc/
01:43:51JdGordon|kugel: sb patch? im getting on the bus in 30min and want to have a looky
01:45:23TheCoolGmanHi I've tried to bring this up before but no one was on IRC, MPEG Player can be such a pain playing this guessing game to see what mpg is going to play. So I was wonder what is the possiblites of making MPEG Player project any video for that target like iPod video projects at 320x240?
01:48:13 Quit Strife80 (Remote closed the connection)
01:49:23CIA-61New commit by mcuelenaere (r22391): Update Dutch translation
01:57:28 Join Strife1989 [0] (
01:58:08 Nick Strife1989 is now known as Strife89 (
02:00:40kugelwpsbuild.cfg is too complex for me :(
02:01:09 Quit Rondom (Nick collision from services.)
02:01:20 Join Rondom [0] (
02:05:56 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
02:09:16 Quit mt (Read error: 104 (Connection reset by peer))
02:15:02 Quit CaptainKwel ("Page closed")
02:15:14 Join HBK [0] (
02:15:31 Join JdGordon_ [0] (i=441bee45@gateway/web/freenode/x-erljrnyaczsjkiju)
02:17:03 Quit daurn ("Leaving")
02:17:41 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
02:18:27 Quit martian67_ (SendQ exceeded)
02:18:37JdGordon_is there a conditional for rtc present?
02:19:01DhraakellianI just got approximately 13-13.5h out of my Fuze
02:19:21pixelmamcuelenaere: as far as I can see, the Player doesn't have a scrollbar so the lang strings should be only for the ones with lcd_bitmap feature (I think)
02:19:40pixelmaJdGordon_: yes
02:19:51mcuelenaerepixelma: hmm then that code shouldn't be compiled for the player either?
02:19:59TruthTacoI havent timed it but i think the rockbox software gives me considerably less battery time than the OFW for the fuze
02:20:01JdGordon_CustomWPS says use %?ca ?
02:20:05 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
02:20:42Dhraakelliannot the most scientific test, of course. (multiple codecs, multiple bitrates, hold button off, backlight on and off a lot due to interaction, DAP turned on and off a number of times, etc)
02:21:07JdGordon_kugel: pixelma: this probably doesnt work but this is what I think the fallback/error statusbar shuld be for all targets... "B:%bl%% V:%pv %?mh<|H> %ar%?ca<%cH:%cM|>%?lh<*|>" see how simple it is? it means its very obvious that the one you tried to load had an error
02:21:10Dhraakelliannot as good as my e260 was getting, but not too horrible
02:21:53kugelJdGordon_: we'll see
02:22:18pixelmaJdGordon_: hmm, no. That is/was only for targets to check if the RTC is set
02:23:24JdGordon_otherwise we need to remove the rtc bits in non rtc bulds (obviously)
02:23:35JdGordon_the rest is target agnostic which makes it very useful
02:23:44JdGordon_the only thing missing is remote hold i think
02:23:48JdGordon_and play status
02:23:58pixelmaIIRC you can use every valid %csomething there but I introduced a RTC present tag, %?cc if I'm not mistaken
02:24:02JdGordon_which could be added a P/p/S ?
02:24:20pixelmaand I thought I put it somewhere (maybe manual)
02:24:30JdGordon_I might have missed it in the wiki?
02:25:21JdGordon_by my count thats les than 30 tokens...
02:27:04 Quit martian67 (Connection timed out)
02:27:18pixelmayes %?cc and it's in the manual but not on CustomWPS
02:27:24JdGordon_actually... the whole play/pause/stop word might even fit
02:28:37pixelmaI always wonder how targets with a real HD active LED deal with the virtual LED tag
02:28:51JdGordon_they display it i assume?
02:29:40pixelmaif nothing's changed it's not compiled for those so I think it's ignored -> display nothing?
02:31:01pixelmanot sure about your suggestion and I need some sleep
02:33:39 Quit TheCoolGman ("CGI:IRC (EOF)")
02:36:33 Quit martian67_ (Read error: 104 (Connection reset by peer))
02:38:35 Join toffe82 [0] (
02:40:02kugelthis is...meh
02:40:20kugelwhy couldn't they use a reasonable script language :(
02:40:49 Quit Strife89 ("Leaving")
02:41:44*kugel avoids that question
02:43:42CIA-61New commit by mcuelenaere (r22392): Reclaim a bit of binsize for non-lcd_bitmap targets
02:45:05 Quit JdGordon_ (Ping timeout: 180 seconds)
02:45:57kugelmcuelenaere: it's only 1 :)
02:46:01kugelstupid charcell
02:46:06mcuelenaereah :)
02:53:33CIA-61New commit by kkurbjun (r22393): M:Robe 500: Increase the max icon size.
02:56:35 Join wooster23y [0] (
03:00:11wooster23yhello rockboxers.. I have sansa e200 and with recent builds under Plugins > Applications > i see a new entry: Remote Control
03:01:03 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
03:01:15wooster23yI have no idea what it is... when I select it I get various entries that appear to be used to control a PC or something (Desktop, Presentation, Media Player)...
03:01:23tmztwooster23y: that's for using the Sansa as a hid remote control
03:01:31tmzt(like a usb device)
03:01:31wooster23ytmzt, ok thanks
03:02:01wooster23yI kinda though that but whenever I plug the sanda into USB, the screen, etc is disabled so I can't access anything
03:02:08wooster23ythe sansa that is
03:02:58wooster23ythe sansa screen is filled with an image of a USB plug, and can't be 'controlled' itself to do any controlling
03:03:00tmztafter selecting the plugin?
03:03:27wooster23yI don't have any HID installed on my desktop , maybe that's the problem
03:03:36wooster23yI've never used HID before ...I'llcheck on that
03:03:55JdGordonwow, that icon change added 83KB!
03:04:06JdGordonno... 166KB!
03:05:00kkurbjunyeah, quite a bit.. unforunately the code reserves a buffer for each screen that is the same size
03:05:11gevaertswooster23y: you have to connect in charge-only mode for that plugin to work
03:05:22kkurbjunso teh remote has support for 35x25 pixel icons that it will never use
03:05:41wooster23ygevaerts, oh, ok, thanks
03:05:53wooster23yI'll try to figure out how to get into that mode :>
03:06:43gevaertsit's in the manual somewhere. Maybe you need to look for "power" instead of "charge" though
03:07:26wooster23yok thanks gevaerts
03:07:48 Quit mcuelenaere ()
03:09:52JdGordonkkurbjun: ah yes, that usual problem
03:12:25kugelno problem, he has still 530k free :)
03:13:46kugelunless we stop JdGordon in his malloc frenzy, icons might get skin_alloc'd or something too soon'ish :)
03:14:13JdGordonthat would be bad? :confused:
03:14:58kugelscreens should probably more often share the same buffer which can sized more dynamically (at compiling time)
03:15:56 Quit kkurbjunW (Remote closed the connection)
03:16:03 Quit DataGhost (Nick collision from services.)
03:16:11 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
03:16:15kkurbjunI think sizing buffers at load times for themes makes sense
03:16:20***Saving seen data "./dancer.seen"
03:16:38 Join kkurbjunW [0] (n=karlk@
03:17:59 Join CaptainKwel [0] (
03:19:40 Quit wooster23y ()
03:26:39 Quit scorche (Connection timed out)
03:28:59 Quit elcan ("leaving")
03:29:08 Join elcan [0] (
03:29:24kugeltime to make the sb patch a bit more tracker'able
03:29:39JdGordonby starting again? :)
03:30:17gevaertsor by ading random chat? :)
03:31:33 Join funman [0] (n=fun@rockbox/developer/funman)
03:32:31kugelby adding a .h ;)
03:37:02 Join Rudy4Pez [0] (
03:38:20Rudy4PezLlorean, Topy44: Another concern about album art .jpg file detection is the "AlbumArt_{32B3752A-B3BC-42B7-81B5-3E33919A6A6E}_Large.jpg" format that WMP uses. I forgot to mention that last night.
03:39:10kugelI don't see how album art.jpg would help there
03:39:11Topy44Rudy4Pez: we actually agreed on front.jpg in the end
03:39:20Rudy4Pez(And also back when I ripped CDs, I used MusicMatch. I'm pretty sure that's the program that generated "album.jpg".)
03:39:43Topy44i have been thinking about the format you just mentioned, but i cant see a way for rb to generate that filename without parsing the dir
03:39:57CaptainKwelHello. May I have write access to the wiki? JasonAYu
03:40:41Rudy4PezTopy44: Really? I thought Llorean would never budge on that, haha. Also, the WMP format is a good argument to have RB parse the dir. I just couldn't remember it last night because I was running on fumes.
03:41:50Rudy4PezBut in the mean time, adding front/album.jpg to the existing list is the simplest solution.
03:42:15Topy44i would ask you to confirm that musicmatch generates album.jpg files by default
03:43:24Rudy4PezOh man, it's been a while since I've used it. I can find out though.
03:44:03funmanCaptainKwel: you should be able to write now
03:44:32 Join dmb [0] (n=Dmb@unaffiliated/dmb)
03:45:24Rudy4PezApparently "dBpoweramp" generates it as well.
03:47:05Topy44never heard of that
03:47:44Rudy4PezNor have I, lol.
03:47:58Rudy4PezBut VMJ uses it too.
03:48:05 Join moos [0] (i=mustapha@rockbox/staff/moos)
03:48:06Rudy4PezAnd I have used that before.
03:50:44Rudy4PezI also found an OS X program called sbtooth that uses it.
03:50:59 Quit dmb (Read error: 60 (Operation timed out))
03:52:15Rudy4PezAs does J.River MediaJukebox/MediaCenter.
03:53:40Rudy4PezIf I recall, I believe Winamp (or a popular album art plugin for Winamp) had it on the list of valid filename formats.
03:53:59Rudy4PezBut it's been a while since I've used any PC-based music programs.
03:54:05Rudy4PezWinamp included.
03:54:14Topy44good enough for me, i'll suggest it... not my decision anyway, but the more programs use it the more likely the stupporn ones like llorean will agree :)
03:54:34CaptainKwelfunman: thanks!
03:55:21*gevaerts wonders how endless discussions about which programs use what album art naming scheme can be called on-topic
03:55:46Topy44its a discussion on wheter or not to include the filename in rb, therefore its on topic imho
03:56:20funmanwhere is the list of filenames checked for album art?
03:56:41Rudy4PezTopy44: Exactly. I personally don't use any of theose programs, but other people obviously do.
03:57:02Rudy4PezGotta think outside ourselves. ;)
03:57:22Topy44funman: its on the docs page, search album art
03:57:47Topy44the routine is in albumart.c line 124 onwards
03:57:59gevaertsAll I see is an endless list of "oh, I foun this obscure program that also uses it", without any reference to what "it" might be, and with a more or less obvious undertone of "take that, Llorean"
03:58:23Topy44no, i do get lloreans point
03:58:43Rudy4Pez"it" is the format of 'album.jpg' as an album art filename.
03:58:49Topy44if one or more relevant pc programs use it its worth consideration - i'm not saying it has to go in, but it should at least be suggested
03:58:51funmanthis list looks quite specific to rockbox, without looking for compatibility with external programs
03:59:02 Join Blue_Dude [0] (
03:59:17Rudy4PezI have a feeling that "relevant" is going to be a point of contention.
03:59:36Topy44yeah :) but you might aswell ask "is mp3 really a relevant file format?" :)
03:59:42Rudy4PezIf the devs have never heard of it, good luck on getting it included.
04:00:04funmanand incomplete, looking at the code (folder.jpg
04:00:17funman"*folder.jpg" isn't mentioned in the list, and ./cover.jpg is loaded afair
04:01:15funmanRudy4Pez: if several programs use a common naming for album art, and other programs are moving toward this scheme, then it's a good hint for rockbox
04:01:51gevaertsfunman: the comment wasn't updated apparently
04:02:13funmanif each program uses a specific filename for storing its informations, then there is little gain for compatibility with "software X album art database"
04:02:20Blue_DudeAll right, I test compiled for Sansa E280, Gigabeat F, and iRiver H120. I don't have the right tools in Cygwin to compile for Onda. I can't think if a thing that I can test on now.
04:02:36Topy44i still believe a parsing algorhyhtm should be considered, but we need to get some core devs discussing that seriously, instead of just dismissing it on the grounds that they dont give a **** :)
04:03:03gevaertsTopy44: stop spreading lies. Nobody dismissed it on those grounds
04:03:14funmanTopy44: discussion is welcome on any topic of course
04:03:21Topy44its ok, i dont want to start another discussion/fight right now :)
04:04:04Rudy4Pezlol "discussion/fight" :P
04:04:35gevaertsalso, don't expect to be taken seriously if the most detailed proposal you can come up with is "a parsing algorithm"
04:04:49funmanif you're not going to discuss about it, then don't mention it. trolls are welcome on #rockbox-community but not here
04:05:05*gevaerts disagrees with funman. Troills are also not welcome there
04:05:20Topy44again, i dont want to start over again now. we already had that discussion, including suggestions on how to implement it, earlier
04:05:32Topy44i'll shut up now :)
04:05:36funmanTopy44: can you point me to a date / irclogs?
04:05:40Topy44sure, sec
04:05:51*gevaerts read the discussions, and he does not remember any suggestions on how to implement this
04:06:39Topy44try yesterday midday GMT+1 (whatever timezone your logs use)
04:08:09gevaertsah yes, the "pick the first random file you see" algorithm
04:08:12Topy44well, there are two ways... either during playlist creation, check for every file added and save the filename in the playlist, or parse dircache on file load
04:08:16Topy44no, not first random file
04:08:26Topy44check for keywords, like album, front, cover, etc
04:08:39Topy44only if none work, use the first file
04:09:02gevaertshow does this interact with looking in parent directories?
04:09:15Topy44in most cases there will be only one file and that will be in some way relevant to the music... of course for people that use large directories with tons of files instead of sorting them by album that will fail
04:09:25funmanfind . -name $my_naming.jpg -exec mv {} cover.jpg \;
04:09:49gevaertswhereas the current system never fails
04:10:14Topy44yeah, but only gets it right if you pre-sorted the collection - didnt i just say i dont want to start this discussion again right now? i am kind of busy
04:10:29Topy44another time
04:10:31Rudy4PezThe current system fails in it's ability to include as wide a variety of filenames.
04:10:47Topy44Rudy4Pez: thanks, thats a good way of saying it
04:10:50funmanRudy4Pez: i think it's not really wanted
04:10:51Rudy4Pez(Wider than 7, at least, or whatever the current number is.)
04:11:06gevaertsRudy4Pez: I have seen exactly zero serious arguments for why that is a real problem
04:11:38Topy44it is by me, and i am sure many others. possibly one should ask other users on their opinion instead of dismissing it on the sole grounds that other devs THINK its not wanted
04:11:39Rudy4PezLast I checked, one of the perks of Rockbox was supposed to be it's versatility..
04:11:53gevaertsFor nsome reason, rockbox does not start playing if I press stop. I occasinally do that. Can't it start playing anyway?
04:12:00Rudy4PezThat's why it's offered on so many platforms.. That's why it supports so many formats.. Right?
04:12:16Topy44gevaerts: sure, just reconfigure your keys :)
04:12:17funmanTopy44: FYI, Folder.jpg is created by windows (WMP?)
04:12:17gevaertsTopy44: wanted or not wanted by users is totally irrelevant
04:12:31Topy44funman: yeah, someone said that previously
04:12:42Topy44gevaerts: is it? isnt rockbox made for, you know, users?
04:12:42Rudy4PezI view supporting more common album art filenames as an extension of that same versatillity.
04:13:01Rudy4PezTopy44: Exactly. I touched on that last night.
04:13:08gevaertsTopy44: ah, there we have the issue. No it's not. It's made for the people who make it
04:13:45Topy44that of course is a whole discussion on its own: do you implement something you dont need yourself, just because another non-dev user might want it? or is it irrelevant because you, the dev in question, doesnt need it
04:13:51Rudy4PezThe more people that Rockbox works for, the more people use it. The more people use it, the more people contribute to it.
04:14:02funmanTopy44: you could work on a patch implementing your idea, that would make your idea more likely to be read & discussed
04:14:26Topy44yes, unfortunately my dev skills arent that far yet. though its not impossible that i might work on that
04:14:33funmanand would help users needing this feature, should the patch be rejected by developers
04:14:45Topy44(FYI, i do a little code, but i havent gone further then tiny changes up to now)
04:15:29Topy44the whole problem with patches that wont get commited is that they need to be maintained all the time to work against current svn
04:16:40Rudy4PezExpansion and versatillity is extremely key to the success and advancement of a communal project such as this one.
04:17:08Rudy4PezThe more people to whom you can appeal, the more people will care. People caring about your project is generally a good thing.
04:17:14gevaertsRudy4Pez: do you have any idea what you're talking about, or are you just typing in random rubbish?
04:17:37Topy44i actually believe he is making a point
04:17:38Rudy4PezNah, I just made that all up. :P
04:17:43Topy44but... i'll shut up now. really. :)
04:17:48funmanTopy44: you can't know if it will be committed before writing it
04:17:56Mode"#rockbox +o funman " by ChanServ (ChanServ@services.)
04:18:12Topy44yeah, but is it right to ignore a suggestion by someone just because he isnt a coder?
04:18:29funmanyes, there is no coding on-demand
04:18:36Topy44seems so :)
04:18:46gevaertsTopy44: we are ignoring the suggestion just because we think it's a bad idea, not because of who made it
04:19:10kugelAnd I do because I just don't care :)
04:19:25Topy44meaby we should agree to disagree and see what other people think once it takes on a more serious form :)
04:19:48Rudy4Pez"Considering a suggestion" and "coding on-demand" are two very different things.
04:19:55*gevaerts also thinks that telling people to work on features they don't want in order to get hypothetical new contributors is probably one of the best ways to make projects fail
04:20:22funmanRudy4Pez: the best way to add a feature to an open source project is to send a patch
04:20:26Topy44its the real difference between open source and commercial software i guess - open source software is made for the people that made it, while commercial software is made for the customers :)
04:22:34 Quit Rudy4Pez ("back to work")
04:26:38Mode"#rockbox -o funman " by funman (n=fun@rockbox/developer/funman)
04:32:18 Quit Llorean (Read error: 104 (Connection reset by peer))
04:32:44 Join Llorean [0] (
04:32:50 Quit Llorean (Client Quit)
04:33:40 Quit kugel (Remote closed the connection)
04:36:22 Join pyro-pitbull [0] (
04:36:37 Part pyro-pitbull
04:49:19UnhelpfulBagder: people who can't svn can git-http :)
04:52:38 Quit Topy44 ()
04:58:12 Join ej0rge [0] (
05:16:25***Saving seen data "./dancer.seen"
05:24:55CIA-61New commit by blue_dude (r22394): FS #10199: Adds limiter DSP function
05:32:01 Join KBH [0] (
05:38:58JdGordonanyone around to test a patch?
05:39:46funmanBlue_Dude: did you see your commit breaks on sh CPUs (undef symbol) and on YH920 (IRAM section is full)
05:40:04Blue_DudeOh yeah. Working on it right now.
05:40:25funmannot sure what happens on yh920, the chessclock plugin shouldn't use IRAM afaiu
05:40:35Blue_DudeThe plugins are going to be pretty easy. I'm not sure yet what to do with the IRAM being full.
05:40:54JdGordongot a good builkd time though :)
05:41:33Blue_DudeOK, I fixed the plugins...
05:41:35funman"rockbox.elf" > I believe make -j* is used and the error messages are not directly following targets
05:42:12funmanI was about to suggest removing the lcd framebuffer from iram for yh920, but it isn't :/
05:43:05Blue_DudeI got Archos player to compile plugins successfully. It needed a swcodec IFDEF
05:45:23CIA-61New commit by blue_dude (r22395): Fix red: plugins on hwcodec targets
05:46:16funmanrecorder/recorderv2 etc are not hwcodec (only the player)
05:47:55Unhelpfulthey're not? i thought all sh1 were...
05:48:01Blue_DudeWell, they choked on the same item. We'll see if it works.
05:48:46Blue_DudeThe bin delta's not as bad as I thought.
05:48:53JdGordonfunman: they are
05:49:01JdGordonall archos/sh1 targets are hwcodec
05:50:04 Quit HBK (Read error: 113 (No route to host))
05:51:06Blue_DudeHm. Just the one target chokes on IRAM. Suggestions?
05:52:09JdGordondisable it on that target?
05:52:23Blue_DudeI'm compiling a local version to work with. I can start removing IRAM tags until it behaves but it's going to affect performance on everything else.
05:52:24JdGordonor on all targets with < X IRAM?
05:52:40funmansorry, i'm mistaken with lcd_charcell
05:53:05funmanBlue_Dude: i'm looking at it (the yh920)
05:53:28Blue_DudeWhat would be an appropriate IFDEF to carve out that target?
05:54:09JdGordoncoolest data abort ever!!!!
05:54:20JdGordonall I see is Da with some pixels
05:54:25JdGordonits written in the last viewport!
05:55:57CIA-61New commit by jdgordon (r22396): it is mildly useful to reset the viewport to the full screen for the panic screen....
05:57:40JdGordonwtf? that didnt fix it?
06:00:16funmanIRAM_LCDFRAMEBUFFER is only respected in 16 bits lcd driver
06:01:45funmanBlue_Dude: a fix would be to define IRAM_LCDFRAMEBUFFER to IBSS_ATTR for *all* targets using drivers/{lcd-2bit*,lcd-1bit*} (in their respective config-*.h), but for the YH920
06:02:12funmanthat would free enough iram (the yh920 build will link with 0x100 more iram bytes)
06:02:28Blue_DudeWhoa. OK, where do I go to fix that?
06:02:43funmanok let me look at it ^^
06:03:06Blue_DudeI haven't dug into the config files at all. Hardware scares me. :)
06:05:45JdGordonBlue_Dude: is this limiter on by default?
06:06:43Blue_DudeWhen settings tries to read in a new value, it should fill with a 0, which is off.
06:07:25Blue_DudeYou have to go to the menu and dial in a value greater than 0 for it to enter the signal chain.
06:08:09 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:08:41Blue_DudeThat's a change I made in the last week or so. Before there was still a residual process in the chain, but now nothing happens with it unless you turn it on.
06:10:08CIA-61New commit by funman (r22397): LCD_DEPTH == 1 : specify if lcd framebuffer is in iram into the model specific config file
06:12:25Blue_Dudefunman: will that fix my problem or is this something else?
06:12:32funmannext commits will
06:12:45Blue_Dudeok, great
06:12:50funmanyh920 is 2bpp
06:16:04JdGordonwhy am i getting a crash on target but not in the sim? :'(
06:16:21Blue_DudeWhat is crashing?
06:17:15JdGordonwps display
06:17:21JdGordoncompelty my code...
06:17:33JdGordonwow, its nice and foobar!
06:17:36Blue_DudeYou're scaring me, man.
06:18:13JdGordonand why the heck is AA not showing in the sim? Unhelpful?
06:18:24JdGordonthere isnt a seperate define is there?
06:23:55CIA-61New commit by funman (r22398): LCD_DEPTH == 2, vertical interleaved : specify if lcd framebuffer is in iram into the model specific config file
06:23:59CIA-61New commit by funman (r22399): LCD_DEPTH == 2, horizontal packing: specify if lcd framebuffer is in iram into the model specific config file
06:24:03DBUGSent KICK CIA-61 to server
06:24:03CIA-61New commit by funman (r22400): LCD_DEPTH == 2, vertical packing: specify if lcd framebuffer is in iram into the model specific config file
06:24:04Kick(#rockbox CIA-61 :Stop repeating yourself) by logbot!n=bjst@rockbox/bot/logbot
06:24:08 Join CIA-61 [0] (n=CIA@
06:24:14JdGordonbrute force it? :)
06:24:42funmandamned, the build system didn't take the 3 commits :/
06:24:50Mode"#rockbox +o funman " by ChanServ (ChanServ@services.)
06:25:03Mode"#rockbox +v CIA-61 " by funman (n=fun@rockbox/developer/funman)
06:25:05Mode"#rockbox -o funman " by funman (n=fun@rockbox/developer/funman)
06:26:39Blue_DudeBuild bot got itself booted. :)
06:29:18funmanhmm why are the commits built one by one ? :(
06:30:55JdGordonthey dont do lots anymore? damn
06:31:17Blue_DudeHey, it worked!
06:31:19funman22400 doesn't seem to be built
06:31:46JdGordonyou made it worse :)
06:31:59Blue_DudeCompiled r22400 on my local copy of yh920 and it went through.
06:32:05CIA-61New commit by funman (r22401): iaudiox5 has a 16 bits display, not a 2 bits. ...
06:32:13funmanBlue_Dude: right, but what about the other :o
06:32:22 Quit CaptainKwel (Remote closed the connection)
06:32:33JdGordonit might take a minute after each round to decide to start again?
06:32:44JdGordonor you killed it.. which means no more builds untill zagor wakes
06:32:52funman22401 is building
06:33:03JdGordonnot on my client
06:33:12Blue_DudeNo there it goes. Skipped r22400 in the build.
06:33:12JdGordonoh yes it is
06:33:14JdGordonsilly screen
06:33:37funmanoh fasty fast build system
06:33:41funmanthanks !!
06:35:51Blue_DudeI'm trying to build on a not-so-new laptop. I'm thankful for the fast build system on here.
06:36:43Blue_DudeThis is like whack-a-mole.
06:36:45CIA-61New commit by funman (r22402): h300 has a 16 bits display, not a 2 bits. ...
06:37:06funmanallright, now it will be all green
06:40:39Blue_DudeOh, pretty green color.
06:41:10Blue_DudeThanks, man!
06:41:24 Join timc [0] (n=aoeu@
06:42:59JdGordonwant to figure out my issue now? :)
06:43:28funman/quit !
06:43:46JdGordonoh good.. a data abort.. at least thats useufl
06:43:58funmanJdGordon: how reproducible is it?
06:44:09JdGordonI dont know what *it* is :p
06:44:53tmztusually unaligned access on arm or similar platforms
06:45:09JdGordonyeah, no i dont tihnk so
06:45:34tmztwhere is it occuring?
06:46:44JdGordonwhen it displays the wps
06:46:47JdGordonnot always though
06:47:42JdGordonit corrupt memory, or NULL pointer
06:47:44JdGordonalmost certainly
06:47:50tmztso rockbox is showing that error?
06:49:51JdGordondefinetly corrupt memory somewhere
06:50:00JdGordonI shoudl take some photos of the different ways this is showing up
06:53:26JdGordonsweet repro in the sim
06:53:35JdGordon[3]+ Floating point exception./rockboxui !
06:53:43Blue_DudeNight all. Thanks again for all the help.
06:54:30JdGordonoh you're joking! this might not even be my bug
06:54:48 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]")
06:55:05JdGordondiv0 error!
06:55:23 Join FOAD_ [0] (
06:55:59JdGordonno, it must be :(
06:57:39 Nick Guest23768 is now known as J-23 (
06:59:10 Quit ehntoo (Read error: 110 (Connection timed out))
07:00:36funmanJdGordon: can you run the sim under valgrind?
07:00:49JdGordoni havnt tried?
07:01:06funman"valgrind ./rockboxui"
07:01:14JdGordonwhats valgrind?
07:01:42funmana multi purpose program to help debugging. in this invocation it will report read/writes out of buffers
07:07:21 Quit FOAD (Read error: 110 (Connection timed out))
07:07:21 Nick FOAD_ is now known as FOAD (
07:13:54 Quit kkurbjunW (Remote closed the connection)
07:14:49 Join kkurbjunW [0] (n=karlk@
07:16:29***Saving seen data "./dancer.seen"
07:31:01CIA-61New commit by jdgordon (r22403): move viewports into the skin buffer, no more viewport limit on the skin
07:31:10*JdGordon really hopes that works!
07:32:35pixelma"really hopes that works" - what happened to good old testing?
07:34:12JdGordongot bored of testing :)
07:34:16JdGordonI mean... morning :)
07:34:30JdGordonna, i tested on my mini and im rpetty sure i got all the bugs
07:35:31JdGordonbugger you archos player! greens across the bopard except there
07:35:58JdGordonpixelma: that should just about even out the previous change's delta.. so you shouldnt have anything to worry about again
07:36:26JdGordonand incase it disappears.... 175s build!!! under 3min!
07:36:44JdGordonall glory to Zagor and gevaerts and rasher and whoever else should get credit
07:37:43pixelmaas if that would be the only thing to worry about ;) I'll yell at you in case my viewport heavy WPS(s) don't work anymore :P
07:38:06JdGordonany chance you could check in the next 90min or so :)
07:38:36pixelmano, probably only after work
07:39:05JdGordonok, so I at least get to sleep a bit before being yelled at :)
07:39:18JdGordonRELAX.... it works!
07:43:33 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:47:57 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
07:48:53 Quit funman ("free(random());")
08:05:18 Quit obo ("No Ping reply in 90 seconds.")
08:05:20 Join obo_ [0] (
08:10:00 Join Rob2222 [0] (
08:15:47 Join flydutch [0] (
08:17:22 Quit Rob2223 (Read error: 60 (Operation timed out))
08:19:22amiconnJdGordon: Sure I won't complain about such small deltas for the players, if it is an overall gain otherwise. While the player has the same amount of ram as the other archoses, it still has the smallest binary of all of them
08:19:39*amiconn doesn't like this attitude of JdGordon and kugel regarding testing
08:20:04JdGordonwhat attitude? :)
08:20:49JdGordonyou dont actually think I'd commit something without testing it first do you?!
08:21:08 Part toffe82
08:23:32amiconnI would not expect anyone doing that in the first place, but when reading your and kugel's statements here, that's not so clear anymore
08:24:10 Join ender` [0] (i=krneki@
08:25:22amiconnOf course it would be a waste to build all builds or even just all target builds before commit - that's something the build system is much better (and faster) at.
08:25:51JdGordonif you're talking about todays "discussion" then thats what we were saying
08:26:23amiconnWhat I do expect though is (1) code being tested and fixed so nothing is knowingly broken, and (2) builds being tested on a representative selection of the affected targets and sims
08:27:03JdGordonI doubt you'll get any arguments from anyone on that
08:27:18JdGordonmaybe from kugel... he doesnt build any targets before ci :)
08:27:22*JdGordon goes to bed
08:27:40amiconn[07:31:10] -*- JdGordon really hopes that works!
08:29:06JdGordon+ sarcasm...
08:29:41 Join Grahack [0] (
08:29:51 Join Lear [0] (i=chatzill@rockbox/developer/lear)
08:30:29LearJdGordon: My viewport wps stopped working...
08:31:55JdGordonthats not a funny joke...
08:32:10JdGordonwas it working before that commit?
08:32:30LearA build from yesterday worked at least...
08:32:51JdGordonzip it up and send it to me.... which target?
08:32:55JdGordonnot working how?
08:33:28LearNope. when I run in the sim, I get this: Invalid parameter list for token 63: "display preloaded image 0"
08:33:48JdGordonyou're trying to display an image which isnt loaded
08:33:55JdGordonthere is no %xl line for it
08:34:05JdGordon3rd dev with that issue now :/
08:34:11LearShould be, as it used to work...
08:34:25JdGordonthe parser was made more strict yesterday
08:36:00LearAnd the complaint is for an empty line, making it hard to find what image, if any...
08:36:24JdGordonyeah, the debug text is on my todo list
08:36:38JdGordonignore that image 0 bit
08:36:57JdGordonpastebin the .wps... ill give it a quick look
08:37:56LearNo need, I acutally had a reference to an image that didn't exist any more. I had moved it to the fixed background. :) Sorry. So the error message was only off by one.
08:38:51*JdGordon goes to bed... for real this time
08:39:10LearAnd I really need to go to work...
08:40:04LearBtw, checkwps doesn't build. No make rule for wps_parser.c...
08:40:55 Quit Lear ("ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]")
08:41:08 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:48:08 Join Thundercloud [0] (
08:56:31 Join mt [0] (n=MTee@rockbox/developer/mt)
08:59:07 Quit elcan (Remote closed the connection)
09:03:27 Join PaulJam [0] (
09:08:10 Join stripwax [0] (
09:12:51 Quit Stephen__ ("Leaving")
09:16:34***Saving seen data "./dancer.seen"
09:17:55CIA-61New commit by jdgordon (r22404): store the image label instead of a number so debug output is actually useful when %xd is used witout a coresponding %xl
09:22:41 Join petur [50] (n=petur@rockbox/developer/petur)
09:24:02 Quit stripwax ("")
09:28:59 Quit Grahack ("Leaving.")
09:31:13CIA-61New commit by nls (r22405): FS #10536 by David Kauffmann, make the manual point to the theme site for additional themes instead of the old wiki galleries.
09:53:22 Quit Thundercloud (Remote closed the connection)
10:01:15 Quit TruthTaco (
10:01:15 Quit courtc (
10:01:15 Quit Torne (
10:01:15 Quit kkurbjun (
10:01:15 Quit Kohlrabi (
10:01:15 Quit scorche|sh (
10:02:23NJoinKohlrabi [0] (
10:02:31 Join scorche|sh [0] (n=scorche@
10:02:40NJoinTorne [0] (
10:02:40NJoincourtc [0] (n=court@unaffiliated/courtc)
10:07:43 Join scorche|1h [0] (
10:07:44 Quit scorche|sh (Killed by (Nick collision))
10:07:49NJoinTruthTaco [0] (
10:07:49NJoinkkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
10:07:49NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
10:07:54 Quit scorche|sh (Success)
10:08:57 Quit PaulJam (".")
10:12:06DEBUGEOF from server (Connection reset by peer) (snapshot: netstuff.c line 545)
10:12:06***Saving seen data "./dancer.seen"
10:12:08***Started Dancer V4.16
10:12:08DEBUGconnect(2) failed on socket 3 (Connection refused) (snapshot: netstuff.c line 150)
10:12:08***Connected to on port 6667
10:12:08***Logfile for #rockbox started
10:12:13***Server message 501: 'logbot :Unknown MODE flag'
10:12:13Mode"logbot :+i" by logbot
10:12:13 Join logbot [0] (n=bjst@gateway/web/cgi-irc/
10:12:13 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
10:12:13 Join TruthTaco [0] (
10:12:13 Join courtc [0] (n=court@unaffiliated/courtc)
10:12:13 Join Kohlrabi [0] (
10:12:13 Join petur [50] (n=petur@rockbox/developer/petur)
10:12:13 Join n1s [0] (n=n1s@rockbox/developer/n1s)
10:12:13 Join ender` [0] (i=krneki@
10:12:13 Join flydutch [0] (
10:12:13 Join Rob2222 [0] (
10:12:13 Join obo_ [0] (n=obo@rockbox/developer/obo)
10:12:13 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
10:12:13 Join kkurbjunW [0] (n=karlk@
10:12:13 Join FOAD [0] (
10:12:13 Join timc [0] (n=aoeu@
10:12:13 Join +CIA-61 [0] (n=CIA@
10:12:13 Join KBH [0] (
10:12:13 Join ej0rge [0] (
10:12:13 Join moos [0] (i=mustapha@rockbox/staff/moos)
10:12:13 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
10:12:13 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
10:12:13 Join Rondom [0] (
10:12:13 Join J-23 [0] (
10:12:13 Join Zarggg [0] (
10:12:13 Join amiconn [0] (i=quassel@rockbox/developer/amiconn)
10:12:13 Join pixelma [0] (i=quassel@rockbox/staff/pixelma)
10:12:13 Join AndyI [0] (i=AndyI@
10:12:13 Join Kopfgeldjaeger [0] (
10:12:13 Join Tristan53 [0] (i=tristan@
10:12:13 Join kachna [0] (
10:12:13 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
10:12:13 Join Xerion [0] (
10:12:13 Join faemir [0] (n=faemir@
10:12:13 Join Lss [0] (
10:12:13 Join z35 [0] (
10:12:13 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-chelptnnsjieyteb)
10:12:13 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
10:12:13 Join intrados_ [0] (
10:12:13 Join alexbobp [0] (
10:12:13 Join antil33t [0] (n=Mudkips@
10:12:13 Join fyrestorm [0] (
10:12:13 Join SUSaiyan [0] (
10:12:13 Join thegeek_ [0] (
10:12:13 Join sinthetek [0] (
10:12:13 Join evilnick_home [0] (
10:12:13 Join suom1 [0] (
10:12:13 Join kadoban [0] (
10:12:13 Join lyngaas [0] (
10:12:13 Join Tuplanolla [0] (n=jani@unaffiliated/tuplanolla)
10:12:13 Join cg_ [0] (
10:12:13 Join AB3JU [0] (
10:12:13 Join Hadaka [0] (
10:12:13 Join jfc [0] (
10:12:13 Join shaggy-h [0] (
10:12:13 Join parafin [0] (
10:12:13 Join bubsy [0] (i=Bubsy@
10:12:13 Join bah_ [0] (
10:12:13 Join w1ll14m [0] (
10:12:13 Join Res1 [0] (
10:12:13 Join markun [50] (n=markun@rockbox/developer/markun)
10:12:13 Join jhulst [0] (
10:12:13 Join gtkspert [0] (
10:12:13 Join Dhraakellian [0] (
10:12:13 Join crwl [0] (
10:12:13 Join sbhsu [0] (
10:12:13 Join goffa_ [0] (n=goffa@
10:12:13 Join freqmod_qu [0] (
10:12:13 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
10:12:13 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
10:12:13 Join at0m [0] (
10:12:13 Join lostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
10:12:13 Join AfterDeath [0] (n=icxcnika@freenode/weird-exception/network-troll/afterdeath)
10:12:13 Join GodEater [0] (n=bibble@rockbox/staff/GodEater)
10:12:13 Join Jester [0] (n=Jester@unaffiliated/jester)
10:12:13 Join Ridayah [0] (
10:12:13 Join Shaid [0] (
10:12:13 Join AlexP_ [0] (n=alex@rockbox/staff/AlexP)
10:12:13 Join tarbo_ [0] (n=me@unaffiliated/tarbo)
10:12:13 Join mortti_ [0] (
10:12:13 Join killan [0] (
10:12:13 Join fred_2 [0] (
10:12:13 Join jordan`` [0] (i=gromit@
10:12:13 Join krazykit [0] (
10:12:13 Join shadearg [0] (
10:12:13 Join redfox [0] (
10:12:13 Join crashd [0] (
10:12:13 Join webmind [0] (
10:12:13 Join aidy [0] (
10:12:13 Join Overand [0] (
10:12:13 Join fxb__ [0] (
10:12:13 Join shodanX [0] (
10:12:13 Join r00s [0] (
10:12:13 Join jon-kha [0] (
10:12:13 Join Slasheri [0] (i=miipekk@rockbox/developer/Slasheri)
10:12:13 Join Zambezi [0] (
10:12:13 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
10:12:13 Join JohnTeddy [0] (i=unstable@tor/regular/sid)
10:12:13 Join Neovanglist [0] (i=Neovangl@
10:12:13 Join aevin [0] (i=eivindsy@unaffiliated/aevin)
10:12:13 Join jvd [0] (i=syscrash@poipu/developer/syscrash)
10:12:13 Join rwong [0] (
10:12:13 Join Guest66974 [0] (
10:12:13 Join rvvs89 [0] (n=ivo@pdpc/supporter/base/rvvs89)
10:12:13 Join rasher [50] (n=rasher@rockbox/developer/rasher)
10:12:13 Join bittin_ [0] (
10:12:13 Join Beta2K [0] (
10:12:13 Join linuxguy3 [0] (
10:12:13 Join maraz [0] (
10:12:13 Join blithe_ [0] (
10:12:13 Join preglow [0] (
10:12:13 Join rphillips [0] (n=rphillip@unaffiliated/rphillips)
10:12:13 Join gibbon_ [0] (
10:12:13 Join dionoea [0] (n=dionoea@videolan/developer/dionoea)
10:12:13 Join Galois [0] (
10:12:13 Join Erant [0] (
10:12:13 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
10:12:13 Join Necos [0] (
10:12:13 Join @ChanServ [0] (ChanServ@services.)
10:12:13 Join bzed [0] (
10:12:16 Join scorche|sh [0] (
10:12:17 Join Torne [0] (i=torne@
10:12:18 Join tmzt [0] (
10:12:18 Join Unhelpful [0] (
10:12:22 Join HellDragon [0] (
10:12:27 Join FlynDice [0] (n=FlynDice@
10:12:29 Join markdrago [0] (n=mdrago@
10:12:33 Join yosafbridge [0] (
10:12:37 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:12:44 Join mt [0] (n=MTee@rockbox/developer/mt)
10:13:04 Join YpsyZNC [0] (
10:13:14 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
10:13:40 Join avacore^ [0] (
10:15:34 Join robin0800 [0] (
10:17:12 Join meermanr [0] (
10:23:17 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
10:25:40 Join pamaury [0] (
10:29:46 Join bertrik [0] (
10:29:49 Quit timc (Connection timed out)
10:30:07CtcpVersion from freenode-connect!freenode@freenode/bot/connect
10:30:25***Server message 477: 'logbot #rockbox :[freenode-info] help freenode weed out clonebots, please register your IRC nick and auto-identify:'
10:37:35 Quit mt (Remote closed the connection)
10:37:44 Quit aevin (Read error: 131 (Connection reset by peer))
10:37:50 Join fml [0] (n=4fd3c348@gateway/web/cgi-irc/
10:39:07fmlI wonder whether we should have separate settings for the width of the scrollbar on the main display and on the remote. Position (Right/Left) may stay the same for both, but I doubt it for the width.
10:40:32fmlAnd one more thing: I'd change the meaning of "scrollbar_pos" so that false means "Left". Then they would apper as "Left, Right" (in this order) in the menu −− more natural IMO. Any objections?
11:10:42 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
11:14:44 Quit kkurbjunW (Remote closed the connection)
11:16:14 Join kkurbjunW [0] (n=karlk@
11:25:19 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
11:38:52 Join einhirn [0] (
11:41:54 Quit Lss (Read error: 104 (Connection reset by peer))
11:44:10*fml pings mcuelenaere because of FS #10538 and FS #10539 and quits
11:44:18 Quit fml ("CGI:IRC 0.5.9 (2006/06/06)")
11:54:34 Join Grahack [0] (
11:59:36 Join mt [0] (n=MTee@rockbox/developer/mt)
12:01:19 Join elcan [0] (
12:11:51 Join mt_ [0] (n=MTee@
12:12:09***Saving seen data "./dancer.seen"
12:31:54 Quit mt (Read error: 113 (No route to host))
12:46:22 Nick mt_ is now known as mt (n=MTee@
13:00:51 Join BlakeJohnson86 [0] (
13:14:43 Join daurnimator [0] (i=daurnima@freenode/staff/daurnimator)
13:24:14 Join FOAD_ [0] (
13:25:34 Quit FOAD (Read error: 60 (Operation timed out))
13:25:34 Nick FOAD_ is now known as FOAD (
13:30:49 Join ehntoo [0] (
13:47:50 Join dfkt_ [0] (
13:48:07 Join AndyIL [0] (i=AndyI@
13:48:52 Quit dfkt (Nick collision from services.)
13:48:56 Nick dfkt_ is now known as dfkt (
13:59:00UnhelpfulJdGordon: which sim? i've used the sim to test AA and PF...
14:01:37 Quit n1s ("Lämnar")
14:02:01 Quit AndyI (Read error: 110 (Connection timed out))
14:02:27 Quit ehntoo (Read error: 110 (Connection timed out))
14:03:44 Quit bertrik (Read error: 110 (Connection timed out))
14:12:11***Saving seen data "./dancer.seen"
14:12:28 Join hd [0] (
14:12:30 Quit hd (Read error: 54 (Connection reset by peer))
14:14:58 Join teru [0] (
14:15:20 Join n1s [0] (n=n1s@rockbox/developer/n1s)
14:20:23 Quit Torne ("reboot, damnit")
14:21:08 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
14:32:54CIA-61New commit by teru (r22406): Update japanese translation. some are not translated though.
14:38:49 Join aaron424 [0] (
14:46:31 Quit LambdaCalculus37 ("This computer has gone to sleep")
14:49:05 Quit moos ("reboot")
14:50:07 Join funman [0] (n=fun@rockbox/developer/funman)
14:51:05 Join jgarvey [0] (
14:53:10 Join moos [0] (i=mustapha@rockbox/staff/moos)
14:53:31 Join daurn [0] (i=daurnima@freenode/staff/daurnimator)
14:55:13 Join wagonbear [0] (
14:55:50wagonbearhello. i've just successfully installed rockbox on my ipod 4g greyscale. it's booted it up and the screen is now quite dark and the text is hard to read. the contrast and lcd mode settings are not really doing anything to help. maybe it's something to do with the theme i've loaded?
15:03:29 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
15:03:29 Quit antil33t (Read error: 104 (Connection reset by peer))
15:03:43 Join antil33t [0] (n=Mudkips@
15:10:50 Quit daurnimator (Read error: 110 (Connection timed out))
15:11:30 Quit LambdaCalculus37 ("Fwump")
15:11:58 Quit kkurbjunW (Remote closed the connection)
15:12:21 Join kkurbjunW [0] (n=karlk@
15:12:57 Quit wagonbear ()
15:25:41 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:25:43 Join bertrik [0] (
15:26:23 Part Grahack
15:26:26kugelamiconn: we were talking about test compiling (to see if it gets reds), not about general testing.
15:33:20CIA-61New commit by lenzone10 (r22407): Updated italian translation.
15:40:10CIA-61New commit by teru (r22408): snake: Clear backdrop before displaying highscores and restore background, foreground and backdrop after displaying highscores. ...
15:42:09 Quit pamaury (Read error: 113 (No route to host))
15:44:28 Part LinusN
15:45:49CIA-61New commit by teru (r22409): snake2: improve redrawing after pause. wrap some long lines.
15:46:09 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.0.13/2009080315]")
15:56:23 Join Valaddiction [0] (
16:00:29 Quit Valaddiction ("byeee")
16:12:14***Saving seen data "./dancer.seen"
16:14:22 Join Torne [0] (
16:43:02 Join mt_ [0] (n=MTee@
16:44:54 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
16:46:26 Join LambdaCalculus37 [0] (n=LambdaCa@rockbox/staff/LambdaCalculus37)
16:55:47 Join Rob2223 [0] (
16:56:05 Join CaptainKwel [0] (i=2669ecc2@gateway/web/freenode/x-lcnwlcwacptdjdzi)
16:58:24 Quit funman ("free(random());")
16:59:05 Join pamaury [0] (
17:00:31 Quit mt (Read error: 113 (No route to host))
17:02:11 Quit Rob2222 (Read error: 60 (Operation timed out))
17:03:43 Quit teru ("Quit")
17:05:06 Join Omlet [0] (
17:08:22 Quit mt_ (Read error: 113 (No route to host))
17:11:52 Quit Omlet ()
17:12:54 Join froggyman [0] (
17:19:58 Join stoffel [0] (
17:24:19 Join Lss [0] (
17:32:27 Quit pamaury ("exit(*(int *)0 / 0);")
17:33:32 Join JdGordon_ [0] (i=460763c7@gateway/web/freenode/x-fjzueyypawizklla)
17:41:26 Quit robin0800 (Read error: 60 (Operation timed out))
17:45:19JdGordon_32bit aligned pointers always have the bottom 3 bits 0 right?
17:45:45bertrikbottom 2 bits
17:46:13JdGordon_there goes that idea then :/
17:46:20JdGordon_2 bits isnt enough to do anything useful with
17:46:55JdGordon_trying to figure out how to turn the tokens[1024] array in the wps struct to something more dynamic
17:47:27Torneyou could make exciting assumptions about the range of pointers and steal bits off the top too? :)
17:47:59JdGordon_that was my other idea.. I dont think that would work for the sim thouhg?
17:48:23Torneit would work on many platforms. :)
17:48:32Torne(i.e. no)
17:48:49JdGordon_doesnt virtual memory addresses throw all those assumtpions out the window?
17:48:54Tornenot entirely
17:48:56bertrikpointers could point anywhere within the 4GB space, on some ARMs IRAM or DRAM is mapped quite high for example
17:49:23Tornebertrik: yes, but they generally won't point to *either half* of the 4GB space, for this particular purpose, on any particular target, i mean
17:49:32JdGordon_I think I'll have to add another 4 bytes to the viewport struct which isnt devastating
17:49:33Tornetokens[] is not going to point to IRAM
17:50:03bertrikit was just an example
17:50:13TorneJdGordon_: probably.
17:50:36 Join mt [0] (n=MTee@rockbox/developer/mt)
17:50:41Torneit would be tricky to do it in a way that would actually work on every platform. :)
17:51:15JdGordon_it was such a bad idea I stopped myself asking that :)
17:51:44bertrikbut maybe the range of possible values for pointers is quite small
17:52:08Tornebertrik: not on all the sims, though
17:52:19JdGordon_I need to store a pointer and an index as cheaply as possible
17:52:26Tornewin32 is probably the smallest and that's 2GB
17:52:30bertrika pointer could be constructed as LARGE_FIXED_OFFSET + small variable integer
17:52:31JdGordon_the range for both would be quite smll
17:53:03Tornehm, i'm not sure how the sim does its ram, actually
17:53:06JdGordon_it needs to be "fast" though also because this code does get run a hell of a lot
17:53:21 Quit petur ("work->home")
17:53:34bertrikoh forget it, I don't really like it myself either
17:54:31 Quit bertrik ("De groeten")
17:54:38JdGordon_I'm thinking of having this token list as a linked list, each list item has a much smaller array inside it, say 32 tokens each
17:54:57JdGordon_maybe even smaller
17:56:14JdGordon_actually.. I could for viewports to have their first toekn at the start of a new list item to make it easier.. slightly more wasteful though
17:57:24JdGordon_keeping track of the last token would be fun though
17:59:06kugelaww, anti-aliased fonts are too nice
18:02:17kugelbtw, I was wondering why x, y, width, and height are ints. screen dimensions are shorts in some places
18:02:24kugel(of a struct viewport)
18:03:48JdGordon_you can change them if you want to get a tiny delta
18:04:42kugelit won't have a huge impact for the view static vps we have, but maybe for skins
18:05:28JdGordon_I tihnk it wouldnt be noticable....
18:06:41*kugel thinks we should add a few typedefs for those type of data (screen dimensions, current tick), for consistency
18:07:04JdGordon_typedefs are bad
18:07:58kugelfor structs/enums/unions, I agree. not so much for int based types that need to be respected
18:12:16***Saving seen data "./dancer.seen"
18:13:18 Join LambdaCalculus3_ [0] (n=rmenes@
18:13:36kugelheh, I just got a data abort where the text is written into a wps vp
18:14:23 Join Riku [0] (i=Lss@
18:14:33 Quit JdGordon_ (Ping timeout: 180 seconds)
18:18:06 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:18:34 Join JdGordon| [0] (n=Miranda@
18:19:20 Quit jgarvey ("Leaving")
18:21:41 Join pamaury [0] (
18:29:38 Join domonoky1 [0] (
18:30:17 Join Sajber^ [0] (
18:30:55 Quit Sajber^ (Read error: 54 (Connection reset by peer))
18:32:17 Quit Lss (Read error: 110 (Connection timed out))
18:46:46 Join Omlet [0] (
18:48:24 Quit domonoky (Read error: 110 (Connection timed out))
18:50:07 Join Strife89 [0] (
18:50:21 Quit antil33t (Read error: 104 (Connection reset by peer))
18:50:30 Join antil33t [0] (n=Mudkips@
18:58:49 Quit BlakeJohnson86 (Remote closed the connection)
18:59:35 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:07:18 Part kkurbjunW
19:08:20 Join AndyI [0] (i=AndyI@
19:08:53 Quit AndyIL (Read error: 104 (Connection reset by peer))
19:10:02 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:12:57 Join Lss [0] (
19:13:01 Join BryanJacobs [0] (n=bryanjac@
19:18:59 Join funman [0] (n=fun@rockbox/developer/funman)
19:31:54 Quit Riku (Read error: 110 (Connection timed out))
19:33:27 Quit kachna (Read error: 60 (Operation timed out))
19:36:53UnhelpfulJdGordon: maybe index1+index2 rather than pointer+index?
19:38:47 Quit LambdaCalculus3_ ("Fwump")
19:41:41Unhelpfulyour dynamic token list... instead of a pointer, and storing some other data in the bits that are unused due to alignment, you could (if the items pointed to are stored together) use an index, of whatever size you think will work, and have the rest of the bits left over.
19:41:48 Quit LambdaCalculus37 ("Leaving")
19:41:49 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:42:19 Join einhirn [0] (
19:42:45 Join mcuelenaere [0] (
19:42:53JdGordon|I'm still confused, but no, I tihnk i'll get around the problem by forcing viewports to start in new list items
19:43:09 Join trisiak [0] (
19:43:44JdGordon|I could just be making it worse than it need be, and just use a linked list for the whole lot of tokens instead
19:44:16 Quit einhirn (Client Quit)
19:44:30JdGordon|that would double the overall ram usage (for the tokens array), but the ram is allocated anyway and not static anymore so its proably a saving
19:45:11UnhelpfulJdGordon: what exactly is tokens an array *of*, at present?
19:45:42JdGordon|an array of wps_tokens
19:46:06JdGordon|which is a wopping 8 bytes i think
19:46:58Unhelpfuland you want a dynamic array of them?
19:47:30JdGordon|the needed size isnt known untill the end of parsing which makes things tough
19:48:06UnhelpfulJdGordon: and it's to be stored where... space chopped from the start of the audio buffer?
19:48:54Unhelpfulcould you grab an unreasonably large amount of space, and return what you don't need?
19:49:24JdGordon|I was going to say we could allocate tokens from the start, and everything else from the end and they could meet in the middle.. but that wont work for bitmaps
19:49:41JdGordon|no, that would fragment the buffer
19:49:47JdGordon|oh hang on!
19:49:50JdGordon|we could do that
19:49:58JdGordon|oh no.. we couldnt
19:50:08Unhelpfulso... we need buffer defrag? ;)
19:50:31kugelI like the blocks of 32 tokes idea, what's wrong with that?
19:50:49 Quit Omlet ()
19:51:42JdGordon|images are loaded at the end of parsing.. so we could do that maybe.. it would look something like this after a few skins are loaded... |tokens skin1|images skin1|tokens skin2|images skin2|−−−−−−−−−−−−|data skin2|data skin1|
19:51:58marazI like the 32 tokes idea, what could possibly be wrong with that?
19:52:05JdGordon|nice and confusing.. but we shouldnt need to debug it so it could work
19:52:31kugel:/ that seems complicated and error prone
19:52:32UnhelpfulJdGordon: why would it need to be a linked list of blocks of tokens? can't you allocate the blocks consecutively?
19:53:13JdGordon|no, we cant because it will alloc a block of tokens, then parse untill that block is full which we assume means more allocation during that time
19:53:35JdGordon|right now allocation always happens from the start
19:54:16BryanJacobswouldn't it be nice if there were some kind of, I dunno, dynamic memory allocator for this sort of thing?
19:54:32JdGordon|that still wouldnt help though
19:54:42JdGordon|the problem is not knowing how big to allocate
19:55:02BryanJacobs... isn't that what you DON'T have to do with dynamic allocations?
19:55:14BryanJacobsrealloc() exists for that reason
19:55:44 Join Sajber^ [0] (
19:55:59JdGordon|realloc means changing everything to use offsets instead of pointers, which is messy
19:56:25JdGordon|I tinhk that scheme I mentioned above is actually the way to go
19:57:04 Quit Sajber^ (Read error: 104 (Connection reset by peer))
19:57:11kugelcan you elaborate why the other ways don't work?
19:57:34JdGordon|the other way being blocks of token arrays? it will work, its just more messy
19:58:05kugel"|tokens skin1|images skin1|tokens skin2|images skin2|−−−−−−−−−−−−|data skin2|data skin1|" looks way more messy
19:58:21JdGordon|viewports/lines/sublines keep track of indexes into the tokens array for where they start/stop...
19:58:23kugelcomplicating detection of "out of buffer" also
19:58:35JdGordon|true, but overall simpler i tinhk
19:58:54JdGordon|parsing only happens once, token traversing happens many times per second
19:59:39JdGordon|and remember, once its in the buffer its static, so as long as its all allocated correctly there shuold be nothing dangerous happening
19:59:56JdGordon|as long as everything plays by the allocation rules
20:00:21kugelyou're still in the duty of explaining why loading the statusbar after the wps doesn't work
20:00:38JdGordon|ha? what?
20:00:39kugelthat looks to me like there's still bugs in there, and you're going to complicate it more
20:01:04JdGordon|I have no idea what you're talking about
20:01:31kugeldon't you remember? when I loaded the statusbar first, the images weren't after the wps is loaded
20:01:43kugel"the order matters", which I don't understand
20:02:18BryanJacobsis there some particular reason why WPSs aren't XML?
20:02:25JdGordon|you didnt show me the diff with that issue... I'm 99% sure that was pebkac :)
20:02:35JdGordon|BryanJacobs: yes! XML is NOT the answer to everything
20:02:46kugeljust swap wps and statusbar loading in settings.c...
20:02:48JdGordon|oh lord.. adding a xml parser would be HUGE
20:03:07BryanJacobsit doesn't have to be a standards-compliant one... it's just a state machine...
20:03:30 Quit GeekShadow ("The cake is a lie !")
20:03:35JdGordon|kugel: if there was an issue remote targets wouldnt work...
20:03:47BryanJacobsOK, technically an XML parser is a pushdown automaton, no?
20:03:51JdGordon|we should go css :)
20:04:04BryanJacobsnow that's complicated
20:04:10*JdGordon| doesnt know the intricacies of xml parsers
20:04:30BryanJacobsI'm pretty sure they only need a stack, not some high-powered computation engine
20:04:56BryanJacobsand they have maybe a dozen states
20:05:43JdGordon|the parser isnt the issue... its the allocation for things like images and viewports
20:06:13LearHrm, resurrecting deprecated strings is naughty... :)
20:09:48JdGordon|kugel: the only possible way that your statusbar thing breaks is if somehow you are resetting the buffer between skins... it works with remote targets therefore the probablem has to be in your change
20:09:54amiconnLear: Originally it wasn't, but things probably changed when target dependent language generation was added
20:11:23kugelJdGordon|: settings.c is the only place where it's resetted? then I didn't do that wrong (reset before any is loaded)
20:11:58 Quit dfkt (Read error: 110 (Connection timed out))
20:12:10Learamiconn: If the old genlang treated an empty translation as no translation, then yes...
20:12:18***Saving seen data "./dancer.seen"
20:12:24JdGordon|I'll have another look tonight if i get a chance
20:14:25 Quit bzed (Remote closed the connection)
20:14:28 Join bzed [0] (
20:15:27 Join Omlet [0] (
20:15:37 Quit Omlet (Client Quit)
20:15:43 Join Grahack [0] (
20:15:51LearNot a new thing, but currently the top three entries in the "Theme Settings" menu are "Browse Theme Files", "Font" and "While Playing Screen". A bit inconsistent...
20:16:02 Join Thundercloud [0] (
20:16:24amiconnLear: I was thinking about the potential id shift problem, not empty translations
20:16:28pixelmathe rename to only font and only WPS is quite new
20:16:41kugelteru really fixed AAF for mpegplayer
20:16:46amiconnThe latter is fixable just by killing the translation from all non-english .lang files
20:16:51kugelany reason not to commit it soon then (anti-aliased fonts)?
20:17:22mcuelenaereLear, amiconn: regarding resurrecting deprecated strings; I have a problem where they show up as empty strings. Any ideas about that? Should I re-deprecate them and add them under an other name?
20:17:23pixelmaIMO it adds nothing than binsize
20:17:47 Quit suom1 (Remote closed the connection)
20:17:50Learamiconn: Yes, but it wasn't done, so I got empty strings for one setting. Not very nice. :)
20:17:51 Join suom1 [0] (
20:18:16kugelLear: that was done on purpose
20:18:58pixelmaempty strings on purpose?
20:19:07kugelbecause wps and font can be preselected, so they behave like real settings, hence the "browse" was thrown away
20:19:38 Join kachna [0] (
20:33:33CIA-61New commit by mcuelenaere (r22410): Fix empty LANG_LEFT & LANG_RIGHT strings.
20:33:52JdGordon|kugel: how does AAF handle non AA fonts...?
20:34:09kugelas before
20:34:21kugeldifferent drawing functions, based on the depth
20:34:52 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
20:35:25JdGordon|so all exsisting fonts will keep working?
20:37:50 Join Zagor [242] (
20:38:39 Quit funman ("free(random());")
20:39:44kugelwtf, my custom sb doesn't show up on target
20:41:27kugel:( skin RAM usage doesn't increase
20:43:02 Quit scorche (Read error: 60 (Operation timed out))
20:48:20JdGordon|which target?
20:49:21CIA-61New commit by learman (r22411): Update Swedish translation.
20:51:56kugelomg logf is empty
20:52:09JdGordon|is it actually loading bmps?
20:52:17JdGordon|its unlelky to change noticably if you arnt
20:52:31rashermcuelenaere: err, did you just break old languages?
20:52:47JdGordon|no, I mean is your .sb doing any %xl's?
20:53:36Learrasher: fixed them, I'd say.
20:54:26rasherHm right, he didn't touch english.lang
20:54:42rasherSo it was just unnecessary
20:55:19Zagorhmm, even the build-id dirs hamper ccache since so many files are shared between builds
20:55:32Learrasher: No. Tried changing scroll bar position using an "unfixed" language?
20:55:44kugelZagor: bluebrother proposed build-<target>
20:55:58Zagorkugel: yes, that's what is running now
20:56:19kugeloh, I didn't that id == target
20:56:23kugeldidn't know
20:56:39 Quit froggyman ("ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]")
20:56:52Zagorit means ipodvideo and ipodcolor will never share cache, which makes new client take longer to get up to speed
20:57:41Zagora quick test shows half of their .o files are identical
20:57:50rasherLear: so someone deprecated LANG_LEFT/RIGHT and then someone else started using it again?
20:57:53evilnickgevaerts has landed, fact fans
20:58:04Zagorand I assume the same is true for many other targets
20:58:07Learrasher: Exactly.
20:58:09evilnickoops, sorry :)
20:59:45rasherLear: Huh, but they're deprecated in English as well
20:59:51rasherI can't make sense of this
21:00:01Zagorrasher: can you add a line saying how many builds were done as scheduled and not?
21:00:25Learrasher: You must look at the wrong version...
21:00:41rasherLear: Yeah..
21:01:24rasherZagor: should be possible
21:01:40rasherZagor: Just count total builds vs dequeued events?
21:02:14Zagorrasher: yeah, that ought to be correct. dequeued and abandoned
21:02:42rasherDoesn't abandon generate a dequeue event?
21:03:25rasherThat makes it much more complicated
21:03:48rasherI'd have to check for abandoned builds whether they were queued builds or not
21:04:08Zagoronly queued builds are ever abandoned
21:04:27Zagoruh, sorry no
21:05:28rasherCan we cheat and only count dequeues vs total? Abandons are rare
21:05:34Zagorall builds the disconnected clients had queued or was running will be marked abandoned
21:05:47Zagorrasher: sure, it's certainly a good start
21:11:03domonoky1is there a link to rashers build-stats page somewhere ?
21:11:21 Quit mt (Read error: 104 (Connection reset by peer))
21:15:18pixelmaforum question: I'd like to move this thread but where should I put it? It's likely a file view problem, so I'd suggest the UI subforum though I'm not sure. It definitely doesn't belong into the RbUtility forum...
21:15:43domonoky1nice :-) we should link to that from the current build page.
21:19:34domonoky1looks like ~20 clients didnt complete any build in the last few build rounds :-)
21:20:38Zagordomonoky1: yes, they performed slower than average. that means other clients overtake them and "steal" their builds
21:22:31 Quit flydutch ("/* empty */")
21:23:59kugelinteresting, it craps out at "if (!wps_data || !buf)"
21:24:28kugelbut wps_data isn't 0 (I dereference it jsut before) and buf isn't 0 (I splash it)
21:26:37JdGordon|what do you mean dereference before that? you actually check its not NULL?
21:27:34 Join hd [0] (
21:27:38 Quit hd (Read error: 104 (Connection reset by peer))
21:29:06kugeldata is null apparently
21:29:24kugelJdGordon: I do "if (wps_data->debug)" just before the splash
21:29:52kugelit seems 0, but it's static struct :/
21:29:59Zagorkugel: that can be true even if wps_data is 0...
21:30:30kugelnormall dereferncing NULL pointers gives a crash
21:30:52Zagoron memory protected system, yes
21:31:05kugelmaybe I remember wrong then
21:32:37JdGordon|do some splashf()-ing
21:34:41kugelawesome, I found it
21:35:05kugelI called sb_skin_init() only in the sim-version of main() ;)
21:37:05kugeland I found the problem why loading sb before wps fails also
21:37:18kugel(yes it was my failure :p )
21:38:02 Quit Dhraakellian (Read error: 60 (Operation timed out))
21:38:57JdGordon|kugel: thats why you should use app_main() and not main()
21:39:08 Quit pixelma (Nick collision from services.)
21:39:08 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
21:39:19 Quit amiconn (Nick collision from services.)
21:39:20 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
21:39:21JdGordon|so, am I now allowed to try that complicated allocation scheme?
21:39:26 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
21:39:26kugelit needs to be before I just put it after gui_syncwps_init
21:39:38 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
21:39:41kugelboth need to be before settings_apply(), which needs to be done early
21:39:50kugelJdGordon|: no
21:39:56kugelI don't like that system, really
21:40:33JdGordon|oh well :) good thing i was being rhetorical
21:41:01 Join sbeh [0] (
21:42:37kugelkeep it simple :( mem allocation is always error prone itself already, complicating even more sounds horrible for me
21:42:48 Join Dhraakellian [0] (
21:43:56sbehhi, yesterday I load new mp3s on my sansa e200, it was ready so I shutted down my ubuntu, but on the display of the sansa stood, that it's stil writing (I think some caching), but the pcs turn just off (why does it do that, I can't unmount the mp3-player while the kernel still flushes to cache?!?!!?!!) nevertheless, the sansa rebootet and the rockbox-bootloader shows 'no partition found' now, do I realy need to unbrick it?
21:53:53*kugel thinks this looks awesome
21:54:54kugelmeh, I'm going to need to kill the skin's default viewport
21:55:21 Quit stoffel (Remote closed the connection)
21:55:37JdGordon|kugel: that will completly break when playback stops thoughg :D
21:56:07*JdGordon| would get the list vp down and right a bit.. but yes, does look good
21:58:51 Quit Strife89 ("Leaving")
22:04:11kugelJdGordon|: why would it break?
22:10:09kugelskin_redraw with WPS_REFRESH_ALL refreshs the default viewport too, which sort of destroys the list
22:10:43 Quit Zarggg ()
22:11:21kugelwe really need to init the id3 and nextid3 structs (fill with sane defaults) for this statusbar to work
22:11:27JdGordon|kugel: because alot of those tags will simple crash if data->id3 == NULL
22:11:41JdGordon|no, we need to make those tags safe
22:12:19kugelmy idea would be simpler :p
22:12:22***Saving seen data "./dancer.seen"
22:12:43JdGordon|simpler != betterer
22:13:28kugelmore complex != the bestest, too ;)
22:14:42JdGordon|well of course not...
22:14:56*JdGordon| 's ideas is the bestest by definition!
22:17:13 Quit Lear ("ChatZilla 0.9.85 [Firefox 3.5.2/20090729225027]")
22:20:43kugelgrrr, this viewport is killing everything :(
22:21:18JdGordon|the fix is to make sure the skin doesnt draw in the list viewport, or not use the full display for the sb skin
22:22:32pixelmaconditional viewport based on the playback status
22:23:07JdGordon|not playback status.... skin/wps "mode" is the way my version did it
22:25:19kugelhow can I force not using the full display in the sb skin?
22:25:38kugelI don't have anyhting outside of my %V
22:26:09 Join jgarvey [0] (
22:26:25JdGordon|you should have a look at how i handled it in my patch...
22:26:33JdGordon|it *might* even still be in sync a bit maybe
22:26:56kugelthe one you posted to custom ui vp?
22:28:07kugeljust never doing WPS_REFRESH_ALL isn't an option I guess
22:28:31JdGordon|you need to add some trickery
22:28:41*JdGordon| suggests shelving sb skinning for a while still
22:28:44JdGordon|we arnt quite ready for it
22:29:41kugelI think we are, it just needs more of the alloc rework
22:31:12JdGordon|hmm... if WPS_REFRESH_ALL only erases the default vp then the statusbar one could set its default to some known area
22:31:18JdGordon|getting around the issue
22:31:48 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
22:31:49JdGordon|of course that is only when the sb is in one rectangle outside the list vp
22:34:40JdGordon|for the sb skin... I tihnk there should be 2 possibilities (both always avilaable)... 1) seperate .sb which is forced into a specified vp, 2) an extension of the .wps which specifies where the list vp is and is careful to not draw in it if its being shown
22:34:50kugelwe could detect if the default viewport is used, and deactivate it entirely if it isn't
22:34:54JdGordon|2 takes precedence over 1
22:35:37 Quit Grahack ("Leaving.")
22:36:00kugeli.e. your first suggestion
22:36:46JdGordon|I think having both as options is the way to go.. i dont really know how to do it properly though
22:37:13JdGordon|I mean, when music is playing it makes sense for a full blown skined bar.. but when there is no music it should be dummed down
22:37:14kugelI don't like 2)
22:37:46kugelthat breaks the interchangability of wps without changing the theme a bit
22:38:03JdGordon|not really..
22:38:19JdGordon|I tinh this holw idea that wps are seperate from themes is rediculous anyway
22:38:32kugelthe .wps doesn't need to care anyway, skin_redraw does
22:39:11kugelI may just add a flag WPS_REFRESH_ALMOST_ALL
22:39:58JdGordon|please dont!
22:41:26JdGordon|just dont call it that
22:41:38 Join Omlet [0] (
22:41:40 Quit BryanJacobs ("Java user signed off")
22:41:48JdGordon|it shouldnt be needed anyway
22:43:27*JdGordon| maanaged to get it all redrawing fine without adding any new redraw flags...
22:43:34JdGordon|you dont want to be outdone by me now do you?!
22:44:23kugelhmm, another problem. if in the wps, who's getting wps_state.do_full_update first? :(
22:44:53kugelJdGordon|: I saw your mode, but I a new flag is much easier and cheaper (I think)
22:45:01JdGordon|they shouldnt both be enabled together
22:45:17JdGordon|umm... hmm...
22:45:39kugelI think that should be possible
22:46:07JdGordon|nup, i dont think so
22:46:22JdGordon|maybe scrapping 2) abovev would good
22:46:22 Nick w1ll14m is now known as Guest50751 (
22:46:28JdGordon|just for simplicity sake
22:47:30kugelthere's nothing complicated about it I think
22:49:05*kugel tries to go with the detecting default vp usage and deactivating
22:51:23kugelwhat is the label of norma %V viewports?
22:52:52kugelhow can you differentiate between all the viewports?
22:53:06JdGordon|all which?
22:53:22JdGordon|only conditional viewpoerrts and the defualt have identifiable labels
22:54:03kugelso I can't find the first defined viewport, basically?
22:55:25JdGordon|you can
22:55:40JdGordon|vp_default_label is #defined to '|' iirc
22:55:53JdGordon|which is illegal for others (enforced by the parser)
22:56:32 Join Zarggg [0] (
22:57:48kugelmeh, I think I'll go with setting the default viewport to the first defined one, and tell the user "you might get a blank a screen once in a while if you don't use viewports at all, so use them"
22:58:03kugelnot using any doesn't really make sense anyway if one wants to see the list
23:02:15sbeh'no partition table' solved by 'touch format.fmt' :(
23:05:51 Join truthtaco_ [0] (
23:06:47Bagder"Guess I'm new around here anyways I'm porting Android OS to my Ipod Nano 1st Gen and My wii I might need some extra help"
23:06:53Bagdernow that's a great quote
23:06:56 Quit sbeh ()
23:07:09 Join ehntoo [0] (n=ehntoo@
23:07:13JdGordon|if only we had some place to group all those quotes...
23:07:34JdGordon|using a precious metal as an adgective :)
23:07:39JdGordon|adjective even
23:07:49*JdGordon| gives up...
23:14:14 Join LambdaCalculus37 [0] (i=41585845@rockbox/staff/LambdaCalculus37)
23:21:12 Quit TruthTaco (Read error: 110 (Connection timed out))
23:21:45 Quit Omlet ()
23:22:08 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
23:23:26 Join BlakeJohnson86 [0] (
23:24:00 Quit LambdaCalculus37 ()
23:24:29 Join Stephen__ [0] (
23:29:11 Join LambdaCalculus37 [0] (i=41585845@rockbox/staff/LambdaCalculus37)
23:32:38kugelJdGordon|: I just copy the second viewport into the default one
23:32:49JdGordon|what what whaaat?
23:33:25JdGordon|how on earth is that suipposed to be correct?
23:33:30JdGordon|(no offence)
23:33:50kugelwhy should it be wrong?
23:34:28JdGordon|what was it for?
23:34:36kugelI really don't want the skin engine to know which skin it is drawing for, that seems worse
23:34:57kugelto kill the default viewport
23:35:14JdGordon|which is why I tinhk one skin per screen at any time is the way to go
23:35:24JdGordon|why are you kiling the default viewport?
23:35:32JdGordon|just set it correctly
23:35:39kugelwell, I do that with the copy
23:35:54kugelI don't kill him, I just give him the values of the second one
23:35:57JdGordon|and the copy comes from where?
23:36:13kugelnot having a second viewport won't work anyway
23:37:01kugelthat's basically your first suggestion ("seperate .sb which is forced into a specified vp") without limiting the viewports further
23:37:48 Quit evilnick ("Page closed")
23:37:57JdGordon|I am so incredibly confused
23:39:12kugelthe default one is the implicit skin vp, which is fullscreen. when this one is redrawn it "kills" the list (and the other ui stuff, until that stuff is redrawn). I now copy the value of the second viewport (the first defined one), so that it is not fullscreen
23:39:36kugelit has at least the size of the statusbar then, but the statusbar may contain more viewports
23:41:44JdGordon|there is nothing wrong with doing a full screen clear *once*.. in fact its probably a good idea...
23:41:48domonoky1kugel: and if you dont have any extra viewports defined ? (think of very simple skins, or old ones)
23:42:18kugeldomonoky1: fail to parse it, i.e. don't support statusbars without at least 1 viewport
23:42:47kugelJdGordon|: what do you need a fullscreen clear for? clearing the whole statusbar is absolutely enough
23:42:47JdGordon|then the parser has to know what its loading.. which we havnt decide if we want
23:43:05JdGordon|when going from wps->menu...
23:43:10kugelthe wrapper does that. after skin_data_load()
23:43:26kugelthe list does that already...
23:43:35mcuelenaereis there some kind of ACTION_MENUITEM_CHANGED for callbacks in apps/menus/* ?
23:43:40JdGordon|no it doesnt... the list only clears the list vp... doesn ti?
23:43:59kugelthe whole screen
23:44:03JdGordon|mcuelenaere: yeah, each menu has a callback for each button press
23:44:10JdGordon|s/has/can have/
23:44:26kugelthat's the GUI_EVENT_REFRESH I committed with custom ui vp
23:44:46mcuelenaereJdGordon|: ok, and what action indicates that an other item is selected?
23:45:14kugelmcuelenaere: I guess you want to stop "live updating" of the scrollbar?
23:45:26mcuelenaerekugel: no, I want it to be correct
23:45:41mcuelenaere(see FS #10541)
23:46:00mcuelenaereor go from width 24 to 3 (or vice versa)
23:46:02 Join roolku [0] (
23:46:17kugelmcuelenaere: then you probably want to add a callback to it. I think that's called before redrawing and actually changing the setting (the new value is passed as parameter)
23:46:27kugelin settings_list.c directly
23:46:36JdGordon|mcuelenaere: really any action could be used... you could be smart and check the current selection and the previous if oyu want
23:46:56JdGordon|you cant force a menu redraw from the item callback
23:47:02 Part domonoky1
23:47:07mcuelenaerecan't you return ACTION_REDRAW?
23:47:15mcuelenaereor send a GUI_EVENT_REFRESH?
23:47:17JdGordon|from the menu callback yes
23:47:22JdGordon|the first is better
23:47:24mcuelenaerethat's what I meant
23:47:26JdGordon|if that works :)
23:47:33mcuelenaereuh, what first?
23:47:38 Quit ehntoo (Read error: 110 (Connection timed out))
23:47:40mcuelenaereah mine
23:47:55kugelI think you could set the setting in the callback (so that it is applied before redrawing)
23:47:57roolkuZagor: in case you are wondering - somebody is running an experiment on the cluster, that's why b??-roolku are slow
23:48:10JdGordon|umm... actually that might not work... you still need to menu to handle the item change, I dont eremember the action handing order
23:48:17Zagorroolku: ok. thanks.
23:48:24roolkuZagor: this is not the case for b22, b38, ro1, ro2, pc and pyros
23:48:31mcuelenaerebut would a GUI_EVENT_REFRESH work?
23:49:01kugelunless you pass a redraw function to it, which isn't quite doable with do_menu stuff
23:49:15roolkuZagor: so I am not sure why they don't complete. I did clear out ccache after your change to the build dirs last night though
23:49:27mcuelenaereI suppose I can't do return action | ACTION_RERAW; ?
23:49:28kugelalso, that is only active when you're using a custom ui vp
23:50:17JdGordon|mcuelenaere: iirc there is some special handling for certain settings when they return
23:50:20JdGordon|in menu.c
23:50:30 Quit bmbl ("Bye!")
23:52:33mcuelenaereJdGordon|: you're sure I can't return an ACTION_REDRAW? (line 404 in menu.c seems to handle this?)
23:52:49JdGordon|I dont remember the ordering
23:52:55JdGordon|and I cant look at the code right now
23:52:59JdGordon|try it and see
23:53:20JdGordon|im at work....
23:53:37JdGordon|if the origional action is passed to the list before the callback then it will work
23:53:55Zagorroolku: the server is also being unfair to them. there's a bug where some clients don't get any builds scheduled, which cause them to sometimes try too big ones first
23:54:09JdGordon|otherwise, if the menu callback has the lists, you could manually send the selection change action to it, then return ACTION_REDRAW
23:54:23kugelok, my way doesn't work
23:54:27Zagorroolku: and since you run so many clients, you are statistically more likely to get hit by this bug (which I'm looking at right now)
23:54:37 Quit LambdaCalculus37 (Ping timeout: 180 seconds)
23:54:52roolkuah okay. :)
23:55:02mcuelenaerehmm the callback doesn't get called when changing the selection..
23:55:35JdGordon|hmm... I guess that sholdnt be a surprise... umm...
23:55:43JdGordon|check the setting handling
23:56:09JdGordon|wait a sec... whats the issue?
23:56:13kugelmcuelenaere: the callback in settings_list.c is called upon changing the selection, but I can't see a way to get a redraw from there
23:56:37kugel(if you add one, that is)
23:57:13mcuelenaereJdGordon|: when changing the scrollbar width, the scrollbar gets drawed before the setting gets set (so you only see the previous scrollbar width, not the current)
23:57:41 Quit moos ("Rockbox rules the DAP world")
23:58:10JdGordon|when going back into the menu you mean?
23:58:16JdGordon|it updates live in the setting
23:58:26mcuelenaereyes, it updates live
23:58:32JdGordon|this is hard to test on the ipod vid... need a huge font to do it :)

Previous day | Next day