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 2012-09-30

00:05:36 Quit Rower85 (Read error: Connection reset by peer)
00:06:07 Join Gallomimia [0] (
00:09:26gevaertsSeems to be a themesite issue. The daily checkwps run doesn't work well...
00:13:18lebelliumhum ok
00:19:09kugelgevaerts: database locket?
00:26:45kugeli just saw that when browsing the sitre
00:28:02gevaertsWe're doing something wrong with database accesses, but I'm not sure yet what exactly
00:37:48*gevaerts tells the database to use WAL instead of the older locking mechanism
00:38:00 Quit Gallomimia (Quit: Gallomimia)
00:39:22 Join Prodicus [0] (
00:43:06 Join perrikwp_ [0] (
00:45:36 Quit perrikwp (Ping timeout: 252 seconds)
00:47:32 Quit perrikwp_ (Ping timeout: 240 seconds)
00:48:01 Join perrikwp [0] (
00:51:44kugelgevaerts: "DFKT Minimum X5 v3 INV Theme" says it "works with current build" on the theme overview (for x5) but "doesnt work with current build" on the theme details page
00:52:15gevaertsYes, that only *looks* like an inconsistency though :)
00:52:54gevaertsThe issue there is that on the overview the target is known, and on the detail page the first found target is used IIRC
00:53:10 Join Gallomimia [0] (~Gallo@
00:53:23kugelwait what? that makes no sense :)
00:53:36kugelthe url has the target string
00:55:19gevaertsI might be wrong, but as far as I remember that's not actually used to pick the right one
00:56:19gevaertsRight. The code seems to agree with that
01:03:09lebelliumI guess I already reported that here some months ago but I think mixing b&w and color themes for targets with the same screen resolution is a very bad idea. The color themes you can download in the H1x0 category (originally made for H10 20GB or X5) are just unusable on the H120 when auto-converted to grey-levels color
01:07:31gevaertsOK, the checkwps run is done, with a lot of handholding...
01:07:53gevaertsH10 and X5 now have the same themes
01:10:05 Join saratoga [0] (123e0cca@gateway/web/freenode/ip.
01:12:05gevaertslebellium: I think I agree, but I'm not sure how to even tell the difference from code...
01:12:34gevaertsBase it on the upload target?>
01:13:50saratogagevaerts: did you get a chance to look into approving new accounts on the forums?
01:15:03gevaertssaratoga: I haven't done anything since last time, really
01:15:33gevaertsI should be able to get to it soon though. The cold that made me a bit inactive seems to be going away
01:18:09saratogaah ok
01:18:47lebelliumgevaerts: Isn't it possible to add a b&w/color field to each target? I assume there is already a touch/non touch field to make the difference between the themes for Cowon D2 and those for iPod video for instance?
01:21:59gevaertslebellium: there isn't, actually
01:22:26gevaertsThe theme site knows about resolution and checkwps results
01:24:44lebelliumhum then why are there 105 themes for the iPod video but only 71 for the D2? I thought it was because only "touch-ready" themes display for the D2
01:24:57lebelliumthis theme website is a bit difficult to understand sometimes :S
01:26:01 Quit ender` (Quit: Our chances of being caught by the RIAA or IFPI are somewhat less than being hit by lightning - or choking on a wasabi-flavoured peanut. -- TheRegister)
01:26:25gevaertsI checked one of those. It had touch areas that checkwps for the D2 doesn't like
01:26:41gevaerts(and which are entirely skipped by checkwps for non-touch)
01:27:48 Quit Poodlemastah (Quit: ZNC -
01:29:04lebelliumok, makes sense now. But that's far from being optimal in my opinion.
01:29:38 Join Poodlemastah [0] (
01:31:15gevaertsI don't think adding a flag every time someone spots a feature difference is a good solution. Maybe we need a bit more information, but not without proper thinking
01:32:56gevaertsYou risk ending up with either unclear heuristics to deduce the flags based on the theme, or asking the uploader to supply lots of extra information that might be wrong
01:35:41lebelliumwell... actually among all current (and future?) targets, I only see 2 relevant infos needed: b&w or color display , and touch or non-touch device. If that info is linked to the target when it is added to the theme website, the guy who uploads a new theme has nothing to do but to select the appropriate target, no?
01:39:47gevaertsWell, first of all, there are monochrome, greyscale and colour targets. It's not as simple as monochrome or colour :) How are these handled? Do colour targets get all themes, or only colour themes?
01:40:51gevaertsSame for touchscreen, do touchscreen targets get all themes (remember, there's grid mode which some people even like!)? Do non-touchscreen targets get all themes?
01:41:01lebelliumthe question is: how many people install a monochrome/greyscale theme on a color target? I'm maybe wrong but I assume almost nobody? :)
01:41:12gevaertsWhy not? I don't know
01:41:57saratogacould we just have people check a "color" box when uploading themes?
01:43:03gevaertsWe could, or we could assume the theme matches the primary target they're uploading for
01:44:13saratogathat would work
01:49:36gevaertsI think touchscreen is a lot harder to do automatically. Simple touchscreen themes will work fine on a non-touchscreen target, but some of those fancy themes won't work well at all. On the other hand, a simple non-touchscreen theme can be made usable on touchscreen with just a few touch areas, while the author primarily focuses on non-touchscreen and will upload for those
01:52:15***Saving seen data "./dancer.seen"
01:56:20lebelliumthe thing is that you don't design a theme for a touchscreen and non-touchscreen device the same way. For a touchscreen device you need bigger icons (play/pause/prev/next etc) to touch. I personnally would not want such a theme on a non-touchscreen device. But probably some people don't care about that
02:04:08 Quit n1s (Quit: Ex-Chat)
02:14:24 Join Armitaage [0] (
02:16:09 Quit lebellium (Quit: ChatZilla 0.9.89 [Firefox 16.0/20120925201946])
02:18:10Armitaagehi there. i have a bit of a problem that i'm not sure if it is answered elsewhere. My Sansa Fuze is behaving really strangely, it refuses to charge and claims that the playlist control file is invalid. i have tried updating the bootloader and rockbox, but the problem persists
02:19:02 Quit perrikwp (Read error: Connection reset by peer)
02:19:06saratogai'd check the file system for errors, and see if there is a playlist control file you can delete
02:19:37Armitaagewhere should i look to do that?
02:20:00saratogarockbox folder i guess
02:21:01Armitaageoh, and one other problem that i noticed. i can't connect mp3 player to comp in rockbox mode, it crashes every time
02:23:19Armitaagesays "data abort at 30051578 FSR 0x8 adress 0xAC5E5DD4 pc:30051578 sp:300A2ED4 bt end"
02:24:55saratogaFS #12184 ?
02:24:57fs-bluebot Fuze V1 locking when transferring files Rockbox 3.9 (bugs, unconfirmed)
02:25:01saratogadepending which fuze you have
02:29:24 Quit Gallomimia (Quit: Gallomimia)
02:32:24 Join Keripo [0] (
02:32:34Armitaageyeah, looks like that is my issue, but i don't see the bug having been resolved. and what i see in the comments is mostly more reports of the bug, but no solutions.
02:35:36saratogayeah been hoping someone who has that problem would tell me which commit broke it for over a year now
02:39:37Armitaagei'm trying by reinstalling rockbox from the ground up, starting with the latest stable release
02:39:57saratogai wouldn't bother
02:40:09saratogaits probably just some file system corruption
02:40:18saratogathat playlist thing anyway
02:40:59Armitaagemy player is completely unusable, and i need to be able to use it, lol. it has all my music on the go
02:41:35saratogadid you check the file system for errors?
02:42:25Armitaagehow would i do that? sorry, not a dev, a user here
02:43:21saratogaassuming your'e on windows, my computer > properties of the disk drive, check for errors
02:43:29saratogaif you're on some other OS, check the documentation
02:48:18Armitaagehmmm...simply deleting the playlist control file seems to have fixed at least the resume playback function....but connecting to the pc is still icky
02:48:39saratogayeah, probably just a corrupted file
02:49:05Armitaagecomputer seems to not recognize the device
02:49:28 Quit Horscht (Quit: Verlassend)
02:50:36 Quit saratoga (Quit: Page closed)
02:53:05 Quit Keripo (Read error: Connection reset by peer)
02:55:12 Join Keripo [0] (
02:58:43 Join saratoga [0] (123e0cca@gateway/web/freenode/ip.
03:02:32 Quit mgottschlag (Ping timeout: 240 seconds)
03:05:51Armitaagewell, it's not really a fix, but i have a way to get to my music, charge, and connect to the comp. but it requires me to mess with things that i shouldn't have to
03:07:05Armitaageeverytime i want to change playlists i have to delete my playlist control file and restart the device, and it might take me several tries before i can access the player via my computer >.<
03:07:58Armitaagethat's pretty sucky, but it's all i've been able to do. thanks for the advice, though ^-^
03:09:29 Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
03:09:46 Join Totalled [0] (
03:14:55 Nick StopMakingSense is now known as PerfectlyNormalB (~fearofmus@unaffiliated/fearofmusic)
03:15:05 Nick PerfectlyNormalB is now known as LittleCreature (~fearofmus@unaffiliated/fearofmusic)
03:19:37 Join amayer [0] (
03:20:50saratogai would just use the sandisk USB mode if you have the problem in the FS task
03:22:27 Join factor [0] (~factor@
03:33:55 Quit Staphylo (Read error: Connection reset by peer)
03:34:10 Join Staphylo [0] (
03:37:42 Quit froggyman (*.net *.split)
03:52:09 Quit eckoit (Quit: eckoit)
03:52:18***Saving seen data "./dancer.seen"
03:53:21 Quit Provel_ (Read error: Connection reset by peer)
03:58:02 Join Provel_ [0] (
04:04:22 Quit Provel_ (Read error: Connection reset by peer)
04:04:44 Join Provel_ [0] (
04:14:46 Join Guinness` [0] (
04:14:46 Quit Guinness (Read error: Connection reset by peer)
04:14:49 Nick Guinness` is now known as Guinness (
04:15:24 Quit amiconn (Disconnected by services)
04:15:24 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:15:24 Quit pixelma (Disconnected by services)
04:15:25 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:15:27 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:15:30 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:16:11 Part amayer
04:22:31 Join froggyman [0] (~froggyman@unaffiliated/froggyman)
04:24:43 Quit factor (Read error: Connection reset by peer)
04:33:39 Join factor [0] (~factor@
04:36:51 Join perrikwp [0] (
04:38:12 Join TheSphinX^ [0] (
04:42:08 Quit TheSphinX_ (Ping timeout: 276 seconds)
04:50:27 Join Guinness` [0] (
04:53:28 Quit Guinness (Ping timeout: 246 seconds)
04:53:28 Nick Guinness` is now known as Guinness (
04:58:10 Quit perrikwp (Read error: Connection reset by peer)
04:59:23 Join perrikwp [0] (
05:01:44 Join Rower85 [0] (
05:21:42 Quit Epicanis (Quit: bedtime)
05:22:25 Quit TheSeven (Disconnected by services)
05:22:35 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
05:41:23saratoga[7]: does the Classic support 48k?
05:41:45saratogai see it defines it in the config file
05:41:57 Quit sciopath (Read error: Connection reset by peer)
05:42:17 Join sciopath [0] (
05:52:21***Saving seen data "./dancer.seen"
05:52:48 Quit zchs (Quit: Leaving)
05:56:13saratogai'm confused by the sample rate code for the ipods
05:57:01 Quit Rower85 (Ping timeout: 246 seconds)
05:57:19 Join Rower85 [0] (
05:57:46saratogawhats the difference between audiohw_set_frequency and pcm_set_frequency ?
05:59:20saratogain wm8975.c audiohw_set_frequency is a null op, but in cs42l55.c it actually changes the rate
06:03:54 Quit gxk (Ping timeout: 260 seconds)
06:10:52 Join gxk [0] (
06:12:15saratogaactually, does anyone have an ipod classic who can test a patch?
06:21:46saratogaif someone would install that build and then run the test_sampr plugin
06:21:50saratogaand see if 48k works
06:22:34 Quit Prodicus (Ping timeout: 246 seconds)
06:26:59saratogaby works i guess plays a tone that sounds right
06:31:47 Join JdGord [0] (~AndChat80@
06:39:56 Quit JdGord (Read error: Connection reset by peer)
06:45:18 Join JdGord [0] (~AndChat80@
06:47:49 Quit JdGord (Read error: Connection reset by peer)
06:59:23 Join JdGord [0] (~AndChat80@
07:00:41 Quit Armitaage (Quit: CGI:IRC (Ping timeout))
07:04:14 Quit JdGord (Read error: Connection reset by peer)
07:19:06 Join eckoit [0] (~ryan@
07:21:47 Quit prof_wolfff (Ping timeout: 245 seconds)
07:44:54 Quit Keripo (Read error: Connection reset by peer)
07:52:24***Saving seen data "./dancer.seen"
08:19:52 Join pretty_function [0] (~sigBART@
08:56:09 Quit pretty_function (Remote host closed the connection)
09:31:02 Join pretty_function [0] (~sigBART@
09:33:39 Quit amiconn (Read error: Connection reset by peer)
09:33:39 Quit pixelma (Remote host closed the connection)
09:34:55 Join pixelma [0] (pixelma@rockbox/staff/pixelma)
09:34:55 Join amiconn [0] (amiconn@rockbox/developer/amiconn)
09:40:41 Join pamaury [0] (~quassel@
09:40:41 Quit pamaury (Changing host)
09:40:41 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:52:28***Saving seen data "./dancer.seen"
09:55:57 Join stoffel [0] (
10:00:15 Quit pamaury (Ping timeout: 246 seconds)
10:27:55kugelJdGordon: ping
10:28:23 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:31:57[Saint]<lebellium> the thing is that you don't design a theme for a touchscreen and non-touchscreen device the same way.
10:32:03[Saint]I'm gonna disagree here.
10:32:15[Saint]Because, that's simply not true.
10:32:46[Saint]If we're talking about an "ideal world" then, maybe.
10:33:00[Saint]If we're talking about cabbiev2, then, certainly not.
10:33:06kugelhaving done themes for both I agree with him
10:33:51[Saint]There's no reason to do so, though, but I think we've established before that your german fingers are siimply too fat :)
10:34:15kugeli think he just meant that you approach differently depending on whether a theme is going to be touch-enabled or not
10:34:36[Saint]But, I don't...though. It can, and should, cater for both cases.
10:34:42kugelI'm not saying you cannot do themes that work equally well on both
10:34:58[Saint]there's no need to have two themes for a target that shares almost everything, except entry type.
10:35:10[Saint]ZYou can detect, and switch conditionally.
10:35:21[Saint]this is why we have a "is touchscreen?" check.
10:36:07kugelby switching conditionally you take a (perhaps only slighly) different design route
10:36:14[Saint]I used to make a habit of coding in touch regions for non-touch targets just in case the user might *one day* find a touch target to use it on.
10:36:42kugelhe said that the design differs, but both design can be in the same theme with a bit of work
10:38:34kugelthe point is: if you make one theme that's specifically for button-targets and one that's specifically for touchscreens, they will be different
10:38:59[Saint]I think I've proved otherwise there, and I stand by that.
10:39:00kugelhowever with compromises and/or hard work you can build both into the same theme
10:39:52[Saint]My 240x320 theme is *exactly* the same for touch/non-touch. Only the larger sizes differ, because they can, and have room to. They don't *have* to, though.
10:40:47[Saint]Anyway, the reason why I brought it up is that I don't think "touchscreen" is a valid flag to determine what target is allowed which theme.
10:41:43kugeleven non-touch-enabled themes work fine in grid mode so it's not unreasonable to show them
10:42:57[Saint]and things like bars automagically make their own touch areas now, so a "simple" theme that includes absolutely no touch regions will at least be partly usable without any user intervention in absolute point mode.
10:43:12[Saint]its a very blurred line.
10:44:00kugelthe automatic touch region for bars dont make it touch-enabled
10:44:08 Quit factor (Read error: Connection reset by peer)
10:44:24kugelit will change to back grid mode automatically
10:45:10kugeleven then, partuly usable would still be unusable if you cannot leave the wps
10:46:06 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
10:48:07 Quit XavierGr ()
11:01:05 Join factor [0] (
11:02:33 Quit pretty_function (Ping timeout: 260 seconds)
11:12:26 Join pretty_function [0] (~sigBART@
11:33:24 Join ender` [0] (
11:45:00 Quit bluebrother (Disconnected by services)
11:45:05 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
11:47:37 Quit fs-bluebot (Ping timeout: 245 seconds)
11:48:52 Join fs-bluebot [0] (
11:52:31***Saving seen data "./dancer.seen"
11:53:04 Join lebellium [0] (
11:53:52 Join bertrik [0] (
11:53:53 Quit bertrik (Changing host)
11:53:53 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
11:58:11lebellium[Saint]: I had the example of my theme in head. For example the 240*320 theme: If we get a 240*320 touchscreen target one day, I don't see how my theme coud be compatible unless you want to touch the small play/pause icon in the top corner. I obviously designed it for a non-touchscreen target
11:58:28[Saint]"one day"?
11:58:41[Saint]There's millions of them. literally.
11:59:25lebelliumI mean if we ever get a 240*320 touchscreen target*
11:59:30lebelliumsorry for my english mistakes
11:59:45[Saint]And, I still meant "one day?"
12:00:02[Saint]Its one the the single most popular smartphone resolutions.
12:00:07[Saint]you're not thinking about RaaA.
12:01:10lebelliumIndeed (for me 800*480 sounds more popular but there are many low-end android smartphones with bad 240*320 screen)
12:01:13[Saint]There shouldn't be any need for a "touchscreen theme" or a "non-touch theme", in my honest opinion, the reason there is such a thing is because theme authors are lazy.
12:01:50[Saint]It is entirely possible for the same theme to handle both cases, and *every* theme should.
12:02:52lebelliumI don't see how that? Most themes have a single play/pause/next/prev/stop/recording icon because you don't need to display them all at the same time for a non-touchscreen device. And that saves space on the screen
12:03:29[Saint]so, display and position the items conditionally.
12:03:46[Saint]this is why we have a "is touchscreen?" condition.
12:05:27[Saint]"play/pause/next/prev/stop/recording" this also isn't a particularly good example. play/pause/stop can share a single space, and recording has its own screen, so the same control space can be used to stop/start recording.
12:06:41[Saint]"play/pause/stop" can easily be the same control space, with three different icons. The touchscreen cabbies all do this.
12:06:58lebelliumIndeed. But you still need to display at least 3 icons: prev, next and play/pause. It's technically possible to doing it conditionally as you said but I don't know where I would get the space on my theme for that 2 more icons without redesigning it completely
12:07:50[Saint]if your theme has no visual representation of those cases for the "non-touch" case, then, (IMO) it is lacking to begin with.
12:10:13[Saint](notice the "IMO", I think a theme should *always* display the current play state via some form on "universal icon" >>/<</|>>/<<|/>/||/[])
12:10:48lebelliumkill the design to support the touchscreen 240*320 targets with the same theme? I won't and I assume other designers won't either. That's why I proposed to make a touchscreen/ non-touchscreen flag.
12:11:29[Saint]or, just....don't design for that resolution?
12:11:44[Saint]the touchscreen/non-touch thing *IS A REALLY BAD IDEA*
12:12:01[Saint]it is totally needless, and sets up more failure cases than it could fix.
12:13:48[Saint]If you can't fit basic user controls onscreen at all times without compromise, though, it seems to me that the theme needs to be re-thought anyway.
12:13:53[Saint]...but, that's just me.
12:14:47[Saint]and, sorry, but I do see play state as a "basic user control" that *needs* a visual representation on a media player. Remember, these devices don't all have a nice hardware button for everything.
12:14:55lebelliumYeah that's a personal opinion. I think the contrary and will never redesign my theme that way. If I ever get a 240*320 touchscreen target I will just make a new theme optimized for it
12:15:55[Saint]well, I look at this from the perspective of the default theme. In Rockbox, we have touchscreen and non-touchscreen targets for 240x320, and I have *one* theme that covers both.
12:16:09[Saint]It wasn'tr easy to do, but, I did it and it works.
12:16:57lebelliumwhat is your theme again?
12:17:17lebelliumI don't know the real names by heart
12:18:04[Saint]I started to rework cabbiev2 for application builds (probably over a year ago now actually, ...heh), It's all in my github.
12:18:40lebelliumso the cabbiev2 in my R0 is the same on a touchscreen target?
12:19:30[Saint]No, kugel took a slightly different approach. My work (with some debate and tinkering) should replace that or merge with it at a later stage.
12:19:58lebelliumah ok because from what I see currently on the R0 there is also a unique play/pause/stop/prev/next icon and I just find that OK
12:20:30[Saint]yeah, that's just...weird.
12:20:43[Saint]play/pause/stop *should* be the same icon, IMO.
12:20:55[Saint]well, not the same *icon*, but the same control space.
12:21:15[Saint]tap to play/pause, long tap to stop.
12:21:17[Saint]simple :)
12:21:34lebelliumthe issue for me is only prev and next that would kill the design
12:22:41[Saint]if your "non-touch" variant included ffwd/rew icons, there'd be no issue.
12:22:46[Saint]you could reuse them.
12:23:48[Saint]tap to skip, long hold to seek. etc.
12:25:24lebelliumindeed. But there is absolutely no more free space on my theme for 2 more icons. And on the 240*320 theme on the R0 I don't see free space either, I would be curious to see how you added the prev/next icons :)
12:25:54AlexP_kugel: Sorry, didin't get to look last night
12:26:11AlexP_And I agree that touchscreen and non-touchscreen themes are different
12:26:40AlexP_but anyway
12:27:49[Saint]AlexP_: I agree they're different too...I seem to be being misinterpreted.
12:28:09AlexP_If you are using the have touchscreen tag then you are essentially designing two themes in one file
12:28:12[Saint]I agree they can be totally differnet, I do *not* agree, however, that there needs to be seperate themes for each case.
12:28:48AlexP_I won't be adding the non touchscreen case to mine, show me a non touchscreen 720x1280 target first :)
12:29:27[Saint]you don't need to, see the beauty? your touch theme "just works" as long as there are HW keys.
12:29:39AlexP_Why would I bother?
12:29:43AlexP_There aren't any
12:29:55AlexP_And it needs lots of conditional casing making it much more complicated
12:30:25AlexP_It is a different design paradigm in many ways, I don't see how trying to force the two together makes sense
12:30:29 Quit pretty_function (Ping timeout: 244 seconds)
12:30:31[Saint]the conditional casing is less complication that two themes for the same target.
12:30:35AlexP_Sometimes it does, and then great
12:30:39AlexP_I disagree with that
12:30:55[Saint]your case is individual, though, as there's no non-touch case for your resolution.
12:31:10[Saint]for shared targets, then, yeah...I think one theme should be handling both.
12:31:17AlexP_One of the worst things you can do re touchscreen design is to try to force touch onto a standard theme, they just demand different things
12:31:30AlexP_And I don't see the advantage of shoving two themes into one file
12:32:13[Saint]it would solve some of the nightmare that is the default theme.
12:32:29AlexP_Not really
12:32:46AlexP_That could be solved equally well, if not better, by having a touchscreen optimised default theme
12:33:16[Saint]...that's what I'm trying to do..
12:33:36AlexP_I meant as a seperate entity
12:33:45[Saint]I just don't think it means I have to stop caring about non-touch cases to do so.
12:33:45AlexP_If you can force both into one file, then great
12:34:07AlexP_No, you ave different ones for each, each optimised for their use case
12:34:22AlexP_Instead of one that (potentially) isn't ideal for either
12:34:30AlexP_No if you can manage that in one, then great
12:34:42AlexP_But anyway, there isn't a right answer and the great thing about freedom of choice is freedom to do what you want :)
12:35:37[Saint]"No, you ave different ones for each, each optimised for their use case" if it was that way, I'd have left it alone :)
12:36:05[Saint]I took the opportunity to work on optimizing some of the ~4 year old theme code too, and saw that I could kill multiple birds, etc.
12:36:40[Saint]I wouldn't call the non-touch cabbie "optimal", is my point. But, I'm trying to get there.
12:36:40AlexP_Whether they are in one file or two, they are different - perhaps to a small extent, perhaps to a large extent
12:36:51AlexP_That is very subjective
12:36:59AlexP_As is pretty much everything about themeing
12:37:10[Saint]obviously so :)
12:37:28AlexP_Right, so you can't say, this is right, that is wrong (a generic you)
12:38:11[Saint]There's brand name me's?
12:38:22AlexP_heh :)
12:38:37AlexP_I meant one, not you personally :)
12:38:43[Saint]I also think you missed the points where I've stated "this is my opinion". Its not a right/wrong thing...I don't care what he does.
12:38:58[Saint]Ohhhhh, wow...I took that wrong, then :)
12:39:25AlexP_You also said things like "that's simply not true."
12:39:33AlexP_Which isn't the case
12:39:37AlexP_But anyway
12:40:13[Saint]Its non-false as well...hmmm.
12:40:20AlexP_It is your opinion
12:40:25AlexP_Truth doesn't come into it
12:40:26[Saint]Heh, yeah. Subjective stuff.
12:42:53 Quit stoffel (Ping timeout: 240 seconds)
12:54:02kugelone design thing that will always differ between touch and non-touch: non-touch shows the playing state (i.e. normally a play icon) while touchscreen shows a play state button (i.e. normally a pause button sicne pressing it pauses playback)
12:55:18*[Saint] should point out that he's not saying that themes should be *identical*, but rather that the design needn't vary simply because one target is touch and one isn't.
12:56:13[Saint]the above is a great example of this actually.
12:56:25kugelit's a design variation
12:56:37 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
12:56:59[Saint]yes, it is...?
12:57:04kugelfor non-touch you design icons as indicators, for touch you design them as buttons
12:57:06[Saint](not sure where that was going)
12:58:08[Saint]It doesn't necessarily mean you design them differently, though, does it...or, it needn't have to.
12:58:19[Saint]I mean, you just used existing icons too...right?
12:58:44kugelyes sure. but the theme design (not the looks of the icons) is different
13:00:30[Saint]so, when you say "but you design them as <foo>"...what does that mean exactly?
13:00:41[Saint]If you're not actually changing them at all, then...?
13:01:00kugelit does not just mean the look of the icons
13:01:28[Saint]sorry, when speaking about icons, I thought it was limited to icons.
13:02:13kugelicons are a subset of a theme design
13:03:09kugelwith "design them as ..." i meant what purpose they are for, and therefore at which times they show up
13:03:26[Saint]sorry...I missed where you were going. You started with "one thing that must differ" (being the icons), and then stated that they in fact remain the same...and now, you're talking about other things...and, I'm lost.
13:04:05kugelI didn't mean that the icon graphics differ, but their purpose
13:04:05[Saint]Oooooh, right. You last post makes the original one a LOT easier to understand.
13:05:03[Saint]Yes, and my theme(s) are doing that too...because it's needed. But, that's the only part that differs (visually) between touch and no-touch.
13:05:22kugelso, even if you use the same graphics; in a non-touch theme icons act as indicator, while on a touch one they usually act as button label
13:07:14kugelthis is what I meant by "approach differently depending on whether a theme is going to be touch-enabled or not"
13:08:13[Saint]and, that's where we part ways I think. As, I don't think a theme should be "touch only" unless it will only ever run on a touch enabled device.
13:08:57[Saint]though, perhaps you mean the touch case, or non-touch case, and not necessarily that it should be seperate themes.
13:09:18kugeli never said they need to be separate themes
13:09:38kugeli said it's doable to build both designs into the same theme
13:10:13kugelbut it requires work and perhaps compromises so not everyone can be bothered
13:13:13[Saint]The entire point of me bringing it up, and the subsequent confusion...and then getting stuck on other topics, was due to me taking issue at the idea of the themesite checking for touch/non-touch, and just in case there's any more room for confusion: "I think it's a really bad idea" :)
13:13:54[Saint]This isn't me saying "all themes should handle all cases", however.
13:15:08lebelliumif it's a bad idea but all themes should not necessarily handle all cases, then people with a touchscreen device will see a lot of unusable themes in their target category
13:15:40[Saint]lebellium: no, they will see exactly no unusable themes.
13:15:47[Saint]Every theme is usable to them.
13:16:15[Saint]You're forgetting about grid mode, which some people actually like. Which some people even use by coice.
13:17:08[Saint]If thetre's no touch tags present in a theme loaded on a touchscreen target the default used to be to switch to grid mode so that the currently loaded theme would still be usable.
13:17:13[Saint]I believe this is still the case.
13:17:56lebelliumIndeed I forgot about that grid mode...
13:18:24*[Saint] just noticed that the *-800x480x16 and *-480x800x16 bitmaps are mostly (all?) identical.
13:18:28[Saint]lol, how did that happen?
13:19:55[Saint]ah, "mostly identical".
13:20:28[Saint]There's a few have been duplicated for seemingly no reason, probably to avoid naming collisions prior to commit(?).
13:20:30lebelliumwhen talking about separate themes for touchscreen and non-touchscreen targets on the theme website I just had the touch tags in mind, I totally forgot about that grid mode, sorry
13:20:47kugelthere could be a checkbox that hides grid-mode themes
13:22:07[Saint]or, do some magic and print something in the blurb like we do fr "works with <foo>"
13:23:00[Saint]Ohhhhh....actually, this is a problem.
13:23:11[Saint]The new theme behaviour breaks this assumption.
13:23:44[Saint]There will now not *ever* be a case of a theme not having a touch tag present if it includes any bar-type tags.
13:24:08kugeldid you read what I wrote above?
13:24:49[Saint]I did, yes. But I'm wondering how the new bar tag behaviour affects this.
13:25:08[Saint]Since they now draw all their touch areas automatically.
13:25:21kugelthe automatic touch region for bars dont make it touch-enabled, it will change to back grid mode automatically
13:25:26[Saint](and it needs to be specifically disabled by the author if this is unwanted)
13:25:35kugelyou are wrong :)
13:26:21kugelthe bar-tag regions are excluded from the logic which determines touch-enabledness
13:26:34[Saint]"the automatic touch region for bars dont make it touch-enabled" Oh, right. Well, that's a good thing. Slightly broken, but, broken in a way that works.
13:27:23kugelwell, it's this way because JdGordon pushed it to be default
13:27:41 Join Horscht [0] (
13:27:41 Quit Horscht (Changing host)
13:27:41 Join Horscht [0] (~Horscht@xbmc/user/horscht)
13:28:35[Saint]the "slightly broken" bit I was referring to was the fact that the theme would have touch tags but still not technically be "touch enabled".
13:28:59[Saint]But, for the record, I don't like the default behaviour of the bar tag touch areas either...and, said so at the time.
13:29:36kugeli know
13:30:08kugelwe're in the same boot in this case
13:38:18lebellium[Saint] I don't know if you did read that but from my tests the %s tag is necessary for the RDS text to properly scroll and that worked correctly in the older builds. Should I open a bug ticket on the website or the right people are already aware of all RDS issues?
13:39:00n1ssaratoga: i could test on my classic, but it has to charge a bit first
13:39:05[Saint]I think "the right people" are aware already, but you should open a bug anyway.
13:43:57kugellebellium: it would be nice if you could find the commit that introduced the problem (heard of git bisect?)
13:45:37kugelbtw, could we tag the branch point? this would allow to bisect issues like "doesnt work in current build but last release" more easily
13:47:50lebelliumkugel: that will be difficult as I don't compile builds myself, I just download builds on the RB website :S For the Clip Zip I used an old April build on my computer to compare with the latest builds but for the R0 lorenzo gave me a recent RDS build. That would be interesting to compile the RDS patch with an old build to see if that still makes my theme crash
13:51:34[Saint]setting up a development environment will take 30~60 mins.
13:51:45[Saint]bisection another ~60mins.
13:52:33***Saving seen data "./dancer.seen"
13:52:49lebelliumI'm using Windows 7, not Linux, I'm not in this development world (yet?) ;)
13:55:13n1slebellium: set up a VM then!
13:56:27lebelliumI already have a VM for Windows XP (necessary to run the Samsung recovery tools) but I don't have a single GB or MB left on my HDD for a new OS, the situation is critical lol
14:03:21 Quit bertrik (Remote host closed the connection)
14:09:20[Saint]couldn't you just run the recovery tools in compatibility mode?
14:09:34[Saint]I thought & was supposed to be entirely backward compatible still.
14:10:45lebelliumThe issue is just the drivers, The Samsung R&D in Korea lives in the past with 32-bit only drivers which can't be installed on any 64-bit OS. Yet, most computers run Win 7 64-bit....
14:11:12[Saint] what world are you talking about?
14:11:20[Saint]"Most computers" run XP.
14:11:52lebelliumI meant most computers sold today run Win 7 64-bit, not Win 7 32-bit except some netbooks
14:16:17AlexP_I would object to having grid mode themes presented as touch themes
14:16:20AlexP_They just aren't
14:16:38AlexP_We would annoy a huge amount of people if we did that, they just aren't proper touch themes
14:29:12kugeli think it depends, there are targets where grid mode is default
14:29:21kugele.g. cowond2 (iirc)
14:30:55kugel[Saint]: according to some sources win7 has surpassed win xp
14:34:26 Join wodz [0] (
14:34:52 Join prof_wolfff [0] (
14:36:03wodzHa, that was easy. Now I can load fully resolved elfs as plugins. I can't figure out how to strip elfs. plugins.make just don't like me.
14:36:16wodzTime to approach relocations...
14:43:04lebellium [Saint]: replacing your 20-DroidSans-Bold.fnt font I use for RDS text by another font like the Rocbox one, the R0 doesn't crash anymore but the RDS text still behaves crazy when scrolling.
14:59:37lebelliumOh! I found something interesting. When displaying RDS name and RDS text in separate viewports, the RDS text is not crazy anymore!
15:05:04 Join Prodicus [0] (
15:10:01lebelliumoh gosh that also fix the crash issue! Oo But why is putting %ty and %tz in the same viewport that bad?! that did not cause any issue with the older builds as far as I know!
15:18:31 Quit wodz (Quit: Leaving)
15:19:57 Join Prodicus_ [0] (
15:20:22n1sBuschel: looking closer at celt_decode_with_ec the freq and X buffers are only used at the same time in denormalize_bands, freq is never used before and X is never used after so i hacked up a patch to share the same buffer for them which saves some iram when allocing them on iram
15:21:07 Join pretty_function [0] (~sigBART@
15:21:16n1sit works but is fairly ugly
15:21:21 Quit Prodicus (Ping timeout: 246 seconds)
15:23:53n1sX is a int16_t and freq is int32_t
15:24:42 Join Prodicus [0] (
15:25:34 Quit Prodicus_ (Ping timeout: 256 seconds)
15:33:21 Quit guymann (Quit: brb)
15:33:49 Join guymann [0] (~c@unaffiliated/guymann)
15:35:18 Join wodz [0] (
15:35:21 Quit guymann (Client Quit)
15:35:42 Join guymann [0] (~c@unaffiliated/guymann)
15:36:55 Quit sciopath (Quit: Quitte)
15:39:18 Quit Provel_ (Ping timeout: 260 seconds)
15:42:13 Join Provel_ [0] (
15:48:22 Join Provel [0] (
15:50:00 Quit Provel_ (Ping timeout: 245 seconds)
15:52:35***Saving seen data "./dancer.seen"
15:55:34 Quit bzed (Ping timeout: 260 seconds)
15:56:59 Quit Provel (Read error: Connection reset by peer)
15:57:17 Join bzed [0] (
15:57:24 Join Provel [0] (
16:22:36 Quit Provel (Ping timeout: 246 seconds)
16:22:46 Join stoffel [0] (
16:23:08 Join Provel [0] (
16:30:25 Quit n1s (Read error: Connection timed out)
16:40:41 Quit stoffel (Remote host closed the connection)
17:11:49 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:19:12 Quit pretty_function (Remote host closed the connection)
17:29:07 Quit mgottschlag (Ping timeout: 255 seconds)
17:30:52 Quit Provel (Read error: Connection reset by peer)
17:31:19 Join Provel [0] (
17:37:13 Join perrikwp_ [0] (
17:39:53 Quit perrikwp (Ping timeout: 257 seconds)
17:41:09 Join stoffel [0] (
17:52:37***Saving seen data "./dancer.seen"
18:04:10 Join Keripo [0] (
18:04:14 Join pretty_function [0] (~sigBART@
18:10:02 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
18:18:08 Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
18:18:11 Join Buschel [0] (
18:18:26 Join Totalled [0] (
18:21:54Buscheln1s: about your freq/X patch. why do you set X to freq[C*N/2]? and not simply to freq[0]?
18:22:08 Nick AlexP_ is now known as AlexP (~alex@rockbox/staff/AlexP)
18:23:15 Quit wodz (Read error: Operation timed out)
18:23:39 Quit Keripo (Read error: Connection reset by peer)
18:24:39Buscheln1s: i do not think this approach is too hacky. i would define a static char-array of 1920*MAX(sizeof(celt_sig), sizeof(celt_norm)) in IRAM and would simply cast it then, e.g. "freq = (celc_sig*) s_buf; X = (celt_norm*) s_buf;"
18:26:30Buscheln1s: could you measure the other parts of the patch (comb_filter, memcpy in celt.c, the index-changes to mdct.c/kiss_fft.c)?
18:32:30 Quit pretty_function (Read error: Operation timed out)
18:32:55Buscheln1s: btw, where does this 1920 come from?
18:33:11 Join Keripo [0] (
18:44:08Buscheln1s: ok, just found out about the index shift when setting X
18:47:05 Quit mgottschlag (Read error: Operation timed out)
18:47:27Buscheln1s: nevertheless we should have the static allocation of two seperate buffers submitted first. and move the "joined" buffer solution in a second submit. this is benefit *if* we should have overssen something and need to bisect some day
18:49:41 Quit guymann (Quit: oops)
18:54:43 Join pretty_function [0] (~sigBART@
18:56:30 Join guymann [0] (~c@unaffiliated/guymann)
18:59:15 Quit Buschel (Ping timeout: 246 seconds)
19:01:51 Quit pretty_function (Ping timeout: 246 seconds)
19:06:04 Join pretty_function [0] (~sigBART@
19:10:01[7]saratoga: sorry, no idea
19:10:08[7]it's been ages since I implemented this
19:10:26saratogano idea what your comment was referring to?
19:11:04[7]well that sampling rate stuff which apparently isn't working
19:11:04saratogaNote: Disable output before calling this function
19:11:24 Quit stoffel (Remote host closed the connection)
19:13:18saratogathat code is called above in audiohw_preinit, so i guess one of the other registers might need to be set before changing that clock
19:13:56saratogaor perhaps cscodec_reset
19:15:14[7]well those comments on the forums seem to be nonsense - there shouldn't be any need to mess with the pcm driver
19:15:57[7]cscodec_reset is a very bad idea
19:16:23[7]it is crucial for audiohw_preinit to be called before audiohw_set_frequency
19:16:47[7]audiohw_preinit will set the sampling rate to 44100Hz
19:17:01saratogadoes sample rate changing even work on the Classic right now?
19:17:13[7]I think it should, but I'm not sure if it has been tested at all
19:17:17saratogaas far as I can tell it shouldn't (but i'm not too familiar with this code)
19:17:26[7]why shouldn't it?
19:17:30 Quit perrikwp_ (Read error: Connection reset by peer)
19:17:43[7]that /* Note: Disable output before calling this function */ is probably copied from another driver
19:17:55[7]not sure if it's strictly necessary for this codec, probably not
19:18:06[7]but it's generally a good idea
19:18:13saratogai think pcm_set_frequency will do nothing since pcm_dma_apply_setting is a NOP
19:18:29[7]that's fine because pcm doesn't even need to know the frequency
19:18:44[7]the cs42l55 is operating as the i2s master and thus controlling the sampling rate
19:18:53saratogaisn't that access point for the plugin API though?
19:18:55[7]i.e. the i2c write in audiohw_set_frequency is the only thing you should need to do
19:19:00saratogaif that is a NOP how would something cahnge the rate?
19:21:46 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:21:50saratogahave you tried the test_sampr plugin on the 6G?
19:28:59 Quit Horscht (Quit: Verlassend)
19:29:44 Join Epicanis [0] (
19:29:57 Join wodz [0] (
19:30:49 Join Buschel [0] (
19:41:59 Join lorenzo92 [0] (
19:44:33 Quit Keripo (Quit: Leaving.)
19:51:01 Quit Rower85 (Quit: Hmmm...)
19:52:39***Saving seen data "./dancer.seen"
19:55:24 Quit ender` (Read error: Connection reset by peer)
19:56:50 Join ender` [0] (
19:57:52 Quit linuxstb (Quit: This computer has gone to sleep)
19:59:54 Join linuxstb [0] (
20:03:23user890104saratoga: i have an ipod classic (120 GB), if you want anything to be tested on that target let me know
20:04:07 Quit the-kyle (Quit: Leaving.)
20:05:18 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
20:06:13lorenzo92kugel: finally tested your patch, A+, works perfectly ;)
20:07:29 Join XavierGr [0] (~xavier@rockbox/staff/XavierGr)
20:08:28 Join the-kyle [0] (
20:08:32lorenzo92kugel: little help with git, if I remove the gpio patch, then need to do again simply git commit? I fetched git freshly, then applied your patchset using the patch url...
20:09:07 Join Rower85 [0] (
20:10:47 Quit Rower85 (Read error: Connection reset by peer)
20:11:02 Join Rower85 [0] (
20:19:08 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
20:20:05lorenzo92pff did a bullshit...sorry you it possible to remove an entry in gerrit? -.-"
20:20:54lorenzo92well it is correct, but not really hehe
20:26:36[Saint]anyone else find it interesting the "bricked" Clip+ was detected as an M200Plus by some lowlevel magic b0hoon managed to trigger?
20:27:21lorenzo92okay, let's ask somehting different... does anyone knows why software keypad lock works in WPS only? Is there a particular reason?
20:29:17 Quit pamaury (Ping timeout: 245 seconds)
20:32:18[Saint]lorenzo92: I believe it's only intention was to prevent interruption to playback.
20:32:38[Saint]It isn't to prevent all arbitrary accidental keypresses.
20:33:03lorenzo92[Saint]: yes, logical, but then why don't enable it also in the FM radio screen?
20:33:32lorenzo92moreover targets with physical lock can be lock in any screen...
20:34:19[Saint]they also have a physical lock to do so with, and don't need to rely on hardware keys that may already be mapped to other functions.
20:35:04[Saint]my suspicion is either: a - no one thought of it, or, b - crowded keymaps and blocking keys made this a nightmare.
20:35:49[Saint]I'm leaning towards the latter, as I know what a nightmare keymaps are.
20:39:28 Join pystar89 [0] (
20:43:50lorenzo92indeed :) but it *may* be a good idea to add the keylock also in FM screen, at suggested by some users
20:48:34lorenzo92updated YP-R0 wiki ;)
20:50:44lorenzo92kugel: why is "Battery charging" in the TargetStatus page labelled as partial for YP-R0? We need definitely work out the USB to make this port stable, I have ideas but rockbox code crashes :p
20:58:18 Quit XavierGr (Read error: Connection reset by peer)
21:03:15 Join XavierGr [0] (XavierGr@rockbox/staff/XavierGr)
21:10:11 Quit pretty_function (Ping timeout: 248 seconds)
21:25:04 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
21:38:39 Quit y4n (Quit: We're fucking 3LN!)
21:42:33n1sBuschel: the 1920 comes from the static mode, and stereo and yeah sure, doing either change separately is probably nicer
21:43:17n1sI'll try to test the other parts of your patch during the week
21:46:10Buschelperfect, i'll let you know if i should find any other optimization
21:52:40***Saving seen data "./dancer.seen"
21:53:48 Join freqmod [0] (
21:54:12freqmod - anything more i can do to get this in to trunk?
22:01:26 Quit lorenzo92 (Quit: ChatZilla 0.9.89 [Firefox 15.0.1/20120907231657])
22:01:40 Join amayer [0] (
22:09:36n1sfreqmod: tell someone to push it :) I have no idea what it's supposed to fix though
22:10:39freqmodaccordik to bertik "* Seeking ported from speex, but fails on some cases (e.g. seek to granule 0)"
22:10:45saratogaalternatively if you're going to be working on opus more, we could probably just give you git write access
22:11:13freqmodwell, i don't plan to work more than bugs i encounter
22:11:18freqmod*than on bugs
22:12:09freqmodbasically it disables my code to guess where to jump to for long seeks (to speed it up for targets with disks), which are a bit flakey
22:12:18freqmodand handles granule 0 as a special case
22:15:06 Quit factor (Quit: Leaving)
22:16:52freqmodbut if you want to give me access, it could be useful to have
22:17:02*freqmod is planning to do a codec2 codec interface
22:17:14freqmodit is finished, but no fixed point of codec2 exists yet
22:18:15freqmodand the bitstream doesn't look to get frozen in the near future
22:19:42 Quit guymann (Quit: exit(EXIT_FAILURE);)
22:20:29 Join guymann [0] (~c@unaffiliated/guymann)
22:22:02Prodicusstill some work to be done on codec2. personally from my experience it seems like his sample files do a lot better than anything I encode myself, which makes me wonder if he's overtrained things.
22:23:33freqmodi have the same experience
22:23:47freqmodbut i have only transcoded mp3's
22:24:39freqmodhowever it does create listenable sounds in about 1 kpbs
22:24:50freqmodso i was thinking about encoding the whole librivox library
22:25:09ProdicusI've worked from lossless originals- recordings of myself and recordings from a conference
22:25:42freqmodhe should probably get some more test samples to tune for
22:26:54Prodicusas you say, for huge audiobook libraries the possibility of getting something at 2kbps is tantalizing and the quality it already hits at 1.4k makes one think it could be good enough to work well at ~2
22:28:16freqmodhe supposedly made a better codec at 3,6kbps a week ago
22:28:44freqmodhowever i think i would still go for 1,2 kbps as the difference between 3,6 and 10kbps opus is much more
22:29:04 Quit benedikt93 (Quit: Bye ;))
22:29:31Prodicusyeah, for a lot of applications the quality improvement from moving to mediumband rather than narrowband is pretty obvious.
22:30:31 Quit guymann (Quit: brb)
22:31:14 Join guymann [0] (
22:31:31 Quit guymann (Changing host)
22:31:31 Join guymann [0] (~c@unaffiliated/guymann)
22:36:39 Join scorche` [0] (
22:36:39 Quit scorche (Disconnected by services)
22:36:39 Quit scorche` (Changing host)
22:36:39 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
22:39:45 Quit Buschel (Quit: ChatZilla [Firefox 15.0.1/20120905151427])
22:46:08 Quit wodz (Quit: Leaving)
22:55:17 Join lebellium_ [0] (
22:57:12 Quit lebellium (Ping timeout: 245 seconds)
22:57:26 Nick lebellium_ is now known as lebellium (
22:58:46 Join Strife89 [0] (~Strife89@
23:03:20 Join factor [0] (
23:12:02 Quit factor (Read error: Connection reset by peer)
23:17:23 Join scorche [0] (~scorche@rockbox/administrator/scorche)
23:17:55 Quit scorche` (Read error: Connection reset by peer)
23:20:29 Join factor [0] (~factor@
23:32:57 Quit n1s (Quit: Ex-Chat)
23:34:33 Quit Strife89 (Quit: Connection reset by deer.)
23:48:10 Quit freqmod (Remote host closed the connection)
23:52:43***Saving seen data "./dancer.seen"
23:53:06 Quit the-kyle (Quit: Leaving.)
23:55:24 Quit Prodicus (Read error: Connection reset by peer)
23:56:04 Join Prodicus [0] (

Previous day | Next day