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

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

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

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2014-01-16

00:00:30[Saint]I don't really see the point in themeing if you'll never be able to guarantee the result.
00:00:55[Saint]it may as well be forced so we know whats there.
00:01:09[Saint]remove the user part altogether.
00:01:10kugelfwiw, falling back to sysfont is also not as easy as you might think
00:01:33[Saint]which is another reason why this seems odd.
00:02:12 Quit sulky (Ping timeout: 252 seconds)
00:02:37[Saint]At this stage it seems simpler to disallow themeing, we'll know exactly whats needed.
00:02:59kugelyea, a black & blank screen
00:03:08kugelthat is terrible
00:03:09gevaertsWell yes. disallowing theming is trivial. Just don't do anything :)
00:03:48 Join sulky [0] (sulky@gateway/shell/
00:04:16[Saint]its just to me, mismatched fonts seems terrible. and it'll look awful, and break intended spacing, and very few targets actually need it.
00:04:55[Saint]so instead of butchering someone's theme, why dont we show our own screen?
00:05:19[Saint]time, date...what else do we need, really? a usb pic.
00:05:37gevaertsSo how do you show time and date?
00:06:19[Saint]those glyphs are pretty much guaranteed to be in RAM, no?
00:06:58[Saint]as it works already. albeit seemingly accidentally?
00:07:01gevaertsWell, according to current thinking, so are he other ones needed for this screen
00:07:13gevaertsWhat works already?
00:07:30[Saint]userfont in usb sbs.
00:07:40kugel[Saint]: mismatched fonts?
00:07:46[Saint]unless that was very recently fixed.
00:08:42gevaertsYou shouldn't be seeing any sbs on the usb screen *at all* with current git. If you do, it's a serious bug
00:08:52kugelthat the USB screen showed something other than the USB logo on a solid black background was actually a bug
00:09:06kugelso if the userfont "worked" it'll still work
00:10:15kugelgevaerts: s/current git/HEAD^^/. the USB screen actually showed the backdrop and sometimes other stuff
00:10:28kugelbut that was due to a bug where the themes got re-enabled partly
00:10:40kugelthe intention was indeed to show usb logo on black
00:10:53gevaertsYes, I know. I did see the commit go by that fixed it :)
00:11:14gevaertsYou might be able to get the built-in status bar, but that's it
00:11:43gevaertsIf you want to show the clock large enough to be readable, you need this patch
00:11:59 Join bthomson_ [0] (sid21436@gateway/web/
00:14:11 Join soap_ [0] (~soap@rockbox/staff/soap)
00:16:11 Join ender [0] (
00:16:26 Quit TheSeven (Disconnected by services)
00:16:41 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
00:16:55 Join Airwave_ [0] (~Airwave@
00:17:06 Quit bthomson (Ping timeout: 245 seconds)
00:17:08 Quit soap (Ping timeout: 245 seconds)
00:17:09 Quit ender` (Ping timeout: 245 seconds)
00:17:09[Saint]errr, sorry, by mismatched fonts I meant the (however supposedly unlikely its suppoed to be) chance of a cache miss. It doesn't sit too well with me that there's no way to guarantee the theme will display as intended.
00:17:09 Quit Airwave (Ping timeout: 245 seconds)
00:17:13 Nick bthomson_ is now known as bthomson (sid21436@gateway/web/
00:17:42[Saint]I might find myself loading a giant bitmapstrip to avoid this.
00:17:53 Quit DormantBrain (Ping timeout: 245 seconds)
00:18:15gevaerts[Saint]: again, there are ways to handle this sort of thing if it turns out to be a problem
00:19:07kugelif you use bitmaps even before the problem shows up then we're going to have a hard time improving it
00:19:26gevaertsNobody is suggesting that this will be the last commit to rockbox :)
00:19:46 Quit Strife89 (Ping timeout: 276 seconds)
00:19:47kugelon the other hand, you're now at least in a situation where you can use bitmaps in the usb screen, whereas previously you couldn't do anything
00:20:27[Saint]I always used to work for me.
00:20:37gevaertsdue to serious bugs, yes :)
00:20:40kugelalso you can fall back to sysfont manually, by %Vd'ing a suitable viewport (%cs support is coming)
00:20:40[Saint]even though I knew it shouldn't.
00:21:30kugelbut automatic fallback to sysfont has a number of problems, and realistically no benefit
00:22:28JdGordonI havnt read the backlig (or patch yet) but why do you think it is unlikey the glyph wont be in cache?
00:22:33kugelproblems are a) poor character coverage, b) changed glpyh geometrics (could destroy carefully crafted themes), c) it's too small for many eyes
00:23:32kugeland probably some others
00:23:56 Quit pamaury (Ping timeout: 246 seconds)
00:24:12kugelJdGordon: because what you have last displayed is in cache
00:24:30gevaertsJdGordon: as I see it, basically because the sbs has been visible all along
00:24:57JdGordonso? changing to the usb viewport or whatever is likely to introduce new strings
00:25:38gevaertsYes. We probably need to make sure that's done *before* the fonts are shut down
00:25:56 Join ildarion [0] (4ef2d722@gateway/web/freenode/ip.
00:26:11JdGordonpotentially with a tag to say which glpyhs you want cached
00:26:35JdGordonI could even see a case for a different font wanted for the usb screen with much larger glphs
00:27:06 Join sprout_now [0] (~RJChesko@
00:27:08saratogaso mentioned this before, but does anyone have any opinion about using the log to disk feature to record more USB errors? right now most error conditions are just logf, which is discarded in official builds
00:27:30saratogain theory it would be simple to change them to logdiskf and then have them flush to memory after USB unmount
00:27:41ildarionNeed some help with my sansa Fuze +, I don't understand exatly what is said: "However the imx_hid_recovery tool is in the SVN is capable of uploading a firmware file in the SbFileFormat based on some reverse-engineering we did. "
00:27:45saratogathis might return some interesting debug info when users report problems
00:29:21gevaertssaratoga: it certainly wouldn't *hurt*
00:29:42saratogado you think it would help?
00:29:51saratogano idea if the error conditions are even interesting
00:30:34 Quit ildarion (Client Quit)
00:31:02gevaertshard to say. It depends on what happens exactly
00:31:24kugelJdGordon: the sbs is updated before the fonts are shown (after usb screen is indicated by %cs), caching all static strings
00:31:42JdGordonkugel: just regadring point b) above... how does putting a blank or fallback square fix that?
00:32:04saratogaat very least it might allow us to classify which general type of failure different users have
00:32:06JdGordonwhat about dynamic strings?
00:32:14JdGordoni;e numbers for a clock/battery remaining
00:32:17*gevaerts nods
00:32:55kugelthe height of the fallback is the same as the font height
00:33:03kugelso the line height doesnt change
00:33:03JdGordonand width?
00:33:37JdGordonhang on, how does the line height change if it falls back to sysfont?
00:33:55JdGordonsurely it draws it at some sensible position without affecting that?
00:34:07JdGordonthats a bug imo if it changes it
00:34:15kugellook at the code
00:34:25kugeland changing that isn't trivial
00:34:27JdGordoncant right now
00:35:20kugelif you're really really worried about cache misses then use bitmaps or an explicit sysfont viewport
00:35:36kugelbut in practice I expect all necessary glpyhs to be cached
00:36:01JdGordonbaseless assertion I think
00:36:17kugelin fact I haven't ever noticed that Rockbox had to load glyphs after cache warmup
00:36:40gevaertsI think right now we don't know, and I think there's only one way to know :)
00:36:45JdGordonme neither, but than only use ~40 glyphs
00:36:54 Quit sprout_now (Quit: TinyIRC 1.1)
00:36:57JdGordonbut most of the world doesnt just use latin letters
00:37:17kugelright, and all those are completely screwed with sysfont
00:37:35kugelwhich is why automatic fallback to it is not an option
00:37:35JdGordonwell, yes
00:37:40JdGordonnot always
00:37:54JdGordonthey are only screwed if sysfont doesnt have the glyph
00:38:08kugelsysfont has only latin letters
00:38:48JdGordoni.e chinese heavy tags have filled the tags, but the user has a skin with "USB connected", those are in sysfont but not in the cache
00:39:17JdGordonfilled the cache*
00:39:27kugelthis case is catched by the initial skin update
00:40:15JdGordonand what about when they want to change it to "Now safe to eject"?
00:40:22JdGordoni.e like apples usb screen
00:41:14kugelyou couldn't do this currently
00:41:48kugelanyway we can add measures to improve the probability or even make guarantees if it turns out to be a problem
00:41:52JdGordononly because the tag isnt there
00:42:08kugelno, our usb core can't do that either
00:42:19kugelright, gevaerts?
00:45:14 Quit ender (Quit: Java is, in many ways, C++--.-- Michael Feldman)
00:47:10JdGordonkugel: back to the linehighet, what is that based on?
00:48:44kugelthe font structure
00:49:05kugelwhich is freed when the font is unloaded
00:49:12JdGordonisnt lineheight some viewport value now?
00:49:22JdGordoni dont understand why drawing one glyph would change it
00:49:36kugelbut that's not the problem. the problem is that the glyphs buffers have different geometrics
00:50:06kugelyou cannot draw a 6x6 bitmap at 12px height because there are not enough bytes in the buffer
00:50:25 Quit TBCOOL (Ping timeout: 252 seconds)
00:50:48JdGordonyou can draw a 6px high glpygh at the bottom of that 12px space though
00:50:56kugellcd_puts* just get the raw buffer for each glyph
00:50:58JdGordonor has fallback literally never worked?
00:51:06kugelwith no width/height information
00:51:30 Join TBCOOL [0] (
00:51:33kugelthe fallback worked but always changed the line height
00:52:05JdGordonthats a bug
00:52:31kugelthe line height isnt a viewport value, not anymore for lists and it never has been for skins
00:54:50***Saving seen data "./dancer.seen"
00:54:57JdGordonso what *is* linehighet?
00:55:27 Join onder`_ [0] (
00:56:09 Quit onder` (Ping timeout: 252 seconds)
00:56:13 Nick onder`_ is now known as onder` (
00:56:39kugelJdGordon: it's a property for put_line()
00:57:13kugelbut the skin engine always uses the line height derived from the font
00:57:39kugelif the font changes the line height does too
00:57:51JdGordonthe font isnt changing, just one glyph needs to
00:58:32kugelwhy just one glyph?
00:58:47JdGordonthe one which needs to use sysfont instead
00:59:17kugelyou're suggesting to use missing glyphs from the sysfont, intermixed with real glyphs?
01:00:28kugelI don't think that's any better. what if sysfont doesnt have that glyph?
01:01:05kugelalso, as I said, the low level code doesn't allow for individual glyphs to have different height than the the font height
01:01:46kugelthis kind of fallback never existed in Rockbox
01:02:08JdGordonwhat did fallback do then?
01:02:39kugelthe usb screen did setuifont(FONT_SYSFIXED)
01:02:58kugelthat is, changing the font entirely, including line height
01:03:15JdGordonno, i mean out of the usb screen
01:03:21JdGordonthere is fallback code there
01:03:25JdGordonwhat does it do?
01:03:35JdGordonmy impression was it would draw one glpygh from sysfont
01:04:13kugelyou're wrong :)
01:04:42JdGordonwhat did it do then?
01:04:46JdGordonchange the whole lines font?
01:04:59kugelwhat is "it"?
01:05:10JdGordonthe fallback mechanism which has been there forever
01:05:16kugelthe only kind of fallback is what I mentioned above
01:05:37JdGordonlook at (i tihnk) get_glyph() or find_glypgh() in font.c
01:06:08kugelJdGordon: I edited this code all day today. there is no such fallback
01:06:54JdGordonthere sure used to be
01:07:12kugelthe fallback you're thinking of just loads all glyphs from sysfont
01:07:40kugelwithout looking the cache (which is also destroyed when the corresponding font is unloaded)
01:09:41JdGordonerr... ok, looks like I was indeed misrememebring
01:09:43JdGordonsorry :/
01:10:00JdGordonit wa fontid fallback, not glpygh fallback :/
01:10:21kugelthe fallback only works for fonts that are in-ram (only sysfont in practice) and you do need to set the viewports font id to that in-ram font
01:10:32kugelit doesn't work like you're thinking
01:11:14JdGordonso do we already show a box for missing glpyghs?
01:11:24*JdGordon will eventually spell that the ame way twice in a row!
01:12:12kugelright. the usb screen sets the font to sysfont, and if that doesnt have the glyphs then the box is shown
01:12:43kugela non-sysfont font is not a thing during usb
01:12:57kugelwell, was. I just changed that
01:13:14*kugel is off
01:18:21 Quit kugel (Ping timeout: 272 seconds)
01:19:09 Join foolsh [0] (
01:19:09 Quit foolsh (Read error: Connection reset by peer)
01:19:24 Join foolsh [0] (
01:23:01 Quit ps-auxw (Ping timeout: 245 seconds)
01:26:14 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
01:27:51 Join Strife89 [0] (
01:32:52[Saint]JdGordon: likely easy to miss as it never worked the way to code intended.
01:33:01 Quit lebellium (Ping timeout: 272 seconds)
01:33:05toehser1head is crashing on boot for me...
01:33:10toehser1in clip zip real target
01:33:12[Saint]I only found out, say, about a year ago, that the sbs was never supposed to work in USB screen.
01:33:25[Saint]It didn't stop it, though.
01:34:07toehser1not sure what theme though - I had been playing with various themes
01:36:13toehser1it was 'sugarbrown'. I've backed out now
01:37:52[Saint]That theme may or may not have included some crazy workarounds for brokenness that may or may not exist anymore.
01:38:07[Saint]But either way, it still shouldn't be able to make the system fall over.
01:40:58toehser1Not sure what my base theme was, something else, that is wps only, so probably didn't make it crash on boot, where it just goes to menu
01:41:11toehser1I'll look into it later, have to go to dinner.
01:42:02toehser1I think the base was "thescoop"
01:42:05toehser1I'll try in sim.
01:44:18toehser1That theme is giving me an "unable to load font" 4x6-L1 message in the sim, but not crashing
01:45:19 Quit APLU (Ping timeout: 252 seconds)
01:45:20[Saint]Its just doing exactly what its supposed to, then.
01:45:34[Saint]If any asset can't be found, it should refuse to load the theme.
01:46:18[Saint]*why* the asset can't be found? Who knows. The themesite certainly shouldn't have allowed it to be uploaded if it called an asset it didn't include.
01:47:23toehser10x0000000000498b2e in search (fcache=0x9e2578 <audiobuffer+96
01:47:23toehser1 p_insertion_point=0x7fffffffd5e8)
01:47:23toehser1 at /home/tom/rockbox/git-tip/rockbox/firmware/font_cache.
01:47:23toehser199 p = lru_data(&fcache->_lru, fcache->_index[mi
01:47:34toehser1After I went and put the font in.
01:47:55[Saint]The theme didn't include the font it itself calls?
01:48:01toehser1Steps: Pick theme "thescoop", then pick theme "sugarbrown" which overrides only the WPS.
01:48:07[Saint]Can you please flag it for admin assistence on the themesite?
01:48:11toehser1right... but... I had it hanging around.
01:48:28toehser1eh... I'm more interested in the segfault :)
01:49:24toehser1after I did the "thescoop" base and "sugarbrown" wps, without font errors (copied the font in), sim is able to get that segfault
01:49:44 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
01:50:25 Join APLU [0] (~mulx@2a01:e34:ee29:12b0::10)
01:51:02toehser10x0000000000498b2e in search (fcache=0x9e2578 <audiobuffer+96
01:51:02toehser1 p_insertion_point=0x7fffffffd5e8)
01:51:02toehser1 at /home/tom/rockbox/git-tip/rockbox/firmware/font_cache.
01:51:02toehser199 p = lru_data(&fcache->_lru, fcache->_index[mi
01:51:39toehser1weird, it truncated the paste...
01:51:56toehser1That variable in the array is "mid" not "mi"
01:52:18toehser1When I get back from dinner I'll see if it fails in 3.13 or if I have a new bug...
02:16:05 Join RJChesko [0] (~RJChesko@
02:16:30 Join SuperBrainAK [0] (~andy@2001:470:8:a61::5f92:59a1)
02:17:36 Quit RJChesko (Client Quit)
02:37:42 Quit shamus (Read error: Connection reset by peer)
02:49:07 Join shamus [0] (
02:54:52***Saving seen data "./dancer.seen"
03:21:04 Quit MarcAndersen (Read error: Connection reset by peer)
03:41:33 Quit Raptors (Read error: Connection reset by peer)
03:42:02 Join Raptors [0] (
04:01:25toehser1head segfaults this way but 3.13 doesn't
04:02:41JdGordonblame kugel
04:02:42[Saint]I suspect you'll want to talk to kugel about this.
04:03:04[Saint]The area he's been working in recently points at this.
04:05:09toehser1The good news is, it doesn't segfault (so far) on the themes I actually use myself in real life... :) so my interest is somewhat more personally detached than it was for the bugs that were hitting me in real life daily use...
04:06:43toehser1Oh - this is cool - corrupted font appearance
04:07:15JdGordonget a creenshot!
04:09:38[Saint]Nice one.
04:10:20 Join theaftermath [0] (
04:11:55toehser1Font corruption survived navigation to menu and back to WPS, but sim reboot left font ok until it segfaulted.
04:13:49toehser1Instructions to reproduce, build head, install "thescoop" and "sugarbrown" fonts, (clip zip sim), install other font that is missing from theme file, switch around between themes and play songs, bam segfault. Or if very lucky, cool font corruption.
04:14:18[Saint]That's odd. Leaving the wps and going back again would ensure the wps was unloaded and reloaded.
04:14:27[Saint]So, apparently its a least semi-reproducible.
04:21:57 Quit Strife89 (Ping timeout: 246 seconds)
04:25:09 Quit theaftermath (Changing host)
04:25:09 Join theaftermath [0] (~theafterm@unaffiliated/theaftermath)
04:35:08 Join Strife89 [0] (
04:46:57 Quit pixelma (Disconnected by services)
04:46:57 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:46:58 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:46:59 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:46:59 Quit amiconn (Disconnected by services)
04:47:02 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:49:53 Quit theaftermath (Quit: ZNC quit)
04:50:53 Join theaftermath [0] (~theafterm@unaffiliated/theaftermath)
04:54:54***Saving seen data "./dancer.seen"
05:05:47 Quit Zarggg (Quit: Zarggg)
05:19:57toehser1I can reproduce the:
05:20:40toehser1Program received signal SIGSEGV, Segmentation fault.
05:20:40toehser10x0000000000498b2e in search (fcache=0x9e2d48 <audiobuffer+964968>, char_code=0, size=8055487,
05:20:40toehser1 p_insertion_point=0x7fffffffd5d8) at /home/tom/rockbox/git-tip/rockbox/firmware/font_cache.c:99
05:20:40toehser199 p = lru_data(&fcache->_lru, fcache->_index[mid]);
05:21:02toehser1I'm willing to believe that the font corruption was just a special case of the same issue.
05:21:26[Saint]Are you able to bisect the issue?
05:22:06toehser1Build from various points seeing which change caused it?
05:22:34[Saint]git has this built in. YEs.
05:23:33toehser1Technically, the answer is, 'yes', I'm 'able' to bisect it...
05:24:47[Saint]Well, I'm assuming you're not doing this all for fun, and you would like the issue fixed.
05:25:01toehser1I have to meditate on the one problem with that... (namely, that the themes I actually use don't have the issue) ...probably my altruism and curiosity will trump my 'hmm, works for me...' mindset.
05:25:23toehser1That is, essentially I _am_ doing it for fun, I sort of like breaking things...
05:26:34toehser1I started paying attention when things broke for me, yes; but you guys have fixed all 3 bugs that affected me... Now I'm just noticing other things.
05:27:48toehser1I'm gradually awakening parts of my brain that remember git and C and gdb. They have been sleeping for a while, since my day job has me immersed in J2EE.
05:28:17[Saint]Well, it would be somewhat nice if you could give back to the community by making us aware of the issues you do find. And given that you're playing with a debugger I am going to hazard a guess that building 3 or 4 binaries isn't going to put you out any.
05:29:08[Saint]Git bisect does ~90% of the work for you anyway.
05:29:14toehser1True enough. The bigger issue right now is that in timezone -5 it is 23:30 and I'm going to bag it for tonight.
05:29:50toehser1Learning git better is necessary for me anyway. The world is using git, not cvs and svn that I'm stuck with at work.
05:30:39toehser1Enough for tonight - I'll do the bisect tomorrow.
05:50:59 Join mudlord [0] (
05:51:15mudlorddoes and its subdomains use Akamai?
05:51:59mudlordthen that doesnt explain my git access issues
05:54:26 Quit [7] (Disconnected by services)
05:54:39 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:55:09mudlordmain reason is because i was interested in tinkering with rockbox
06:00:42mudlordfor instance, i was wondering why the mikmod plugin ain't a codec
06:06:15 Quit mudlord (Remote host closed the connection)
06:14:10 Quit cmhobbs (Ping timeout: 260 seconds)
06:30:37 Quit shamus (Read error: Connection reset by peer)
06:31:08 Join shamus [0] (
06:54:58***Saving seen data "./dancer.seen"
06:59:46 Join Rower [0] (
07:01:27 Nick SuperBrainAK is now known as DormantBrain (~andy@2001:470:8:a61::5f92:59a1)
07:06:37 Nick DormantBrain is now known as SuperBrainAK (~andy@2001:470:8:a61::5f92:59a1)
07:09:54 Nick SuperBrainAK is now known as DormantBrain (~andy@2001:470:8:a61::5f92:59a1)
07:24:39 Quit n17ikh (Ping timeout: 246 seconds)
07:25:48 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
07:28:55 Join saratoga_ [0] (123e1c65@gateway/web/freenode/ip.
07:29:26saratoga_mudlord: (for the logs) because the format isn't streamable and so can't use the buffering engine
07:49:16*[Saint] wonders if saratoga_ is banning again...
07:49:36saratoga_yeah, banned another hosting company
07:58:01 Join fragilematter [0] (~fragilema@unaffiliated/fragilematter)
08:01:41 Join wodz [0] (
08:02:22 Quit saratoga_ (Quit: Page closed)
08:05:10 Quit ps-auxw (Ping timeout: 260 seconds)
08:18:48 Join ender` [0] (
08:20:10 Join einhirn [0] (~Miranda@2001:638:605:4:ec:9e41:1042:2a9f)
08:21:47 Join Strife1989 [0] (
08:23:48 Quit Strife89 (Ping timeout: 248 seconds)
08:26:12 Quit Tomwi (Read error: Connection reset by peer)
08:26:25 Join Tomwi [0] (
08:26:44 Quit Tomwi (Read error: Connection reset by peer)
08:26:56[Saint]You go get 'em, Tiger.
08:28:09wodzI wonder for how long DEBUG builds were broken for ipods unnoticed :-)
08:29:36[Saint]I built one around 2 years ago...
08:30:12[Saint]It doesn't /really/ surprise me it would go unnoticed.
08:30:32[Saint]We don't really have a lot of active development wrt: iPod(s) anymore.
08:30:57wodzis DEBUG and SIM mutually exclusive?
08:31:26 Join Tomwi [0] (
08:31:47[Saint]I don't _think_ so.
08:35:10wodzhmm, so how should one do such build?
08:38:15wodzApparently if you do configure a -> d -> s DEBUG define IS NOT set.
08:39:21[Saint]run configure, select advanced, select debug, pres enter again, select simulator, enter again, build?
08:39:32[Saint]It accepts the value, at least.
08:39:53[Saint]I can't say I ever noticed if it built correctly or not, though.
08:40:15[Saint]And config is *reaaaaaaaaally* non-obvious the way it just sits there waiting on a blank prompt.
08:41:22[Saint] <−−- there we go, it let me configure it.
08:41:36[Saint]How the build turns out may be a different story altogether.
08:43:50wodzIt accepts the value BUT DEBUG define is not set. It is possible to build sdlapp in DEBUG mode and this seems to work
08:45:38 Join lebellium [0] (
08:47:20wodzYeah I am 100% sure DEBUG in sim simply doesn't work. Now the question if this is intended
08:55:00***Saving seen data "./dancer.seen"
08:58:56 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
09:00:51 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
09:10:10 Join kugel [0] (
09:10:10 Quit kugel (Changing host)
09:10:10 Join kugel [0] (~kugel@rockbox/developer/kugel)
09:10:40kugelwodz: sim and DEBUG should work together
09:10:53kugelin fact, originally DEBUG should be automatically defined for SIM
09:11:00kugeli think some commit broke that
09:11:54 Quit bertrik (Read error: Operation timed out)
09:12:54 Quit einhirn (Ping timeout: 245 seconds)
09:19:27 Quit kugel (Ping timeout: 246 seconds)
09:20:36 Quit wodz (Quit: Leaving)
09:21:02 Join petur [0] (5bb7304d@rockbox/developer/petur)
09:27:44JdGordonany complaints with hardcoding the current usb screen layout into the default sbs?
09:28:52JdGordonjust needs a _inbuilt__usblogo_ for the %xl tag
09:32:04JdGordonand new tags for the HID mode and keyboard name
09:34:45 Join kugel [0] (~kugel@rockbox/developer/kugel)
09:35:34kugelJdGordon: yes, let's not do this yet. i would like to stabilize things first
09:36:14kugelwe don't even know for sure whether USB still causes random corruption
09:39:23JdGordonwell, either way, if the sbs is being enabled there is no harm doing it
09:39:41JdGordonI really dont like the viewport_theme_stuff anymore
09:44:30 Quit kugel (Ping timeout: 272 seconds)
10:13:30 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:16:15 Quit fragilematter (Quit: Leaving.)
10:16:42 Join wodz [0] (
10:20:15wodzkugel: buflib integrity checks went in. Thanks for support.
10:20:24gevaertskugel, JdGordon: right now the USB code has no "safe to eject" detection. Last time I tried adding that (back in 2008) I got the distinct impression that various devices and OSes handle this differently, and I couldn't get it to work
10:20:55 Join kugel [0] (~kugel@rockbox/developer/kugel)
10:22:24wodzgevaerts: Yes, that is not standardized. However I'd expect most recentish OSes do the right thing
10:23:48gevaertsWhat *is* the right thing?
10:25:56 Quit kugel (Ping timeout: 245 seconds)
10:26:01pamaurythere isn't really a right thing to do, on most OS the primary goal of "safe to eject" is to flush buffers
10:26:06wodzgevaerts: I linked not so long ago thread from some embedded dev. forum about what is considered 'right way' and what OSes do
10:26:15pamauryonce all data have been sent to the device, you are safe to eject
10:27:06pamauryof course there is the trick of the removable data so you somehow get informed of the state but what use would you make of it ?
10:27:15pamaury*removable media
10:27:41wodzpamaury: have you seen g#725 ? I am not entirely sure it will work as imx233 has slightly different crt0.S then anything else
10:27:43fs-bluebot3Gerrit review #725 at : 3imx233: fix to properly support INIT_ATTR by Marcin Bukat (changes/25/725/1)
10:29:57wodzAnyone feels like testing g#718 on irivers and/or iaudios? Or should I just commit and wait for complaints if any?
10:29:58fs-bluebot3Gerrit review #718 at : 3coldfire: Implement HAVE_INIT_ATTR magic by Marcin Bukat (changes/18/718/2)
10:31:06 Quit fs-bluebot (Remote host closed the connection)
10:31:22 Join fs-bluebot [0] (
10:42:57pamaurywodz: yes I saw, I just didn't have time to touch rockbox stuff since last week end
10:43:07pamauryand it's not a priority :)
10:47:55 Quit pamaury (Ping timeout: 265 seconds)
10:55:02***Saving seen data "./dancer.seen"
11:13:50coppermy themes are getting downvoted again (0/10 and 1/10)
11:14:15coppernot sure if someone's mad at me or if there is something terribly wrong with the themes, in which case I wish someone reported it
11:14:40PurlingNayukiG#726 fixes some translation in chinese-trad, please commit.
11:14:42fs-bluebot3Gerrit review #726 at : 3 by Purling Nayuki (changes/26/726/1)
11:15:42 Quit mc2739 (Ping timeout: 272 seconds)
11:15:52 Join fragilematter [0] (~fragilema@unaffiliated/fragilematter)
11:17:21 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
11:19:17PurlingNayukiDon't worry I just voted it a 10 star
11:19:22copperand no-one ever fixed the rating of
11:19:29copperPurlingNayuki: that's not the point
11:20:16PurlingNayukiWhat I'm trying to say is that there're always people don't like your theme
11:20:31gevaertsPurlingNayuki: in this case it goes further than that
11:20:43PurlingNayukiSystem errors?
11:21:01copperPurlingNayuki: someone cheated the system to downrate those themes with extremely low ratings that were out of bounds
11:21:03gevaertsFor some unexplained reason, *someone* has been downvoting (some of?) copper's themes in creative ways
11:21:33PurlingNayukiCreative ways :(
11:21:46gevaertsYes, negative values and things like that
11:22:00copperI also find it strange how quick "they" were to downrate my themes
11:22:09copperlike they visit the site every day, or… they're here
11:22:23PurlingNayukiCan G726 be committed?
11:22:24fs-bluebot3Gerrit review #726 at : 3Fix chinese-trad translation by Purling Nayuki (changes/26/726/1)
11:22:49PurlingNayukiAny IP address check or other limit when getting rates?
11:23:00gevaertsThe bug that allowed negative ratings should be fixed now, but we should still reset the ratings for the affected themes. I'm not *entirely* sure how to see which ones those are though, although "all of copper's themes" is probably a good start
11:23:25coppermy ratings were reset when I updated all my themes a couple days ago
11:23:25gevaertsI don't want to reset *all* ratings, because we tend to get lots of angry theme designers if we do that
11:23:30gevaertsAh, ok
11:24:00copperbut "the guy" is at it again
11:24:05coppernot sure what he wants
11:24:14PurlingNayukicopper: BTW can you port this to clipzip?
11:24:16copperwould this be the forum to report theme issues?,24.0.html
11:24:33copperPurlingNayuki: someone else could do it
11:25:10PurlingNayukiI was just... being lazy
11:25:53gevaertscopper: yes
11:26:03copperyay for unhelpful thread titles: "Themes don't seem to be working."
11:26:48copperanyway, nothing there about my themes
11:27:38copperoh well
11:27:50copperI guess someone hates my guts
11:27:54copperlet them downrate
11:27:59copperwho cares
11:29:01PurlingNayukiAnd, I think G#412 is also good to commit
11:29:02fs-bluebot3Gerrit review #412 at : 3 by Nick Peskett (changes/12/412/1)
11:29:11gevaertscopper: for the record, there's also no report about any of your themes on the "admin" section (i.e. nobody clicked on "report theme")
11:29:29gevaertsI suspect we'll need to add some logging to the rating code
11:31:02copperyeah that's all I should care about: if there's a bug in my themes, that needs to be fixed
11:33:55copperbut if no-one reports them…
11:36:17copperPurlingNayuki: the Clip Zip is so tiny, I would hardly know what to do
11:36:52pixelmaabout the "ok to remove" for USB earlier, my impression was that there isn't a standardised way. E.g. the cowon loader of my M5 is able to show the notice when on XP but not on Windows 7
11:40:04wodzpixelma: when you are around - could you test g#718 on one of your coldfires please?
11:40:06fs-bluebot3Gerrit review #718 at : 3 by Marcin Bukat (changes/18/718/2)
11:42:38pixelmaI only have one ;) Let's see
11:43:33PurlingNayukiI Just sold my codefire :(
11:51:58pixelmawodz: could you compile a test build for me? That would be an advanced build for the M5 with radio mod. Otherwise I think I won't have time for this before the weekend
11:52:57wodzok, I'll do the build
12:05:26[Saint]copper: ...just a thought, but, have you considered that the same person(s?) may be upvoting their own themes in the same ridiculous fashion?
12:05:55copperyes, that has crossed my mind
12:06:00[Saint]Look for some themes with suspiciously high ratings (in the same device class, as I assume this is "competition"), and low download counts.
12:06:25copperI'm not sure that's a foolproof metric
12:06:36[Saint]Oh, definitely not.
12:06:38copperdownloading and rating are two completely different behaviors
12:06:53[Saint]I make two assumptions, which are /likely, but not foolproof:
12:07:06copperif a theme is polarizing, it can attract lots of high ratings while getting comparatively fewer downloads
12:07:11[Saint]1 - the author is maintaining a competing theme actively, look for recent themes.
12:07:23[Saint]2 - the author won;t be able to resist upvoting their own shit
12:08:26[Saint]From that, however flawed it may be, it may be possible to produce a few likely candidates.
12:09:08[Saint]...TBH, I really wish I could match the IP that issued the rating against the form user ID they almost certainly have, and then ban them.
12:09:21[Saint]Because, frankly, this type of shit is terrible.
12:09:38[Saint]s/form user/forum user/
12:10:21copperthe weird thing is: it's not all my themes that get aggressively downrated
12:10:29coppermy name is shown on all of them
12:11:13copperso maybe it's just someone who really doesn't like the one or two themes, or didn't manage to make them work on their setup, or something like that
12:11:32copperor it's someone just messing with me
12:12:22[Saint]I sincerely doubt its personal.
12:12:34[Saint]Well, anywhere past "that guy makes a theme".
12:12:49[Saint](or several, as it were)
12:13:14copperboth highly rated, no downrating that I can see
12:13:43copperat this point I just want to know if there is something terribly wrong with the Googley themes
12:14:00coppersome use case that I won't run again because my usage is completely different
12:14:07[Saint]Its really not hard to associate your name with your various other handles, and you don't get any abuse via any Rb channels (mailing list, forums, IRC, etc.) do you?
12:14:28copper[Saint]: I did on IRC
12:15:05copperand my point was, the name showing on my Googley theme is the same as the one showing on my other themes
12:15:18copperso they could have downrated my other themes, but they didn't
12:15:27[Saint]I would hate to think that anyone in this community is that much of a dick.
12:15:31[Saint]But its always possible.
12:15:56copperI should have added something in the theme description to encourage users to report theme issues
12:16:04[Saint]Its well and good to just not like somebody. Can't get on with everyone. But...this is silly.
12:16:21[Saint]I sincerely hope its not strictly personal.
12:16:28copperbecause as it is now, I'm not sure every theme users knows how to report theme issues
12:16:35copperor where
12:17:10[Saint]That's largely due to the fact that no such place exists, I guess.
12:17:44[Saint]I would recommend creating a forum thread, for all Googly variants, or possibly all your themes, period. ANd directing users to report issues there.
12:17:44coppergevaerts: I noticed only now that the report form isn't just for copyright issues, but also "does not work" issues
12:18:22copperit's way down the page though, maybe users don't even see it
12:18:31[Saint]I'm not sure that's useful to authors, though.
12:18:40[Saint]"does not work" is kinda...vague.
12:19:03[Saint](and you wouldn't see it anyway, I don't think)
12:19:03coppera rating of 0 is definitely a "does not work" issue :P
12:19:26copperI know, I'm just replying to gevaerts who mentionned that there were no such reports for my thems
12:19:32[Saint]No. Its a "ratings are quite obviously broken, and users are assholes, or at least one is" issue. :)
12:19:47[Saint]Well, gevaerts is technically correct.
12:20:02[Saint]That "does not work" is of absolutely no use to you for this purpose.
12:20:11[Saint]Nor would you see it if anyone clicked it (I think).
12:20:59copperbut surely someone here would have told me
12:21:21[Saint]If it was one of a few select people, maybe.
12:21:33[Saint]It usually just gets hidden, and you'll get an email.
12:21:58[Saint]A very nonspecific one, I believe, just saying what happened, and not why.
12:23:05 Join Strife89|lappy [0] (
12:24:34[Saint]I think there's only a very limited number of people that would be able to connect your real name and handle immediately.
12:25:18copperwhat's your point?
12:25:18[Saint]You would most certainly find out if your theme was removed. But I'm not sure how personal it would be.
12:25:58coppersorry I don't understand what my nick and real name have to do with it?
12:26:16 Quit Strife1989 (Ping timeout: 272 seconds)
12:26:56[Saint]Sorry. When you said you assumed someone would tell you about it, I thought you meant perhaps seperately from the action being taken.
12:27:06[Saint]Like a "Hey, copper, just so you know...this happened"
12:27:20copperno I meant someone would tell me if my theme got a report from the report from
12:27:23[Saint]As opposed to an email one day telling you a theme was removed because it fails to parse or whatever.
12:27:28copperah ok
12:27:59[Saint]Sorry. I messed that up. Yes. You would be notified.
12:28:36gevaertsNot automatically
12:28:45 Join ildarion [0] (4ef2d722@gateway/web/freenode/ip.
12:29:02[Saint]Eeek. I thought I said that. Gah.
12:30:04ildarionHey !
12:30:21[Saint]When I was still serious about maintaining themes, my tactic was to use a forum thread in the theme section to both promote the theme(s) and facilitate user feedback.
12:30:29[Saint]You should probably consider this.
12:30:40[Saint]But I relize its a PITA to update `20 themes.
12:30:51[Saint]especially when its not a functional change.
12:30:59copperactually I've done that on ABI
12:31:31copperthat's where I'm active, that's where people know me, and that's where DAP users hang out
12:31:51copperbut, there is no link to that thread in my theme descriptions
12:32:10[Saint]Right. But then bug reports go there - and you become the middle-man for passing bug reports between ABI and us.
12:32:19[Saint]Further fragmenting the userbase.
12:32:55*[Saint] isn;t particularly fond of Rockbox development stuff ending up in ABI, but understands why it happens.
12:33:25[Saint]The main problem is that we can't fight disinformation effectively.
12:33:36copperI understand
12:33:53[Saint]And people start to think of it as an official source, and can't tell between binaries from Jow Random there, and official ones from "us".
12:34:17copperyeah I remember the iPod Classic builds shenanigans on Head-Fi
12:34:43[Saint]But there are a few of us that hang out in both camps, like yourself, and saratoga, and soap(?), that do a great job pointing users here from there when necessary.
12:34:52[Saint]But they can't be everywhere at once. Sadly.
12:36:31copperI really wish I could upload my * archives on the theme site
12:36:56ildarionCan I ask for little advice about sansa fuze + ?
12:37:08coppernot only is updating 14 themes annoying and time consuming, it also takes up a lot of real estate on the theme site
12:37:10[Saint]I _think_ Ben Brown, or Alex Mayer did some preliminary work in that regard.
12:37:16[Saint]It was definitely on the ToDo.
12:37:25[Saint](its a very sane thing to expect)
12:37:40[Saint]ildarion: You certainly can.
12:37:41copperildarion: go ahead and ask your question
12:38:46[Saint]copper: It likely wouldn't be too hard to parse the .zip file for "sub .zip" files, and treat those as separate uploads.
12:39:14[Saint]But in that case everything would need to share the same name, and description (well, not really, but it would be easier).
12:39:17copperthat wouldn't allow for different theme descriptions
12:39:34coppermy themes are really just one meta-theme
12:39:44copperdifferent colors for different tastes
12:40:47[Saint]It would be bloody nice if you could just select the primary color with a color picker widget plugin thingo, and have a single theme.
12:41:01copperyeah I've thought about that too
12:41:01[Saint]Or if it adapted according to the foreground color set.
12:41:13[Saint](and set a different font color via hardcoding in the theme)
12:41:26ildarionOk, so my sansa Fuze + is deap corrumpt, I read i can fix this with the "Recovery mode ", I search the software tool for acces to the device in "ROM Recovery":
12:41:52copperbut the color picker allows for setting RGB colors
12:41:54[Saint]like, if the fg color == <specifi value>; load <set_of_colored_images>
12:41:58coppertoo many combinations
12:42:27[Saint]That would make for one *massive* theme, though.
12:42:33[Saint]...but it sure would be a cool idea.
12:42:42[Saint](and technically possible, if I'm not mistaken)
12:43:17coppernot really practical
12:43:40copperthe simplest way is really just selecting a different .cfg file, which is how it is now
12:44:05coppera "smart" meta theme would only increase the complexity of the markup
12:44:23[Saint]Dramatically so.
12:44:33[Saint]...but I like fun things like that. :)
12:44:50copperildarion: what do you mean by "deap corrumpt"?
12:44:52wodzildarion: You need to catch pamaury
12:46:45ildarioncopper, i formated, and the OF is corrupt, impossible to boot on him : "Database error please format the device !", "Sansa Will now shutdown."
12:47:07ildarionSo impossible to install a fresh bootloader
12:50:45ildarionI want to try this first
12:50:56[Saint]copper: getting pretty out there only have a few, solid colored elements that differ in each "Googly", no? What if we generated those colors using viewport background colors, using an overlay with a negative transparent cutout section to provide icons?
12:51:12[Saint]Then Googly could be any valid RGB combo.
12:52:09[Saint]We can abuse draw order by drawing directly into the backdrop buffer, then drawing an overlay on top with a transparent relif section to provide colored icons.
12:52:47[Saint]<−− Muhfuggin' theme-philosiphising-ninja right here, yo
12:53:02copperthere's no alpha channel support, is there?
12:53:13[Saint]Indeed there is.
12:53:46copperactual alpha channel support, not just transparency?
12:53:59*[Saint] nods
12:54:44[Saint]see, for instance, stock color iconset.
12:55:04***Saving seen data "./dancer.seen"
12:55:47*[Saint] wishes to ad that he has no idea how feasible relying on draw order may be for certain operations, but the idea sure sounds bad ass.
12:56:11[Saint]*abuse of draw oder that doesn't technically exist
12:58:29[Saint]PS: I apologize in advance, profusely, for any crazy theme related pursuits this conversation may or may not cause. :)
13:01:32coppersince when does Rockbox support 32 bit BMP with alpha channel?
13:01:49 Quit Zagor (Ping timeout: 252 seconds)
13:01:57[Saint]Oooohhhh...2 or 3 years now I think.
13:02:10[Saint]Perhaps closer to 2? I forget.
13:02:27[Saint]"a while" covers it.
13:02:32copperthere's the matter of font color though
13:03:08[Saint]that's FG, we would only be dealing in BG.
13:03:14ildarionpamaury is regularly online or is better to send him a mail ?
13:03:19copperit's white with an alpha channel of 80% on my themes
13:03:31[Saint]Ooooh. Right.
13:03:52wodzildarion: he is regular online albeit he is very busy
13:04:01copperstraight up #ffffff would pop too much
13:04:08*[Saint] nods
13:04:55copperanother problem is usability
13:05:14copperit would be difficult for users to use the color picker to set the theme color
13:05:35copperunless the theme color gets updated as the user changes values, in real time
13:05:47[Saint]If they set a bullshit value, too dark, or zero contrast with the font, things could get really fun. Yeah.
13:06:13copperand, I suspect a lot of users don't really want to have to search for the right color themselves
13:06:23copperthey'd rather have a few "presets" to choose from
13:06:25 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
13:06:51[Saint]Perhaps this type of theme is better as a philosophical excercise, but dammit, I sure am tempted to try it to see exactly how possible it is.
13:08:30[Saint]I am currently trying to get my head around a nice tabbed UI system with touchscreen (I have been for ages actually, its something I pick up from time to time), and that has been getting my little ROckbox attention time lately.
13:10:02[Saint]One of the problems I have is that I'm really not an artist. Not a digital one, at least. I can create the backbone of great things, but when it comes to giving them a face to greet the public, I spew forth shite.
13:27:15pixelmasomeone "stole" my theme idea... except I haven't talked about it yet
13:30:14 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
13:30:15[Saint]pixelma: me?
13:30:24[Saint]Like, just now?
13:31:13pixelmayes and also " ;) " I also had the idea about a graphical theme adapting to the user's foreground and background colour settings. If you are philosophical we already have one that does just that - which is the fallback one
13:32:04[Saint]Oh, I knew you weren't being entirely serious. Don't worry. No smiley required. :)
13:32:18copperhey, we've ALL had the idea :P
13:32:19[Saint]If nothing else it proves that the idea isn't entirely insane.
13:32:30[Saint]Or, we're both insane...
13:32:48copperyou don't think I thought of that 6 months ago when doing my theme variants? :P
13:33:12[Saint]It doesn't matter. No one will believe you now. :P
13:34:53copperI thought about it, but didn't get into implementing it, because I believed what I still believe: it's not very practical
13:35:20[Saint]On the contrary. I believe its quite practical indeed.
13:36:48[Saint]I can foresee several ways foreground, background, and the use of relief cutting "negative" overlays could be (ab)used to create a striking effect.
13:38:27[Saint]I would be happy to help you brainstorm and generally fart around on such an experiment.
13:40:01 Join kugel [0] (~kugel@
13:40:01 Quit kugel (Changing host)
13:40:01 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:40:45[Saint]kugel: toehser1 found some segfault shennanigans you might be vaguely interested in
13:40:55[Saint]it seems to gel with your recent work
13:41:47[Saint](approximately 14 hours ago in the logs)
13:41:54copperI meant from a user's point of view
13:42:25kugellook at git log: :-)
13:43:46kugelfwiw, we could define a magic color for bmps that means use the foreground color
13:44:45kugelwe already support that for non-alpha bitmaps actually volute it's not yet implemented for those with alpha channel
13:44:46copperwhat about text alpha channel
13:45:22[Saint]a what now?
13:46:46pixelmawe already had that for a while and that was "magic" cyan
13:47:20[Saint]for text?
13:47:34pixelmamagic colour for BMPs which meant "foreground"
13:47:41[Saint]Ah. Right. Yes.
13:48:06pixelmadidn't work with alpha channel last I tried though :P
13:48:26pixelmathat's been a while, I admit
13:48:38[Saint]kugel just said that a few minutes ago. :)
13:49:10pixelmatrue, should have read that up to the end
13:49:43wodzpixelma: <- M5 build with FM mod
13:49:47coppertext alpha channel, like CSS: RGBA
13:49:55copperred, green, blue, alpha
13:50:12coppere.g. "white with 80% transparency"
13:51:05[Saint]that is possible, in theory.
13:51:14[Saint]perhaps even trivial.
13:51:33copperI'm sure it's possible, I'm asking if it's already implemented ;)
13:51:59[Saint]If it was, I wouldn;t be talking about the possibility. :)
13:52:13kugelwhat are you asking for? it confuses me that you say text
13:53:57pixelmawouldn't that already be possibly if you created a font file that way?
13:55:27PurlingNayukiI was just looking at Onda VX898s :(
13:55:30kugelyou want text to be, say, 50% transparent?
13:55:34[Saint]pixelma: I imagine convttf would strip anyhting it didn't understand out.
13:57:11kugelwell, that's certainly possible, though it doesn't seem useful to me
13:57:28copper[Saint]: how do you draw a background rectangle with text on top of it? %VB() %dr()?
13:57:44kugelyes you could potentially craft fonts this way but it'snot ideal
13:57:57copperkugel: I would need it for my themes if I made them use fg and bg colors
13:58:19kugelwhy that?
13:58:22[Saint]copper: technically speaking, it wouldn;t /need/ to be drawn to the backdrop buffer.
13:58:34[Saint]The text would just need to come after the rectangle.
13:58:49copperkugel: because my top bar text isn't 100% white, it's 80% white
13:58:51[Saint]draw order is dictated by order parsed.
14:00:17copperkugel: in order to get the right text color, I need alpha channel support, or else I need to set a custom text color that goes with the background color
14:00:28kugelbut there is no text that's 80%
14:00:36copperwith the second solution, it's not possible to have the user set their own fg and bg colors
14:00:50copperkugel: yes there is
14:01:00copperthe volume and battery indicators
14:01:30copperin the top bar
14:01:31kugelhow do you make it 80%? using suitable vp colors?
14:01:51copperwait, what's vp
14:02:49copperI flaten my Gimp image and pick the resulting color, and use that color with %Vf() in my markup
14:03:18copperwhich is the equivalent of (white, 80%) but as RGB, without an alpha channel
14:03:34copperagainst the background color
14:03:56kugeli see
14:03:59coppera light shade of gray wouldn't really work
14:04:05coppervisually speaking
14:04:19copperthe nuance is slight, but it does make a difference
14:04:30copper(between 80% white and 100% white, I mean)
14:04:57copperanyway, there's still the issue for users to find the right color for them
14:05:07copperit's difficult enough on a real PC with The GIMP or Photoshop
14:05:42[Saint]the fg/bg color pickers have a realtime view.
14:07:07copperok, it's similarly difficult then :P
14:07:11kugelthere are other problems too, like gradients that can look shit because images aren't dithered (unlike the backdrop)
14:07:52copperone of the points of a "product" is to make sensible choices
14:08:19[Saint]Lucky we're not making a product then. ;)
14:08:25coppermy themes are products
14:08:43copperregardless of what you think Rockbox is ;)
14:09:19wodza box which rocks obviously
14:09:55copperbut, alpha channel support could make my themes lighter on bitmaps, and perhaps more easily modifiable (as far as colors go)
14:10:18coppersince I'm not using gradiants or non-rectangular shapes
14:10:22wodzpixelma: any estimated timeframe for the test on M5?
14:10:24[Saint]wodz: Hmmm, I thought it was...what was it, "a bag of shit"?
14:10:56wodz[Saint]: sounds like, yes :-P
14:11:23[Saint]That's something we can all be proud of. :D
14:14:52toehser1Personally, I would rather have a 'basic theme language' (the existing WPS format) and an 'advanced theme language' (based on Lua) with the ability to add new settings to a user settings menu and to implement more advanced things - in concept - to allow 'meta-themes' that can work for many targets or color choices... in concept... But given that my themes already just 'use the system colors', I don't care that much, and it would be a lot of work.
14:15:08toehser1Did I see something implying the bug I found is already fixed?
14:17:39 Quit kugel (Ping timeout: 246 seconds)
14:20:04 Join cmhobbs [0] (~cmhobbs@fsf/member/cmhobbs)
14:23:48[Saint]toehser1: you did.
14:26:24copper[Saint]: I can't display a rectangle below text
14:26:38PurlingNayukiCool, I just noticed that Clip Zip can now mount when being connected to compter
14:26:53coppercan't make it work: either the rectangle doesn't show, or nothing shows
14:27:48[Saint]I'm about 80% certain that /used/ to work.
14:29:02copperI tried putting the rectangle viewport before and after the text viewports
14:30:01[Saint]Ooooohhhh, fuck, of course.
14:30:15[Saint]the text viewport is bringing its own BG to the party.
14:30:47[Saint]And there's no way I'm aware of to specify a transparent viewport BG.
14:31:10 Quit Rower (Ping timeout: 260 seconds)
14:31:22[Saint]Dammit. That was dense of me. ANd rather short sighted.
14:32:27[Saint]Oh, well, there *is* a way to specify a transparent viewport BG, but its only when there's a fullscreen backdrop loaded.
14:32:45[Saint]And its not on an individual basis. Its all of 'em.
14:32:48copperthat's the case with my themes
14:33:08copperthey all have a fullscreen backdrop
14:33:13[Saint]Oh? There is a fullscreen backdrop? Hmmm.
14:33:35[Saint]Then yeah. As far as I'm aware this /should/ work.
14:34:24 Join einhirn [0] (
14:34:31 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:34:46 Join nick_p [0] (
14:35:22[Saint]No, megenta isn't a "magic" color for viewport foreground/background.
14:35:34[Saint]Just bitmaps.
14:36:18copperuh, I can't even make %Vb() work at all
14:36:49nick_pPurlingNayuki: If you've tested G#412, please update it with your findings
14:36:50fs-bluebot3Gerrit review #412 at : 3Option to use the lighter hinting algorithm (FT_LOAD_TARGET_LIGHT). by Nick Peskett (changes/12/412/1)
14:36:53[Saint]You need to specify a fullscreen viewport before any other viewport declarations.
14:37:14 Quit cmhobbs (Ping timeout: 260 seconds)
14:37:30PurlingNayukinick_p: What findings?
14:37:40PurlingNayukiI just find it works.
14:38:09[Saint]"I found it works" ;p
14:38:33nick_pPurlingNayuki: Well that'd be good in there, you don't have to be a dev to review patches
14:38:35kugelcopper: %Vb works only without backdrop
14:39:06[Saint]Ah, crapballs.
14:39:20[Saint]I confused %Vb with the "draw to backdrop buffer" tag.
14:39:43PurlingNayukiLight hinting is useful for small size asian characters so I think that will be useful.
14:40:43PurlingNayukiThat works, just means It's fine with me to commit this so more people can get to know it
14:40:54 Join amayer [0] (
14:44:15nick_pPurlingNayuki: I meant, insted of commenting here, review the patch and put your comments there, for posterity ;)
14:45:06 Join pretty_function [0] (~sigBART@
14:45:19PurlingNayukiI wasn't on computer so I didn't;-)
14:47:06kugelnick_p: do you have a before-after pic?
14:47:08PurlingNayukiDon't worry I know how to gerrit;-)
14:48:38nick_pkugel: Not at hand, but I suppose I could put one together a bit later...
14:52:59copperkugel: ugh, is that written anywhere?
14:54:42copperalso, "Viewports cannot be layered transparently over one another. Subsequent viewports will be drawn over any other viewports already drawn onto that area of the screen."
14:55:08***Saving seen data "./dancer.seen"
14:55:34copperso that's that
14:56:26toehser1What are all the SIGUSR1s that I have to handle nostop?
14:56:32copperforget about alpha channel, there's not even basic support for transparent layering
14:57:35toehser1Haven't managed to segfault it again so far. Seems fixed. I'll have to find another way to break it :)
15:10:10kugelit looks more blurry to me
15:10:49nick_pI was expecting bigger, not blurry!
15:10:55kugeland obviously more space between characters but there are already options to achieve that
15:11:25nick_pWhatever, works for me
15:11:30PurlingNayukiAnd me
15:12:19copperkerning is a bigger problem IMO
15:12:37kugelthe character spacing is horrible in the before case but apart from that i find the before case better
15:14:12nick_pNo-one's forcing you to add the -L option when you convert fonts, if you don't you get the results you would have without the patch applied
15:14:21kugelthe before isn't really bold, as intended, perhaps that's the problem
15:15:11kugelconsidering it should be bold the after is indeed better
15:15:21nick_pI prefer more anti-aliasing, which the lighter hinter algorithm seems to provide...
15:15:40[Saint][02:54:44] <copper> also, "Viewports cannot be layered transparently over one another. Subsequent viewports will be drawn over any other viewports already drawn onto that area of the screen."
15:15:46[Saint]tha's actually bullshit.
15:16:35[Saint]Layering is certainly possible. Unpredictable, but possible.
15:17:23[Saint]...well, even that's not true.
15:17:30[Saint]Its /reasonably/ predictable.
15:17:43wodzkugel: do you know the revision you are certain DEBUG was possible in SIM build?
15:18:56copper[Saint]: feex eet!
15:19:13 Quit wodz (Quit: Leaving)
15:22:28kugelwodz: i tried this combo this morning and it worked
15:22:48kugeljust select a,s,d in configure
15:23:06pixelmahmm, would be cool if there was a way to use a third (or even fourth) user selectable colour out of the line selector colours. While %St(line selector start color) would give me the value there doesn't seem to be a way to use that in viewport settings (a .wps with %St there seems to get rejected, guess it's not allowed to have other theme tags in %Vf etc.)
15:24:09 Quit kugel (Quit: Yaaic - Yet another Android IRC client -
15:24:15 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
15:28:30 Quit pretty_function (Remote host closed the connection)
15:30:09 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
15:48:54[Saint]No, the %V* tags can't be used that way unfortunately.
15:59:33 Join mudlord [0] (
16:01:33mudlordsaratoga: read your response in the logs
16:01:46mudlordwhich is odd, because otherwise SID won't be a codec
16:01:53mudlordunless I am missing something
16:03:23mudlordsame with the current MOD codec
16:07:09 Join pretty_function [0] (~sigBART@
16:08:44toehser1For both multi-color-theme themes and multi-target themes, isn't what we want some separation of the markup from the styles? And, wouldn't it be nice to be able to apply author-created dynamic behavior? It seems that some analog could be made of wps(markup) wps(styles) wps(lua) the way a browser does html/css/javascript... A theme could have 50 'stylesheets', of which the user can pick from the ones that match the current target... The viewports are a
16:10:17 Quit nick_p (Quit: Leaving)
16:15:33 Join tertu [0] (~tertu@
16:20:03 Quit [Saint] (Remote host closed the connection)
16:21:52 Join [Saint] [0] (~saint@rockbox/staff/saint)
16:23:54 Quit ildarion (Quit: Page closed)
16:24:16 Join ildarion [0] (4ef2d722@gateway/web/freenode/ip.
16:24:58ildarionHey pamaury ! Are you busy ?
16:27:03 Quit Strife89|lappy (Ping timeout: 252 seconds)
16:29:32pamauryildarion: ping, somewhat, I can follow a discussion
16:35:14 Quit pretty_function (Remote host closed the connection)
16:36:36ildarionpamaury: When I ask for help, they told me to do with you. My Sansa Fuze + system is deep corrupt. I formated serveral times buts don't work, when I start on the OF, I get : "Database error please format the device !", "Sansa Will now shutdown.". So I can't reinstall with a good bootloader. I see some light here : Where I can get the "tool" ? Idon't find it.
16:38:04pamauryah the tool was renamed, it's known as sbload, in our trunk in utils/imxtools/sbtools
16:38:30pamauryhowever if you reformated and I still get errors I don't really know what else you could do
16:38:40pamaurymaybe the internal storage is broken
16:41:00pamaurydoes the internal storage actually works ? Did you try to put some file on it, unplug, replug and check if the file is still there ?
16:43:02ildarionThe problem appeared when I formatted Sansa to install over the "windows 7 installer utility" (I had no other usb key), using my backup of Rockbox, it works but not the OF
16:43:22ildarionSo I don't think there are an hardware issue
16:43:58pamauryah, it probably messed with the partition table
16:44:22pamaurydo you have a Linux at hand ?
16:45:00mudlordspeaking of bugs, i noticed a heisenbug the other day with my sansa clip zip on a git revision of rockbox
16:45:28mudlordunable to reproduce it again, but it caused a lockup
16:45:52mudlordi was browsing the database, which had plenty of entries, and then it locked up
16:46:55pamaurymudlord: if you don't have any more info or a way to reproduce it, it's pretty much useless
16:47:46mudlordas I said, heisenbug
16:48:46toehser1There was a lockup-causing issue fixed in the last few days, so, maybe it was that one.
16:49:23mudlordso maybe its time to update again :)
16:49:49ildarionpamaury: I have an old laptop on lubuntu
16:51:00pamauryildarion: could you plug your device and give me the output of dmesg ?
16:51:42pamauryildarion: and could you do so both with Rockbox USB mode and Sansa OF USB mode ?
16:54:12toehser1Old lubuntu laptop is fine, it will have things like fdisk and mkdosfs and lsscsi and lsusb.
16:55:11***Saving seen data "./dancer.seen"
16:59:42ildarionI think i was doing the good command, what do you exatly in the results ?
16:59:58ildarionexatly want*
17:00:00pamauryildarion: it's an issue of logical sector size
17:00:13pamauryI think, so I want to check that
17:00:45 Quit fragilematter (Quit: Leaving.)
17:02:47 Join Baranko [0] (
17:03:44Barankoguys, maybe someone can explain some feelings on rockbox
17:03:52Barankowhlie switch tracks
17:04:11Barankoevery new track on first swc i hear some sounds
17:04:19Barankolike mouse somewhere
17:05:46Barankoalso you may hear this while moving pc mouse, and headphones connected to soundcards
17:06:05Barankoon original fw fpr clip+ it's almost none
17:06:48Barankomaybe there is some tip for missing first second on track?
17:07:06Barankoor this sounds just by design?
17:07:40gevaertsThat shouldn't happen
17:08:51Barankosome guys on russian forums suggest that it's some interference with sd reader in sansa
17:09:08Barankois there any workaround?
17:09:27 Join ildarion_laptop [0] (4ef2d722@gateway/web/freenode/ip.
17:10:14Barankoand i also made some tests with dev builds:
17:10:25Barankois there anything alse that i can do for project?
17:10:33ildarion_laptopOn rockbox : 361.709151] usb 1-1: New USB device found, idVendor=0781, idProduct=74e1 [ 361.709162] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 [ 361.709168] usb 1-1: Product: Rockbox media player [ 361.709173] usb 1-1: Manufacturer: [ 361.709178] usb 1-1: SerialNumber: 10000000100001B400002DCDF27F49B40
17:11:00pamauryildarion: pastebin please, I need the part about mass storage mostly
17:11:04 Quit kugel_ (Ping timeout: 252 seconds)
17:11:06pamauryildarion_laptop: ^
17:13:05PurlingNayukiG#727. Good night guys :)
17:13:06fs-bluebot3Gerrit review #727 at : 3Fix incorrect language string reference in eq_menu.c by Purling Nayuki (changes/27/727/1)
17:19:44ildarion_laptoppamaury: on rockbox and ROM Recovery :
17:26:16pamauryildarion_laptop: I don't need the ROM recovery but the Sansa OF USB mode (ie hold volume down when plugging)
17:27:11 Quit ildarion_laptop (Ping timeout: 272 seconds)
17:27:15pamauryildarion_laptop: don't you get more info ? the dmesg output is clearly missing some mass storage bits here
17:27:55pamauryildarion: ^
17:28:29pamauryyou need to wait a bit once you plugged the device for mass storage to load
17:28:52ildarionmy command is: dmesg | grep -i usb this is correct ?
17:29:32pamauryno, I need all parts, grep for usb is too strong
17:30:07pamaurywhen you plug a mass storage device, some part are logged by usb and some by other subsystems like scsi
17:31:19toehser1Won't lsusb and lsscsi give you what you want without having to fish it out of the dmesg?
17:31:29pamauryI'm not so sure
17:32:20pamauryplus lsscsi is non standard
17:32:40pamaurywhereas the logical sector size is printed by the kernel in dmesg
17:35:11pamauryildarion: pastebin the whole dmesg output in doubt
17:35:37ildarionyeah, i just find the good command^^
17:37:19 Quit petur (Ping timeout: 272 seconds)
17:37:56 Join krabador [0] (~krabador@unaffiliated/krabador)
17:38:05 Quit Zagor (Read error: Connection reset by peer)
17:41:50 Quit ildarion (Quit: Page closed)
17:45:09 Join ildarion2 [0] (4ef2d722@gateway/web/freenode/ip.
17:51:04ildarion2pamaury: so on rockbox:
17:55:15ildarion2On sansa :
17:56:15ildarion2I'm not sur the sansa plug really work
17:56:24pixelmawodz (logs): I wanted to have a look today in the evening or tomorrow. Is there anything specific I should try out?
17:57:34 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
17:58:18pamauryildarion2: are you sure it was sansa OF in the second log ?
17:58:28pamauryit seems like it's rockbox again
18:00:19pamauryyou need to hold volume down when plugging USB to boot into Sansa OF
18:00:38ildarion2I know. I think on sansa plug doesn't work
18:00:48pamaurywhat does it do ?
18:01:34ildarion2I need to keep puching volum down or i can releas it after few secs ?
18:01:58pamauryyou need to keep for a few second
18:02:30pamauryto be clear, the procedure is: power down the device, hold volume down (device is still off), plug usb and keep holding volume down for a few seconds, then release
18:03:25ildarion2when i do that, once volum down release, rockbox USB screen pop
18:05:45pamauryI see, that's annoying, the problem is the logical sector size here. I think there is a way to solve it by editing the disk by hand, but it's tricky. Otherwise one would need to convince the kernel to change the logical sector size
18:11:48pamauryildarion2: maybe we can try something by hand, but you need to follow my instruction exactly
18:12:08pamauryfirst, can you pastebin the output of fdisk (ask it to print the partition layout)
18:12:55pamauryand select sector unit !
18:13:05pamauryildarion2: ^
18:14:22pamaury(I think sector is the default in most version of fdisk but we never know)
18:21:00mudlordoh, nice to see those usb bugs are being worked on :)
18:22:35 Join pretty_function [0] (~sigBART@
18:22:52 Quit Scromple (Read error: Connection reset by peer)
18:23:01 Join Scromple [0] (~Simon@
18:24:19pamaurymudlord: technically speaking these are not usb bugs, those are theme bugs ;)
18:25:58pamauryildarion2: still here ?
18:26:35ildarion2yes, I'm learning linux command in the smae time
18:27:03 Quit Scromple (Read error: Connection reset by peer)
18:27:12pamauryI think in this case it's "fdisk" and then hit "p"
18:27:16mudlordTheme bugs causing crashes? xD
18:27:22 Join Scromple [0] (~Simon@
18:27:37pamaurymudlord: yes, memory corruption
18:27:50mudlordnow that makes sense
18:29:09pamauryildarion2: sorry, you need to specify the disk, which is probably *not* /dev/sda in this case, more likely /dev/sdb (check the output of dmesg). So run "fdisk /dev/sdb"
18:30:11 Quit Scromple (Read error: Connection reset by peer)
18:30:40ildarion2Ok, and after tip P ?
18:31:05pamauryyes, and pastebin the output
18:31:51 Join Scromple [0] (~Simon@
18:34:30pamaurylooks weird, /dev/sdb is probably not the device, can you post the output of dmesg to check which drive is your device ?
18:35:03 Quit krabador (Quit: Sto andando via)
18:35:49ildarion2seem to be my device, 16GO is my sansa default memory
18:36:13pamauryok, it has weird partition
18:37:30pamauryno wonder the OF doesn't work with this
18:38:03pamauryso now, run again "fdisk /dev/sdb"
18:38:05 Part theaftermath ("Leaving")
18:38:52pamaurypress "d" (for delete)
18:39:00pamauryit should ask for which partition, select "1"
18:39:15pamaurythen print "p" again and pastebin the output (don't quit the tool !)
18:42:47gevaertsI think you need to clear the MBR first on that one. It's seen as a sun disklabel partition table, and that's not going to work
18:43:30pamauryyeah clearly, hit "o"
18:44:02pamaury(this will clear the table)
18:44:09pamaurythen "n" (for new), then "p" (for primary)
18:44:19pamaurythen "1" for first entry
18:44:29pamauryfor the first sector, type "2048"
18:44:48pamauryfor last sector just hit enter
18:45:28pamaurythen hit "t" to change label, then "c" (for W95 FAT32 (LBA))
18:45:41pamaurythen hit "p" and pastebin the output
18:46:40ildarion2label is the name ? (so no important)
18:47:01gevaertsIt's the partition type. If that's wrong, expect problems
18:47:39gevaertsWell, linux won't care, but rockbox will, windows will, and the OF might
18:48:32ildarion2So "t" => Code Hexa: "c",...
18:49:08 Nick DormantBrain is now known as SuperBrainAK (~andy@2001:470:8:a61::5f92:59a1)
18:51:03pamaurygood, now type "w" to write and exit
18:51:22 Join b0hoon [0] (
18:52:21pamauryildarion2: then run "mkdos.fs /dev/sdb1"
18:53:34mudlordreinstalling rockbawx :3
18:53:44mudlordupdating to newest fw
18:53:58mudlordthis theme fix sounds good
18:54:07mudlordand i found a workaround to my net problems
18:54:15ildarion2after i was hit "w" i get this wierd answer
18:54:22ildarion2Sorry this is in french
18:54:37ildarion2this is a regular answer or ?
18:54:50gevaertsildarion2: in that case, first unplug and replug
18:55:15***Saving seen data "./dancer.seen"
18:55:28 Quit K1773R (Ping timeout: 252 seconds)
18:55:29pamauryyeah unplug and replug and only after run the mkdosfs command
18:56:01 Quit AnLe (Ping timeout: 252 seconds)
18:56:25 Join AnLe [0] (
18:56:34 Quit Karrde (Ping timeout: 252 seconds)
18:58:09 Join Karrde [0] (
18:58:33pamauryildarion2: is it done ?
18:58:55ildarion2"command mkdos.fs not find" maybe mkdosfr ?
18:58:57 Join rela [0] (~x@pdpc/supporter/active/rela)
18:59:21pamauryyeah there is no dot , sorry
18:59:34 Join krabador [0] (~krabador@unaffiliated/krabador)
19:01:03ildarion2seem to have been done
19:01:30ildarion2it was supposed to do what ?
19:01:33pamauryok, first thing first, can you try to unplug the device and boot the OF ?
19:01:53pamaury(normally you won't be able to run rockbox because it won't find the firmware)
19:01:58 Join K1773R [0] (~K1773R@unaffiliated/k1773r)
19:02:22ildarion2same issue
19:02:38ildarion2"Database error please format the device !", "Sansa Will now shutdown."
19:02:41pamauryok, replug in rockbox usb mode
19:02:50pamauryI was expecting the error but wanted to make sure
19:03:14pamauryis it plugged ?
19:03:36pamaurynow for the weird stuff, run "fdisk -b 4096 /dev/sdb"
19:04:33pamaurythen "d", then "n" (new) "p" (primary) "1"then "256" (first LBA) then just enter, then "t" then "c" then "p" and pastebin (don't quit)
19:07:17pamauryok good
19:07:21pamauryhit "w"
19:07:33pamauryunplug the device, replug and pastebin the output of dmesg
19:10:07 Join krabador_ [0] (
19:10:38 Quit krabador (Ping timeout: 260 seconds)
19:13:06pamauryhum, didn't work as expected, can you run "fdisk /dev/sdb", hit "p" and pastebin the output ?
19:13:49 Quit mudlord (Remote host closed the connection)
19:15:46pamaurythe pastebin is *after* you replugged the device ?
19:15:53pamaury(I mean the dmesg one)
19:16:22ildarion2yes it's after replugg
19:16:50ildarion2now you say hi, i not 100% sur :p
19:16:59ildarion2say it*
19:17:16ildarion2but yeah I unplugged
19:17:33pamaurycan you pastebin again the output of the current dmesg ? Just to be sure
19:18:18ildarion2i can qui the procedure about fdisk ?
19:19:40pamaurymy bad, ok my fault i'm stupid ^^
19:20:13ildarion2do you want dmesg N
19:20:15pamauryI told you to run "fdisk -b 4096" but it was "fdisk -b 2048" :-/
19:20:20 Join bertrik [0] (
19:20:20 Quit bertrik (Changing host)
19:20:20 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:20:45ildarion2i do with 2048 so
19:20:47pamauryso can you redo this now:
19:20:47pamaury"fdisk -b 2048 /dev/sdb"
19:20:47pamaurythen "d", then "n" (new) "p" (primary) "1"then "512" (first LBA) then just enter, then "t" then "c" then "p" and pastebin (don't quit)
19:20:54 Quit Scromple (Read error: Connection reset by peer)
19:21:12 Join Scromple [0] (~Simon@
19:26:01ildarion2pastebin free limit reach so :
19:27:22pamaurygood, now hit "w", unplug and replug and paste the output of dmesg after having plugged (and wait for a second)
19:27:35*pamaury didn't know there was a limit
19:27:42 Quit Raptors (Read error: Connection reset by peer)
19:28:14 Join Raptors [0] (
19:28:29 Quit alucryd (Ping timeout: 252 seconds)
19:32:33pamauryah good, now can you try to boot the OF ?
19:33:54 Join alucryd [0] (~alucryd@archlinux/trusteduser/alucryd)
19:34:17ildarion2Seem to be block on the loading screen of sansa, I think is a good sygn
19:35:16ildarion2normaly, if I put a it will be install no ?
19:36:58pamauryalso I suggest you reformat (again) the drive using windows, because with the fews tricks we did, there might be some inconsistency
19:38:09ildarion2can you explain to me what the issue was ?
19:38:49ildarion2for leave an anwer on the forum for other people^^
19:42:01 Quit krabador_ (Quit: Sto andando via)
19:42:22 Join krabador [0] (~krabador@unaffiliated/krabador)
19:42:41pamaurythe problem is the "logical sector size"
19:42:58pamaurythat is: when the OS reads the number for the MBR, those are units of "sectors"
19:43:24pamaurymore precisely, it's a logical sector, but the problem is difference software might have different notion of sector size.
19:43:41pamauryThe OF has been compiled to use a logical sector size of 2048 bytes, which is quite unusual
19:43:52pamauryOn the other hand, Rockbox tries to guess it at runtime
19:44:05pamauryAnd finally Linux uses the value provided by the USB SCSI layer
19:44:31pamauryBy default the OF reports 2048 and since Rockbox guesses right, it reports 2048. So both Linux and Windows use that figure. Everyone is happy
19:45:06ildarion2Ok, so when the windows utility USB installer has done theyre shit on sansa, it has changed the value
19:45:23pamauryWhat you did completely erased the partition table and when you plugged it again, OF couldn't boot so you had to use Rockbox
19:45:53pamauryand Rockbox guessed 512 because that how the USB installer thing put on it
19:46:19pamauryAnd then you couldn't do anything because both Linux and windows would helplessly use 512 instead of 2048
19:46:45pamauryThe trick to solve it is tricky: you first create a partition table using 512 sector size and put the partition at sector 2048.
19:46:46fs-bluebot3Gerrit review #512 at : 3AMSv2 sd: restore switch sd into high-speed mode by Gunbuster (changes/12/512/2)
19:47:06pamaurythen you create the partition at this sector
19:47:49pamauryThen you rerun fdisk using the -b option to convince him it's actually a 2048 logical size and you recreate the partition at sector 512. And since 512 * 2048 bytes = 2048 * 512 bytes, the partition didn't move, but the figures use different units
19:48:13pamauryThen you unplug and replug, now Rockbox guesses right: 2048 bytes and you are saved
19:48:45 Quit Baranko (Remote host closed the connection)
19:49:08pamauryI guess the conclusion is: the OF is stupid but we knew that before ^^
19:49:38pamauryneed to go, be back later
19:50:00ildarion2the conclusion is "Do not use your MP4 for dark thing like instal the windows 7 usb installer" :p
19:51:07gevaertsUnless you know how to recover :)
19:51:53ildarion2the time spend to recover is not a great deal
19:52:32ildarion2I really need to buy a flash drive 4go+ :p
19:54:19 Quit pamaury (Ping timeout: 276 seconds)
19:56:41 Quit pretty_function (Remote host closed the connection)
19:57:33 Quit ildarion2 (Quit: Page closed)
19:58:15 Quit rela (Read error: Connection reset by peer)
20:01:02 Part b0hoon ("GTG... Bye.")
20:22:57 Join n1s [0] (~n1s@rockbox/developer/n1s)
20:32:37 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
20:37:16 Join wodz [0] (
20:38:30wodzpixelma: The patch touches such essential things that it would either just work or it will hang on boot (so be prepared to use paper clip if I screwed something up)
20:55:19***Saving seen data "./dancer.seen"
20:58:04wodzholly crap: "Sorry wodz, you are banned from using this forum! This ban is not set to expire."
20:58:33wodzgevaerts, saratoga: ^
20:58:51*gevaerts asks saratoga to be more careful :)
21:01:24wodzpamaury: fs#12926 is exactly the behavior of rk27xx usb driver (except that rk27xx survives a few rounds of lsusb -v)
21:01:25fs-bluebot 3usb connection hangs after lsusb -v (bugs, unconfirmed)
21:02:45gevaertssaratoga: I think we should be a bit less ban-happy. I don't think it helps much, and accidents *do* happen
21:09:31 Quit y4n (Read error: Connection reset by peer)
21:09:39gevaertswodz: you should be fine again
21:10:27 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
21:11:16wodzgevaerts: yes, thanks
21:13:00pamaurywodz: I thought you needed to spam it to crash it ?!
21:14:24wodzpamaury: yes I need to run it more then once. Single lsusb -v kills usb on OF though
21:15:58 Quit krabador (Quit: Sto andando via)
21:28:05copper[Saint]: I went ahead and replaced the storage / hold switch indicators with 32 bit RGBA images
21:28:11copperworks well
21:33:11 Quit bertrik (Ping timeout: 272 seconds)
21:40:37 Quit wodz (Quit: Leaving)
22:02:38 Join listener [0] (4dfecb2f@gateway/web/freenode/ip.
22:04:27saratogahow did i ban wodz?
22:04:50pamaurydid you ?
22:06:58 Join Rower [0] (
22:10:52 Join wodz [0] (
22:11:42 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
22:13:52wodzhuh, chello is pretty big ISP in Poland banning whole domain is not good idea
22:14:57wodzOn another topic - anyone considered using kconfig-frontends instead of our hand crafted configure?
22:20:22pamaurywhat would it bring ?
22:23:09wodzmanuconfig should be more intuitive
22:27:10 Quit Rower (Quit: Hmmm...)
22:27:12 Quit wodz (Quit: Leaving)
22:30:47 Quit bluebrother (Disconnected by services)
22:30:52 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
22:34:07 Quit fs-bluebot (Ping timeout: 252 seconds)
22:35:25 Join fs-bluebot [0] (
22:44:11 Join kugel [0] (
22:44:11 Quit kugel (Changing host)
22:44:11 Join kugel [0] (~kugel@rockbox/developer/kugel)
22:44:41kugelwodz: yeah i considered kconfig
22:47:02pamaurykugel: I'm working on usb simulator
22:47:15pamaurythere is a huge cleanup work to make it apply and compile again !
22:47:38kugelnice. does the mainline vhci work for this purpose?
22:50:48pamauryI don't think so, I looked at it briefly and it seems to be rather different
22:51:09pamauryanyway, the actual usb driver is only a small part of the overall patch
22:55:20***Saving seen data "./dancer.seen"
23:00:06pamauryomg, it still uses the wakeup data structure ^^
23:05:30 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
23:08:03 Quit kugel (Read error: Connection reset by peer)
23:08:26 Quit n1s (Quit: Ex-Chat)
23:14:48 Quit amayer (Quit: Leaving)
23:21:12 Join Icedman [0] (~Icey@unaffiliated/icedman)
23:22:34IcedmanI am wondering what's better the clip+ or clip zip. Currently reading forum posts about the experiences
23:24:46bertrikI use my clip zip more than my clip+, the zip has a full colour screen (OLED, a bit crap) and has a radio chip that supports RDS
23:25:28bertrikthe clip+ has only a bicolor OLED, but is more 'crisp'.
23:26:39Icedmanhow are the buttons?
23:36:03 Quit scorche|sh (Ping timeout: 272 seconds)
23:38:21toehser1The buttons are fine on both until the power button on the clip+ breaks off the circuit board... I'm up to *3* failed Clip+ players in my family (between my daughter and I, mostly, as we're the heavy users). They work fine by plugging them into USB to turn them on, though... So far, no Clip Zips have failed this way.
23:38:45toehser1The other difference is the micro USB vs. mini USB - the Zip is compatible with more currently common cables...
23:39:06bertrikI have a clip+ with a lame button, still works, but doesn't click anymore
23:39:39pamaurykugel_: appartently the usb vhci that I was using was not really maintained and doesn't compile out of the box with kernel 3.12, so maybe I'll switch to usbip
23:40:05pamauryunfortunately that means reimplementing there network protocol more or less, except if they have a library for it
23:41:24pamaurykugel_: do you know if it's possible to build it as a module ? I would prefer to avoid recompiling a whole kernel just for one driver
23:42:34Icedmanthanks for your replies
23:42:47IcedmanI will get the zip then
23:43:04kugel_pamaury: yes, and usbip and vhci-hcd are separate modules
23:43:41pamauryoh cool, debian unstable builds staging my default so I have them on my box :)
23:43:44 Join sprout_now [0] (~RJChesko@
23:43:51pamauryyeah but i can't find anywhere the vhci-hcd protocol documented
23:43:57toehser1One other difference as long as we're fleshing them all out - the zip has the microphone on the front top bezel, the + on the back by the clip I think. Neither is necessarily better, but something to be aware of if you record.
23:47:56 Quit sprout_now (Client Quit)
23:47:56 Join dfkt_ [0] (OxO29A@unaffiliated/dfkt)
23:49:14pamaurykugel: also I think it's not clear you can control the vhci hcd stuff without usbip because the dependency is reverse: vhci-hcd depends on usbip_core
23:49:21 Quit dfkt (Ping timeout: 272 seconds)
23:55:30 Nick SuperBrainAK is now known as DormantBrain (~andy@2001:470:8:a61::5f92:59a1)
23:56:25kugel_pamaury: that dependency might be only to make the build system happy
23:56:42pamauryno I checked the code
23:57:03*kugel_ didn't
23:57:27pamauryyou *have* to use the tcp/ip protocol
23:57:31kugel_the menuconfig items is a bit misleading then
23:57:51pamauryit's just that the usbip stuff is cut into two parts: a common part which is mandatory
23:58:08pamauryand the usbip host part which exports physical devices over ip
23:58:23pamaurythe vhci_hcd does the reverse: it presents exported devices to the local system

Previous day | Next day