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 2013-08-17

00:01:35jlbiasinipamaury: I've started here:
00:03:38 Join grawity [0] (
00:04:56grawityI see the site still says USB is broken for nano2g? :(
00:05:36 Quit lorenzo92 (Remote host closed the connection)
00:09:51 Quit bertrik (Remote host closed the connection)
00:14:42 Quit ender` (Quit: Bulk trash will be disposed of in incinerators, not compactors. And they will be kept hot, with none of that nonsense about flames going through accessible tunnels at predictable intervals. -- Evil Overlord List #63)
00:19:04lebelliumgrawity: it doesn't look like it will be fixed anytime soon
00:20:08grawityHm, so I'd better not plan upgrading
00:20:19grawityoh well, the ipod itself is not going to last until next year anyway :D
00:20:32lebelliumwhy? dead battery?
00:29:30grawityno, the buttons are ... "worn out" if that's the right term; I have to press them really hard to react
00:30:00grawityon the other hand, rockbox seems to be doing wonders to battery use >_>
00:44:45 Quit jlbiasini (Quit: jlbiasini)
00:46:12 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
01:04:14***Saving seen data "./dancer.seen"
01:18:14 Join lebellium_ [0] (
01:21:06 Quit lebellium (Ping timeout: 260 seconds)
01:21:09 Nick lebellium_ is now known as lebellium (
01:35:31 Nick SuperBrainAK is now known as DormantBrain (
01:37:28 Quit n1s (Quit: Ex-Chat)
01:40:13pamaurysomebody wants the performance graphs for the fuze+ ?
01:52:17TheSevenhm, how many devices do we support that require a (software) FTL?
01:52:25TheSevenand how many of them haven't been fully figured out so far?
02:22:04 Join wowaname [0] (
02:23:15 Quit lebellium (Quit: ChatZilla [Firefox 24.0/20130812173056])
02:33:29pamauryTheSeven: not a lot and a lot
02:34:11pamaurynot figured out (at least): zen xfi2, rk27xx (at least two versions), many samsung devices
02:35:08pamauryi rhink the nano2g is the only one we have
02:50:26 Nick DormantBrain is now known as SuperBrainAK (
03:04:16***Saving seen data "./dancer.seen"
03:42:22 Quit Gareth (Remote host closed the connection)
03:49:55 Quit wowaname (Ping timeout: 264 seconds)
03:55:57TheSevenhm... I wonder how feasible it would be to just replace the FTL with something generic...
03:56:43TheSevenif we had MTP I'd say YAFFS2/Direct, but with MSC we would need a translation layer
04:07:26pamauryyou mean just drop compatibility with the of ?
04:07:33pamaurymtp is clearly doable
04:08:06pamaurywe have the code, just needs some refreshing
04:15:24pamauryTheSeven: what are the alternatives for flash filesystems ?
04:16:02pamauryand what is the difference between yaff2 and yaff2/direct ?
04:16:38 Join dewlap [0] (~dewlap@2001:c08:3700:ffff::11ab)
04:19:25TheSevenpamaury: yaffs2/direct is basically yaffs2 for OS-less systems
04:19:34TheSeveni.e. a standalone implementation of it with minimal dependencies
04:19:47TheSeventhe on-flash format is the same
04:20:17TheSevenhm, the memory usage of yaffs2 might be a problem
04:20:38TheSeven1MB of RAM per gigabyte of storage is probably a bit too much for our purposes...
04:20:49TheSevenbut maybe that can be optimized
04:21:10TheSevenIIUC it was developed with flashes from tens to hundreds of megabytes in mind
04:22:52pamauryyes that's a bit high
04:24:44pamaurywhat about using a generic FTL, how does that compare with such a solution ?
04:26:11TheSeventhat might be more efficient memory-wise (the N2G FTL needs ~500K of RAM for 8GB of flash)
04:26:21TheSevenand it's probably easier to implement
04:26:46TheSevenbut I don't particularly like the idea of FTLs at all
04:28:27TheSevenfrom a "clean design" point of view: why emulate a fixed-size reliable device on top of an unreliable device, if you could just use an FS that can deal with unreliable media? an FTL has to reserve a significant amount of blocks (~5-10%) as spares to compensate for failing blocks, which both wastes space and makes it stop functioning once it's worn out to some degree
04:28:51TheSevena flash file system just shrinks in size as the flash wears out
04:29:13pamauryindeed, the issue is to choose one which suits our needs
04:29:24TheSevenmight not be terribly important with these 10K erase cycle flashes that we usually deal with, but in the light of modern flashes with more like 300-3000 cycles, this may get more important
04:31:11TheSevensamsung builds SSDs with TLC flash these days (yes, 3 bits per transistor, "tri-level" is actually nonsense, it's eight-level from a technical point of view)
04:32:45TheSevenI wonder how their reliability and endurance will be like in practice (I'm talking about 840 basic series)
04:33:43TheSevenso except for the memory usage, YAFFS2/direct seems like the best choice... for an FTL there's basically only yaftl, and I'm not sure how stable that is
04:33:54TheSevenwe could of course just use the nano2g FTL ;)
04:34:08pamaurywell, I prefer not to ;)
04:34:23TheSevenit's actually quite well-engineered
04:35:51pamaurymaybe if were to choose the FTL solution that would make sense, we have it and it works
04:36:17TheSevenmight possibly need to remove some page size and ECC data position hard wiring
04:36:40TheSevenand rockbox would need to pull in emCORE's nandfsck code :)
04:36:56TheSevenright now emCORE (or the apple bootloader) takes care of cleaning up after an unclean shutdown
04:36:58pamauryyaff website says the memory usage can be reduce with larger pages, I'm not too familar with NAND in general, what is the typical page size these days ?
04:37:16TheSeven2K to 4K
04:37:21TheSeven(bytes, not bits)
04:38:02 Quit amiconn (Disconnected by services)
04:38:02 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:38:04 Quit pixelma (Disconnected by services)
04:38:04 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:38:05 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:38:07 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:38:32pamauryI guess yaff also requires a memory allocator
04:40:22TheSevenapple's does as well
04:40:51TheSevenin apple's firmware, they had a separate one-time allocator just for the FTL, not used by anything else
04:41:00TheSeventhey just reserved 600K or so of memory for it
04:41:28TheSevenI think my implementation of it gets away without one, but that actually ends up wasting space for non-maximum-size flashes
04:41:48TheSevenoh, and the unclean shutdown cleanup operation will need far more memory once again
04:42:14pamauryas long as you don't waste too much that's fine, or if you can estimate the size in advance, you bufalloc it once for all
04:42:26TheSeven(I didn't reverse engineer that but instead designed and implemented that from scratch based on my understanding of the structures... it ended up being seemingly more reliable than apple's ;)
04:43:07pamauryagain for cleaning that fine: use a lot of memory at mount then release, that can work
04:43:47TheSevenhm, I guess yaffs, being basically a log file system, needs to keep a file index tree in memory to be able to do any directory enumerations with reasonable performance
04:44:51TheSevenhm, that Tnode tree looks like it could easily be compressed
04:45:01TheSevensimilar to how I compress my hard drive bad block tables on the classic :)
04:45:15TheSevenwhich is in turn very similar to how ARM page tables work
04:45:32TheSeveni.e. encode large consecutive chunks adequately
04:47:16pamauryso if I understand correctly, yaff basically keeps the whole directory file in memory ?
04:47:58TheSevenat least mostly
04:48:29TheSeventhis used to be generated while mounting, reading a lot of pages, which could take minutes
04:48:36TheSevenrecent versions changed that
04:49:18TheSevennow it basically stores a snapshot of that index on the flash while unmounting to speed up the next mount
04:50:16TheSevenone could possibly make it rely on that on-flash index during normal operation...
04:51:07TheSevenor maybe we "just" have to design an even better flash file system from scratch...
04:51:50roxfanyes, we clearly don't have enough of them
04:51:51pamauryexactly my thought, I had some spare between 30pm and 32pm every day ;)
04:52:08pamaury*spare time
04:52:54pamaurythere are not that many of them
04:53:12pamaurymostly jaff, logfs, jffs and the upcoming f2fs
04:53:33pamauryor maybe one of them was superseeded by ubifs
05:04:20***Saving seen data "./dancer.seen"
05:04:22pamauryit seems to me that filesystems in general, (and this even more true for mp3 players) basically have two types of files: big binary blobs which are basically written at once and read many times are random or more likely sequentially; and small textual files that can get changes here and there randomly. You really want to take advantage of this if you can
05:06:55pamauryand if you are using something like mtp this is perfect: you know the sizes of the files in advance and the type of the file, you can take good decisions when you are actually writing because most of writing occurs via usb for us
05:08:13pamauryanyway, time to sleep
05:16:52 Quit pamaury (Ping timeout: 246 seconds)
05:20:21 Join kevku [0] (~kevku@2001:0:c38c:c38c:30ac:9a96:3d69:bee7)
05:22:59 Quit TheSeven (Disconnected by services)
05:23:08 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
05:25:51 Join rela_ [0] (
05:29:07 Quit rela (Ping timeout: 246 seconds)
05:31:55 Join wowaname [0] (~wow@
05:32:29wowanameis there any way to rewrite the header to the ipod's disk
05:32:45wowanamei can view files in nautilus but can't copy anything to it
05:32:53wowanameand i can view files on the ipod
05:35:48wowanameD: i dont want to back up everything and format it
05:37:02wowanameok goodbye
05:37:05 Quit wowaname (Quit: <3)
06:56:12 Join dv__ [0] (
06:57:17 Quit dv_ (Ping timeout: 268 seconds)
07:04:24***Saving seen data "./dancer.seen"
07:04:59 Quit ps-auxw (Ping timeout: 245 seconds)
07:10:42 Quit Provel (Read error: Connection reset by peer)
07:11:08 Join Provel [0] (
07:23:53 Nick dv__ is now known as dv_ (
08:00:12 Quit pystar89 (Ping timeout: 276 seconds)
08:12:06 Nick SuperBrainAK is now known as DormantBrain (
08:19:31 Join ender` [0] (
08:49:29 Join rela [0] (~x@pdpc/supporter/active/rela)
08:51:55 Quit ender` (Quit: How many SEO experts does it take to change a light bulb, lightbulb, light, bulb, lamp, lighting, switch, sex, xxx, hardcore)
08:52:42 Quit rela_ (Ping timeout: 264 seconds)
09:04:27***Saving seen data "./dancer.seen"
09:15:33 Quit dewlap (Read error: Connection reset by peer)
09:23:28 Quit rela (Quit: Leaving)
09:24:48 Join rela [0] (~x@pdpc/supporter/active/rela)
09:40:26 Join melmothX [0] (~melmoth@unaffiliated/melmothx)
10:00:11 Join pystar89 [0] (
10:02:18 Quit rdn (Remote host closed the connection)
10:19:03 Quit fyrestorm (Ping timeout: 264 seconds)
10:35:26 Join pretty_function [0] (~sigBART@
10:35:44 Quit pretty_function (Remote host closed the connection)
10:53:15copperit would be really nice if it was possible to use %ss() in a conditional
10:56:56copperoh, it is!
10:57:06copperI must have been doing it wrong before
10:57:17 Join n1s [0] (
10:57:17 Quit n1s (Changing host)
10:57:17 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:58:21copperhmmm, no, I remember now: what doesn't work is: %ss(0, 2, %St(font))w
10:58:27copperhmmm, no, I remember now: what doesn't work is: %ss(0, 2, %St(font))
10:59:25copper%?if(%ss(0, 2, %St(font)), =, 14)<fourteen|something else>
11:02:21pixelmaa font is just a remembered path not one of a set range of values (like "yes", "no" etc.). Does %St(font) give you anything back?
11:04:30***Saving seen data "./dancer.seen"
11:08:09pixelmajust thinking that it may give you a full path if at all and then you would need a different part of the substring
11:08:55 Join SovonHalder [0] (SovonHalde@
11:12:08copper%St(font) returns the font's name
11:12:26copperit just doesn't seem to work in a %?if() conditional
11:12:41coppernot the font path, the font name
11:13:06copperlike "14-Nimbus"
11:13:39copperwithout the .fnt file extension
11:14:15copperusing %ss() with %St(font) would allow me to make conditionals for font sizes instead of full font names
11:14:20 Join Sovon [0] (SovonHalde@
11:15:15 Quit SovonHalder (Ping timeout: 246 seconds)
11:15:53SovonIs there any way I can set iPod like this? when I play something from Database > Album > track, then from the while playing screen if I press MENU button, it will take be back to the actual previous Album folder in databse, not the rockbox menu ?
11:16:47Sovonjust like Apple firmwire did
11:23:36Sovonanyone please−−-a little help here
11:24:17n1sdon't think there's such a feature for the db, it's there for the file browser
11:25:54Sovonisn't it possible to just go back to previous location instead of root rockbox directory ?
11:28:43pixelmaSovon: what happens if you press the Select buttpn?
11:28:47pixelmabutton too
11:31:13pixelmacopper: maybe it treats the 14 in your example as number and then can't compare substring to number? I don' know that much about the %?if tag though and don't know if there's a way around it
11:31:18Sovonholy sit! thank you
11:31:29Sovonpilexma: thank you man
11:33:02SovonIt's about experience. I'm beginning to realist that more & more each day
11:34:47pixelmathank our fine manual too (you could use the one for the Ipod Video since it's similar enough and there doesn't seem to be one for the Classic yet) ;)
11:36:39SovonI didn't know there was a manuall..
11:36:45SovonI'll read it now
11:53:15 Join bertrik [0] (
11:53:15 Quit bertrik (Changing host)
11:53:15 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
11:55:19*copper is tempted to remove one metadata line on his Fuze+ theme, at the risk of further displeasing people
11:56:08copper versus
11:56:32copperI feel like the former is visually more balanced than the latter
12:02:56 Quit Sovon ()
12:07:50copperdecisions, decisions…
12:07:56 Join lebellium [0] (
12:13:41copperI could do a hack
12:14:11copperI could duplicate 14-Nimbus.fnt, rename it to something like 14-Googley.fnt
12:14:44copperand then I could do conditionals based on the chose font
12:14:55copper14-Nimbus: display what I like personally
12:15:10copper14-Googley: display what I like personally
12:15:21copper14-Nimbus: display what people want
12:17:32copperthe number of conditionals would be insane
12:26:41copperhmm, not that insane
12:50:50copperthe font conditional doesn't seem to work for the backdrop
12:52:51 Quit sakax (Remote host closed the connection)
12:53:45copperdamn it
12:53:54copperit was a neat idea
12:54:12copperI can work around it
13:04:32***Saving seen data "./dancer.seen"
13:09:48coppernope, I can't
13:29:11 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
13:40:46copperok, screw it
13:41:06copperI have decided what I want for my themes, everyone else is free to modify them
13:41:26copper"Perfection is Achieved Not When There Is Nothing More to Add, But When There Is Nothing Left to Take Away"
13:41:42copperthat's my philosophy here
13:42:10 Quit Topy44 (Ping timeout: 260 seconds)
13:43:02 Join Topy44 [0] (~Topy44@
14:15:54 Join dewlap [0] (~dewlap@2001:5c0:1000:a::101)
14:43:13 Join jlbiasini [0] (
14:50:20 Join prof_wolfff [0] (
15:04:34***Saving seen data "./dancer.seen"
15:10:05copperis the radio signal strength unit really 'dBu'?
15:10:14copperIsn't it actually 'dBµV'?
15:11:05copperah, no
15:11:12copper"u" stands for "unloaded"
15:11:21 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
15:11:55copperpamaury: can you (or anyone else) give me an idea of dBu ranges for weak / good / strong?
15:13:15lebelliumwhy do you need that?
15:13:17lebelliumjust do %?tr<%xd(Va)|%xd(Vb)|%xd(Vc)|%xd(Vd)|%xd(Ve)|%xd(Vf)|%xd(Vf)>
15:13:20lebelliumwith text instead of pic
15:13:56pamaurycopper: not really, ask bertrik
15:14:04copperbertrik: ^^
15:14:29copperlebellium: what?
15:15:17copperI have no idea how that line would be interpreted (what image/text will show up with what dBu value)
15:15:47lebelliumjust do %?tr<weak|strong|very strong>dBu or something like that
15:15:51lebelliumyou don't need actual values
15:16:54copperI assume that would only make rockbox display "weak" for the first third of the scale, "strong" for the second, and "very strong" for the third, without any regard to the actual strength of the signal
15:17:37coppermy point is that I don't know that rockbox uses a scale that actually represents that
15:17:41lebelliumthe values may change from one player to another
15:17:43lebelliumso don't use them
15:17:46coppermaybe "strong" is actually already "very strong"
15:17:49lebelliumlet rockbox scale for you
15:19:05lebelliumcopper: my post here may interest you:
15:22:15copperI see, thanks
15:22:30copperI was wondering if that last %xd(Vf) above was a typo
15:24:06lebelliumIt's not written clearly on the wiki
15:24:18lebelliumso I did not know before that the last one is only there for 100% state
15:25:53coppernothing is written clearly on the wiki!
15:27:27lebelliumneither in the user manual. That's why I disagree when [Saint] wants to kill the CustomWPS wiki page in favor of the user manual :P
15:45:29bertrikcopper: 70 dBuV is very strong, around 25-30 dBuV is approximately the switching point between mono and stereo (stereo),
15:46:39bertrikthe scale doesn't go below 0 dBuV
15:47:38bertrikweak/ok/strong/very strong is all a bit subjective ofcourse
17:02:29jlbiasiniHi all, I'm completely stuck on g#532. I even tried to write an empty touchscreen_to_pixels function with only "return BUTTON_NONE" in it. Even so the simulator seems to work just fine... Does anybody know how to make the touchscreen driver silent with a switch?
17:02:32fs-bluebotGerrit review #532 at : [touchscreen] touchdev_disable and wakeup implementation by Jean-Louis Biasini (changes/32/532/4)
17:04:35***No seen item changed, no save performed.
17:32:44coppergoddamn rockbox busg
17:32:59 Join rela_ [0] (
17:35:53 Quit rela (Ping timeout: 264 seconds)
17:37:41 Quit rela_ (Ping timeout: 264 seconds)
17:38:57 Join stoffel [0] (
17:40:49 Join simabeis [0] (
17:47:37 Quit simabeis_ (*.net *.split)
18:11:18pamauryjlbiasini: I don't understand what is your issue
18:11:44pamaurymy guess is that the simulator works very differently than the device
18:18:38pamauryjlbiasini: which device are you using in the simulator to check this ?
18:24:33fs-bluebotBuild Server message: New build round started. Revision a56f1ca, 217 builds, 19 clients.
18:28:51 Quit melmothX (Quit: @)
18:31:58 Quit stoffel (Ping timeout: 260 seconds)
18:32:17fs-bluebotBuild Server message: Build round completed after 464 seconds.
18:38:28 Quit jlbiasini (Ping timeout: 248 seconds)
18:50:56 Part grawity ("Bye.")
19:03:50 Nick DormantBrain is now known as SuperBrainAK (
19:04:36***Saving seen data "./dancer.seen"
19:07:01 Join krabador [0] (~krabador_@unaffiliated/krabador)
19:17:30 Join jlbiasini [0] (
19:18:23jlbiasinipamaury: I tested with the ondav777
19:22:38pamauryok, i'll try to have a look
19:45:01jlbiasiniIs there someone around with a touchscreen device? It would helps if someone could test g#568
19:45:04fs-bluebotGerrit review #568 at : [touchscreen] Test disable touch device while locked by Jean-Louis Biasini (changes/68/568/1)
19:45:37jlbiasinijust compile, install and see if the touchscreen still react while key are lock
19:46:13 Join ender` [0] (
20:04:37jlbiasinipamaury: same with M:Robe 500
20:09:33pamauryI have the zen xfi2 but it's unstable and touchscreen driver is really suboptimal. I'll try though
20:09:49pamauryI'm not surprised it doesn't work in the sim: I'm pretty sure this function is not called by the simulator
20:10:11jlbiasinias long as it used touchscreen.c that should be enough
20:11:18pamauryhave you check that this function is actually called ?
20:12:50jlbiasinino but I totaly wiped out the touchscreen_to_pixels to return always button_none And the simulator still did react to key event
20:13:31pamaurythat's what I thought: it is not used by the simulator
20:14:30jlbiasinithis is strange because the keymap of the simulator for ondav777 doesn't map touchscreen key
20:14:56jlbiasiniso I assumed they where taken from the touchscreen code
20:15:30jlbiasiniobvviously not
20:15:47pamaurywhat do you mean by "doesn't map touchscreen key" ?
20:17:22jlbiasiniwell if I understood correctly simulator use keymaps from uisimulator/buttonmap but uisimulator/buttonmap/onda-vx777.c only map the power_button
20:18:04jlbiasinistill the keypad can be used to control the simulator
20:18:41jlbiasiniI suppose the keypad is automaticly mapped for touchscreen on those devices simulator
20:19:41jlbiasiniwhich would explain why they don't call touchscreen_to_pixels to get keys
20:20:33jlbiasinithat's also make sense as the purpose of this function is to convert touch coordinate into a button
20:21:14jlbiasiniwhich is not posible to input from the simulator
20:21:32jlbiasiniso short version: I need a test on a device itself... :D
20:22:45pamauryno no, the simulator uses the normal keymap. The "keymap" in uisimulator/buttonmap/ is some simulator hack to map computer keys to physical keys (it's kind of replacement for the button drive)
20:23:04pamaurythe touchscreen is handled directly: a clic is like a touscheen touch
20:23:21pamaury*button driver
20:25:43jlbiasiniso I have to test it with moouse?
20:34:31 Join stoffel [0] (
20:38:05copperDoes "tuner tuned" (%tt) mean that it's pretty much the best signal the FM radio can get for that station?
20:41:06pamaurycopper: if it matches the intuitive meaning yes, it means the frequency is fixed and won't change. But you can get a better signal latter if you move for example, since the headpohnes act as an antenna
20:41:44copperyeah I mean the best signal at this particular place and time
20:42:53copperI think I'm done for today
21:01:05jlbiasinipamaury: I just tested with the mouse, still no luck
21:04:39***No seen item changed, no save performed.
21:06:11jlbiasiniand with empty touchscreen_to_pixels the mouse still work as input in the sim. I must be missing something... :/
21:13:13 Quit bluebrother (Disconnected by services)
21:13:18 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
21:15:46 Quit fs-bluebot (Ping timeout: 260 seconds)
21:16:12 Quit akaWolf (Ping timeout: 246 seconds)
21:17:10 Join fs-bluebot [0] (
21:18:04 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
21:27:55 Quit y4n (Quit: MOTHER EUROPA CALLING ME!)
21:38:06 Quit krabador (Quit: Sto andando via)
21:38:16 Join krabador [0] (~krabador_@unaffiliated/krabador)
21:38:19 Join krabador_ [0] (
21:42:54 Quit krabador_ (Client Quit)
21:42:54 Quit krabador (Client Quit)
22:00:23 Join rdn [0] (
22:06:49copperdid someone change CPU frequency and / or boosting on the Classic? It seems like scrolling metadata is a lot slower
22:10:18bertriktuner tuned means that we detected an actual station at a specific frequency IIRC
22:11:17 Quit jlbiasini (Ping timeout: 264 seconds)
22:17:22 Quit stoffel (Ping timeout: 260 seconds)
22:18:11 Quit soap (Ping timeout: 276 seconds)
22:26:42 Quit kevku (Ping timeout: 260 seconds)
22:30:07 Join soap [0] (
22:30:07 Quit soap (Changing host)
22:30:07 Join soap [0] (~soap@rockbox/staff/soap)
22:45:18 Join jlbiasini [0] (
22:51:05jlbiasinipamaury: touchscreen implementation is very complicate anyway. It seems there is even some device that are unlocking softlock with their touchscreen. Impossible to implement a touchscreen disable while locked on those. I think that we should start only with touchpad devices.
22:57:04 Quit Beta2K (Ping timeout: 264 seconds)
23:00:39pamauryjlbiasini: as I said the sim works differently, it doesn't use touchscreen_to_pixels
23:04:35jlbiasiniok, but still, the fact that softlock is handled by the touchscreen is a problem...
23:04:43***Saving seen data "./dancer.seen"
23:05:07pamauryI don't follow, where do you see that softlock is handled by the touchscreen ?
23:08:14jlbiasinipamaury: there is some ACTION_TOUCH_SOFTLOCK in action.h apps/gui/skin_engine/skin_touchsupport.c apps/gui/skin_engine/skin_parser.c
23:10:07pamaurysoftlock is more than action_touch_softlock, I'm confused, I really don't understand what is the issue. Is something in the actual trunk not working or is it only with your patches ?
23:12:16jlbiasiniif some ACTION_TOUCH_SOFTLOCK is used on some device, then to unlock the device you use the touchscreen. That means that if we disable the touchscreen while locked it will lock the device forever...
23:12:43jlbiasiniOr did I get something wrong?
23:15:12pamaurythe way I see it is that it should work like with phones: if the screen goes off and touchpad/screen is disabled, you need to a press a physical button to enable it again
23:18:46pamauryand obviously you need to design a keymap which doesn't lock but that's not specific to touchscreen, you can easily make you device unsuable with a broken keymap
23:19:05jlbiasiniwell, I don't really know how the touchscreen interface is working. But if some devices are already using touchscreen for softlock that will be very difficult to implement because we will have to change their keymyaps one by one
23:21:24pamauryI don't even know what this ACTION_TOUCH_SOFTLOCK is but I don't think we have a such a thing as using the touchscreen to unlock and if we do that's completely stupid then
23:21:49jlbiasinitrue that!
23:22:39pamauryup to my knowledge, the action to softlock is ACTION_STD_KEYLOCK and it is only allowed in context with ALLOW_SOFTLOCK
23:23:14pamauryyou want to check what this ACTION_TOUCH_SOFTLOCK is because usually ACTION_TOUCH_* are internal things. Maybe it means "no button and i'm softlocked"
23:23:24pamauryand thus display a "i'm softlocked" message
23:24:08jlbiasiniok so I will have to wait for someone to test my patch on a device anyway and in the meantime I will have a look to the other touchpad target
23:27:34 Quit jlbiasini (Quit: jlbiasini)
23:41:39 Quit n1s (Quit: Ex-Chat)

Previous day | Next day