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 2013-08-27

00:01:48 Quit jlbiasini (Quit: jlbiasini)
00:19:55 Nick SuperBrainAK is now known as DormantBrain (
00:45:40 Quit bertrik (Remote host closed the connection)
00:49:14 Quit thomasjfox_ (Quit: Konversation terminated!)
00:57:46 Quit ender` (Quit: #include <unistd.h> | #define Bork for | #define bork fork() | int main() { Bork(bork ;; bork) bork; })
01:07:04 Quit pamaury (Ping timeout: 245 seconds)
01:22:27[Saint]lebellium: I admit, that behavior is non-obvious.
01:22:52[Saint]I guess some RTC will present a bullshit time/date if RTC is non-set.
01:23:08[Saint]For instance: 01:01:1970 00:00
01:23:19[Saint]^ a fairly common one.
01:23:45lebelliumI guess I already got −−:−−
01:23:47[Saint]Apparently when <target_in_question>'s RTC is non-set, its simply ":"
01:23:56[Saint]...which is kinda annoying.
01:24:24[Saint]But, anyway, yes...I did forget about the last tuple of that tag.
01:24:52lebelliumyes sure. But then this tag is not what I wanted. I wanted to add a clock for RTC targets (Archos Recorder) and display another thing on my Ondio
01:24:59[Saint]Its (iirc) <rtc is set|rtc non-set|I have no fucking idea what you're talking about> :P
01:25:44[Saint]errr, sorry, s/set/present/
01:26:15lebellium%?cc<yes|no|maybe> works
01:26:20lebelliumthat's what I wanted
01:26:24*[Saint] nods
01:26:40[Saint]yet more things I need to make clearer in the documentation.
01:26:52lebelliumthen I can use "maybe" as "there is no RTC in hardware"
01:27:01[Saint]There are other tags that use the last tuple similarly.
01:27:08[Saint]like volume and battery, for example.
01:28:14[Saint]%?bc<mute|1|2|3|4|5|...|"I don't even know man....I give up">
01:28:35lebelliumthe last one is not 100%?
01:28:45lebelliummax volume*
01:29:02lebelliumwho knows that? :D
01:29:07[Saint]The very last tuple _should_ be "indeterminable"
01:29:24lebelliumI thought this was only valid for battery and FM strength
01:30:15[Saint]Not being able to determine the current volume is incredibly unlikely, but iirc, the tag does behave this way.
01:30:30lebelliumShould I update 10 themes? :S
01:31:15[Saint]No, its not really an issue, more of a "safety net" I believe. Conditions will usually work happily if one or more tuples are stripped from the end of it.
01:32:01[Saint]the volume tag usually catches people out compared to how the think it works.
01:32:37[Saint]<mute|1|2|3|4|5|...|line level|ranges above line level|indeterminable>
01:32:45lebelliumI thought I used cabbiev2 as example for volume. But my 1st theme was back to 2011, so I don't remember exactly
01:33:33[Saint]It wouldn't surprise me at all if cabbie didn't include this, as not knowing the current volume is incredibly unlikely.
01:34:20[Saint]I notice lots of user-built themes forget to include tuples for line level and ranges above line level.
01:34:41[Saint]so their volume gauges only go as high as -1dB
01:35:18lebelliumthat probably means the documentation is not clear enough or they copied a bad theme :D
01:35:47[Saint]Or both. At least I can do something about the former.
01:35:55[Saint]The latter is harder to combat.
01:46:18***Saving seen data "./dancer.seen"
01:47:01*[Saint] plays with his new 240GB iPod Video
02:00:14 Join __jae__ [0] (
02:02:03 Quit DormantBrain (Ping timeout: 245 seconds)
02:04:00 Join DormantBrain [0] (~andy@
02:04:21 Nick DormantBrain is now known as SuperBrainAK (~andy@
02:04:31 Quit __jae__ (Ping timeout: 240 seconds)
02:09:45 Quit lebellium (Quit: ChatZilla [Firefox 24.0/20130822154523])
02:33:55 Quit habys (Quit: WeeChat 0.4.1)
02:42:41 Quit crose (Quit: Leaving)
02:43:11 Join habys [0] (
03:20:21 Nick jmspeex_ is now known as jmspeex (
03:35:40polemon__hey guys, what is the status of the USB issue for the Ipod Nano 2G?
03:46:19***Saving seen data "./dancer.seen"
03:58:51Jinxvedos: interesting... thanks for sharing the information
03:59:06Jinxat least apparently you can boot to original FW without reflashing?
04:25:32 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:25:42 Quit amiconn (Disconnected by services)
04:25:49 Quit pixelma (Disconnected by services)
04:25:50 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:25:52 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:28:21 Quit uwe_ (Read error: Operation timed out)
04:30:39 Join uwe_ [0] (
05:12:57 Quit TheSeven (Disconnected by services)
05:13:07 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
05:46:23***Saving seen data "./dancer.seen"
05:46:42 Quit [Saint] (Remote host closed the connection)
05:47:51 Join [Saint] [0] (~saint@rockbox/user/saint)
06:21:15 Join advcomp2019__ [0] (
06:21:15 Quit advcomp2019__ (Changing host)
06:21:15 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
06:24:27 Quit advcomp2019_ (Ping timeout: 256 seconds)
07:33:42 Join saratoga [0] (123e1c32@gateway/web/freenode/ip.
07:34:08saratogawho has access to the gerrit config? changing the launch page to ",n,z" would prevent showing the sandbox entries
07:34:16saratoga,n,z rather
07:38:44 Nick SuperBrainAK is now known as DormantBrain (~andy@
07:46:25***Saving seen data "./dancer.seen"
08:03:46 Quit soap (Ping timeout: 260 seconds)
08:07:30[Saint]Holy crap the iPod Video boots fast with nothing but rockbox in the OSOS
08:07:37[Saint]No dual-boot, but...meh.
08:07:55[Saint]Its about 4~5 seconds faster than stock Rockbox install.
08:09:45*[Saint] is very much liking his new iPod Video
08:09:56[Saint]...if only it had a better CPU
08:11:32[Saint]I was lucky enough to get one of the 64MB RAM models. The pawn show I got it from apparently just looked at the serial number to determine the specification. It was listed as being 80GB, but it has the 240GB aftermarket Toshiba disk in it, and an aftermarket 800mAh battery.
08:11:48[Saint]All for $70 NZD, quite the score.
08:15:32[Saint]I would've been happy if it was stock, as it is in nice condition.
08:15:41copperthe HDD in my Classic makes it uncomfortably slower than my Fuze+ though
08:15:49[Saint]The aftermarket parts were a very welcomed surprise, though.
08:16:23[Saint]ALso, yes...HDD targets do seem very slow when you're used to flash storage.
08:16:27copperseems like I'm always waiting for the HDD to spin down
08:16:46[Saint]But, I don't mind the sacrifice in speed for the additional capacity it brings.
08:17:00[Saint]y'know there's a spin-down setting, right?
08:17:09[Saint]set it to 3 seconds.
08:17:41[Saint](iirc, that's the lowest selection available)
08:18:04copperI meant it's always loading stuff
08:18:08copperno idea what
08:18:56copperand it feels like I have to handle it like a fragile child with a bone disease or something
08:19:14[Saint]Mine behave fine.
08:19:19*[Saint] shrugs
08:19:44[Saint] to transfer 240GB at ~6MB/s
08:19:55coppermine goes to 20!
08:20:22[Saint]My Classic does too, this Video doesn't seem terribly speedy.
08:34:54[Saint]Booting two iPod Videos next to each other, one with a completely stock install, and the other with Rockbox is OSOS alone makes for quite the dramatic comparison.
08:35:21[Saint]The OSOS install has booted and spun down the HDD before the stock install finished booting.
08:36:54copperwhat's OSOS?
08:40:07*[Saint] forgets the exact acronym
08:41:03[Saint]Have a look at
08:42:21[Saint]The installation method I use is "4) OSOS contains only Rockbox"
08:44:40 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
09:05:44 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
09:09:55 Join ender` [0] (
09:13:46 Quit bertrik (Remote host closed the connection)
09:16:39 Join soap [0] (
09:16:39 Quit soap (Changing host)
09:16:39 Join soap [0] (~soap@rockbox/staff/soap)
09:19:07 Join petur [0] (~petur@rockbox/developer/petur)
09:25:53 Join LinusN [0] (
09:34:58 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:46:29***Saving seen data "./dancer.seen"
10:08:01 Quit saratoga (Quit: Page closed)
10:27:59 Join einhirn [0] (
10:32:38 Quit einhirn (Ping timeout: 256 seconds)
10:49:41 Quit pamaury (Ping timeout: 264 seconds)
11:07:06 Join krabador [0] (~krabador_@unaffiliated/krabador)
11:17:55 Quit bluebrother (Disconnected by services)
11:18:02 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:19:09 Quit fs-bluebot (Ping timeout: 246 seconds)
11:20:26 Join fs-bluebot [0] (
11:21:50 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:26:23 Join wodz [0] (
11:28:42 Join jlbiasini [0] (
11:30:09 Join __jae__ [0] (~jae@
11:30:34wodz[7]: ping
11:34:37 Quit __jae__ (Ping timeout: 240 seconds)
11:45:31pamaurywodz: yes?
11:46:31***Saving seen data "./dancer.seen"
11:47:35jlbiasinipamaury: we should wait for kugel to have a look to it too, but if you want to have a look yourself, I redid my g#569...
11:47:38fs-bluebotGerrit review #569 at : touchpad: Disable touchpad on softlock. by Jean-Louis Biasini (changes/69/569/30)
11:48:44wodzpamaury: which crt0.S you were referring to when talking about rk27xx hwstub port?
11:49:28wodzpamaury: And another question, how to define structure in IDA which contains another structure as a member?
11:50:23 Join lebellium [0] (
11:54:09 Join crose [0] (
12:56:19 Quit rdn (Remote host closed the connection)
12:58:24jlbiasinipamaury: could you test g#577 on your device? Also if you know something about the question given as comment that would help...
12:58:26fs-bluebotGerrit review #577 at : touchscreen: softlock touch settings implementation by Jean-Louis Biasini (changes/77/577/5)
13:18:07kugeljlbiasini: what about the simulator?
13:18:31copperhow would he test touchpad sensitivity in the sim?
13:18:56copperI misread that
13:25:17 Join __jae__ [0] (
13:30:03 Quit lebellium (Quit: ChatZilla [Firefox 24.0/20130822154523])
13:31:25pamaurywodz: the crt0.S in hwstub/stub/, there is only one...
13:31:41pamaurywodz: to create substructures, create a field (with 'd') and then press Alt+Q
13:33:25jlbiasinikugel: thanks very much for your comment! The simulator exclusion is beccause the driver part doesn't get compilated in the simulator leading to a compilation error because the _device doesn't get declared/defined pamaury told me to exclude it to solve this but pehraps do you have a better idea?
13:34:37wodzpamaury: wait but this crt0.S only setups stack, clear bss and jumps to main so what do you really commented out?
13:35:42pamauryhum, really ? that must be on my working copy then...but I thought I pushed it to the trunk
13:41:18kugeljlbiasini: should be properly simulated
13:46:32***Saving seen data "./dancer.seen"
13:47:02pamaurywodz: the stub relocates itself in HEAD
13:47:57kugeljlbiasini, pamaury: fwiw, I would like to discuss why/if there's a need for a setting for this
13:48:21wodzpamaury: will look today evening
13:48:28pamaurykugel: what do you suggest, make it the default behaviour ?
13:48:39kugelsince it only affects the fuze+, an unstable port, changing existing behavior is acceptable
13:49:21kugelyes, the new behavior makes more sense than the because a touchpad is a lot more sensitive than tactile buttons
13:50:00jlbiasinikugel: why not but the discussion already occured. But as I'm trying to expand that behaviour to touchscreen target too (see g#577) it would also change behaviour of those target
13:50:03fs-bluebotGerrit review #577 at : touchscreen: softlock touch settings implementation by Jean-Louis Biasini (changes/77/577/5)
13:50:12pamaurythat's fine for me, but what about touchscreen then ?
13:50:52kugeldo we have touchscreen targets wihtout hardware lock button?
13:51:04jlbiasinikugel: yes
13:51:12jlbiasiniat least
13:51:37kugelthe onda's aren't even unstable
13:52:00kugelwe have no stable touchcreen targets, we're pretty free to experiment I'd say
13:52:12jlbiasinigreat then
13:53:18jlbiasiniI ws of the idea that a setting was useless in the first place, but the do-not-change-usual-behaviour kind of scare pamaury and me
13:53:36kugelobviously we need to make sure there's still a way to unlock
13:54:13pamaurythat's the keymapper's job ;) But yeah I need everyone wants the device to behave this way nowadays, and you're right, there is no really stable touchscreen targert
13:54:47jlbiasinianyway there is also another setting comming soon: the lock only touchdev on softlock
13:55:41kugeljlbiasini: why would you want that?
13:55:51jlbiasiniand I would like to have touchpad sensitiivity enlarge to get used by touchscreen if possible. I think it maks sense to have those general touch setting together
13:55:52kugellet me guess: use volume buttons while screen off
13:56:39kugelthat's an old problem, we ideally want a generic solution to that. sansa clips and other targets have this issue as well
13:57:29jlbiasinithere is a lot of difference betwenn touchpad and hard key. So people often want to have their touchpad lock and still want to use hard button. As it really depend on people/situtation a setting is mandatory on this
13:57:40kugelI liked the behavior of CM a few versions ago: when the screen is off the volume buttons still work and a long press on those skips tracks
13:59:28jlbiasiniI think that we should start be the obvious hard and virtual keys have to be seperated
13:59:47jlbiasinithen we can add a different keympas on hold
14:00:11kugelthe clip have only hard keys
14:00:41pamaurythe right solution is probably a different keymap on hold, virtual vs hard is not sufficient
14:00:46kugelI think this has very little to do with touch or not
14:00:58jlbiasiniso it's another problem how do you call that, othogonality :p :D
14:01:27jlbiasinithis could be handle with a special keymaps on hold
14:01:46pamauryso let's go back to the current patch to implement the lock: what do we need to make that work on the sim and to be as clean as possible and to handle touchpad and touchscreen ?
14:01:47jlbiasiniI was also thinking about something like that before
14:02:31jlbiasiniyes because I'm not familliar at all with simulator so i'm a bit lost on this...
14:03:00pamauryI don't really like the touchpad_filter_enabled of the last commit, I don't get why it is there
14:03:01kugela separate keymap (in particular one that doesnt depend on the current context/screen) when locked seems sensible to me
14:03:57jlbiasinipamaury: it an imitation of the touchscreen_to_pixel implementation
14:04:44pamauryyes and I really dislike this one, this has always make the simulator even more tricky to work with touchscreen
14:05:03jlbiasiniit filter the driver output based on BUTTON_TOUCHPAD and thus provides a generic way to silent touchpad evenn on touchpad without a knowned power saving function
14:05:57jlbiasinivery usefull actually but pehraps there would be a better way to implement it
14:06:45kugelas I understand it it basically there to filter out BUTTON_TOUCHPAD (a bitmask) buttons and still let tactile keys through
14:06:46pamauryI don't like it because it puts more logic in the low level driver, which you then have to emulate in the simulator
14:07:13jlbiasiniafter all I'm only rewriting the whole stuff for the second time :D so if there are better idea on the implementation please, tell me...
14:08:51jlbiasiniI agree that not being able to test the implementation in the simulator is really a pull down
14:09:17jlbiasiniespecially when it's something that would impact about 10 different targets...
14:09:39 Join einhirn [0] (
14:10:28pamaurywhat about this: in action.c, we call touchpad_enable_device (#ifdef TOUCHPAD) and touchscreen_enable_device(#ifdef TOUCHSCREEN)
14:10:53pamauryfor native targets, we implement touch{pad,touchscreen}_enable_device (with power management for fuze+ and boolean for others)
14:11:09pamauryfor hosted, we add them in firmware/hosted/sdl/button.c
14:11:23jlbiasiniI have a question about drivers actually: why doesn't android build have a target-button.c? And where am I suppose to declare my touchscreen_enable_device then?
14:11:34pamauryand for this to work, we need a define of the touchpad buttons to filter them out in the sim
14:11:49kugelpamaury: a wrapper in button.c would be ok...that does compile for the sim doesnt it?
14:12:05pamauryyes I think it does
14:12:30jlbiasiniI totally agree with that except I don't get it at all...
14:12:38pamauryin any case, we will need some define to tell the simulator which buttons are the touchpad and which one aren't...
14:12:39kugeljlbiasini: it has a button-android.c somewhere
14:12:57jlbiasiniyes but no button-android.h
14:13:06kugelpamaury: right, that's in button-target.h which is seen by the sim
14:13:51kugeljlbiasini: it's called button-target.h
14:14:29jlbiasinibut there are none in the android dir I think, I didn't find any
14:14:36pamauryok, so we keep the BUTTON_TOUCHPAD define in button-target.h, drop the filter function and implement touchpad/screen_enable_device in hosted/sdl button driver which in turn uses a boolean and the BUTTON_TOUCHPAD define
14:14:45pamauryand for android, well I don't know
14:16:01kugeljlbiasini: it's in the app subdir
14:16:23pamauryfor android I guess we can filter touchscreen in hosted/android/button.c:button_read_device
14:16:28kugelpamaury: I think so
14:16:43pamaurydoes the android port has a softlock mechanism ?
14:16:58pamauryor does it rely on the OS ?
14:17:23jlbiasiniyes but I thinks this one is more app related and doesn't get include in button_android.c
14:17:26kugelnot that I'm aware of
14:18:28jlbiasiniin sdl there is button-sdl.h button-sdl.c apps/button-target.h app/button-application.c
14:18:37pamaurykugel: jlbiasini: let me have a try at this, I'll post the patch in a 5 minutes
14:21:24jlbiasinikugel: actually button-android.c include buttonmap.h not very consistent but we can put it there
14:22:21kugeljlbiasini: button-target.h is always included, implicitely via button.h
14:22:44kugel(if not its a bug)
14:24:58jlbiasinipamaury: what I don't get is how you are going to get generic silent touchpad when no power saving is avalaible, but I guess that I would just wait to see your implementation before complaining
14:28:06pamaurywe implement it for every target which has touchpad/touchscreen, ie no generic
14:29:31jlbiasinithe point of the filter was precisely to have a generic function... If not I can implement it myself it's not really dificult...
14:30:40jlbiasiniat least we could have a filter in action.c
14:31:00pamauryI don't really like it, from the point of view of someone not knowing how it works, if you split it in many places it gets hard to understand. I think it's just simpler this way: for a new port just implement touch*_enable_device, that's one boolean for a start so very easy and for touchpad just add an extra define for the simulator
14:31:52jlbiasiniI was thinking of describing the code in the wiki after having implemented it
14:32:06pamauryjust wait for my code and then comment ;)
14:32:08[7]wodz: pong
14:33:38wodz[7]: pamaury answered my question already (about structs in ida)
14:33:43jlbiasinibut I think the generic way is really important else It would basically be a f+ functionnality and there not need to paly around with all kind of ifdef everywhere
14:34:30 Join robin0800 [0] (
14:38:00 Join amayer [0] (
14:41:03*jlbiasini starts to understand what is pamaury is talking about...
14:47:33pamaurymay I suggest something at the same time ? It requires to answer one question: what is a touchpad ?
14:48:02pamaurybecause at the same time we do this, we could also report the touchpad absolute position in button data
14:48:19pamauryit will be unused for now but if one day someone wants to implement gestures...
14:49:31pamaurywe could even move the button interpretation from the drive to touchpad.c, using touchpad areas just like with touchscreen
14:50:26jlbiasinipamaury: for the absolute report I was also thinking about something like that
14:51:07pamaurykugel: any thoughts ?
14:51:54jlbiasinifor the second part it seems to me difficult because most touchpad device at least till now are at least partly button oriented, even the fuze+ that is probably the more advanced touchpad target we have for now
14:52:51jlbiasiniie having generic mapping like on touchscreen doesn't really make sense IMO
14:53:16pamaurynot generic, generic parser
14:53:38pamauryeach target defines regions for buttons and touchpad.c lookup this mapping to find the button
14:53:51pamauryit is how it works on the simulator btw
14:56:08 Quit kugel (Remote host closed the connection)
14:56:16 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:57:39jlbiasinipamaury: yes you are right this definitely makes sense but for what I understand f+ is for now the only real touchpad device. Others are merely a combination of separated touch element
14:58:08jlbiasiniso I don't know if it's worth to implement such a thing for only one target
14:59:40 Quit robin0800 (Quit: KVIrc 4.3.1 Aria
14:59:47pamauryI have some potential targets in mind which have a touchpad: samsung YP-Q2, YP-T10 too
15:02:03jlbiasiniok then it would makes sense...
15:02:42pamaurymaybe that should be part of another commit but it's better to think about this ahead
15:02:52pamauryso you're right, a filter like function it probably better
15:04:35jlbiasiniI guess that touchpad.c is going to be even closer to touchscreen.c so?
15:05:57pamaurywhich worries me a bit
15:06:24jlbiasiniusine gaz spirit :D
15:07:04copperwhat's the english equivalent of that btw?
15:07:23copperthe kitchen sink?
15:07:43pamaurythe thing is that touchscreen.c is mostly screen agnostic except for grid mode
15:08:06pamauryand I can't think of a device with both touchpad and touchscreen
15:08:39jlbiasiniwhat do you mean by "screen agnostic"
15:09:12pamauryit doesn't care that it is a touchscreen or a touchpad
15:09:39pamauryit's mostly in apps/ and button.c that the fact that a touchscreen is a touchscreen is used
15:10:42jlbiasinias a matter of fact a lot of touch function could be shared between all touch device
15:16:18jlbiasinipamaury: I think we should first poost on the mailing list on that and wait because else it's goning to end up with people coming it the last minute to say that it ain't right
15:16:37jlbiasinigee! this think is never to end up in the tree...
15:17:17 Quit michaelni (Ping timeout: 264 seconds)
15:18:43 Quit jlbiasini (Quit: jlbiasini)
15:23:47 Join maruk [0] (
15:28:58pamauryjlbiasini: kugel: what about this: pamaury/6353512">
15:29:56 Join michaelni [0] (
15:34:09wodzpamaury: ok, it is no surprise that updated crt0.S doesn't work on rk27xx. This core doesn't implement copro interface so you likely end up with data abort or something.
15:34:27wodzor undef instr
15:34:37pamauryah, ok, I thought this
15:34:50pamauryso that's just a #ifdef to add then
15:35:01pamaurymaybe I'll have time to look more in depth at the usb problem later today
15:36:11wodzI will have some time this evening so maybe we could colaborate
15:46:35***Saving seen data "./dancer.seen"
15:50:44 Quit krabador (Quit: Sto andando via)
16:02:11 Quit Zagor (Quit: Clint excited)
16:31:00 Quit pamaury (Ping timeout: 248 seconds)
16:55:18 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
16:56:39 Quit shamus (Ping timeout: 240 seconds)
16:56:46 Join shamus [0] (
16:57:31 Join n1s [0] (
16:57:31 Quit n1s (Changing host)
16:57:31 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:12:49[Saint]Hmmmm...yah, ...nah. iPod Video definitely needs UI boosting.
17:13:53[Saint]I guess its the relatively large LCD and relatively small CPU making the difference.
17:15:13[Saint]I thought maybe it was worth seeing if I could get away with disabling it, but, nope.
17:16:26wodzit also gets quite a few irqs from scrollwheel
17:18:17[Saint]I got away with disabling it on the Classic, initially - early on in the port - it was basically unusable without boosting on UI interaction, now however - similar to the Fuze + - the CPU almost never needs to boost under normal circumstances.
17:18:49[Saint]I don't think this adds up to very much power saving, though, as that SoC is quite efficient.
17:19:23copperscrolling lines are kinda slow on my Classic with my themes
17:21:36 Quit Poodlemastah (Quit: ZNC -
17:22:34[Saint]well, yay, glad that worked.
17:23:10[Saint]I now have an ipod bootloader that doesn't care about the Apple OS or IPL anymore.
17:24:27[Saint]Perfect for OSOS install - since booting the Apple OF isn't actually possible.
17:24:45[Saint]...and IPL is dead.
17:28:41 Quit wodz (Quit: Leaving)
17:32:44 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:34:26 Quit petur (Quit: Nettalk6 -
17:39:43 Join madcat1990 [0] (~madcat199@
17:44:26 Quit Zagor (Quit: Leaving)
17:45:03madcat1990Well, I was gonna report a bug, but apparently it has been fixed in Dev builds
17:46:19madcat1990I don't know what it is that is doing it, but here's how to reproduce it each time in 3.13 stable, and what it shows :
17:46:32madcat1990Seeing that it is fixed in Dev, I don't know if I should report it or not :X
17:46:36***Saving seen data "./dancer.seen"
17:47:20[Saint]If it is fixed in git head, don't report it, no.
17:53:05 Nick DormantBrain is now known as SuperBrainAK (~andy@
17:53:20madcat1990Fair enough, I thought the same. Just wanted to make sure in case you guys needed it documented
17:54:15madcat1990Btw, have I thanked you guys for the awesome work you do, today?
17:54:40[Saint]Not /today/... :)
17:54:58[Saint]But I have seen you do so in the past. Its appreciated, I'm sure.
17:56:09n1smadcat1990: btw, the debug menu is not intended for users and bugs there should be expected
17:56:38[Saint]I don't think bugs should ever be expected.
17:56:50[Saint]Its not intended for users, sure, but it should work.
17:57:00madcat1990[Saint], I agree wholeheartedly. It is my duty, as tester, to report any findings
17:57:28madcat1990Also, I'm waiting on my next paycheck to make another donation
17:57:46madcat1990>.< it has now reached the point where I find other Firmwares ugly, and dysfunctional.
17:58:42[Saint]One should expect to find bugs in the git head builds, possibly, and that's kinda the point of them...any reported bug is better than one that goes unnoticed, even if the user isn't technically supposed to be poking around there. :)
17:58:57[Saint]Anyhoo - its fixed, so - yay. :)
17:59:15madcat1990I always poke around there. Especially the buffering thread one. I love to see how my device is buffering
17:59:56madcat1990And the debug scrollwheel when I had my iMini2G, loved it too, just through that one screen, I found it that the scroll wheel is nothing more than a touchpad
17:59:57[Saint]I only find it useful myself to check buffer allocations from the skin engine.
18:00:03madcat1990not ugly magic
18:00:19madcat1990Yeah =/ that engine is breaking USB on my device
18:00:25n1swell, i'm pretty sure there are known bugs in there that noone has cared to fix but i don't remember any specific one right now
18:00:25madcat1990Default themes work fine in rockbox though
18:00:36[Saint]you and a lot of other people.
18:00:52[Saint]and cabbie working, where others don't, makes no sense at all.
18:01:02[Saint]I *really* wish I knew what was going on there.
18:01:16[Saint]like, really...really really.
18:01:19madcat1990This might be a stupid suggestion, but would the size of the files make a difference? or what they contain?
18:01:29madcat1990there *HAS* to be something the others use that cabbie doesn't
18:01:38madcat1990or something they use more
18:01:44madcat1990Thusly, breaking the skinning engine
18:01:56[Saint]It very well may. I don't know. My best guess is bad alignment of memory allocations.
18:02:09madcat1990But that's not done through the WPSes, is it?
18:02:25[Saint]Why some themes slide though, even when they are vastly larger than cabbie and far more complex code, I do not know.
18:02:35[Saint]Additionally, simpler, smaller themes also fail.
18:02:39[Saint]It is baffling.
18:02:48madcat1990Then its not a thing of size
18:02:56madcat1990its a thing of, a parameter being used that breaks it
18:03:23[Saint]No, even very simple themes that are essentially exact clones of cabbie can break USB.
18:03:35[Saint]Its not, afaik, any one tag or combination thereof.
18:03:40[Saint]It is *truly* baffling.
18:03:49[Saint]I can't find a definite trigger.
18:04:24[Saint]It has excaped the attempts of about half a dozen people's serious attempts at debugging.
18:05:07[Saint]lots of bugs have been caught and fixed whilst looking for the cause of this one, though, so that's great.
18:05:09madcat1990Has anyone tried REPLACING cabbie with another theme and attempting to replicate it?
18:05:16madcat1990maybe chunks of them are being loaded into memory, and left there?
18:06:08[Saint]I don't think so. The way memory allocation works shouldn't allow for this. And I have looked at the bufflib allocations whilst changing themes and there is no indication of leftovers.
18:06:20[Saint]I honestly have NFI what is happening.
18:06:56madcat1990You know what you guys need? a visual representation of a device's memory
18:07:25madcat1990something along the lines of this :
18:11:42 Join polemon_ [0] (
18:12:11 Join rdn [0] (
18:14:41 Quit polemon__ (Ping timeout: 240 seconds)
18:18:49 Join Strife89 [0] (~Strife89@2602:306:250f:2289:10e6:66d5:5350:6cbc)
18:21:09 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
18:21:32 Join wodz [0] (
18:22:41wodzmadcat1990: This is awersome. You need full system emulator though. We don't have such thing for any of our targets. Writing one is HUUUUGE task.
18:23:32madcat1990I know. But it'd solve most of our debugging problems. :)
18:24:00wodzNot really.
18:24:25wodzthere are things you cant relibly emulate
18:25:23Bagderthere are?
18:27:08wodzok, 'can't' is unfortunate 'are hard to' is better
18:27:33TorneWell, there's *lots* of things we can't emulate because we don't know how they actually work
18:27:49Tornewriting a full system emulator for a system that you don't ahve docs for requires assuming that your reverse-engineering is correct :)
18:28:08Bagdertrue, but it would also help the reverse-engineering!
18:28:34madcat1990Bagder, if anything it'd prove ours wrong.
18:28:53Torneobviously there are classes of bugs you can still catch even then, but some of the time you are just begging the question :)
18:30:16madcat1990I really doubt there's an easy way to emulate ALL these devices.
18:30:43madcat1990Best thing to do is replicate Rockbox's behavior on them, but that won't help us much
18:30:55madcat1990Though, if by some chance, we can make a memory dump through USB
18:30:58madcat1990and see what's happening?..
18:31:16wodzmadcat1990: basically pamaury is working on this :-)
18:31:51madcat1990So me and one of the devs are in the same wavelength. I'm not a complete idiot! yay
18:32:34wodzBUT transfering mem dump through usb means that you cannot live test rockbox usb stack (although you can take a snapshot at particular event)
18:34:27madcat1990Might not live test USB stack, but you can see when the skinning engine is being naughty
18:35:12 Join eyfour [0] (
18:35:34wodzIf the bug is reproducible on f+ I think we might ask pamaury to try.
18:36:31 Join Poodlemastah [0] (
18:37:00madcat1990wodz, I can prolly ask a buddy of mine who bought one of those to test it
18:39:23 Quit maruk (Quit: Leaving.)
18:39:55[Saint]It is.
18:40:07[Saint]...reproducible on Fuze +, I mean.
18:45:50pamaurymadcat1990: emulating a device like the fuze+ is really a lot of work, the documentation is very vague on a number of (crucial) points, so that probably wouldn't help much. What will help is live debugging, that we are currently trying to implement
18:47:01madcat1990pamaury, I realise this. Which is why I'm saying, live USB Debugging will allow us to poke at the memory, but not live USB monitoring. But it will at least allow us to check what the skinning engine is doing
18:47:04madcat1990and what's causing it to misbehave
18:49:50wodzpamaury: can we try to insert hwstub into rockbox and trigger it on usb connect? That way we will be able to take mem snapshot at arbitrary time
18:50:05madcat1990And dump it to the SD?
18:50:13wodzno to PC :-)
18:56:35 Join pretty_function [0] (~sigBART@
18:57:46 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
18:57:49 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
19:00:28wodzIf I have struct foo_t { char a, char b, short c, char d} how it will be padded in array? I mean struct foo_t ble[100]? I guess compiler will try to guarantee halfword alignment of member c in every array item so should I expect trailing padding byte for every item?
19:01:35pamaurywodz: that's already more or less possible with my hwpatcher tool, although not very practical. But memory dump won't be easy to analyse either
19:03:14wodzI don't say it will be easy
19:03:47wodzpamaury: btw. what is the status of your amsv2 rom disasm?
19:10:15 Nick SuperBrainAK is now known as DormantBrain (~andy@
19:13:37 Quit eyfour (Quit: WeeChat 0.3.7)
19:16:43 Quit shamus (Ping timeout: 240 seconds)
19:22:07 Join shamus [0] (
19:26:21 Quit [Saint] (Remote host closed the connection)
19:26:51 Part LinusN
19:27:47 Join [Saint] [0] (~saint@rockbox/user/saint)
19:37:17pamaurywodz: haven't work on this for a while, can send you my latest progress
19:39:45wodzpamaury: I was just curious
19:46:39***Saving seen data "./dancer.seen"
19:48:48 Join polemon__ [0] (
19:50:56 Quit polemon_ (Ping timeout: 245 seconds)
19:57:47 Quit bertrik (Remote host closed the connection)
20:51:42 Quit thomasjfox (Remote host closed the connection)
20:54:30 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox)
20:59:18 Quit pretty_function (Remote host closed the connection)
21:09:13 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
21:46:41***Saving seen data "./dancer.seen"
21:57:45 Quit madcat1990 (Quit: Leaving)
22:03:18 Quit kevku (Ping timeout: 260 seconds)
22:32:16 Join bertrik [0] (~quassel@2001:610:76a:0:749c:5212:ba7c:37ba)
22:32:16 Quit bertrik (Changing host)
22:32:16 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
22:32:33 Join thomasjfox_ [0] (~thomasjfo@rockbox/developer/thomasjfox)
22:32:44 Join jlbiasini [0] (
22:33:12jlbiasinipamaury: ping
22:34:31 Quit thomasjfox (Ping timeout: 276 seconds)
22:34:58 Nick thomasjfox_ is now known as thomasjfox (~thomasjfo@rockbox/developer/thomasjfox)
22:48:44jlbiasinipamaury: I just tried your patch it apply cleanly but gave me a lot of bug: (this is just the beginning I cuted the following as it was endless...
22:59:44 Join einhirn [0] (
23:00:18 Quit Strife89 (Quit: Heading out.)
23:08:43 Nick DormantBrain is now known as SuperBrainAK (~andy@
23:16:03 Join dewlap [0] (~dewlap@2001:5c0:1000:a::ad7)
23:17:32 Quit bertrik (Remote host closed the connection)
23:24:30 Quit n1s (Quit: Ex-Chat)
23:26:29 Quit amayer (Quit: Leaving)
23:30:43pamauryjlbiasini: as I said I didn't try to compile it, you have to learn how to fix this ;) In button.h I wrote #define "touchpad.h" instead of #include
23:31:33jlbiasiniok i'll have a look tomorrow then...
23:32:04pamaurytrying to compile now...
23:32:23pamauryok that does the work for touchpad, now touchscreen...
23:33:02wodzpamaury: If you find time could you check if hwstub gets past setup request at all?
23:33:15pamaurywodz: last time I tried it didn't
23:34:02pamauryjlbiasini: I have updated the patch at pamaury/6353512">
23:34:31wodzpamaury: So at least we have the same behavior
23:35:16pamaurylet me have a try now
23:40:14pamaurywodz: you should definitely fix the usage() of tools/scramble.c for rockchip
23:40:26pamauryeach time I use I have to lookup the source to realise I need the -modelnum=
23:41:02wodzpamaury: sure, will fix it
23:42:28pamauryhum, I get undefined instruction if I try to run it from microsd with scramble
23:43:34pamauryapparently the relocate code is broken on rk27xx
23:44:20jlbiasinipamaury: your patch work on f+ I'm givin a try on the ondav777 simulator to see on a touchscreen target
23:45:21jlbiasinisome more compilation error, I'll see tomorrow
23:45:25 Quit jlbiasini (Quit: jlbiasini)
23:46:44***Saving seen data "./dancer.seen"
23:58:21 Join petur [0] (~petur@rockbox/developer/petur)

Previous day | Next day