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 2013-07-07

00:03:10 Quit y4n (Quit: Do you like hurting other people?)
00:09:33 Quit kevku (Ping timeout: 245 seconds)
00:27:26 Quit melmothX (Quit: #.#)
01:35:04 Quit n1s (Quit: Ex-Chat)
01:51:09***Saving seen data "./dancer.seen"
01:52:52 Quit pamaury_ (Ping timeout: 264 seconds)
01:54:57 Quit lebellium (Quit: ChatZilla 0.9.90 [Firefox 23.0/20130703181823])
02:06:06 Quit shamus (Read error: Connection reset by peer)
02:06:19 Join shamus [0] (
02:55:43 Quit ender` (Quit: It's amazing how the the human mind does not process the the fact I used the the word "the" twice each time.)
03:38:16 Quit Rower (Quit: Hmmm...)
03:51:12***Saving seen data "./dancer.seen"
04:02:15[Saint]Bwahahahaha....yeah, ...nah, that's not gonna happen.
04:03:09[Saint]buflib is *so* entrenched in the heart of Rockbox, it is basically impossible to cast the bugger out without a LOT of backbreaking.
04:04:00[Saint]I would be very curious to see if some of these strange issues just magically fix themselves if it were to no longer be present, as I suspect they might.
04:05:42[Saint]There's no "cut buflib out" its basically "go back in time to pre-buflib, and then massage back in a shit-tonne of commits, many of which rely on it"
04:05:49[Saint], nope! :)
04:07:41[Saint]I have a sneaking suspicion that if one were to core out either buflib, the theme engine, or both, many active flyspray tasks would become irrelevant.
04:08:22[Saint]I like these types of suspicion, though, I get to have an opinion whilst knowing that it would be a massive amount of work for anyone to prove me wrong ;)
04:54:31 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:54:32 Quit pixelma (Disconnected by services)
04:54:34 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:54:39 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:54:40 Quit amiconn (Disconnected by services)
04:54:42 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
05:14:06 Quit [7] (Disconnected by services)
05:14:11 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:23:32 Quit belak (Ping timeout: 264 seconds)
05:28:29 Join belak [0] (~belak@facebook/engineering/belak)
05:37:27 Quit prof_wolfff (Ping timeout: 268 seconds)
05:51:14***Saving seen data "./dancer.seen"
05:57:12 Quit Guest11127 (Ping timeout: 245 seconds)
06:18:35 Join Guest11127 [0] (
06:32:14 Nick yosafbridge` is now known as yosafbridge (
06:33:20 Quit Epicanis (Quit: to sleep, perchance to dream)
06:43:23 Nick SuperBrainAK is now known as DormantBrain (
06:52:01 Join Zarggg [0] (
07:08:46 Join shai [0] (
07:10:38 Quit shai (Client Quit)
07:17:52 Join kevku [0] (~kevku@2001:0:c38c:c38c:1046:9510:3d69:bea5)
07:27:54jhMikeS[Saint]: I know what to do. No need to kill buflib. I just don't like stepping all over everything in sight, even if I want to.,
07:28:46[Saint]So we can test "the old ways", by disconnecting buflib, and leaving it in place?
07:29:01[Saint]I didn't see that as being possible, or rather, more work than just gutting it out.
07:29:27[Saint]also, agreed, either way involves touching pretty much everything.
07:30:45jhMikeSI think tweaking the design is the way. Many problems were there before it that are still there but it does introduce its own difficulties.
07:51:16***Saving seen data "./dancer.seen"
08:17:05 Quit sakax (Read error: Connection reset by peer)
08:27:21 Quit bluebrother (Disconnected by services)
08:27:26 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
08:34:54 Join Guinness` [0] (
08:36:49 Join kiwicam [0] (~quassel@
08:39:12 Join derf_ [0] (
08:39:28 Join uwe_ [0] (
08:42:55 Join lahwran_ [0] (~lahwran@python/site-packages/lahwran)
08:43:44 Quit Guest11127 (*.net *.split)
08:43:44 Quit user890104 (*.net *.split)
08:43:44 Quit lahwran (*.net *.split)
08:43:45 Quit uwe__ (*.net *.split)
08:43:46 Quit derf (*.net *.split)
08:43:46 Quit Guest83379 (*.net *.split)
08:43:47 Nick derf_ is now known as derf (
08:44:44 Join sakax [0] (
08:44:46 Quit lahwran_ (Excess Flood)
08:48:46 Quit kiwicam (Remote host closed the connection)
08:56:16 Join kiwicam [0] (~quassel@
09:07:33 Quit copper (Quit: ZNC -
09:12:52 Join user890104 [0] (
09:41:56 Quit belak (Quit: belak)
09:51:18***Saving seen data "./dancer.seen"
10:02:25 Join belak [0] (~belak@facebook/engineering/belak)
10:04:21 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
10:13:58 Join copper [0] (~copper@unaffiliated/copper)
10:14:43 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
10:41:35kugel[Saint]: where did your buflib rant tonight come from?
10:42:40[Saint]I'm just sick of seeing all these random issues that came to light (perhaps, by chance) at approximately the same time.
10:42:56[Saint]Its hard to see this as what seems to be an accepted issue now.
10:43:37[Saint]Its nice to know it isn't /just/ me that noticed the timing.
10:44:08[Saint]The rant came from how insanely hard it would be to actually confirm this.
10:45:19copperthe timing with regards to what?
10:45:30coppersome recent change?
10:45:49[Saint]primarily the adoption of buflib by the theme engine.
10:45:54[Saint]this isn't a recent thing.
10:46:32kugel[Saint]: which issue?
10:46:57[Saint]"themes cause <insert batshit crazy behavior here>"
10:47:23kugelcan you point to a specific fs task or forum post?
10:47:25copperis that over a year old?
10:48:32[Saint]lebellium seems to be the master of crafting themes that manage to break USB.
10:48:38[Saint]This seems to be the primary issue.
10:48:55kugelI know about the radioart and rds text bugs (the first is related to, but not directly caused by, buflib)
10:49:07[Saint]Another user was in here a few days ago trying to help uys with the N2G USB stuff, and changing themes affected the behaviour there as well.
10:49:44[Saint]the primary issue is "themes somehow break USB"
10:49:51[Saint]and its been popping up more and more recently.
10:50:04kugelwould be worth trying if my fix for the radio art also helps for USB bugs
10:50:15[Saint]Oh, certainly.
10:50:39[Saint]FWIW, I'm not blaming you. So, don't read into that. I appreciate the work, immensely.
10:50:52[Saint]I just think, in hindsight, it may have gone a little too quickly.
10:51:40[Saint]The problem for those trying to look into it is that it is quite complicated, and touches basically everything.
10:51:49copper[Saint]: how much of that (themes break USB) is people repeating what they read here?
10:52:06copperi.e. a rampant rumor
10:52:39kugel[Saint]: buflib itself is not the problem (it's pretty solid in my experience), but using it properly indeed adds complexity
10:52:53[Saint]copper: I couldn't say, really. But a lot of it is "confirmed" by the user with trial and error. Many users I have not seen here before, although, it may be under a different name.
10:53:19[Saint]ie. "USB works with theme A, but not with themes B, C, and D"
10:53:30kugelthe playback code is using it improperly which causes radioart problems (seemingly unrelated, but the code is connected through the global buffer management)
10:53:35[Saint]which, just should happen. It shouldn;t be possible.
10:54:33copperwell it seems that my Fuze+ theme download count increased dramatically after I rewored the theme files to potential fix USB breakage
10:55:11[Saint]I wanted to attempt to go back in tiome to pre-buflib, and see if I could massage the work til now back into it...but that is an incredibly daunting task.
10:55:12copperI wish there were stats
10:55:25[Saint]So much so I looked at it for a few minutes and burst out laughing.
10:56:06kugelthinking about it, yes my radio art fix could very well fix the usb problem as well
10:56:19copperwhat is radio art?
10:56:32 Join n1s [0] (
10:56:33 Quit n1s (Changing host)
10:56:33 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:56:33[Saint]station art for FMS
10:56:49copperwhere does the art come from?
10:57:02[Saint]ie. "when tuner_freq == foo; display bar"
10:57:08[Saint]theme assets.
10:57:17kugel[Saint]: the skin engine should be easy to disconnect from buflib
10:57:43[Saint]you and I have varying degrees of ease, apparently ;)
10:57:51kugelit has the classic allocation mechanism still in, e.g. checkwps doesnt include buflib
10:57:57coppermy theme went from 100+ over a few months, to 900+ in like a month
10:58:13[Saint]I thought about just the theme engine alone, but I couldn't discount it just being broken in general across the board.
10:58:28[Saint]so I wanted to cut it out completely.
10:58:53kugelno, that's not easy, in the same way it wasn't easy to add it to the codebase :)
10:59:51kugel[Saint]: please point people with theme/usb problems to
10:59:59[Saint]I understand how my earlier rant may have seemed like "Gah! Bloody kugel...bloody breakin' things!", but that wasn't the case, sorry. I was just frustrated.
11:01:57 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:02:18[Saint]buflib seems to be the new tagcache. :)
11:02:31[Saint]"Errrr....I'm not touching that!"
11:03:08 Quit akaWolf (Quit: my exit)
11:05:21 Join polemon [0] (
11:05:28[Saint]jhMikeS seemed as though he may have an idea of what is going awry, but he is often quite secretive/vague about what he is working on until we see the WIP, so I dunno.
11:06:19[Saint]After many combined hours of staring at it, I have only a very basic understanding of it myself.
11:06:52 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
11:07:00 Join melmothX [0] (~melmoth@unaffiliated/melmothx)
11:07:08 Join olspookishmagus [0] (~pookie@
11:08:41kugelgah, stupid whitespace-only changes :/
11:13:11kugel[Saint]: can you reproduce the usb problems?
11:14:37[Saint]I have been unable to, sadly. I suspect *nix Vs. Windows may be at play here.
11:15:07[Saint]I attempted to the other day (this was N2G specific), but I couldn't reproduce what OP was seeing.
11:16:00[Saint]That is another difficulty. Assuming it *is* buflib related, it may not /just/ be specific to the theme, but also the exact revision of the build, and whatever else resides in the buffer at the time.
11:16:51[Saint]...a way to dump the current device state and play with it in an emulator would be *awesome*.
11:18:05[Saint]It is quite a nasty cycle. Those that see the problem usually can't debug it, and those who can debug it can't see it. :-S
11:20:16 Quit akaWolf (Read error: Connection reset by peer)
11:21:38 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
11:25:56 Join lebellium [0] (
11:26:36lebelliumI read the logs :)
11:26:48lebelliumso I should try your patch kugel?
11:28:01[Saint]Since you are a 5th Dan Blackbelt in creating themes that break USB, yes. :)
11:28:57lebelliumHehe ;)
11:30:08lebelliumI'll try with Clip Zip after my breakfast. But I already see the same issue as usual: I just plugged my Clip Zip to the PC and this time it works while last times I tried it froze with my theme. Random random.... :(
11:32:09 Join Rower [0] (
11:33:06lebelliumfrom the main menu with no music playback it just worked. But now I try again while playing music and from WPS and now it freezes as usual. Nice. I'll see if this patch helps
11:34:01[Saint]That seems to confirm my thinking that whatever is currently in the buffer has an effect on the behavior.
11:35:10pamauryhaha, good news! I'm 99% sure i've identified the reason why the fuze+ locks up after bootloader USB mode :)
11:35:25[Saint] is bufib?
11:35:29*[Saint] runs...
11:35:50pamaury;) not really :p
11:37:14pamauryit's related to power management, the init sequences changes the cpu speed before calling power management. But on imx233, changing the frequency requires power related changes, and something goes wrong apparently
11:40:02kugellebellium: does it freeze on connect or on disconnect
11:46:44lebelliumkugel: both? I mean it won't connect (sometimes starting to get recognized in Windows Explorer then fails) and when I disconnect it, it's stuck, I have to reset the device
11:47:27lebelliumI try your patch in a few minutes
11:50:23polemonhey guys
11:50:44polemondid you manage to work on the USB issue?
11:51:22***Saving seen data "./dancer.seen"
11:53:13kugellebellium: can I download the theme?
12:01:37copperrayboradio on the Classic is a known offender
12:02:14lebelliumkugel: lebellium%20Samsung-like&searchtype=name">
12:02:58copperlebellium: nice link
12:04:07 Quit onder` (Read error: Connection reset by peer)
12:04:20copperyou've been busy!
12:05:17[Saint]Someone likes Rammstein *way* too much ;)
12:05:36[Saint]that, and VBR mp3...
12:05:39lebelliumthat's marketing. To identify any of my theme in one second :D
12:06:57lebelliumbut now a common joke on some forums is that I only have one album on my mp3 players ahaha
12:07:26[Saint]There's no proof you have the rest of the album ;)
12:12:52 Quit pamaury (Ping timeout: 248 seconds)
12:15:44kugellebellium: another question, since you maintain a theme for so many targets: What would help to reduce your workload and enable sharing a single code base for many/all themes?
12:15:59 Join onder` [0] (~onder@
12:17:12 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
12:17:29kugelhm, the mouse navigation doesnt quite work in the clipzip sim
12:17:44lebelliumI use more or less the same code for any of my themes. But I'm not sure it's really clean, easy to understand. It's not the cabbiev2 quality standard :)
12:18:01[Saint]kugel: nor does trying to reproduce USb errors in the SIM, I'm afriad.
12:18:15[Saint]I have *never* been able to get USb to behave anywhere near the real thing.
12:18:30*[Saint] wonders why he always manages to type USb
12:22:25copperis the theme/buflib thing the ONLY known cause of USB problems?
12:22:56[Saint]I'm not even sure anyone can say with absolute certainty it is the cause.
12:24:19copperin other words: have people reported USB problems with cabbiev2?
12:25:06lebelliumkugel: the patch doesn't fix the issue, sorry
12:25:21kugellebellium: okay, that's unfortunate
12:25:26[Saint]copper: yes.
12:25:47kugelI'm not aware of reports with cabbiev2
12:25:55[Saint]but I'm not sure what that has to do with your prior question, nor how it could be another wording thereof.
12:26:19[Saint]cabbie still uses the theme engine/buflib just as any other theme does.
12:26:25copperwell I thought it was a given that cabbiev2 didn't exhibit any of the theme-related issues
12:26:44 Quit melmothX (Read error: Connection reset by peer)
12:26:48copperpast conversations
12:26:56copperI guess I got it wrong
12:27:19[Saint]It does happen. I guess it relates to the rather small footprint cabbie has.
12:27:43kugelI assume it doesnt happen with cabbie until it's on FS
12:28:05kugelI'm not hunting bugs based on rumors
12:28:51lebellium*PANIC* playlist_resume (); 00M when disconnecting it. That's a new one for me :D
12:29:06 Join melmothX [0] (~melmoth@unaffiliated/melmothx)
12:29:12[Saint]polemon was playing with this the other day.
12:29:34[Saint]He initially stated that cabbieV2 *didn't* display any issues, but then later discovered it in fact...did.
12:29:34kugellebellium: so it didn't connect but also didnt freeze? then this panic occured on disconnect?
12:30:16kugeldoes it connect fine with other themes?
12:30:39kugel[Saint]: I don't even know who polemon is
12:30:44lebelliumkugel: this time it connected properly but when disconnecting, instead of getting stuck on the USB screen, it directly displayed this PANIC message
12:30:55[Saint]logged channel.
12:31:06kugellebellium: okay, that's an improvement :)
12:31:28kugelit means the buffer is still owned by someone else when the playlist code tries to grab it
12:31:50kugelwithout the patch two pieces of code would access the buffer at the same time
12:32:40[Saint]polemon actually discovered something that isn't so easy to see on many targets, that in many cases, USB behaved fine if USB was inserted *before* booting the device.
12:33:18lebelliumyes there may be some improvements with this patch but unfortunately I cannot check that very well since it's totally random. Over the last 15 minutes I connected it to the computer 10-15x and sometimes it works, sometimes not, in the same conditions :(
12:33:55[Saint]Yes, that's about as far as we got the other day "Oh! Ohh! I think there's a pattern...oh, wait, nope...shit."
12:34:18kugellebellium: can you add that diff on top?
12:34:58kugel[Saint]: please log such observations on the relavant FS tasks (or open new ones)
12:35:26[Saint]Isn't that basically heresay?
12:35:35lebelliumI'm not sure what I am supposed to do with that kugel, I'm still a git/linux/rockbox newbie
12:35:40[Saint]I can only direct those seeing the problem to do so.
12:36:01kugelyou can't expect people to read all backlogs when they (re-)visit such stuff later on
12:36:44kugel[Saint]: yes, I meant to say that
12:36:51[Saint]I don't. But is it not trivial to grep logs when given the information to do so?
12:37:14kugelno it isnt, it also isnt trivial to get a local copy of all logs
12:37:25[Saint]I certainly don't read all the backlog. No way I expect anyone else to.
12:38:08lebelliumkugel: how do I add this diff? just "cd Rockbox" and paste that text in the terminal?
12:38:15kugeljust saying that these observations have to be logged where they're relevant and can be found easily
12:38:22kugelirc logs aren't suitable for this
12:38:41kugellebellium: download as raw and use patch to apply
12:38:57kugelor edit the file manually, it's a one-liner :)
12:39:08[Saint]lebellium: download as raw, save as "name_here>.patch; apply with patch
12:39:20[Saint]or, apply it by hand (more difficult)
12:39:40kugel[Saint]: no need to repeat me :P
12:39:43lebelliumfunny, I would say changing a line in a file is easier for me than applying a patch in the terminal :p
12:40:01[Saint]I was just going to correct that, usually it isn't a single line ;)
12:40:10[Saint]apparently I'm also lagging.
12:48:23kugelif what I suspect is true than this is a stone-age old bug (older than buflib)
12:51:04kugellebellium: does that line help?
12:51:06[Saint]Do tell, now my connection seems to have sorted itself out.
12:51:51lebelliumkugel; I'm trying to mount Clip Zip to install the new build, but somehow even with cabbiev2 it won't mount now :)
12:52:17[Saint]OF, rolo from sdcard?
12:52:31[Saint]There's options man. Options! :P
12:53:00lebelliumOF means 20 min database update. Never lol
12:53:55kugelwhen I last used the OF it didnt update when you boot with the cable inserted and you could hard-reset it between unmounting/safe removal and unplugging the cable
12:55:04kugelthe clipzip is as3525 like older clips right?
12:57:30funmankugel: amsv2 like clipv2 and clip+
12:57:44funmanso as3525 + tiny changes
12:57:52kugelah yea, I remember
12:58:12kugelit's a long time ago since I was into that :)
12:58:21funmanbiggest change is as3543 PMU + different USB and SD controllers
12:58:29funmanme too but it's printed deep in my mind :P
13:00:25kugeldifferent USB controller?
13:00:34funmanamsv2 controller is the same than nano2g
13:01:02lebelliumkugel: next time I don't listen to you :P "Refreshing your media" and it takes forever with 8 + 32GB
13:01:15kugelfunman: aren't there problems with that controller (or the driver)?
13:01:29kugellebellium: up on boot?
13:01:44 Quit polemon (Ping timeout: 256 seconds)
13:02:20kugelhm, don't know. all samsa OFs I've used skipped that when the cable was inserted at boot
13:02:22funmankugel: yes it crashes sometimes but works most of the time
13:02:58kugelperhaps the connect problems lebellium reports are caused by that rather than buffer/buflib/skin engine issues?
13:03:24kugelalso [Saint] said the nano2g has similar issues?
13:04:02[Saint]don't they share the same USB controller?
13:04:27[Saint]Its shared with at least one other target, iirc.
13:04:40[Saint]N2G USB being broken has been a known issue for some time, though.
13:10:23 Quit kevku (Ping timeout: 245 seconds)
13:12:30kugellebellium: still refreshing? ::(
13:12:48lebelliumupdated RB build and now "refreshing media" while disconnecting. Arghhhh I'll kill Sansa people
13:14:01kugelsaratoga: can you please refer to the original task when you close ones as duplicate?
13:14:57kugellebellium: the rolo-from-sd idea would have been quicker
13:15:13lebelliumprobably. If I new what that means :D
13:15:43kugelput the new rockbox.sansa on the sd and "play it" from the old rockbox installation
13:15:59kugelor copy it over to the .rockbox folder and reboot
13:17:21lebelliumok good to know
13:19:10kugel(you could also have removed the sd while booting the OF so the refresh doesnt last as long)
13:24:17lebellium1st try: worked. 2nd try: doesn't mount, and PANIC playlist_resume while disconnecting
13:25:08 Join kevku [0] (~kevku@2001:0:c38c:c38c:2445:cf4b:3d69:be3e)
13:25:20kugellebellium: can you retry with that line changed to "sleep(HZ*10);"?
13:26:57kugelthat's a 10s sleep so don't be surprised if the UI doesnt do anything for that long after disconnect
13:43:47 Quit olspookishmagus (Quit: All for nothing)
13:47:43lebelliumkugel; 1st try: 10 seconds then OK, 2nd try: 10 seconds then Panic playlist_resume
13:47:47lebelliumwhile disconnecting
13:48:02kugeldid it connect?
13:48:10lebelliumyes both times
13:48:27kugelgevaerts: ping
13:49:06lebelliumand now 3rd time: won't connect "undefined instruction". I take a break, these USB issues make me crazy
13:49:17kugelhm, I'm surprised it connected in the 2nd case
13:49:38kugelI can see how it panics when it didnt properly connect, but if it connected it shouldnt panic
13:50:11kugelundefined instruction? Now that's really weird
13:50:41lebelliumand sometimes it replaces my status bar with the failsafe status bar to connect
13:50:57lebelliumit's another behaviour every time I connect it
13:51:25***Saving seen data "./dancer.seen"
13:53:35lebelliumI confirm it may mount properly but when disconnecting, it displays panic playlist_resume after 10s
13:53:39lebelliumI got it a 2nd time
14:03:55kugellebellium: alright, I need the help of gevaerts now, thanks for your testing
14:06:19lebelliumyou're welcome. But given the fact it's a highly random issue, I don't think it's possible to fix it, unless a dev has a Clip Zip and faces the same issue.
14:08:29lebelliumOnly [Saint] talked about that until now. All other reports are from normal users like me, not devs.
14:09:06[Saint]Of course its possible to fix.
14:09:10kugelthere aren't so many devs anymore :(
14:09:36[Saint]It never used to happen, which means we were either A: getting it right, or B: not pissing it off as much as we are now.
14:09:46[Saint]the trick is getting it back to one of those states :)
14:10:12copperbisecting over a year's worth of commits?
14:10:13[Saint]Also, sadly, yes.
14:10:15coppershould be fun!
14:10:29[Saint]In terms of active committers, this project is in reasonably bad shape.
14:10:39kugelbisecting random behavior isnt useful
14:11:04[Saint]We need new blood. But DAPs are no longer fun.
14:11:16copperthey're not?
14:11:31lebellium[Saint]: Rockbox has been ported to Clip Zip since late 2011 now, almost 2 years. I've been reported the theme issues since then. Nothing new on themes breaking USB for 2 years. I really think that if a dev faced the issue, that would be easier to fix.
14:11:39[Saint]Seeing as how a phone can do everything, and more, no..not really.
14:12:20coppermany smartphones are still low capacity, and many have bad output
14:12:31copperlow capacity without sdcard expansion*
14:12:48[Saint]data plans+cloud storage
14:13:09copperstreaming is impractical in many situations
14:13:10kugelcopper: most phones are good enough
14:13:26copper(if not most situations)
14:13:58copperI mean there are many reasons people still buy DAPs
14:14:13coppereven though that includes Android and iOS DAPs
14:14:23[Saint]Right. But those people are A: aging, and/or B: enthusiasts.
14:14:32[Saint]look around your various hifi communities.
14:14:41[Saint]Those that frequent them are aging.
14:16:16kugellebellium: I think it's expected that the default statusbar is shown
14:16:50lebelliumkugel: AFAIK, no. A
14:16:58kugeliirc skins are disabled during usb because they require access to the storage (those with custom fonts at least) which cannot be done
14:17:10lebelliumskins are not disabled during USB
14:17:59copper[Saint]: indeed, I guess that younger people also have lower storage needs
14:18:00[Saint]it _should_ dump what it needs into the buffer when the screen changes.
14:18:12[Saint]No need to access the disc for the theme during connection.
14:18:14kugellebellium: okay, but the fonts are unloaded
14:18:23kugelno idea how that works if the skin uses that font
14:18:34[Saint]It seems to "Just Work"
14:18:43[Saint]My themes maintain userfont during USB.
14:18:59lebelliumI don't know how it works but with any of my theme, only the menu list is replaced by the rockbox connection logo. The background and my status bar (with fonts) is still there
14:19:08kugel[Saint]: the code clearly unloads all fonts
14:19:22[Saint]Yay for bugs, then.
14:19:45kugeland it calls skin_unload_all(), guess what that does
14:20:04[Saint]well it certainly doesn't clear the .sbs...I know that much :)
14:20:10[Saint]perhaps it _should_.
14:20:14[Saint] doesn't, though.
14:20:21kugelthat's at least the intention of the code
14:21:18 Quit AlexP (Remote host closed the connection)
14:21:28[Saint]Why does it need to dump the font? We know *exactly* what needs to be displayed, couldn't it just keen N many glyphs in the buffer?
14:21:33[Saint]Or, all of them, for that matter?
14:21:44[Saint]I don't understand why it would need to access the disc.
14:22:13kugelyou can't determine if the glyhps are in cache
14:22:24pamauryyeah the number of active devs has decreased recently, mostly because we have no new devs. I guess people just stopped using mp3 players
14:22:36kugelthe skin can display arbitrary strings, you cannot know if the glyphs needed for that will be accessible
14:22:38[Saint]git hurt us a lot, too.
14:22:39 Join AlexP [0] (~alex@rockbox/staff/AlexP)
14:22:43[Saint]which is *fucking* sad.
14:22:47[Saint]...pardon my French.
14:23:03kugelgit has nothing to do with it
14:23:19[Saint]I argue otherwise.
14:23:39[Saint]Several very active commiters fell right off the map, as soon as we switched.
14:23:47 Quit sakax (Quit: Leaving)
14:24:07 Join sakax [0] (
14:24:12pamauryhum, who did ?
14:24:12kugellebellium: I loaded samsung like onto my fuzev2
14:24:30[Saint]kugel: skins in general can, but the USB screen can't. We know exactly what will be displayed there.
14:24:33kugelthere is some full screen image that overlays the menu
14:25:01kugel[Saint]: if the sbs is visible during the usb screen it it can display what it wants
14:25:19[Saint]and if it is supposed to disallow the .sbs, it makes it even easier to predict what is required.
14:25:26[Saint]heh, timing.
14:26:32lebelliumkugel: and That's how it connects randomly
14:27:02[Saint]you can clearly see userfont there.
14:27:12kugelyes, I can see that on my fuzev2 as well
14:27:16[Saint]and ( need to say this) the .sbs
14:27:26[Saint]So, something is awry.
14:27:36[Saint]I had no idea it was supposed to dump userfont.
14:29:09kugellebellium: what's this full screen image is flashing over the menu?
14:29:40[Saint]smells like improperly layered viewports fighting with screen updates.
14:30:29copperisn't it just the display's low refresh rate?
14:30:56copperlike when you photograph a 60Hz CRT display
14:31:04lebelliumThe main difference between my themes and almost any other is that I use %?mp<%VI(Z)|%VI(X)> and a "mini-player"
14:31:48kugelmy fuzev2 has frozen upon screendump :\
14:32:00[Saint]copper: kugel is talking about on-device, not the visible tearing in the images supplied.
14:32:53kugellebellium: is the fuzev2 version out-of-date or something?
14:33:06lebelliumno, it works perfectly but look at the description
14:33:10lebelliumyou need a recent build
14:33:14lebelliumdue to the high amount of fonts loaded
14:33:21lebelliumgavaerts changed the limit for me
14:33:28kugeloh a reboot helped against the flashing
14:34:22[Saint]lebellium: many did you need? 0_o
14:35:00lebellium[Saint]: it's because of the iRiver remote! H320 and Fuze share the same themes
14:35:12kugelwell, I cannot reproduce the connect-failure or panic on disconnect on this device
14:35:34[Saint]if you're just using it to display numerics and the occasional symbol, it may have a much smaller footprint if you used bitmaps.
14:36:17 Quit AlexP (Remote host closed the connection)
14:36:20[Saint]by "how many did you need" I meant "how many fonts, per theme, do you require in the worst case?"
14:36:21lebelliumwell I don't remember exactly but feel free to look at my code and see how crazy it is :D
14:37:06[Saint]I remember doing an audit of your theme(s) many, many moons ago. But a lot has changed since then.
14:37:17[Saint]And I would wager it has been re-written a few times since.
14:38:00lebelliumI guess I have about 10 fonts loaded for the Fuze theme
14:38:51 Join AlexP [0] (~alex@rockbox/staff/AlexP)
14:39:14kugeltwo fonts are loaded
14:39:15lebelliumIt's like 2 themes
14:39:27lebelliumRemote theme = CLip Zip them
14:39:33lebellium+ the main theme
14:39:41lebelliumClip+ *
14:40:05[Saint]So...10 fonts per screen, or, in total?
14:40:11lebelliumin total!
14:40:13lebelliumnot per screen
14:40:15[Saint]the latter makes it a lot easier to imagine.
14:40:43*kugel still counts 2
14:40:50lebelliumkugel: heh?
14:41:03[Saint]the worst case look to be the .fms, which is using 3, unless I'm mistaken. That's not so bad.
14:41:24lebelliumFMS uses big %Fl(8,34 Ubuntu [Bold].fnt)
14:41:26lebelliumthat doesn't help
14:41:28[Saint]are you limiting the glyphs loaded?
14:41:32kugel4 after entering the fms
14:41:41[Saint]Ah, I missed one.
14:42:44lebelliumBut indeed USB connection works better with my theme for fuze than Clip Zip. So I don't think it's related to all fonts loaded
14:43:36[Saint]are you using the 'glyphs' tuple for the %Fl command?
14:43:45kugelseems like I need a clipzip to reproduce it
14:43:54lebelliumI don't know what it is [Saint]
14:44:45[Saint]That prevents the theme from trying to load *every* glyph the font has. It keeps only N amount of glyphs cached.
14:45:14*lebellium googling "glyph"
14:45:15kugelevery glyph is never loaded
14:45:24copperglyph = character
14:45:30kugelby defaults it's capped to whatever fits into 10k or so
14:45:38[Saint]So, if you were only displaying the chars Aa-Zz, you could safely limit it to 56 glyphs
14:45:48[Saint]its 200 or so by default, iirc.
14:45:58[Saint]or some size limit.
14:46:02[Saint]whatever comes first.
14:47:28 Quit jhMikeS (Ping timeout: 264 seconds)
14:47:33[Saint]With large fonts, this can reduce the memory usage by a reasonably amount. Not terribly important for some targets, crucial for others.
14:48:00lebelliumI'll have a look when I have some time
14:48:45[Saint]If you know, for example, that you will only be displaying the numbers 0-9 and the symbols : and %, you could cap it at 12.
14:49:14coppermy Fuze+ skin uses 309KiB
14:49:16lebelliumfor battery percentage I could do that indeed
14:49:39[Saint]Otherwise it may waste memory better used for the audio buffer caching glyphs that will never be displayed.
14:50:39[Saint]On the Classic, ...meh. On some targets, with 2MB of RAM total, every bit counts. :)
14:51:10[Saint](there's other targets to add to the "don't care about RAM" list...but, ...anyway)
14:51:40copperthe Fuze+ seems to have about 60MB like the Classic
14:53:15coppernot sure if those are MiBs
14:59:34 Join dfkt [0] (dfkt@unaffiliated/dfkt)
15:01:21pamaurythe fuze+ has 64MiB of ram
15:02:34gevaertskugel: pong
15:09:02kugelgevaerts: there seems to be some usb problems on the clipzip
15:09:38kugelit struggles with connecting, and it appears that usb_storage_disconnect() isn't always called
15:10:54kugelhow can it happen that this isnt always called (but usb_storage_init_connection() is)
15:12:39 Quit Rower (Quit: Hmmm...)
15:13:01 Join kaputnik_ [0] (
15:14:20 Quit kaputnik__ (Ping timeout: 256 seconds)
15:15:21gevaertskugel: the only way I see is that usb_thread() never sees USB_EXTRACTED
15:16:31 Join dan991199 [0] (
15:16:33kugelor it sees it but usb_state != USB_INSERTED
15:17:20 Part dan991199
15:18:18gevaertsHm, yes
15:18:27 Join Rower [0] (
15:18:58gevaertsI don't see how that could happen though
15:19:02kugelbut that doesnt look possible
15:20:08kugellebellium reported that even when usb storage connects fine then he can see the panicf() I inserted into playlist.c (it triggeres when it cannot get the audio buffer which is last-owned by usb_storage.c)
15:21:18pamauryhum, there is something fishy on the Zen X-Fi2: it has the same Mobile DDR chip as the Zen X-Fi3 but it fails when the ram is run at 133MHz whereas it works on the Zen X-Fi3 :-/
15:21:38kugellebellium: can you apply that patch as well (on top of the others): ?
15:22:19kugelpamaury: different modules
15:24:19 Quit Rower (Quit: Hmmm...)
15:26:08 Join Rower [0] (
15:26:14 Join prof_wolfff [0] (
15:27:20pamauryno :( same parameters, everything is the same
15:27:42kugelgevaerts: if it doesnt see USB_EXTRACTED how can the usb screen see SYS_USB_DISCONNECTED?
15:29:40kugelbecause that's what happens, the usb screen thinks disconnected and calls playlist_resume() where the panic occurs
15:30:17gevaertsDid you push those changes not to call audio_get_buffer() yet?
15:30:40kugelnot to master, but lebellium runs that (from gerrit)
15:31:23kugelwith that a call to usb_storage_disconnect() is required to release the audio buffer
15:32:10gevaertsIs usb_core_exit() called?
15:32:23kugelI can't tell from here :p
15:32:42*gevaerts asks lebellium instead :)
15:33:02*kugel doesnt think he can easily tell that either
15:34:49lebelliumkugel: I can apply that patch but later this afternoon or evening. I need a break for now USB makes me crazy :)
15:35:08gevaertsThe thing is, the old usb_storage_disconnect() doesn't do anything, so we don't really reliably know that it ever got called properly
15:35:49kugelhowever it seems to be the case on my fuzev2
15:36:35gevaertsSo it's likely not to be an obvious bug in usb_core.c
15:36:39lebelliumBTW kugel: I also have issues with USB connection and my theme on Fuzev1 (tried today) while I thought USB was more reliable on V1 than V2
15:36:46 Quit Rower (Quit: Hmmm...)
15:38:55 Join Rower [0] (
15:46:26kugelgevaerts: the implementation of usb_enable() is the same
15:49:36kugel[Saint]: have you tried lebellium's theme on one of your nano2g?
15:51:29***Saving seen data "./dancer.seen"
16:01:46 Quit Rower (Quit: Hmmm...)
16:07:32 Quit akaWolf (Ping timeout: 248 seconds)
16:10:40 Join Rower [0] (
16:17:43 Quit Rower (Quit: Hmmm...)
16:19:59 Join Rower [0] (
16:21:28 Join lorenzo92 [0] (
16:27:59 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
16:33:32 Quit n1s (Ping timeout: 276 seconds)
16:43:51 Quit Rower (Quit: Hmmm...)
16:47:37 Join Rower [0] (
16:47:51 Nick dv__ is now known as dv_ (
16:51:38 Quit melmothX (Quit: #)
16:55:45 Join ender` [0] (
17:05:58 Quit lorenzo92 (Ping timeout: 256 seconds)
17:14:52 Quit Rower (Quit: Hmmm...)
17:17:43 Join Rower [0] (
17:21:43 Quit Rower (Client Quit)
17:23:01 Join Rower [0] (
17:27:01 Quit Rower (Client Quit)
17:28:22 Join Rower [0] (
17:31:21 Quit Rower (Client Quit)
17:32:39 Join Rower [0] (
17:51:31***Saving seen data "./dancer.seen"
17:51:46pamauryTo all fuze+ users: could you try the latest build and tell me if you still have the freeze after usb bootloader mode ?
17:52:18 Quit Rower (Quit: Hmmm...)
17:53:20copperpamaury: what's that? Plugging it while powered off, then powering it on?
17:54:06pamauryyes, power off, plug usb then unplug, it should boot and most probably freeze. I just committed a potential fix
17:54:19 Join Rower [0] (
17:59:21copperit turns on while plugged, displays "bootloader usb mode"
17:59:30copperwhen plugged in*
17:59:57copperit stays on that screen while plugged in
18:00:22pamauryyeah, then unplug it
18:00:32coppernow it boots
18:00:38copperand seems to be working fine
18:00:56pamaurynow redo the same except that you should only let it plugged a few seconds
18:01:16copperturn off, plug in, wait 3 seconds, unplug?
18:02:11copperworks fine
18:02:51pamauryare you using the latest build ?
18:03:02copperbuilt myself
18:03:15pamaurynow can you retry with an older one ?
18:03:16copperI ran git pull −−rebase
18:03:35pamaurysay 10 commits before head
18:04:44copper130605 good enough?
18:06:00copperyup, freezes on the boot screen with the logo at the top and the build version at the bottom
18:06:33pamaurygood, so that's indeed a fix, thanks
18:07:42pamaurythis has a been a long running issue, it's good that I finally fixed it
18:08:44coppergood job :)
18:26:26 Join lorenzo92 [0] (
18:26:32 Quit lorenzo92 (Client Quit)
18:29:23 Quit liar (Ping timeout: 248 seconds)
18:30:29 Join liar [0] (
18:36:43 Quit liar (Ping timeout: 264 seconds)
18:44:16 Join lorenzo92 [0] (
18:44:45lorenzo92I'm rewriting the packer for YP-R0 (R1 too) to be fully open (and much more fast too ^^)
18:45:10lorenzo92once I figure out how to mess with a cramfs on windows, i will be able to insert the player into rbutil :)
18:47:02pamaurywhy do you need to mess with cramfs ?
18:47:50 Quit Scall (Read error: Connection reset by peer)
18:48:19 Join Scall [0] (~chat@unaffiliated/scall)
18:48:37lorenzo92because I need to modify some data in the cramfs of the player in order to boot rockbox
18:49:15 Join liar [0] (
18:49:25lorenzo92i found an interesting option for cramfs tools but I don't understand what's its porpouse, it correctly appends a previous image but i don't know if it's intended to *not* doing any merge...
18:49:33lorenzo92in fact, it doesn't
18:50:43pamauryI guess a cramfs image is compressed ?
18:51:34lorenzo92hum yeah
18:51:37lorenzo92it should be
18:52:12lorenzo92the problem are permissions on windows. alternatively, if possible, a small ram disk would be the solution ^^
18:54:07 Quit liar (Ping timeout: 264 seconds)
18:56:18 Join liar [0] (
19:22:53 Quit kevku (Ping timeout: 245 seconds)
19:25:29 Quit Rower (Quit: Hmmm...)
19:27:16 Join Rower [0] (
19:28:11 Quit Scall (Read error: Connection reset by peer)
19:30:43 Quit pamaury (Ping timeout: 248 seconds)
19:31:13 Quit Rower (Client Quit)
19:31:18 Join Scall [0] (~chat@unaffiliated/scall)
19:36:08 Quit onder` (Ping timeout: 256 seconds)
19:36:22 Join kevku [0] (~kevku@2001:0:c38c:c38c:3d:4be4:3d69:beaa)
19:41:16 Join joshumax [0] (
19:41:25 Join Rower [0] (
19:42:17 Join onder` [0] (~onder@
19:48:43joshumaxoh hi
19:49:09joshumaxJust a quick question, does anyone remember the old "zvue" media player?
19:50:00lebelliumkugel: I just tried your latest patch. I can't reproduce the Panic resume_playlist issue when disconnecting, but I don't know if it's just a coincidence or if it fixed something :)
19:50:58lebelliumBut now I have more often "*Panic* stkov usb" when connecting. Maybe just a coincidence too
19:51:35***Saving seen data "./dancer.seen"
19:52:21kugelhm interesting
19:54:16kugelI guess that also explains the undefined instruction error you saw earlier
19:56:11kugellebellium: please add this patch and see if it does anything
19:57:14lebelliumok. I'll do it in about 40min, brb
19:57:54kugelstkov is evil
20:13:14 Quit Scall (Ping timeout: 276 seconds)
20:14:47 Join Scall [0] (~chat@unaffiliated/scall)
20:31:06 Quit kevku (Ping timeout: 260 seconds)
20:42:13 Quit Rower (Read error: Connection reset by peer)
20:43:01 Join Rower [0] (
20:43:53 Join kevku [0] (~kevku@2001:0:c38c:c38c:1cea:1955:3d69:beaa)
20:59:43 Join krabador [0] (~krabador@unaffiliated/krabador)
21:05:04lebelliumkugel: I just got a "undefined instruction" while connecting but no more stkov usb so far
21:09:34 Nick DormantBrain is now known as SuperBrainAK (
21:19:19 Quit Guinness` (Ping timeout: 264 seconds)
21:33:15 Quit Rower (Quit: Hmmm...)
21:34:41 Join Rower [0] (
21:35:41 Quit lorenzo92 (Ping timeout: 256 seconds)
21:40:16 Quit kiwicam (Remote host closed the connection)
21:44:45 Join z180 [0] (
21:45:06 Join einhirn [0] (
21:47:21kugellebellium: really strange
21:47:31kugellebellium: does the panic still happen on disconnect?
21:48:36lebelliumI haven't seen it for some time now.
21:49:10kugeldoes it still freeze?
21:50:02kugelon disconnect
21:50:10lebelliumsometimes but becoming rare. Disconnecting works better overall. But connecting is still quite unstable
21:50:44kugelbut it's not fixed yet?
21:50:56kugeli.e. the freeze is still there in some cases
21:51:00 Quit liar (Read error: No route to host)
21:51:37***Saving seen data "./dancer.seen"
21:52:16lebelliumAfter applying your last 2 patches, I guess it still froze 2/xx times when disconnecting. I don't know if I can say it's fixed yet. But looks better to me overall
21:53:22lebelliumand IIRC the 2 times it still froze, it did not mount well before
21:54:00kugelat least the panic is gone
21:54:11lebelliumseems to be the case indeed
21:54:12 Join liar [0] (
21:58:37 Quit Scall (Read error: Connection reset by peer)
21:59:29 Quit kaputnik_ (Ping timeout: 256 seconds)
21:59:48 Join Scall [0] (~chat@unaffiliated/scall)
22:10:41 Join Guest11127 [0] (
22:13:55 Join kaputnik_ [0] (
22:22:58kugelhm, who's the clipzip maintainer (if any)?
22:23:25lebelliumand/or funman
22:24:21kugelfunman: ping
22:29:11 Quit y4n (Quit: Today is the perfect day for a perfect day.)
22:30:19 Join jhMikeS [0] (
22:30:19 Quit jhMikeS (Changing host)
22:30:19 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
22:51:32 Join lorenzo92 [0] (
23:12:32 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
23:26:06 Quit mrtux (Quit: gtg)
23:26:38 Join mrtux [0] (~mrtux@unaffiliated/mrtux)
23:27:00 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
23:30:57 Quit Rower (Quit: Hmmm...)
23:31:03 Join musiklover [0] (58d7727e@gateway/web/freenode/ip.
23:34:28musikloveri would like to get a music-player that can connected via USB to the pc where is acts as a connected mass storage device, that i can mounted it's FS, copy/delete music. When playingin music i would like to browse on a folder baysis. The player should ignore id-tags. is there such a player (like meizu m6 was one) that can run rockbox and does rockbox offer the mentioned features?
23:39:43 Quit Zarggg (Ping timeout: 264 seconds)
23:44:42musikloverare there chances that craetive zen player will get rockbox?
23:47:16 Quit bertrik (Ping timeout: 246 seconds)
23:51:40***Saving seen data "./dancer.seen"
23:53:26lorenzo92pamaury: the xor key is a little different with respect to the z5 one, now I understand why it wasn't working ^^
23:57:02 Quit kevku (Ping timeout: 245 seconds)
23:57:58musikloverIs there a rockbox compatible device in a formfactor like this: or this

Previous day | Next day