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

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

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

#rockbox log for 2012-12-02

00:00:24 Join KiwiCam [0] (~quassel@
00:03:29 Join webguest36 [0] (
00:03:51saratoga_i don't understand why pushes don't implicitly update the source they pulled from
00:04:19kugelI think recent git version do this
00:04:19 Quit webguest36 (Client Quit)
00:05:34Tornesaratoga_: you mean, why the push url is different?
00:05:40Tornethat'sjust an optimisation for our server
00:05:48Torneyou can use the saem url for both
00:05:54saratoga_i mean, why do i have to tell it theres a push URL?
00:05:55Torneas long as it's one of the authenticated protocols
00:06:00Tornebecause the url you cloned from is read only
00:06:29Torneif you just cloned the other url to start with you don't have to do it
00:07:50Tornebut the read-only one is faster, and doesn't require you to have already created a gerrit account
00:08:02Torneso it's what we have at the start of the instructions
00:09:04[Saint]Is requiring a gerrit account to clone the repo that big a deal? If they have no intention of contributing back, don't we have a github they can pull from?
00:09:40 Quit lebellium (Quit: ChatZilla 0.9.89 [Firefox 18.0/20121128060531])
00:09:44Torneit works fine cloning from the git:// protocol
00:09:52Torneand it's faster
00:09:57derf[Saint]: Yes, it is that big a deal.
00:10:20[Saint] tell.
00:11:19gevaerts[Saint]: what would requiring a gerrit account to clone actually buy us?
00:11:33derfAnything you do to raise the activation energy required will cause significant fractions of people not to bother.
00:11:43***Saving seen data "./dancer.seen"
00:12:04*gevaerts nods
00:12:13derfYou know why things like "Login with Facebook" are so popular?
00:12:29[Saint]gevaerts: nothing, and I wasn't implying it would. I just wondered why it was mentioned specifically.
00:12:36derfBecause site owners know that whenever they ask users to make an account just for their site, most of them simply won't.
00:13:00[Saint]yeah, thankfully...there's, yeah, no.
00:13:08[Saint]not for gerrit at least.
00:13:23derfI mean, this is largely the reason that oauth exists.
00:13:58[Saint]Argh, yes. I parsed that incorrectly.
00:14:44[Saint]I just thought that since the main idea of gerrit (only idea?) it review and contribution that creating an account for it wouldn't be a big deal.
00:14:46[Saint]That's all.
00:14:56[Saint]*is review.
00:15:09gevaerts[Saint]: we *do* require an account for gerrit
00:15:18gevaertsWe don't require one for git cloning
00:15:48[Saint]I know, but, it's not like there's nowhere else to clone from.
00:16:04derf[Saint]: This is not how human psychology works.
00:16:25derfWhen someone has written a patch and is looking for a place to upload it, they will happily make an account.
00:16:37derfWhen they want to browse some code or run git blame... not so much.
00:16:52derfAnd unless those "other places" are the _first_ place you're pointing to, they won't go looking for them.
00:16:59gevaertsI don't see the point of shutting down services for *zero* advantage if there are reasons to keep them, even if you think those reasons aren't very strong
00:22:06jnctrying to follow the backlog of conversation here
00:23:30[Saint]Don't try too hard. It's fairly non-interesting. :)
00:24:03jnc[Saint]: should 'badblocks -n' work on Rockbox in usb mode?
00:24:59[Saint]<speculation hat on> It _should_.
00:25:20jncokay hmm, it bailed on my Fuze v1 with 3.8.1
00:25:58saratoga_in MSC mode the device is just a normal hard disk, so you can do normal hard disk things
00:26:42jncthe suspicion that 3.8.1 is "good" for Fuze v1 and USB may prove false
00:26:53[Saint]And while it _should_ work, 3.8.1 is a whole barrel full of ancient, so I'm not too sure how meaningful it not working there is.
00:27:04[Saint](without knowing if it is functional in HEAD)
00:27:04jncah, copy.
00:27:26jncanyone with a more or less known working USB mode care to try vs. head, and vs 3.8.1 ?
00:27:43jncI'm trying to hunt down where Fuze v1 USB fail comes from
00:27:46saratoga_i don't know if USB ever worked perfectly on the fuze
00:29:05jncif it can't be fixed for next stable release, then perhaps disable it for stable?
00:29:33saratoga_is there any point in doing that?
00:29:41jncprevent data loss
00:30:42jnctransferring data from host to rockbox on sansa fuze v1 tends to corrupt existing unrelated files
00:30:54saratoga_for some people
00:31:06[Saint]That suggests that the FS would already be hosed, to me.
00:31:31jncI have 3 different Sansa V1's to test with
00:34:42jncbefore each test, I use the oem firmware to "format", load rockbox, sync and unmount
00:35:08jncthe only thing that would hose the filesystem is data corruption from rockbox usb transfer
00:35:24saratoga_do you really need to format?
00:35:59jncno, but it remmoves doubt of filesystem corruption; to your point
00:45:30saratoga_would it be easier to just DD something over and then read it back and see if it matches?
00:46:57saratoga_that way you don't need to care about the file system
00:47:34 Quit melmothX (Quit: bau)
00:55:30n1swhen copying lots of data to the sd card via usb my fuzev1 almost always hangs at some point, i think it'sw always donr that
00:55:51n1si just use the c200 as a card adaptor (when i remember)
00:56:10n1sbut it's never hosed the filesystem worse than fsck.vfat being able to fix it
00:56:50n1snot sure if it applies to internal storage too, unzipping an rb build usually works fine and i don't use it for music
01:00:02*[Saint] just got grilled for spending $500 on a GFX card.
01:00:48[Saint]Raedeon HD7950
01:01:15*[Saint] remembers being one of the first in NZ to get an AGP FX5200, lol.
01:01:23the-kyle1Now you're playing with power. <grins>
01:01:29 Nick the-kyle1 is now known as the-kyle (
01:01:54[Saint]shit. this, is the wrong channel entirely...
01:06:38 Quit Robin0800 (Read error: Connection reset by peer)
01:12:18 Quit ender (Quit: I strap on the body armor, which feels tough enough, but closes with Velcro strips. I know this is state-of-the-art gear, but I'd feel more confident if it wasn't held together with the same stuff they use to fasten kids' sneakers. -- Richard Kadrey: Sa)
01:17:34 Join nateloaf [0] (
01:32:15jncfuze v1 usb mode 3.6 does endless reboot, interesting
01:32:46[Saint]I don't think that's that interesting at all.
01:32:47 Quit pamaury_ (Ping timeout: 265 seconds)
01:32:57[Saint]iirc, that's a modern bootloader combined with old USB behaviour.
01:33:08jncah yes
01:33:28jncso old bootloader would be needed to test old 3.6?
01:34:07jncsorry I am so naive
01:34:09[Saint]I _think_ so, but, an AMS guru really needs to weigh in here.
01:34:27[Saint]I was never really quite sure exactly how the bootloader tied into the USb behaviour in the first place.
01:35:08jncbetter question would be, should I update the bootloader to something other than HEAD when testing older versions of rockbox, how old, and so on?
01:35:11[Saint]But I do distinctly recall a big mess got created when we stopped relying on the OF for USB, and I seem to recall the bootloader having a lot to do with this.
01:35:14 Quit bertrik (Remote host closed the connection)
01:36:39gevaertsBack then there wasn't any usb support
01:36:50gevaertsSo there's no point in trying to test it
01:37:20jncaye so ... what point release was usb included?
01:37:50[Saint]release notes should tell you that.
01:38:06gevaertsI don't know, but it definitely was one that works with a current bootloader :)
01:38:16jncthanks gevaerts
01:38:59gevaertsThe issue is that a current bootloader boots rockbox when it sees USB, and old rockbox reboots (to get to the OF) when it sees USB
01:39:42[Saint]Ah, it's less technical than I thought.
01:39:44jncokay I see, that does make sense it would reboot like I just obvserved then
01:40:49gevaerts[Saint]: it's *very* simple. The difficult bits appear when you try to figure out migration plans between OF-for-USB and rockbox-for-USB
01:41:23jnclooks like USB rockbox started with 3.7
01:43:09 Quit n1s (Quit: Ex-Chat)
01:50:27 Join breakfastsquid [0] (
01:50:59 Quit brkfstsqd (Ping timeout: 250 seconds)
01:58:42saratoga_ g#363 looks good to me
01:58:45fs-bluebotGerrit review #363 at : tagnavi: Add clause for comparison of multiple tags with same value in DB search by Eliran Levi (changes/63/363/1)
01:58:55saratoga_anyone want to review it? i'm not too familiar with the DB but it looks safe and sensible
02:01:37 Quit Rower85 (Quit: Hmmm...)
02:11:45***Saving seen data "./dancer.seen"
02:33:57 Quit sakax (Read error: Operation timed out)
02:35:45 Quit ruskie (Quit: ...)
02:36:52 Join sakax [0] (
02:49:45 Join ruskie [0] (ruskie@sourcemage/mage/ruskie)
02:53:31 Quit sentriz (Ping timeout: 252 seconds)
03:06:46 Quit efyx__ (Ping timeout: 245 seconds)
03:14:55 Quit Clear_runway (Ping timeout: 265 seconds)
03:20:28 Join efyx__ [0] (~efyx@
03:29:18 Quit the-kyle (Ping timeout: 246 seconds)
03:32:36 Quit Gallomimia (Read error: Connection reset by peer)
03:33:15 Join Gallomimia [0] (
03:44:32 Quit [Saint] (Remote host closed the connection)
03:50:57 Join sentriz [0] (~Senan@
03:55:09 Quit scorche (Disconnected by services)
03:55:13 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
04:10:51 Quit sentriz (Quit: Leaving)
04:11:47***Saving seen data "./dancer.seen"
04:13:56 Quit Poodlemastah (Quit: ZNC -
04:17:57 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:17:57 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:17:57 Quit amiconn (Disconnected by services)
04:17:57 Quit pixelma (Disconnected by services)
04:51:07 Join TheSphinX_ [0] (
04:54:11 Quit TheSphinX^ (Ping timeout: 245 seconds)
05:31:09 Quit TheSeven (Disconnected by services)
05:31:17 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
05:32:36 Quit GodEater (Ping timeout: 264 seconds)
05:33:09 Join GodEater [0] (
05:33:09 Quit GodEater (Changing host)
05:33:09 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
05:40:39 Join [Saint] [0] (~saint@rockbox/user/saint)
05:43:58*[Saint] wonders about the USB requirement for "stable" anyway...
05:44:09[Saint]saratoga_: ^
05:44:47[Saint]I mean, if the iPod mini is stable, and has no support for firewire, isn't that fairly similar?
05:47:16[Saint]I'm unsure, as my mini is currently non-functional and I have a bad memory...but I seem to recall it doesn't reboot to the OF on Firewire plug either and just sits there and charges.
05:47:47[Saint]Meaning it'd be in pretty much exactly the same position as the Nano2G if we disabled Rockbox USB.
05:49:48[Saint]gevaerts: (log) did you ever get any further with USB reboot on detect for the N2G?
05:50:09[Saint](I'm aware of what exists on gerrit currently)
05:54:38[Saint]Ahhhhhh, herp-derp. No Firewire != No USB. Yeah, ...whoops.
06:11:51***Saving seen data "./dancer.seen"
06:34:34 Quit breakfastsquid (Remote host closed the connection)
06:50:58jnc[Saint]: so far, I see failure mode from "badblocks" with firmware 3.8 which was reportedly okay in the fuze v1 bug report
06:51:23jncdoing some more thrashing of 3.7 which has completed okay w/ badblocks a few times yet
07:02:12 Quit Gallomimia (Read error: No route to host)
07:02:28 Quit prof_wolfff (Ping timeout: 246 seconds)
07:03:24 Join Gallomimia [0] (
07:30:58 Quit [Saint] (Remote host closed the connection)
07:33:09 Join [Saint] [0] (~saint@rockbox/user/saint)
08:11:55***Saving seen data "./dancer.seen"
08:13:54jnc[Saint]: how to know, what git revision corresponds with a point release (say 3.7)
08:23:06 Join _jhMikeS_ [0] (~jethead71@
08:23:07 Quit _jhMikeS_ (Changing host)
08:23:07 Join _jhMikeS_ [0] (~jethead71@rockbox/developer/jhMikeS)
08:23:07 Quit jhMikeS (Disconnected by services)
08:23:07 Nick _jhMikeS_ is now known as jhMikeS (~jethead71@rockbox/developer/jhMikeS)
08:29:59 Quit Prodicus (Read error: Connection reset by peer)
08:30:26 Join Prodicus [0] (
08:35:00[Saint]jnc: I'm fairly sure there's easier ways (or ways to do this more directly that aren't necessarily any harder nor easier), but you can look up the revisison from which the release was tagged from fairly trivially.
08:35:09[Saint];a=tags for example
08:35:23*[Saint] hand typed that url, so he hopes he got the syntax right.
08:36:56[Saint]Then you can do "git reset −−hard HEAD~XXXXXXX where XXXXXXX == the revisions unique hash
08:37:21[Saint]*"git reset −−hard HEAD~XXXXXXX"
08:38:52[Saint]If you have any local changes you need kept, stash them first with "git stash" and then re-apply them (if desired/needed) after the new checkout with "git stash pop".
08:39:25*[Saint] suspects someone will now chime in and rail me on my horrible git workflow :P
08:40:40[Saint]Now, I'm even less sure about this one, but, I'll try:
08:42:15[Saint]If the revision you need to check out was before the git transition and you know the specific svn revision you can do "git log 1 −−grep=@rXXXXXX" where rXXXXXX == the known subversion revision.
08:42:48[Saint]errr, "git log -n −−grep=@rXXXXXX", rather.
08:43:15[Saint]That should return the commit description and the hash identifier.
08:43:36[Saint]You would then pass the has identifier to the command listed above.
08:47:14[Saint]So, for example, you could look at;a=tags and see that the 3.12 release was tagged from 2ee456528104451945c2e8370eea513e1831f55b by clicking on the commit description.
08:48:08jncbefore your suggestion, I was grepping through git log
08:48:17[Saint]We'd then pass the first seven digits (this is almost always enough for a unique identifier) of that hash to the "git reset −−hard HEAD~XXXXXXX" command.
08:48:23jncmuch better to do it the right way now :)
08:49:18[Saint]So, "git reset −−hard HEAD~2ee4565" would checkout the 3.12 release over you current repo.
08:50:11 Join stoffel [0] (
08:50:26[Saint]you should also be able to grep the git log for "Tag Release"
08:52:24jnc[Saint]: "Tag Release" not appearing in git log? I don't know
08:53:33jncbrowsing the tags on the web tool looks better to me
08:55:46[Saint]Hum, grepping the log for 'Tag Release' was just a guess I suspected /may/ work, sorry.
08:56:07[Saint]I assumed grepping the log would include the commit description
08:56:51jncit does say some things like "bumping to version 3.7" but not about the final tag and point release
08:57:09jncI am glad you explain to me how to find out the real commit relation
08:57:12*jnc :)
08:57:45[Saint]iirc, the commit description for tagging a release should always be "Tag Release *.*"
08:57:50*[Saint] shrugs
08:58:11[Saint]or Tag *.* Release
08:58:16[Saint]...something like this.
09:01:27 Join melmothX [0] (~melmoth@unaffiliated/melmothx)
09:03:27[Saint]But, yeah. I shouldn't have just guessed it'd work like that, sorry. I know enough in git to do the things I need to do, but I'm certainly no git guru, heh. :)
09:07:11 Join the-kyle [0] (~kyle@
09:18:12 Quit sakax (Ping timeout: 264 seconds)
09:20:58 Join kevku [0] (x@2001:470:28:773::3)
09:25:56 Quit Prodicus (Ping timeout: 248 seconds)
09:30:09 Join sakax [0] (
09:35:12 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
09:37:55 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
09:38:04 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
09:38:40 Quit akaWolf (Client Quit)
09:57:54 Quit the-kyle (Read error: Connection reset by peer)
09:58:20 Join the-kyle [0] (~kyle@
09:59:15 Quit Gallomimia (Quit: I am likely going to change locations)
10:05:06 Join lorenzo92 [0] (
10:07:44 Join pamaury [0] (
10:07:44 Quit pamaury (Changing host)
10:07:44 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:11:56***Saving seen data "./dancer.seen"
10:15:18 Join n1s [0] (
10:15:18 Quit n1s (Changing host)
10:15:18 Join n1s [0] (~n1s@rockbox/developer/n1s)
10:15:24 Quit lorenzo92 (Quit: ChatZilla 0.9.89 [Firefox 17.0/20121120062532])
10:27:19 Join ender [0] (
10:30:07JdGordonanyone run on OSX recently?
11:02:49 Join B4gder [241] (~daniel@rockbox/developer/bagder)
11:04:09 Quit Bagder (Read error: Operation timed out)
11:04:16 Join Horscht [0] (~Horscht@xbmc/user/horscht)
11:05:41pamaurysaratoga_: you are the one who looked at the RMMA results of the fuze+ last time right ?
11:09:15 Quit sakax (Remote host closed the connection)
11:13:43 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
11:22:54 Join lebellium [0] (
11:24:25 Quit sciopath (Quit: Quitte)
11:30:08 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
11:31:40 Quit B4gder (Read error: Connection reset by peer)
11:31:47JdGordonanyone know how to build gcc-sh on osx? getting errors building stmp-multilib (i tihnk)
11:33:53 Join Bagder [241] (~daniel@rockbox/developer/bagder)
11:41:59 Quit akaWolf (Quit: my exit)
12:07:22 Join sciopath [0] (
12:10:37 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
12:11:58***Saving seen data "./dancer.seen"
12:12:39 Quit sciopath (Ping timeout: 260 seconds)
12:14:16 Quit [Saint] (Read error: Connection reset by peer)
12:16:39 Join [Saint] [0] (~saint@rockbox/user/saint)
12:17:19 Join Topy44|2 [0] (
12:17:28 Quit stoffel (Remote host closed the connection)
12:20:32 Quit Topy44 (Ping timeout: 252 seconds)
12:22:02 Join sciopath [0] (
12:30:03 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
12:31:28 Quit fs-bluebot (Ping timeout: 246 seconds)
12:32:40JdGordonhmm, why cant my build client build the checkwps? really old (or too new??) gcc?
12:32:46 Join fs-bluebot [0] (
12:33:39 Quit bluebrother^ (Ping timeout: 260 seconds)
12:34:59 Quit Provel (Read error: Connection reset by peer)
12:36:09gevaerts[Saint]: historically we haven't treated USB as a strict requirement for stable
12:36:37gevaertsAnd the USB detection change on nano2g somewhat depends on USB working at all...
12:37:27gevaertsAlso, I suspect you're never going to see tags and branches in regular git log. They don't change the code
12:47:40 Quit the-kyle (Quit: Leaving.)
12:47:40 Join the-kyle1 [0] (
12:52:43 Quit jhMikeS (Ping timeout: 256 seconds)
12:53:41JdGordoncan the build clients be told to not build checkwps?
12:53:46JdGordonapparently my gcc is busted
12:54:51 Join Rower85 [0] (
13:23:16 Quit [Saint] (Remote host closed the connection)
13:33:04 Join Guest44874 [0] (
13:33:52 Quit Rower85 (Ping timeout: 256 seconds)
13:47:48pixelmasaratoga_: just as a note - the Iaudios do have dual boot for a while now (since DevCon IIRC). I believe the Gigabeats still don't have though
13:53:45lebelliumif that's the case, all documentation on rockbox website is outdated
14:11:59***Saving seen data "./dancer.seen"
14:15:40 Join prof_wolfff [0] (
14:18:08 Join Poodlemastah [0] (
14:24:51AlexPThat is no surprise
14:37:21[7]saratoga_: I don't really see a roadblock on making a classic RB bootloader
14:37:33[7]there's no real difficulty there
14:37:41[7]the problem is getting that bootloader installed
14:38:06[7]we currently use the emcore kernel for the bootloader because we currently need the emcore kernel to handle the installation of the bootloader
14:38:17[7]because the bootloader needs to be flashed to the NOR on the classic
14:38:51[7]we can only boot from the firmware partition with very old apple bootloader versions on the first generation classic that still have a buffer overflow vulnerability
14:39:36[7]sure, we could make the emcore installer install a rockbox bootloader
14:39:57[7]but I don't see a way to get completely rid of emcore without much effort
14:40:33[7]we'd not only need a bootloader, but also an installer, because we cannot directly flash a classic, but instead have to write a flasher that we boot on it
14:40:49[7]sure, that could be done in a much more simple way than emcore does it
14:42:00[7]and then we still need rbutil to handle booting that flasher somehow, i.e. it needs to instruct the user how to enter DFU mode, and connect to DFU (which requires either itunes to be installed, but several itunes processes/services to be stopped, or a custom driver that conflicts with itunes)
14:42:59[7]Within the amount of time that I have available I'd do my best to support whichever route you choose
14:43:26[7]but tbh I think that time would be better spent on nailing down this nasty usb breakage
14:44:04*gevaerts agrees
14:45:38gevaertsIf the problem is "emcore/freemyipod might go away", we could just mirror the relevant bits. We could even provide our own "official rockbox" emcore builds if we want to
14:48:35[7]it might make sense to make a "rockbox bootloader" app on top of emcore, which replaces the (more fancy, but more bloated) emcore bootloader
14:48:57 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
14:49:50[7]that app would probably be a 5-liner or something like that :)
14:49:50[7]because the kernel and libboot handle most of what it needs
14:50:36[7]basically removing the whole graphical UI and boot options, making it try to boot rockbox if possible, and fall back to a flashed rockbox if not
14:50:50[7]if the flashed rockbx is missing as well for some odd reason, just bail out on the console
14:51:42[7]that would remove the complexity that the boot menu introduces, but has some disadvantages as well: you have no way to access storage except through rockbox
14:51:42*gevaerts hasn't seen emcore in action, so he doesn't really know *how* fancy the emcore bootloader is
14:52:28gevaertsI think we should try to keep the storage bits
14:52:28[7]so by dropping in a broken rockbox binary you could screw things up badly
14:53:16gevaertsAnyway, that still requires emcore, so let's start by making emcore acceptable for rockbox
14:53:25[7]emcore doesn't do usb mass storage either, the emcore bootloader merely has options to manually boot the flashed rockbox fallback image, and to nuke the database and settings in case those prevent rockbox from booting
14:53:39soapscorche`, gevaerts: To expand upon gevaerts query on the 26th asking if there was a way to require active moderator approval for someone's first posts and scorche` replying that moderation appears to be board based not user group based: Couldn't we just block ALL normal boards from the "New User" group and require an obviously human post from someone in the "New User" group to whatever board we give them access to before they get manually added to the "
14:53:39soapRegular User" group?
14:54:26gevaertssoap: that's actually outdated. We *can* require moderator approval for first post (or any number of first posts) if we want to
14:54:32[7]fyi, emcore shares the (old) usb lowlevel driver, as well as the storage/fat subsystems with rockbox, but has a different kernel below that (preemptive multitasking, memory allocator)
14:55:18gevaerts[7]: so basically USB storage means booting the flashed rockbox image?
14:55:46gevaertsOK. I assume detecting a button to do that wouldn't be too hard
14:56:09[7](you can also access storage through the emcore kernel debugging interface, but that's not targeted at end users)
14:56:36[7]sure, that can be done, but there's still the problem with a corrupted config/database/whatever making the fallback image lock up during boot
14:56:41[7]I've seen that happen several times
14:57:15[7]ideally we'd just fix dualboot and boot into disk mode when some button is pressed
14:57:31gevaertsYes, of course, but that depends on unknowns
14:57:52[7]it depends on figuring out what on earth we're doing that locks up the OF's i2c driver
14:58:08[7]probably not all that complicated, but debugging that is a bitch
14:58:44soapand sorry for interrupting again, but so bans still bog down the forums, but not for quite as long.
14:59:26gevaertsI'd say one realistic option would be to make a traditional rockbox-style bootloader that does USB on button or failed boot, have that as the flashed image, and always boot that from emcore
15:00:44 Join TheSphinX^ [0] (~briehl@
15:00:49 Quit TheSphinX_ (Read error: Operation timed out)
15:01:09gevaertssoap: I haven't noticed that, but then I've only banned one account in the past week..
15:01:26gevaertsI'd say that as long as we're at the current banning rate, it's not a problem
15:02:09soapI agree with that unless the fix is as simple as adding an index to the DB as Zagor postulated.
15:02:37[7]gevaerts: shouldn't be too hard to flash that directly, and just rely on emcore loader to do the HW init and install it through an emcore installer
15:02:38soapAnd, to be clear, the lag is less, but the test I just ran froze forum access for 50 seconds.
15:03:35gevaertssoap: ok. I'd say that as long as scorche hasn't found time to look into that, it's not a problem :)
15:04:57[7]currently the boot process looks like this: bootrom loads emcore loader from nor and sigchecks it, emcore loader escapes through an exploit. then emcore loader initializes i2c, pmu, sdram, lcd and spi, and loads the emcore kernel from NOR
15:06:07[7]then the emcore kernel parses its embedded boot configuration, and locates the file to be booted on the nor, which is the emcore boot menu. that in turn tells the kernel to load libui, libpng and libboot, and displays the boot UI. once the user has selected an option, the binary for that is loaded from NOR or HDD, and executed
15:06:56[7]if we jump in at the emcore boot menu stage, we can make a 5 liner that loads something else and runs it like rockbox
15:07:44[7]but actually booting the emcore kernel is already somewhat similar to booting rockbox, so we could just drop in a rockbox bootloader instead of the emcore kernel
15:08:09gevaertsSo we'd only borrow the emcore loader
15:08:18gevaerts(and installation stuff)
15:08:55[7]yes, and the installation process
15:08:55[7]and IMO the installation process is currently the most annoying part from a user's perspective, not the boot menu... so maybe we should tackle that first
15:09:34[7]the emcore loader is much smaller than emcore, it's some selfcontained assembly blob
15:10:26gevaertsOn the other hand, the boot stuff would (I think) allow us to get rid of the "unstable" stigma, which might help find more people to work on the rest
15:14:43[7]well, but that's a purely artificial problem
15:15:22 Join Rower85 [0] (
15:15:44gevaertsIt is, yes
15:15:56[7]I don't see how having a flaky, complicated, non-automated method of installation is any better than using a well-working 3rd-party bootloader
15:16:05 Quit Guest44874 (Ping timeout: 256 seconds)
15:16:35[7]you could of course write your own flasher to get rid of emcore during installation as well
15:17:08[7]isn't that complicated, but would drop the on-device installation confirmation and status UI. the question is whether we want that or not.
15:17:24[7]which would remove all freemyipod stuff except for the emcore loader
15:17:34gevaertsI think any solution that involves actually rewriting stuff isn't a good idea
15:18:02[7]well it isn't that much stuff that needs to be rewritten. it's more like throwing out 90% of the existing stuff
15:18:05*gevaerts personally doesn't think relying on freemyipod stuff is that much of a problem
15:18:45[7]well I don't view it from that aspect at all, I'm looking at it from a bloated vs. simplified aspect
15:19:11[7]because 90% of what happens during installation isn't really required at all and only related to UI
15:19:21[7]the actual flashing is like 50 lines of C code
15:19:51[7]and there isn't really anything that can go wrong during this process, so we don't really need status feedback
15:20:06[7]emcore just seems overkill for that purpose
15:21:25[7]the emcore installer could handle unpacking a rockbox zip to the hdd during installation, but we currently don't use that feature either
15:28:10 Quit ender (Read error: Operation timed out)
15:31:00 Join Prodicus [0] (
15:33:29 Join ender [0] (
16:01:43n1si agree that it's the installation process, rather than the bootloader that's the bigger problem
16:11:49 Quit nateloaf (Quit: Leaving.)
16:12:03***Saving seen data "./dancer.seen"
16:23:03[7]but I don't see how we could make the installation process much easier
16:23:41[7]it's fairly easy on linux but the usual driver nightmare on windows - maybe write some tool that uses setupapi or DIFx to automate driver installation
16:24:19[7]then again umsboot is causing a lot of issues on some pcs, we might want to get rid of that
16:24:46[7]if we decide to go for an emcore-less setup, we could probably squeeze it all into the DFU binary and skip that step
16:25:20[7]but OTOH it would be better to finally nail down the root cause of this usb problem...
16:28:50copper<[7]> it's fairly easy on linux but the usual driver nightmare on windows
16:28:58copperyeah I didn't manage to do it on Windows
16:29:09copperI tried many times
16:30:04[7]copper: did you fail at 1. bringing the device into DFU mode, 2. getting the tool to recognize the ipod, or 3. getting UMSboot to mount to copy the installer?
16:30:42copper1) succeeded but nothing was happening in Windows
16:30:53[7]with or without itunes installed?
16:30:58copperand the tool didn't find the device
16:31:23copperwithout iTunes, the "device" driver always failed to install properly
16:31:50copperbut then I had a crappy Asus Windows install at the time
16:31:56copperI didn't try again on my clean Win 7 install
16:34:18copperthe Asus Windows install was Windows 7 as well, just loaded with a shitload of crapware
16:38:50 Join TheSphinX_ [0] (
16:40:32 Quit TheSphinX^ (Ping timeout: 245 seconds)
16:41:48 Join Ward [0] (~Mirandaha@
16:42:12 Nick Ward is now known as Guest38219 (~Mirandaha@
16:42:35 Quit XavierGr (Ping timeout: 265 seconds)
16:42:51 Join pedro_angelo [0] (~pedro_ang@
16:47:50 Join TheSphinX^ [0] (
16:48:59 Join SuperBrainAK [0] (
16:49:40 Quit TheSphinX_ (Ping timeout: 248 seconds)
16:54:21 Join TheSphinX_ [0] (
16:57:11 Quit TheSphinX^ (Ping timeout: 264 seconds)
17:00:08 Quit lebellium (Ping timeout: 256 seconds)
17:13:09 Join LjL [0] (~ljl@unaffiliated/ljl)
17:14:26 Quit ender (Quit: 9% of lawyers give the rest a bad name.)
17:14:48LjLhello. i was looking at the Sansa Clip Zip (i have a Clip+ currently), and while i gather that RockBox is still unstable, i don't seem to find very detailed information, like there are for other players, on what works and what doesn't. i seem to get the general impression it's marked unstable but, basically, everything works fine, is this accurate?
17:14:52 Join ender [0] (
17:15:43bertrikyes, it's about as stable as the clip+
17:15:55bertrikwhich is quite similar in hardware too
17:16:49bertriksome themes make the clip zip behave weirdly, but that likely not a player-specific problem
17:17:11 Quit ender (Client Quit)
17:17:19bertrikI think also some colours are a bit off for some clip zips with a particular type of display
17:18:05bertrikand perhaps one of the snake plugins doesn't quite work, but I can't remember if this was fixed already
17:18:59LjLwell, seems like nothing serious. the one thing i've read on the wiki is that the USB may have problems, with two bug reports about file corruption... but it doesn't exactly look like those two reports are filled with comments confirming the issues
17:19:07 Join ender [0] (
17:35:51 Join sentriz [0] (~Senan@
17:51:29 Join lazer [0] (~lazer@
18:05:38 Join lebellium [0] (~chatzilla@
18:08:08 Quit Rower85 (Ping timeout: 265 seconds)
18:09:19 Join Rower85 [0] (
18:12:06***Saving seen data "./dancer.seen"
18:14:36 Quit Buglouse (Read error: Connection reset by peer)
18:33:49 Quit Guest38219 (Read error: Connection reset by peer)
18:40:06 Quit SuperBrainAK (Ping timeout: 260 seconds)
18:41:25 Join trees_ [0] (~orionaai3@
18:42:14trees_wondering about something
18:43:23trees_what format would you recommend to get best audio-quality vs compression; nothing that sounds weaker than a 256mp3 but ideally nothing as big as flac files I have already
18:43:51copperVorbis -q 5 (~160 kbps) is generally transparent
18:43:56trees_I have only got experience working with mp3 and FLAC (well and shit like wav n stuff)
18:44:40trees_vorbis eh? Never used it, is free/easy to acquire codecs needed for conversion?
18:44:40copperthere's lossyFLAC if you're anal about being 100% sure that you won't get any artifacts - they're about half the size of FLACs (~440 kbps)
18:44:53copperyes Vorbis is free
18:44:58copperFree as in Speech
18:45:02copperand as in beer
18:45:02trees_I use f2k for conversion so am assuming it'd be easy
18:45:35trees_lossyFLAC eh? Sounds a bit redundant
18:46:05trees_thank you, am looking forward to pretending my 1Gig ipod nano has enough room for music
18:46:24copperis it rockboxed?
18:46:33trees_rockbox to the end
18:46:46copperyou could maximize space by using Opus at 96-128 kbps
18:46:51copperwith a dev build of rockbox
18:46:53 Quit radio23 (Remote host closed the connection)
18:46:55trees_if any devs are here btw, this big ugly cap-locked THANKU is for you
18:47:00 Join radio23 [0] (
18:47:09trees_\umm... 96-128 sounds aweful tho
18:47:15coppernot with Opus
18:47:28copperOpus is the latest lossy codec
18:47:31trees_Never even heard of Opus
18:47:47trees_Ill check it out.... thanks :>
18:50:02n1sOpus is not very fast on rockbox yet so it will suck more battery than mp3/vorbis
18:51:01trees_this is relevant
18:51:23trees_as this ipod is a pos and the battery is more deteriorated than not at this point I get 1hr MAX listening time
18:51:43trees_what it comes down to is that I need to get a sansa w the SD expansion slot
18:53:38 Nick the-kyle1 is now known as the-kyle (
18:53:51 Quit Horscht (Quit: Verlassend)
18:54:00copperMusepack is the most battery efficient lossy codec I think
18:54:26copperbut it's tuned for transparency at about 170 kbps
18:54:29trees_hmmm.... Musepack eh? hows the compression quality?
18:54:52copperat the default quality setting (~170), it's great
18:55:13trees_as it stands all that I have loaded on the rockbox is 320mp3s so that would be on par?
18:56:09*gevaerts thinks that all of this is subjective enough to make testing one's own music a requirement
18:56:47trees_am thinkin the same thing. Or that I should just upgrade hardware and get somethin w SD expansion
18:56:53soapif you can distinguish musepak at 170 from MP3 @ 320 from MP3 @ half that average bitrate I'll give you a 4GB Nano. This really is drifting far from #rockbox
18:58:10trees_'from mp3 at half that average bitrate'.... meaning somethin like 160mp3?
18:58:30soapassuming VBR, yes.
18:59:15trees_okay well that is something of a moot point but I can take the hint. One last question tho
18:59:20 Quit sentriz (Ping timeout: 252 seconds)
18:59:42trees_Ive been needing to upgrade hardware for quite some time now... what SD card supporting rockbox supporting player do you advise?
18:59:48soapYou mentioned foobar already. Use the ABX plugin for it. I'll be shocked if you can successfully ABX a 320kbps MP3 from a -V5 (~128kbps VRB) MP3. There. I more than doubled your storage for you.
19:00:20soapAny which takes SD cards and is "supported" is good-to-go.
19:01:12trees_okay soap not gonna debate with you but for what its worth 128s sound weak and dull compared to 320s. Particularly when theres organic sounding basslines
19:01:43trees_anyways, thanks for the info guys. Take care and fare well...
19:01:46soapThere is no basis for that claim. basslines are the easiest thing to encode. Blind test with the ABX comparator.
19:01:59soapplacebo is a powerful thing.
19:03:04trees_there is no basis for opinion and hearing. We all hear differently, I wouldnt be clogging up what little storage I have w clunky 320s if I didnt mind hearing my tunes in 128 but like I said I dont want to get into a debate about codecs so lets agree to disagree. Ciao.
19:03:13 Quit trees_ (Quit: pz)
19:17:12copperhe didn't get your message
19:17:53copperexpectation bias and ABX tests are surprisingly hard to explain to people
19:20:01 Join Ward [0] (
19:20:26 Nick Ward is now known as Guest20990 (
19:21:59copperit's very hard for people to let go of the complete trust they have in their own ears
19:22:26 Quit KiwiCam (Remote host closed the connection)
19:42:16 Quit Guest20990 (Quit: Blarglarg)
19:52:02jnccopper: I've been mucking about with Opus, and I cannot tell the difference (with headphones) between my source 44.1KHz/16-bit music and the Opus encoded output ; while I do notice differences with say Vorbis or MP3 even at higher bitrates
19:52:23jncfor me it sounds good enough
19:52:48copperby doing ABX tests?
19:53:32copperwhat's you're threshold on MP3/Vorbis?
19:54:01jncwell I go to 320kbps and MP3 is almost the same, but I still somehow always pick it out
19:54:16copperp value < 0.05?
19:54:19jncand Vorbis has a few corner cases where it does not encode well, and I can pick that out
19:57:02jncdon't know about p value, did it with a shell script and compared notes after 5 runs
19:57:14jncprobably not statistically interesting :)
19:57:43copperyou need proper ABX software and like 12 runs
19:58:15copperfoobar2000 with the ABX comparator component on Windows, squishyball on linux
19:58:37 Join webguest95 [0] (
19:58:38jncinteresting, will take note of "squishyball"
19:58:47 Join pretty_function [0] (~sigBART@
19:59:14webguest95anyone looking for a cheap Samsung Z5?
19:59:34 Join Gallomimia [0] (
19:59:34 Quit webguest95 (Client Quit)
20:00:45jnccopper: do you know, how should I proceed to find out what broke between v3_7 and v3_8 ? git bissect won't work
20:01:20copperI'm not a Rockbox developer
20:01:24jncoh okay
20:02:01copperbtw, 5 good guesses out of 5 is the required minimum
20:02:06copper4 out of 5 doesn't cut it
20:02:30copperbut 12-16 runs is best
20:03:19lazerAlexP: now it works... :)
20:03:20gevaertsjnc: in what way doesn't git bisect work?
20:03:40AlexPlazer: Glad to hear it
20:05:05 Join Buglouse [0] (~Buglouse@unaffiliated/Buglouse)
20:05:10lazerI had to install the new bootload v7pre4 and to install the the firmware to rom and ram
20:06:14lazerI can now boot from rom or ram but not from compact flash... but it works. When the player runs it detects the compact flash.
20:08:09jncgevaerts: it's saying that v3_7 is a branch, and head is not; something like that
20:08:21jncgevaerts: 'merge base'
20:08:26gevaertsOK, so you basically need to find the branch points
20:09:09jncgevaerts: yes, could use a little help understanding what that means. I researched and found a picture that explains, but I don't know what to do about it
20:09:45jncalso I'm unfamiliar with how rockbox git and svn correlate "back then in time"
20:10:33gevaertsjnc: f1fd602 is the last revision before the 3.7 branch, 48b1a2d is the last one before 3.8
20:10:59gevaertsThere's a better way I can't remember, but I found those using git log −−graph −−oneline −−all
20:11:38*jnc pokes keyboard some more
20:12:08***Saving seen data "./dancer.seen"
20:12:45jncgevaerts: FYI though 3.7 final was okay and I can't make it crash in usb mode on fuze v1 ; maybe a fluke? badblocks is pretty good at triggering failure mode
20:13:51gevaertsHard to say. You never know...
20:22:00 Join SuperBrainAK [0] (
20:26:18jncgevaerts: is the theory that all 3.7 branch should have been merged in before making 3.8 branch?
20:26:41gevaertsNo. The 3.7 branch shouldn't have anything that's not in master
20:26:47jncah okay
20:35:29saratoga_we basically just take all the master, declare ti stable and branch for releases
20:35:33saratoga_nothing really gets merged
20:36:11saratoga_then maybe a couple fixes go in before we release the binary
20:36:44saratoga_pamaury: yes, but i have no idea whats with those RMAA graphs you posted, never seen that before
20:37:28saratoga_Bagder, Zagor: could one of you update the website with the Clip Zip stable?
20:37:30pamauryit's weird that the frequency response has such a behaviour in the 10K range
20:37:33saratoga_i pushed it last night
20:37:58copperpamaury: link?
20:38:29melmothXis it stable?
20:38:31pamaurysee the fuze+ forum thread or this
20:39:49gevaertssaratoga_: should we declare it stable outside of a release? There's no 3.12 for it...
20:39:55gevaertsThat might confuse people
20:40:19saratoga_do we usually wait for a release? i don't remember doing that
20:40:39pamaurycopper: are you familiar with such graphs ?
20:40:43gevaertsI *think* we do, but I'm not sure now
20:41:23copperpamaury: I'm familiar with RMAA yes
20:42:45copperand what's the forum thread?
20:43:58copperthose noise levels are pretty bad
20:44:35pamaurywe are comparable to the OF though
20:44:45 Quit Gallomimia (Read error: No route to host)
20:44:49 Join wodz [0] (
20:44:56 Nick ukleinek_ is now known as ukleinek (
20:45:05copperI just wonder if it's the Fuze+ or the sound card
20:45:19 Join Gallomimia [0] (
20:45:25copperI'm guessing it's the soundcard
20:45:35wodzpamaury: Where to look for usb_state problem?
20:46:11pamaurycopper: this for the test with my sound card as reference
20:46:27pamaurymight come from my sound card
20:46:53pamaurywodz: wait a minute, I'll give you pointers in a sec
20:47:15copperso the Fuze+ is really that bad??
20:47:28saratoga_copper: those noise levels are very good
20:47:40pamauryI don't call it bad
20:47:47saratoga_mostly because thats an unloaded test though
20:47:54pamauryyeah of course
20:48:06 Quit akaWolf (Ping timeout: 265 seconds)
20:48:18 Quit pretty_function (Ping timeout: 260 seconds)
20:48:23copperthey're way higher than the Clip+
20:49:10saratoga_3 dB difference
20:49:35saratoga_probably just due to differences in the volume level between the tests
20:50:13copperI was looking at the OF figures
20:50:20saratoga_again those are both unloaded tests though, so basically meaningless
20:51:07copperIIRC a sine wave at-70 dBFSS is audible
20:51:28copperhmmm my term screwed up
20:51:34pamauryI mostly did the test because there was a noticable noise on the fuze+ which revealed a problem, now that should be better
20:51:41saratoga_-70 dBFS isn't really a volume
20:51:57saratoga_so its hard to evaluate what that statement means
20:52:15saratoga_you'd need to know what FS is
20:52:39coppera sine at -70 dBFS played back at my usual listening volume
20:52:53soapInteresting that before the changes FR was the same between OF and RB, as was crosstalk. Noise THD went way down with the changes, crosstalk improved, and FR got a bit wonky.
20:53:32pamaurywodz: so you said that usb_state stays as USB_EXTRACTED ?
20:54:28 Join cela [0] (~cela@
20:55:14 Join sakax [0] (
20:55:24wodzpamaury: yes
20:56:30pamaurywodz: you can either remove #define USB_STATUS_BY_EVENT is config.h for rk27xx target OR implement status by interrupt in usb-drv-rk27xx.c by call calling usb_status_event(USB_INSERTED); on connection and usb_status_event(USB_EXTRACTION); on disconnection. I think interrupt 7 of the udc is connection status (even though it's marked as reserved)
20:57:08celaI get a crash after I unplug the usb on Fuze+ when in bootloader transfer mode, my bootloader states version 1.0. Is there an update to fix this issue?
20:58:35jnccela: commonly reported problem, looking at flyspray reports
20:59:15pamauryunfortunately it's not clear that the bootloader in the trunk is much better w.r.t to this issue, we could release a new version though
21:00:35wodzpamaury: How do you know about int 7? Have you tested this on your rk27xx?
21:00:55pamauryiirc yes and I had a look at the rom to notice that
21:01:06pamaurybut check my yourself, I'm not 100% sure
21:01:58wodzInteresting, will check. Ok the other route is to remove USB_STATUS_BY_EVENT define. Do I need something special in this case?
21:02:06celaIs there a newer bootloader version than 1.0 for Fuze+ , the main Fuze+ port itself is much more stable than it used to be , thank you for all the good work.
21:03:22 Part cela
21:04:03pamauryno there is none but I want to do a release (yes I've said this for a long time I known).
21:04:13wodzpamaury: ^
21:04:13pamaurywodz: iirc no, it will poll usb_detect in a thread
21:04:18wodzah ok
21:04:33 Quit y4n (Quit: Today is the perfect day for a perfect day.)
21:05:09wodzint 7 is triggered on connection, don't know if it retriggers on extraction.
21:10:25wodzpamaury: according to SDK this irq should trigger both on connect and disconnect - will check
21:11:20 Quit Gallomimia (Read error: Connection reset by peer)
21:11:51*pamaury leaves for a moment
21:12:07 Join Gallomimia [0] (
21:15:20 Join Robin0800 [0] (
21:17:09 Quit Gallomimia (Ping timeout: 260 seconds)
21:21:05 Quit n1s (Quit: Ex-Chat)
21:34:45 Join Gallomimia [0] (
21:37:16wodzok this int triggers both on usb connection and disconnection
21:39:57 Quit lazer (Read error: Connection timed out)
21:40:36 Join lazer [0] (~lazer@
21:41:04wodzhmm, wrong it trigger on connect only
21:41:26pamaurywodz: perhaps there is one for disconnection ?
21:42:45pamaurywodz: manual states it is triggered on connection (and says nothing about disconnection)
21:43:03wodzhmm, it seems to trigger for disconnection but I can't distinguish it from connection.
21:43:05pamauryin this case you should use the usb_detetc mechanism
21:43:26wodzok, lets see
21:48:46wodzI am too tired to analyze this properly but with USB_STATUS_BY_EVENT commented out it is not any better
21:49:22 Quit wodz (Quit: Leaving)
21:50:27 Join sentriz [0] (~Senan@
21:50:41pamaurywodz (logs): perhaps usb_detect needs to return different values, i'll have a look later
21:55:52 Quit saratoga_ (Ping timeout: 245 seconds)
22:08:17 Quit lazer (Quit: Verlassend)
22:12:12***Saving seen data "./dancer.seen"
22:14:17 Quit Buglouse (Ping timeout: 245 seconds)
22:37:00 Join Buglouse [0] (~Buglouse@unaffiliated/Buglouse)
23:01:29 Quit Gallomimia (Read error: Connection reset by peer)
23:11:03 Quit Rower85 (Quit: Hmmm...)
23:12:13 Join Gallomimia [0] (
23:15:27 Join Rower85 [0] (
23:33:58 Quit sentriz (Ping timeout: 252 seconds)
23:39:20 Join jhMikeS [0] (~jethead71@
23:39:20 Quit jhMikeS (Changing host)
23:39:21 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
23:42:11 Quit Belzebub (Ping timeout: 264 seconds)
23:53:16 Quit lebellium (Quit: ChatZilla 0.9.89 [Firefox 18.0/20121128060531])
23:53:23 Quit bertrik (Remote host closed the connection)
23:55:55 Join Guest44874 [0] (
23:58:33 Quit Rower85 (Ping timeout: 256 seconds)

Previous day | Next day