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 2011-09-23

00:00:13[Saint]If you could post your thoughts/findings to FS #12254 it would be very much appreciated.
00:00:14fs-bluebothttp://www.rockbox.org/tracker/task/12254 RaaA cabbiev2 [Full Touch Application Port] (patches, new)
00:01:52 Quit n1s (Remote host closed the connection)
00:01:56pixelmawell, I wasn't sure if I missed something or not (regarding the scrollbars). It's definitely more usable than SVN cabbiev2 but I'd like to try it a bit more
00:02:56[Saint]Sure, the more testing the better. Preferably with album art, and without.
00:03:16pixelmaI'm currently a bit annoyed by possibly using the latest (working) incarnation of my own theme
00:03:29 Join mudd1 [0] (~cmertes@ip-78-94-202-227.unitymediagroup.de)
00:03:36pixelmaerr... s/using/losing
00:03:50[Saint]Oh...bugger.
00:03:53[Saint]No backups?
00:04:16[Saint]Sorry if I caused this, in some roundabout fashion. :-S
00:04:41pixelmanah, my fault in the end
00:04:45polemon[Saint]: so the whole iPod_Control can go?
00:04:59[Saint]It can, yes.
00:05:09polemongreat news
00:05:12[Saint]There's no real reason to delete it, though.
00:05:19polemontakes up space
00:05:29[Saint]It'll just get recreated given the correct circumstances.
00:05:30polemonespecially that "iTunesControl" file
00:05:39Tornehuh?
00:05:46Torneit's, like, 2MB when empty ;)
00:05:47polemonok, well, then it'll get recreated
00:05:58Tornethat's not really an amount of space worth reclaiming
00:06:04polemoniPod_Control is around 50MB right now
00:06:14polemon20 of those is a binary file called iTunesControl
00:06:22Tornedid you previously have itunes-synced music on there?
00:06:30[Saint]that's stiall a very trivial amount, if you ask me...but, "meh".
00:06:30polemonyes, but the music is gone
00:06:31Tornealso, 50MB is also not an amount of space worth reclaiming, really :)
00:06:35TorneYou deleted the music?
00:06:40polemonyes
00:06:48TorneWell, that's the itunes database for the music you deleted, then
00:07:01TorneIf you delete it, it will be recreated, but it will be much, much smaller
00:07:07Tornesince it won't have any database entries :)
00:07:51polemonI just deleted all folders that were there prior to RockBox installation
00:08:15 Quit petur (Quit: Leaving)
00:08:37Torneanyway, deleting them is fairly useless as they get recreated reasonably easily
00:08:50Tornejust set the folders hidden and they won't show up in the file browser in rockbox by default
00:09:27polemontoo late, they're already gone
00:09:51polemonok, I'll reboot the iPod now, and see what I can do...
00:10:09 Quit mudd1 (Ping timeout: 245 seconds)
00:10:55pixelma[Saint]: found a backup but it's failing to load the WPS currently, probably because of the repeat_press etc. change. (Yes, I haven't updated in a while)
00:11:35[Saint]Ah, that's good.
00:15:37 Quit liar (Quit: hallowed are the ori!)
00:15:41pixelmathat's what it was
00:22:18 Join mudd1 [0] (~cmertes@ip-78-94-202-227.unitymediagroup.de)
00:23:03***Saving seen data "./dancer.seen"
00:26:00 Join milk [0] (~milk@94-193-93-226.zone7.bethere.co.uk)
00:29:26 Quit ender` (Quit: I like work: it fascinates me. I can sit and look at it for hours. -- Jerome K. Jerome)
00:36:17 Quit mudd1 (Ping timeout: 260 seconds)
00:40:49 Join Scromple [0] (~Simon@115-64-195-104.static.tpgi.com.au)
00:46:07 Join user890104_ [0] (~Venci@6bez10.info)
00:46:39 Join soap [0] (~soap@rockbox/staff/soap)
00:47:49 Quit user890104_ (Remote host closed the connection)
00:52:29 Join Scr0mple [0] (~Simon@115-64-195-104.static.tpgi.com.au)
00:55:10 Quit Scromple (Ping timeout: 248 seconds)
01:00
01:07:24 Join funman [0] (~fun@rockbox/developer/funman)
01:16:02 Quit MethoS- (Remote host closed the connection)
01:19:19 Quit [Saint] (Ping timeout: 245 seconds)
01:33:33 Quit funman (Quit: leaving)
01:45:54-->"ods 100" received from cgthayer (~chatzilla@107.43.121.86)
02:00
02:14:50 Quit dre (*.net *.split)
02:14:50 Quit Farthen (*.net *.split)
02:20:32 Join dre [0] (dre@gateway/shell/xzibition.com/x-aagulhyylwfnqgjb)
02:20:32 Join Farthen [0] (~Farthen@2a01:4f8:101:2a4:0:bc28:b2e1:9)
02:23:06***Saving seen data "./dancer.seen"
02:23:49 Quit dre (*.net *.split)
02:23:49 Quit Farthen (*.net *.split)
02:31:12 Join Sophiie [0] (~Sophiie@pdpc/supporter/active/sophiie)
02:31:54 Part Sophiie
02:34:20 Join dre [0] (dre@gateway/shell/xzibition.com/x-aagulhyylwfnqgjb)
02:34:20 Join Farthen [0] (~Farthen@2a01:4f8:101:2a4:0:bc28:b2e1:9)
02:34:59 Quit simonlnu (Quit: brb)
02:35:26 Join simonlnu [0] (~simon@unaffiliated/simonrvn)
02:41:38 Quit mamarley (Read error: Operation timed out)
02:41:54 Quit factor (Read error: Connection reset by peer)
02:43:23 Quit dre (*.net *.split)
02:43:23 Quit Farthen (*.net *.split)
02:45:58 Join mamarley [0] (~quassel@Lee-15-155.rh.ncsu.edu)
02:47:36 Quit mamarley (Read error: Connection reset by peer)
02:48:53 Join mamarley [0] (~quassel@Lee-15-155.rh.ncsu.edu)
02:51:50 Quit mamarley (Read error: Connection reset by peer)
02:53:08 Join mamarley [0] (~quassel@Lee-15-155.rh.ncsu.edu)
02:58:08 Join factor [0] (~factor@74.197.205.204)
03:00
03:00:47 Quit mamarley (Read error: Connection reset by peer)
03:03:18 Join mamarley [0] (~quassel@Lee-08-199.rh.ncsu.edu)
03:25:13 Join kadoban [0] (~kadoban@ip98-165-177-158.ph.ph.cox.net)
03:27:32 Quit kevku (Ping timeout: 244 seconds)
03:58:15polemonanybody still here?
03:58:42polemonwhen selecting the line selector, how do I make the lines being indented?
04:00
04:23:07***Saving seen data "./dancer.seen"
04:24:06polemonor is there some documentation on how to make themes?
04:27:31JdGordonCustomWPS on the wiki for that
04:27:50JdGordoni tihnk lines will be indenteded as long as the title is enabled
04:28:09JdGordonthe scrollbar might need to be on the left also
04:28:19polemonI think it's the scrollbar
04:28:24polemonit indents the lines
04:28:24JdGordonyou can now use thmes to draw the lists any way you like
04:28:30polemon, even when not shown
04:28:53polemonwell, I want to make some older themes work again
04:29:38polemonthings like nClix and ElectricBarsOfColour
04:30:12polemonmaybe I'll do another one in the future, but it seems there isn't many people using old iPod nanos these days
04:30:32polemonalso, how is the support for things like the iPod touch and the latest iPod nanos?
04:42:04 Join Keripo [0] (~Keripo@eng389.wireless-resnet.upenn.edu)
04:42:53 Quit amiconn (Disconnected by services)
04:42:53 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:42:53 Quit pixelma (Disconnected by services)
04:42:55 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:42:57 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:43:15 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:47:53 Quit [7] (Disconnected by services)
04:48:26 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:52:10 Quit kadoban (Quit: bye)
04:52:28 Join kadoban [0] (~kadoban@ip98-165-177-158.ph.ph.cox.net)
05:00
05:01:19 Quit tchan (Read error: No route to host)
05:01:42 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
05:23:17 Join Rob2223 [0] (~Miranda@p4FFF2C0F.dip.t-dialin.net)
05:26:05 Quit Rob2222 (Ping timeout: 252 seconds)
05:36:50 Quit Horschti (Quit: Verlassend)
05:45:46 Nick shai_ is now known as shai (~Shai@l192-117-110-233.cable.actcom.net.il)
05:55:37 Quit shai (Quit: Leaving)
06:00
06:02:23 Quit Keripo (Ping timeout: 260 seconds)
06:23:09***Saving seen data "./dancer.seen"
07:00
07:10:26 Join n1s [0] (~quassel@rockbox/developer/n1s)
07:20:50 Quit powell14ski_ (Quit: powell14ski_)
07:30:26 Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
07:36:25 Join Buschel [0] (~chatzilla@p54B66928.dip.t-dialin.net)
08:00
08:05:25 Join nick-p [0] (~nick@82-69-105-120.dsl.in-addr.zen.co.uk)
08:08:14 Join JdGord [0] (~AndChat@106.70.193.73)
08:14:00BuschelJdGord: quite some traffic in FS #12273 :)
08:14:02fs-bluebothttp://www.rockbox.org/tracker/task/12273 use buflib for fonts (patches, new)
08:15:39 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:20:52 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 6.0.2/20110902133214])
08:20:54LinusNLalufu: i didn't write that code, i just committed a patch. however, i think the idea is to send periodic info to the ipod.
08:21:44*Zagor wants more commits today!
08:22:12 Join ender` [0] (~ender@foo.eternallybored.org)
08:23:13***Saving seen data "./dancer.seen"
08:25:57n1sZagor: i can give you a whitespace cleanup one if you like :)
08:26:05Zagorhehe
08:26:50 Part Zagor ("Client exiting")
08:26:55 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:29:18LalufuLinusN: yes, the called function eventually is used to send info from the ipod to the device. I was just wondering why this specific method was chosen.
08:30:05LalufuOr maybe I'm trying to misuse some functions.
08:30:30 Quit ej0rge (Ping timeout: 276 seconds)
08:30:34LalufuIf I want an event to fire roughly every 100ms, is is a valid choice to create a timeout for that?
08:34:02 Quit JdGord (Ping timeout: 260 seconds)
08:35:51LinusNyou can use a tick task to send an event, or you can use sleep() and TIME_AFTER() in a thread. it depends on the application.
08:36:40 Join JdGord [0] (~AndChat@106.70.193.73)
08:44:41 Join jdgord_ [0] (~AndChat@106.70.193.73)
08:44:42 Quit JdGord (Read error: Connection reset by peer)
09:00
09:00:59 Quit jdgord_ (Quit: Bye)
09:01:33LalufuLinusN: ticks are used now, but those are very high frequency, so the tick handler spends most of it's time just returning.
09:01:43Lalufuthat looks like a bad use of CPU time to me.
09:02:08Lalufuthus the idea to register a timeout with a base frequency of 100ms or thereabouts
09:02:47LalufuBut kernel.c says: "be mindful this is not really intended for continuous timers" about the timeout_* functions
09:09:20LinusNthe timeout functions work in the same way as the iap_task
09:09:51LinusNbut the implementation is slightly more efficient than the iap_task
09:11:24LinusNit doesn't really matter how you implement it, since there has to be a comparison to the tick counter somewhere
09:11:29 Join kevku [0] (x@2001:470:1f15:1d8f::)
09:11:51LinusNwhether you sleep() or add a tick task
09:12:29LinusNsleep() is a little more gentle though
09:17:01Lalufuthe iap handler does not really have a thread that could sleep in a useful manner, it's entirely triggered by outside events
09:20:04 Join petur [0] (~petur@rockbox/developer/petur)
09:21:16LinusNthen i think a timeout callback might the best choice
09:22:13LinusNas long as it is only activated when you attach a peripheral, of course
09:24:40kugelLalufu: triggered by outside events seemsperfect for a thread
09:25:22kugelyou can queue_wait_w_tmo() with timeout to wait for incoming events (via queue_post() or queue_send())
09:26:07kugelusing timeout_* forperiodic stuff in favor of tick task is a but strange because theey're also implemented via tick tasks
09:26:32Lalufukugel: the current code has two main entry points: iap_task() (which is registered via tick_add_task()) and iap_getc() which is called directly from firmware/serial.c
09:26:39Lalufufirmware/drivers/serial.c even
09:27:35Lalufuiap_task() basically just counts up an internal counter, to queue an internal message every 100 or 500 ms.
09:27:43 Quit Scr0mple (Read error: Connection reset by peer)
09:27:58Lalufuthe latter seemed kind of duplicate to me, as timeout_* does the same thing and already exists.
09:28:21 Quit scanf (Quit: leaving)
09:29:40kugelyou can use timeout just fine
09:29:46LalufuOK, thanks
09:31:16kugelbut when you say queue an internal message it sounds like more than one message could be in the pipeline. tick_tasks/timeout aren't very suitable for this as you need to implement the queueing
09:33:40Lalufuiap_task() posts SYS_IAP_PERIODIC to the button queue
09:34:41Lalufuwhich eventually leads to iap_periodic() being called
09:35:17Lalufumultiple messages being in the queue for that is not really critical
09:43:20 Join pamaury [0] (~quassel@vit94-1-82-67-248-70.fbx.proxad.net)
09:43:21 Quit pamaury (Changing host)
09:43:21 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:51:56 Join [Saint] [0] (~st.lasciv@203.184.50.187)
10:00
10:03:11pamauryjust seen the discussion on tick and timeout, I think most devices also have a generic timer_* implementation so you can register a timer at the frequency you want (and it is *not* based on tick timer)
10:08:10amiconnThe user timer (timer_*) is not meant for permanent use in the core
10:08:29 Join ej0rge [0] (~alhaz@207.135.137.71)
10:08:44amiconnThis is because many SoC only have two hardware timers. One is already taken by the kernel tick, the other is made available through timer_*
10:09:27amiconnThe latter has a priority framework so that a plugin can "steal" the timer from another module.
10:10:05 Quit milk (Quit: baaaiiii)
10:10:44amiconnOne example is backlight fading on H1x0. It uses the timer, but if the greylib is running (which needs the timer) the timer is taken and the backlight doesn't fade
10:11:04 Join B4gder [241] (~daniel@rockbox/developer/bagder)
10:11:15 Quit kadoban (Ping timeout: 276 seconds)
10:12:14 Join dre [0] (dre@gateway/shell/xzibition.com/x-aagulhyylwfnqgjb)
10:12:14 Join Farthen [0] (~Farthen@2a01:4f8:101:2a4:0:bc28:b2e1:9)
10:20:14 Quit mikroflops (Ping timeout: 252 seconds)
10:23:16***Saving seen data "./dancer.seen"
10:30:28pamauryiap is not permanent, on the contrary it has a very restricted usage
10:33:20 Join mikroflops [0] (~yogurt@h-34-156.a238.priv.bahnhof.se)
10:34:10 Join Jak_o_Shadows [0] (~hayden@CPE-144-136-211-121.sa.bigpond.net.au)
10:59:53 Quit [Saint] (Ping timeout: 260 seconds)
11:00
11:04:03 Join mudd1 [0] (~cmertes@ip-78-94-202-227.unitymediagroup.de)
11:30:12 Quit factor (Ping timeout: 260 seconds)
11:37:42Lalufuis there an explanation somewhere how the public interface to the tagcache works and how one uses it?
11:44:14 Quit mudd1 (Ping timeout: 248 seconds)
11:45:04pamaurynot really, have a look at how tagtree uses it, that's your best chance
11:45:10 Join factor [0] (~factor@74.197.205.204)
11:59:29nick-pHas anyone spoken with sideral in the last few weeks?
12:00
12:15:06 Join dfkt|a [0] (~dfkt@unaffiliated/dfkt)
12:16:19 Join pondlife [0] (~Steve@rockbox/developer/pondlife)
12:19:46 Quit dfkt|a (Remote host closed the connection)
12:23:13 Join dfkt|a [0] (~dfkt@unaffiliated/dfkt)
12:23:18***Saving seen data "./dancer.seen"
12:39:25 Join dfkt|a_ [0] (~dfkt@chello062178002170.1.11.univie.teleweb.at)
12:39:25 Quit dfkt|a_ (Changing host)
12:39:25 Join dfkt|a_ [0] (~dfkt@unaffiliated/dfkt)
12:39:26 Quit dfkt|a (Read error: Connection reset by peer)
12:45:27 Quit dfkt|a_ (Ping timeout: 260 seconds)
13:00
13:00:55 Quit nick-p (Quit: Leaving)
13:19:16 Join MethoS- [0] (~clemens@134.102.106.250)
13:22:43 Quit soap (Quit: Leaving)
13:35:58 Join Topy [0] (~Topy44@f048142056.adsl.alicedsl.de)
13:40:04 Quit T44 (Ping timeout: 260 seconds)
13:49:19 Quit sinthetek (Ping timeout: 260 seconds)
14:00
14:06:42 Quit pondlife (Quit: Leaving.)
14:09:40 Quit factor (Read error: Connection reset by peer)
14:17:23 Join sinthetek [0] (~sinthetek@cpe-174-111-016-189.triad.res.rr.com)
14:17:23 Quit sinthetek (Changing host)
14:17:23 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
14:18:36 Quit Jak_o_Shadows (Remote host closed the connection)
14:23:22***Saving seen data "./dancer.seen"
14:27:14 Join factor [0] (~factor@74.197.205.204)
14:27:59 Join nick-p [0] (~nick@82-69-105-120.dsl.in-addr.zen.co.uk)
14:53:34 Join wodz [0] (~wodz@iwl138.internetdsl.tpnet.pl)
14:54:18wodzpamaury: Do you have a clue if f+ uses internal codec block or external one through i2s?
14:55:15polemonis there any wor underway to port RockBox to the later generation iPods?
14:55:42polemonversion 3,4,5 would be nice.
14:55:46polemonI don't know aboutr
14:55:52pamaurywodz: it uses an internal block
14:56:15wodzsome work on the 4g nano is started I believe but it is on early reverse engineer stage
14:56:19polemonthe latest iPod, supposedly, they got things right in the latest version
14:56:40polemonis the 4g and 5g anywhere similar?
14:56:54wodzpamaury: so pcm driver should be easy - dedicated dma channel pushing samples to the fifo :-)
14:57:02n1spolemon: probably
14:57:15polemonby the looks of it, it's pretty much the same shit in a new wrapper
14:57:21*gevaerts suspects that polemon means ipod *nano*, since the 3rd, 4th, and 5th gen ipods are supported
14:57:30polemoniPod nano it is
14:57:49polemonthe normal iPods are called "classic" for some reason
14:57:59gevaertsThey aren't
14:58:06n1si don't think the exploit used on the other samsung ipods work on the nano 5g though but i'm not sure
14:58:19pamaurywodz: yes, and you also need to setup the zillions things to make it work, but I don't have much time during the week now. And I also need to implement batteryc harging and touchpad otherwise the device is not usable
14:59:16n1son the 4g nano i think it's flash/ftl that's blocking
14:59:35wodzpamaury: yep datasheet is a bit scary with all this settings to simply play samples
15:00
15:00:43 Quit wodz (Quit: Leaving)
15:03:47 Quit B4gder (Quit: Konversation terminated!)
15:12:59user890104polemon: about the 3g, 4g, etc nanos please see http://www.freemyipod.org/wiki/Status and http://www.freemyipod.org/wiki/Todo_list
15:13:37user890104if these tasks are completed, the resulting code could be easily ported to rockbox
15:22:05 Quit Zarggg (Quit: Rebooting client...)
15:37:57 Join stoffel [0] (~quassel@p57B4A411.dip.t-dialin.net)
15:41:30*n1s wants a button to exit usb mode without disconnecting the cable
15:42:55LalufuI'd buy that.
15:45:36pamauryI think most device have a button to *prevent* usb mode on connexion
15:46:02n1syeah, this would be convenient in some cases though
15:47:21ZagorI propose, uh, "left"
15:48:23n1sit will be tricky with the HID buttons though
15:50:41Zagorhow do you mean?
15:51:19n1sleft is in use in some of the HID maps
15:51:19gevaertsIf you use HID, you have to assume that people want *all* buttons I think
15:52:16Zagorbut... if you use hid you don't want to be stuck in the usb mode screen anyway do you?
15:52:16*Zagor is confused
15:53:25gevaertsNot necessarily, but that's the way it works now
15:54:35pamauryin usb mode you can't use the device for anything basically so it's natural to bind all buttons
15:54:38CIA-14New commit by nls (r30583): libtremor: small whitespace cleanup.
15:56:41CIA-14r30583 build result: All green
15:57:20Zagorpamaury: yeah, but it would be convenient to exit usb disk mode but keep charging
15:57:31n1sexactly
15:58:41gevaertsThis should be rather easy to implement, but you need to either take buttons away from HID or only support it without HID
15:59:28n1sperhaps a long press of the "switch hid map" button
16:00
16:00:17gevaertsor have a "no actual HID" map in the list where you can exit?
16:00:27Zagoronly without hid is the best imho.
16:00:27 Quit antil33t (Read error: Connection reset by peer)
16:00:41gevaertsIt's easiest anyway :)
16:00:52 Join antil33t [0] (~antil33t@203-100-223-143.callplus.net.nz)
16:01:50CIA-14New commit by nls (r30584): libtremor: arm asm for ff_vector_fmul_window_c on v4 and v5 cores, speeds up c200v1 and fuzev1 by ~0.4MHz and ipod classic by ~1MHz.
16:01:52ZagorI think wanting to be able to exit disk mode while in HID mode is a pretty strained use case
16:02:41n1scan you get out of hid while in the usb screen?
16:02:47 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
16:02:54ZagorI guess not
16:03:35CIA-14r30584 build result: All green
16:03:46Zagoralthough since there are mode select buttons in hid mode, having exit buttons is possible too
16:04:03n1syeah
16:04:06 Join WalkGood [0] (~4@unaffiliated/walkgood)
16:04:31n1si'd like a regular exit button most but i don't use hid so i might be biased :)
16:04:43*Zagor does the 103 second build dance
16:05:00gevaertsZagor: that's still more than a minute!
16:05:28Zagoradd more machines! we need more power, scotty!
16:06:30*TheSeven adds more build targets to compensate for the additional machines
16:06:52Zagorhehe
16:07:13n1swe also need more odd game plugins!
16:07:40 Join dfkt|a_ [0] (~dfkt@chello062178002170.1.11.univie.teleweb.at)
16:07:41 Quit dfkt|a_ (Changing host)
16:07:41 Join dfkt|a_ [0] (~dfkt@unaffiliated/dfkt)
16:11:44CIA-14New commit by zagor (r30585): Enable mysql_auto_reconnect
16:12:14 Quit dfkt|a_ (Ping timeout: 256 seconds)
16:12:36CIA-14New commit by zagor (r30586): Also sort builds based on number of able clients.
16:13:57 Join wodz [0] (~wodz@87-206-240-131.dynamic.chello.pl)
16:15:40wodzn1s: I guess you should update CodecPerformanceComparison wiki after massive vorbis optimizations
16:21:37 Quit petur (Quit: *plop*)
16:23:26***Saving seen data "./dancer.seen"
16:27:02 Part LinusN
16:29:21 Join y4n [0] (y4n@unaffiliated/y4ndexx)
16:44:46 Join dfkt|a_ [0] (~dfkt@chello062178002170.1.11.univie.teleweb.at)
16:44:55 Quit dfkt|a_ (Changing host)
16:44:55 Join dfkt|a_ [0] (~dfkt@unaffiliated/dfkt)
16:49:52 Quit dfkt|a_ (Ping timeout: 260 seconds)
16:53:30 Join powell14ski [0] (~powell14s@c-174-51-194-6.hsd1.co.comcast.net)
16:54:45 Quit antil33t (Read error: Connection reset by peer)
16:55:08 Join antil33t [0] (~antil33t@203-100-223-143.callplus.net.nz)
16:56:51 Quit nick-p (Quit: Leaving)
16:57:01-->"ods 100" received from cgthayer (~chatzilla@107.37.47.206)
16:59:56 Quit ender` (Read error: Connection reset by peer)
17:00
17:00:00 Join ender` [0] (~ender@foo.eternallybored.org)
17:00:15 Quit mystica555 (Ping timeout: 258 seconds)
17:00:39 Quit mystica555_ (Ping timeout: 260 seconds)
17:02:19 Part Zagor
17:03:25 Quit ender` (Read error: Connection reset by peer)
17:03:30 Join ender` [0] (~ender@foo.eternallybored.org)
17:07:32 Join dfkt|a_ [0] (~dfkt@chello062178002170.1.11.univie.teleweb.at)
17:07:32 Quit dfkt|a_ (Changing host)
17:07:32 Join dfkt|a_ [0] (~dfkt@unaffiliated/dfkt)
17:08:33 Quit stoffel (Ping timeout: 276 seconds)
17:09:41 Quit dfkt|a_ (Read error: Connection reset by peer)
17:14:27 Join mystica555 [0] (~Mike@71-211-199-174.hlrn.qwest.net)
17:16:25 Join ender1 [0] (~ender@foo.eternallybored.org)
17:17:03 Quit WalkGood (Quit: CaminaBien)
17:17:04 Join dfkt|a [0] (~dfkt@unaffiliated/dfkt)
17:17:19 Quit ender` (Read error: Connection reset by peer)
17:22:43 Join dfkt|a_ [0] (~dfkt@chello062178002170.1.11.univie.teleweb.at)
17:22:43 Quit dfkt|a_ (Changing host)
17:22:43 Join dfkt|a_ [0] (~dfkt@unaffiliated/dfkt)
17:22:43 Quit dfkt|a (Read error: Connection reset by peer)
17:25:45 Join Strife89 [0] (~Strife89@207.144.201.128)
17:34:30 Quit ender| (Ping timeout: 260 seconds)
17:34:33 Quit ender1 (Ping timeout: 276 seconds)
17:34:55 Join stoffel [0] (~quassel@p57B4A411.dip.t-dialin.net)
17:43:56 Quit dfkt|a_ (Remote host closed the connection)
17:46:02 Join kadoban [0] (~kadoban@ip98-165-177-158.ph.ph.cox.net)
17:57:52 Join bertrik [0] (~bertrik@ip117-49-211-87.adsl2.static.versatel.nl)
17:57:52 Quit bertrik (Changing host)
17:57:52 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
17:59:31 Quit factor (Read error: Connection reset by peer)
18:00
18:00:59 Quit TheSeven (Remote host closed the connection)
18:01:03 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
18:09:31 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:10:21 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
18:10:31 Quit jordan` (Ping timeout: 244 seconds)
18:12:20 Join GermanMushroom [0] (~c@s5146db6a.adsl.wanadoo.nl)
18:17:12 Join factor [0] (~factor@74.197.205.204)
18:23:27***Saving seen data "./dancer.seen"
18:24:26 Join ender| [0] (~ender1@foo.eternallybored.org)
18:28:46 Join ender` [0] (~ender@foo.eternallybored.org)
18:33:15 Join Horscht [0] (~Horscht@p5DD57212.dip.t-dialin.net)
18:33:16 Quit Horscht (Changing host)
18:33:16 Join Horscht [0] (~Horscht@xbmc/user/horscht)
18:35:05 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
18:35:58 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
18:47:25 Quit benedikt93 (Quit: Bye ;))
18:51:26 Quit GermanMushroom (Ping timeout: 248 seconds)
18:56:27 Quit ej0rge (Ping timeout: 260 seconds)
18:57:07 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
18:58:59 Quit pamaury (Remote host closed the connection)
19:00
19:00:15 Join GermanMushroom [0] (~c@s5146db6a.adsl.wanadoo.nl)
19:00:57 Join pamaury [0] (~quassel@vit94-1-82-67-248-70.fbx.proxad.net)
19:00:57 Quit pamaury (Changing host)
19:00:57 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
19:03:27 Join mudd1 [0] (~cmertes@ip-78-94-202-227.unitymediagroup.de)
19:13:33 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
19:19:13 Join Jerom [0] (~jerome@2a02:8420:216:a800:f66d:4ff:fe45:790f)
19:26:37 Join ej0rge [0] (~alhaz@207.135.137.71)
19:29:02 Join Buschel [0] (~chatzilla@p54B666DA.dip.t-dialin.net)
19:29:26 Join Horschti [0] (~Horscht@xbmc/user/horscht)
19:30:37 Join Buschel_ [0] (~chatzilla@p54B666DA.dip.t-dialin.net)
19:33:07 Quit Horscht (Ping timeout: 255 seconds)
19:33:29 Quit Buschel (Ping timeout: 245 seconds)
19:33:34 Nick Buschel_ is now known as Buschel (~chatzilla@p54B666DA.dip.t-dialin.net)
19:34:33 Join Xerion [0] (~xerion@5419A766.cm-5-2c.dynamic.ziggo.nl)
19:59:37 Join mystica555_ [0] (~mike@c-76-25-32-116.hsd1.co.comcast.net)
20:00
20:00:39 Join webguest988 [0] (~cfe6de72@www.haxx.se)
20:00:44webguest988hi
20:01:10webguest988is anyone here
20:02:44webguest988how do i request for my model
20:03:35 Join Zarggg [0] (~zarggg@24.229.139.169.res-cmts.sm.ptd.net)
20:03:35 Quit webguest988 (Client Quit)
20:09:28 Quit bluebrother (Disconnected by services)
20:09:29 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
20:10:48 Quit antil33t (Read error: Connection reset by peer)
20:11:14 Join antil33t [0] (~antil33t@203-100-223-143.callplus.net.nz)
20:11:23Buscheln1s: in total your tremor-changes results in a minor speedup on PP5022 (27.4 -> 27.8 MHz) and a huge speedup on S5L8701 (49.1 -> 43.3 MHz). so on nano2g vorbis does now decode faster than mp3 ;)
20:12:31Buschelas you mentioned the target with slower RAM benefits more
20:12:56Buschel(test file was our vorbis_256)
20:13:02 Quit fs-bluebot (Ping timeout: 248 seconds)
20:13:38amiconnLooks like you swapped the values for PP5022
20:14:10 Join fs-bluebot [0] (~fs-bluebo@g226068072.adsl.alicedsl.de)
20:14:26amiconnBtw, vorbis is faster than mp3 on PP too if you take into account that mp3 uses both cores
20:14:27Buschelno, I wrote awrong number. should be: 27.4 -> 26.8 MHz
20:15:07amiconnHmm, I should probably test on PP5002 (expected speedup is larger than on PP502x)
20:15:51wodzWhat is our policy with filenames when porting some external code? espeak has source files with .cpp extensions which is a bit weird since this are plain C files
20:16:05amiconnIf it uses dram, that is
20:18:16 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
20:18:42Tornewodz: renaming it is a trivial enough transform that it seems like a good idea
20:18:58n1samiconn: the larger block sizes get the largest gain on targets with small fast iram and slow dram
20:19:08kugelwodz: does it compile cleanly with cc?
20:19:25n1sour test_files have no such file but you can make one with oggenc using q-1
20:19:26kugelgcc will use g++ if it sees .cpp I think
20:19:58Torneunless you've actually tried compiling it with it renamed to .c don't assume it's plain C
20:20:06Torne"not using classes or templates" is not sufficient.
20:20:14n1sour toolchains don't include g++
20:20:24amiconnn1s: Yeah, and dram on PP5002 is slow-ish because of the broken cache
20:21:00Tornewodz: anyway, just rename them
20:21:04n1sas i wrote in the commit message, such a file got a 20MHz speedup on mcf5049
20:21:09TorneIt's trivial for someone updating espeak later to just rename the nw version as well
20:21:17n1s65MHz > 45MHz
20:21:46*amiconn should probably mass-recharge his targets
20:21:55amiconnAnd update, of course
20:22:39n1swodz: yeah, i'll try to update it for the tragets i can when i feel i'm done :)
20:22:41 Quit Jerom (Quit: Leaving.)
20:23:09*n1s has one more thing that helps fast iram slow dram targets he wants to do
20:23:30***Saving seen data "./dancer.seen"
20:23:32 Quit GermanMushroom (Read error: Connection reset by peer)
20:23:33n1sand one or two more odd tweaks
20:23:54 Quit Buschel (Ping timeout: 245 seconds)
20:26:03amiconnDo you have a PP5020 target? (not 502x with x>0)
20:26:19n1sno, my only pp is 5024
20:26:26n1sc200v1
20:26:51n1sor was that 5022?
20:27:50wodzespeak compiles cleanly with gcc (not g++)
20:28:26Tornegcc ignores you
20:28:31Zagorwodz: try parameter -x c to ensure c not c++
20:28:33Tornegcc still compiles it as C++ if it's .cc or .cpp
20:28:46TorneIt just doesn't link libstdc++ for you
20:29:00Torneyeah. -x c will override the extension
20:30:14n1swe configure gcc with −−enable-languages=c does it still do c++?
20:30:37Zagorn1s: I assumed wodz is using the host compiler
20:30:59n1saha
20:31:46amiconnn1s: PP5022 and PP5024 are the same except the latter has an integrated AS3514
20:32:11n1samiconn: right
20:32:12n1sbtw, there's a lot to do if someone wants to help tremor's memory consumption, it basically uses 32 bit types for everything
20:36:19wodzright, with -x c it throws errors
20:37:51Tornethen it's C++ :)
20:39:35wodzfirst error is about S_IFMT not defined so this is header problem rather
21:00
21:06:17 Join Strife1989 [0] (~Strife89@207-144-19-39.cstel.net)
21:09:09 Quit Strife89 (Ping timeout: 256 seconds)
21:12:37 Quit pamaury (Remote host closed the connection)
21:17:06 Join pamaury [0] (~quassel@vit94-1-82-67-248-70.fbx.proxad.net)
21:17:06 Quit pamaury (Changing host)
21:17:06 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
21:18:37wodzok with minor modifications (mainly C type casting) espeak compiles with gcc -x c
21:20:16 Quit MethoS- (Remote host closed the connection)
21:35:03pamauryI have to decide what I do for the fuze+ bootloader: header to load at the right location or bootloader copies itself
21:38:13wodzheader is probably cleaner, self relocation is more common approach in bootloaders
21:39:20kugelheader is easier to implement isn't it?
21:39:50kugelyou can even use load_code() api for this
21:40:22wodzself copying is ~4 asm lines + some changes in linker script
21:40:42wodzI think in terms of work both approaches are equal
21:41:35pamauryno, the header thing is because the final bootloader file uses the SB file format, with instructions like "load <bloc> at xxx", so I need the xxx otherwise I must hardcode it. On the other hand, the self copying code is a bit ugly because I already use linker tricks because the bootloader setups the mmu
21:41:44kugeloh I thought we talk about more complex relocation than copying
21:41:53pamaury(so the header will disappear from the final binary, it's just a temporary thing)
21:44:45pamauryok, let's go for the copy approach
21:45:20 Part ironzorg
21:46:33CIA-14r30584 build result: All green
21:46:59pamaurydoes mov rx, pc puts the real pc value into rx or is the an offset due to pipelining ?
21:47:20TorneARM?
21:47:26Tornethat gives you two instructions later
21:48:02pamauryso if the first instruction is at X and I do mov r1, pc, r1 will be X+8 ?
21:48:11pamauryfor all ARMs ?
21:48:12TorneYes
21:48:15TorneWell
21:48:19TorneX+8 in ARM code
21:48:21TorneX+4 in thumb
21:48:44TorneNot sure about thumb2. I suspect dragons.
21:48:51pamaurythat will be arm :)
21:49:04pamauryso the self-copying code will be more than 4 lines ;)
21:49:12TorneConveniently this gives you the return address, for branches
21:49:23Tornemov lr, pc; bx r0; nop
21:49:31Tornewill return to the nop, from the routine in r0
21:49:44Tornevital on ARMs that are too old to have a blx instruction
21:50:35Torneanyway, yeah, it doesn't matter how long the chip's pipeline actually is
21:50:50Tornethe ARM programmer's model pretends to have a three-stage fetch, execute, memory pipeline, always
21:51:19kugeladd rx, pc, #8 would do the same (if pc was the current ins), so just slightly more convinient
21:51:35kugelnot actually "vital" :)
21:51:49Torneokay, not vital
21:51:56Tornewell, it's less typing. and i call that vital :)
21:52:15pamauryso I will have to do -8 anyway if I want to get the address ?
21:52:16TorneI have wanted to know the address of the instruction two instructions further on more often than the address of the current instruction :)
21:52:28kugelthat's asuming you do asm :)
21:52:35Tornepamaury: yah. just do sub r0, pc, #8 to get the current instruction address.
21:52:49kugelbut most of the time the compiler does that for you
21:52:54TorneIf you stick that as the first instruction in your binary then that just gives you the load location
21:53:00pamauryok thanks
21:53:36Tornekugel: if yo uneed to know the address of where you are right now you are probably doing asm :)
21:54:22kugelexcept in the "mov lr, pc; bx r0" case
21:54:44Tornekugel: right, but *you* don't need to know that, only the compiler does :)
21:58:04 Quit factor (Ping timeout: 245 seconds)
21:59:06pamaurylet's see if it works :)
22:00
22:01:09pamaurynot quite /o\
22:04:11 Quit y4n (Quit: PÆNTS ØLF!)
22:17:18pamaurythe f+ linker script is becoming more and more ugly
22:22:30 Join lovasoa [0] (~lovasoa@78.251.5.255)
22:23:31***Saving seen data "./dancer.seen"
22:30:32ukleinekthe linker script is currently a topic in ARM-Linux land, too. Ugly is that.
22:33:52ukleinek(if you're interested: http://mid.gmane.org/20110919184027.GZ16381@n2100.arm.linux.org.uk)
22:34:20wodzukleinek: Did you managed to recover your archos?
22:34:29 Quit stoffel (Remote host closed the connection)
22:35:09ukleinekwodz: not sure. If I only pack say 1G of data it works fine. If it gets more some corruption occurs.
22:35:38ukleinekI already thought about doing some tests to find the broken blocks and mark them bad in the fs
22:35:55ukleinek(not sure that will work, need to read some docs for that)
22:36:03ukleinekanyhow, bedtime here
22:37:28pamaurymanaged to make it work \o/ As usual, linker script proves is not your friend
22:37:35 Quit Stummi (Quit: Bye!)
22:37:35pamaury*proves that it is
22:39:10 Join Buschel [0] (~chatzilla@p54B66838.dip.t-dialin.net)
22:40:53CIA-14New commit by pamaury (r30587): imx233/fuze+: rework crt0 and linker script to be able to load at any address and self-copy at the right one
22:42:35CIA-14r30587 build result: All green
22:43:06bertrik100 seconds, nice!
22:43:58 Quit wodz (Quit: Leaving)
22:44:13pamaurywodz: the header solution would have been easier to implement :) But at least the self-copy one is more general
22:44:46 Quit mamarley (Remote host closed the connection)
22:46:08*Strife1989 wraps things up and heads home.
22:46:58 Quit Strife1989 (Quit: Vamoose!)
22:51:51 Quit pamaury (Remote host closed the connection)
22:59:21 Join hilbert [0] (~hilbert@adsl-89-217-205-236.adslplus.ch)
23:00
23:00:49 Quit ender` (Quit: Computers are terrific - they beget terror.)
23:01:51 Join factor [0] (~factor@74.197.205.204)
23:02:47 Quit factor (Read error: Connection reset by peer)
23:03:41 Join TheLemonMan [0] (~LemonBoy@ppp-165-44.26-151.libero.it)
23:08:19 Join fml [0] (~chatzilla@manz-590f2c92.pool.mediaWays.net)
23:09:19fmlCan I load a WPS file (by playing it) which is e.g. in the root dir, i.e. not in the wps folder? I can't do that with the latest sim.
23:12:49 Quit bieber (Read error: Connection reset by peer)
23:13:07 Join bieber [0] (~quassel@162-78.97-97.tampabay.res.rr.com)
23:28:18 Part Zagor
23:32:25 Quit fml (Quit: ChatZilla 0.9.87 [Firefox 6.0.2/20110902133214])
23:33:15 Quit dre (*.net *.split)
23:33:15 Quit Farthen (*.net *.split)
23:37:14 Join pamaury [0] (~quassel@vit94-1-82-67-248-70.fbx.proxad.net)
23:37:15 Quit pamaury (Changing host)
23:37:15 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
23:37:47 Join dre [0] (dre@gateway/shell/xzibition.com/x-aagulhyylwfnqgjb)
23:37:47 Join Farthen [0] (~Farthen@2a01:4f8:101:2a4:0:bc28:b2e1:9)
23:40:41 Quit mudd1 (Quit: Ex-Chat)
23:50:15 Quit mystica555_ (Ping timeout: 276 seconds)
23:54:21 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 6.0.2/20110902133214])

Previous day | Next day