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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2012-01-22

00:02:37 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
00:39:23 Quit saratoga (Ping timeout: 258 seconds)
00:39:52 Quit ukleinek (Ping timeout: 272 seconds)
00:40:20 Quit curtism (Quit: Live Long and Prosper)
00:43:55 Join dreamlayers [0] (~bgjenero@rockbox/developer/dreamlayers)
00:51:29 Join ukleinek [0] (~ukl@2001:6f8:1178:2:21e:67ff:fe11:9c5c)
00:51:56 Quit pamaury (Remote host closed the connection)
00:55:47 Quit lebellium (Quit: ChatZilla 0.9.88 [Firefox 10.0/20120118081945])
01:06:50***Saving seen data "./dancer.seen"
01:07:18 Join Amqui [0] (
01:12:17 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
01:30:15 Part dreamlayers
01:32:43 Join curtism [0] (
01:38:28 Join Zarggg [0] (
01:43:06 Join PaulJam [0] (
01:56:56 Quit yosafbridge (Quit: ERC Version 5.3 (IRC client for Emacs))
01:57:13 Join yosafbridge [0] (
02:14:58 Quit PaulJam (Ping timeout: 252 seconds)
02:15:28 Join beslayed [0] (
02:25:15 Quit beslayed (Read error: Connection reset by peer)
02:31:28 Quit bertrik (Ping timeout: 252 seconds)
03:06:51***Saving seen data "./dancer.seen"
03:08:55 Quit AlexP (Remote host closed the connection)
03:19:49 Join CaptainKewl [0] (
03:28:23 Join beslayed [0] (~user@
04:05:24 Quit Slasheri (Read error: Operation timed out)
04:07:53 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
04:17:01 Join [Saint_] [0] (~Saint]@
04:17:01 Quit [Saint_] (Changing host)
04:17:01 Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
04:18:45 Quit [Saint] (Ping timeout: 252 seconds)
04:37:30 Quit Topy (Read error: Connection reset by peer)
04:37:34 Join T44 [0] (
04:45:01 Quit beslayed (Remote host closed the connection)
04:49:30 Join dys` [0] (
04:53:23 Quit dys (Ping timeout: 252 seconds)
04:53:35 Quit [7] (Disconnected by services)
04:53:51 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:54:44 Quit amiconn (Disconnected by services)
04:54:44 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:54:55 Quit pixelma (Disconnected by services)
04:54:58 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:55:00 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:55:06 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:55:22 Join JdGord [0] (~AndChat@
04:59:15 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
04:59:24 Quit JdGord (Read error: Connection reset by peer)
04:59:42 Join JdGord [0] (~AndChat@
05:00:00 Quit anewuser (Read error: Connection reset by peer)
05:05:01 Join Rob2222 [0] (
05:06:52***Saving seen data "./dancer.seen"
05:09:27 Quit Rob2223 (Ping timeout: 260 seconds)
05:20:05 Nick dys` is now known as dys (
05:45:35 Quit [Saint] (Ping timeout: 240 seconds)
05:46:01 Join [Saint__] [0] (~Saint]@
05:46:01 Quit [Saint__] (Changing host)
05:46:01 Join [Saint__] [0] (~Saint]@unaffiliated/saint/x-8516940)
05:47:52 Quit [Saint_] (Ping timeout: 252 seconds)
06:00:45 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
06:15:43 Join perrikwp [0] (
06:43:56 Join froggyman [0] (~froggyman@unaffiliated/froggyman)
06:55:29 Quit t0rc (Quit: WeeChat 0.3.6)
07:00:03 Quit [Saint] (Remote host closed the connection)
07:02:49nosa-jwhat do i look for to customize the usb tranfer screen?
07:02:55nosa-jjust wondering
07:03:20nosa-ji can find the images for the rest of the theme but i can't locate that
07:04:56 Quit JdGord (Quit: Bye)
07:06:55***Saving seen data "./dancer.seen"
07:16:29 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
07:22:01 Join y4n [0] (y4n@unaffiliated/y4ndexx)
08:03:05 Join leavittx [0] (~leavittx@
08:09:16 Join funman [0] (~fun@rockbox/developer/funman)
08:11:29 Quit funman (Client Quit)
08:11:47 Join funman [0] (~fun@rockbox/developer/funman)
08:22:33 Join stoffel [0] (
08:23:03 Quit leavittx (Ping timeout: 245 seconds)
08:23:23 Join robin0800 [0] (~robin0800@
08:30:29 Quit CaptainKewl (Ping timeout: 248 seconds)
08:39:34 Quit robin0800 (Quit: Leaving)
09:02:35 Join bertrik [0] (
09:02:35 Quit bertrik (Changing host)
09:02:35 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
09:05:10 Join kevku [0] (x@2001:470:28:773::)
09:06:58***Saving seen data "./dancer.seen"
09:22:17 Join robin0800 [0] (~robin0800@
09:23:37 Quit Keripo (Quit: Leaving.)
09:29:10 Quit KiwiCam (Quit: leaving)
09:29:24 Join KiwiCam [0] (
09:30:06 Quit robin0800 (Ping timeout: 252 seconds)
09:47:57 Join robin0800 [0] (~robin0800@
09:53:32 Quit robin0800 (Ping timeout: 260 seconds)
09:57:40 Quit KiwiCam (Quit: leaving)
09:58:20 Join KiwiCam [0] (
10:05:07 Join Amqui1 [0] (
10:06:15 Quit Amqui (Ping timeout: 240 seconds)
10:10:36 Quit stoffel (Remote host closed the connection)
10:13:46 Join pamaury [0] (
10:13:47 Quit pamaury (Changing host)
10:13:47 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:40:31 Quit factor (Read error: Connection reset by peer)
10:58:01 Join factor [0] (~factor@
10:58:46 Join stoffel [0] (
10:58:58 Join wodz [0] (
11:02:25wodzdesowin: speaking about TCP/IP stack for rockbox - lwIP is probably the only option and should be fairly easy to port. Much more work would be to implement something useful with it. And moreover wifi part is quite complex I recall.
11:02:54JdGordon[Saint]: did you test the patch?
11:07:02***Saving seen data "./dancer.seen"
11:12:40funmanwodz: for which device ?
11:13:22wodzfunman: sansa connect
11:13:36 Join TomColler [0] (
11:14:15wodzanyway it would be quite fun to have some web page served by http server running on rockboxed dap :P
11:14:56bertrikwhat's next, port java?
11:16:06[Saint]I was thinking of all sorts of fun things, like playlist recommendations, opening up your device so that "friends" can suggest tracks and control playback, streaming for
11:16:27[Saint]*streaming from "friends" connected devices...
11:18:58[Saint]"Fun Stuff™"
11:20:22[Saint]The device itself needn't have web capability. Just be plugged into something that does. #dreaming
11:20:28 Join robin0800 [0] (~robin0800@
11:21:25 Quit stoffel (Ping timeout: 255 seconds)
11:30:40 Join PaulJam [0] (
11:31:30 Quit wodz (Quit: Leaving)
11:38:43 Quit bluebrother^ (Ping timeout: 252 seconds)
11:43:43*[Saint] is getting topic reply emails from threads he has *never posted on*.
11:43:57[Saint]W. T. F?
11:48:03 Quit fs-bluebot (Ping timeout: 240 seconds)
11:52:12 Join Bretos [0] (
11:55:48 Join lebellium [0] (
11:59:04 Quit robin0800 (Quit: Leaving)
12:10:59 Quit pamaury (Remote host closed the connection)
12:12:16 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
12:15:38[Saint]Bretos: need to wait/ask for approval. If you have a (Rockbox related) question, just ask away...if someone can answer, they'll do so.
12:19:04[Saint]If you're looking for a social channel however, #rockbox-community may be more suited, though I can't promise any stimulating conversation :)
12:22:03BretosPassion for knowledge isn't the reason I came here. Thanks for information about #rockbox-community channel, I though it is the only one rockbox channel.. idk why ;) thanks again, and bye
12:22:36 Quit Bretos (Quit: Leaving)
12:27:44 Nick [Saint__] is now known as Guest63240 (~Saint]@unaffiliated/saint/x-8516940)
12:28:40 Quit Guest63240 (Disconnected by services)
12:29:08 Join Guest63240 [0] (~Saint]@
12:32:07 Quit Guest63240 (Client Quit)
12:44:49 Join TheLemonMan [0] (
12:47:35 Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
12:53:54 Quit [Saint_] (Quit: Quit)
13:07:06***Saving seen data "./dancer.seen"
13:24:32 Quit TBCOOL (Quit: maintenance..)
13:29:30 Join anewuser [0] (~anewuser@
13:29:30 Quit anewuser (Changing host)
13:29:30 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
13:31:35 Join fyre^OS [0] (
13:34:14 Quit fyrestorm (Ping timeout: 252 seconds)
13:52:30 Join leavittx [0] (~leavittx@
13:55:01gevaertsnosa-j: the USB screen can't be themed. We'd like it to be but the fact thet there can be absolutely no disk access from the rockbox side makes it tricky
14:19:15pamauryI would like other developpers opinion. I would like to handle the imx233 and the stmp3770 with the same codebase since there are pretty much the same except for a few things here and there. What would be the best option ? use imx233 and #define HAVE_STMP3770 ?
14:22:19pamauryor have a #define IMX233_VARIANT with two possible values: VARIANT_IMX233 and VARIANT_STMP3770 ?
14:24:29 Join bluebrother [0] (
14:24:29 Quit bluebrother (Changing host)
14:24:29 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
14:24:36gevaertspamaury: that sounds similar to PP5020/PP5022 I think
14:24:46pamauryhow is it handled ?
14:24:52 Join fs-bluebot [0] (
14:25:06 Join stripwax [0] (
14:25:10gevaertsCan I say I can't remember? ;)
14:25:52gevaertsAh, yes
14:26:06gevaertspamaury: see config.h around line 534
14:26:31gevaertsNot sure if that really helps
14:27:00 Quit stripwax (Client Quit)
14:27:11pamauryhum, not quite because lots of things rely on CONFIG_CPU=IMX233
14:27:47pamauryI could introduce this CPU_STMP3770 things perhaps
14:31:14 Join TBCOOL [0] (
14:36:33 Quit leavittx (Ping timeout: 252 seconds)
14:38:46 Join robin0800 [0] (~robin0800@
14:42:29 Join dfkt [0] (
14:42:31 Quit dfkt (Changing host)
14:42:31 Join dfkt [0] (dfkt@unaffiliated/dfkt)
14:44:46pamauryhaha, I got the sd working on the creative zen style m300
14:49:28pamaurygevaerts: what about a #define CONFIG_CPU_SUB or CONFIG_SUB_CPU or CONFIG_CPU_VARIANT ?
14:52:40gevaertsThat could work. I'm not sure if it's a good idea to do that for one case without at least having plans for doing it elsewhere though. On the other hand, this is a bit of an area where I'm not too involved
14:53:42pamauryyeah, that's true. Then perhaps a HAVE_STMP3700 will do
14:53:50 Quit TheLemonMan (Quit: WeeChat 0.3.6)
14:54:03 Join TheLemonMan [0] (
14:54:11 Join {phoenix} [0] (
14:55:43 Quit {phoenix} (Remote host closed the connection)
14:57:29 Quit TheLemonMan (Client Quit)
14:57:46 Join TheLemonMan [0] (
14:58:22 Quit TheLemonMan (Client Quit)
14:58:36 Join TheLemonMan [0] (
14:58:58 Quit maraz (Quit: brb, fbi.)
14:59:13bertrikpamaury, congrats on the sd
15:01:35bertrikthe start of a new port, getting things to work for the first time, is usually the most fun (at least for me)
15:02:31pamaurythe biggest problem to me is that the creative zen style m300 does not seem to have external memory, only 512Kb of internal sram
15:03:20bertrikdo we have PCB scans of it somewhere?
15:04:14pamauryyes, I have photos of it, only of side of the PCB though, the other one is hardly reachable but from what I saw there is not chip. See the wiki: akismet
15:04:32pamauryI made some nice boxes :)
15:05:49 Join maraz [0] (
15:06:07pamauryi might be wrong of course but the OF doesn't initialize the emi from what I saw and it has a very weird architecture. I suspect some parts of the code work like codec, they are loaded on demand
15:07:08***Saving seen data "./dancer.seen"
15:08:05bertrikhm, 512 kB isn't nearly enough to run rockbox on it, I guess. What part do you have working now, lcd and sd?
15:10:11pamauryonly sd yet because I have a generic "recovery util" that I can load and poke at the ssp, I've not yet really started to look at the OF code but I'm really embarassed because the OF file has some "data section" that actually contain code and I'm not sure how it works
15:10:46pamaurythere is still a slight chance that I missed the dram on the board and haven't found the emi init yet
15:11:53bertrikdoesn't the clip v1 have 2 MB of memory inside the SoC package?
15:23:18kugelpamaury: stmp37xx is a soc similar to imx233?
15:23:31pamauryyes, it's an older version, more or less
15:24:28TheLemonManpamaury: i have some bsp code adapted to the stmp37xx register layout, if you're interested in it just whistle :)
15:24:34kugelah so it's comparable to PP indeed
15:24:52pamauryTheLemonMan: I produced a pdf file of the difference between imx233 and stmp37xx
15:25:43TheLemonManyep, i could spare you some time adapting the code
15:26:03TheLemonMandont have enough free time to actively develop the port :( at least till summer
15:26:41gevaertsscorche: if I make more forum changes, do you prefer those as patches on top of the previous one, or on top of pristine akismet/smf?
15:28:46 Join n1s [0] (~n1s@rockbox/developer/n1s)
15:32:12pamaurywhat I'm afraid of is that the OF could use paging and page fault handling to simulate a larger memory using the sd as a backend, that would make the OF analysis pretty hard
15:34:02pamauryTheLemonMan: for the pdf file about the differences
15:36:37 Quit Sundiver (Ping timeout: 240 seconds)
15:38:32 Join Sundiver [0] (
15:47:10kugelfunman: did you have a look?
15:51:27 Join lorenzo92 [0] (
15:52:49 Quit robin0800 (Ping timeout: 252 seconds)
15:52:53lorenzo92kugel: hey :) something new regarding R0? ...i'm sorry I cannot help atm cause I'm very busy hehe
15:53:08kugelI'm close to enabling asm
15:53:45kugellorenzo92: I noticed something strange. sometimes it has 100% cpu usage, most of it spend in the kernel (sys time also 100%
15:54:17lorenzo92kugel: yep. This is due to bad filesystem... -> do a fsck.vfat and you will see that it is solved!
15:54:25lorenzo92kugel: anyway, great as always :D
15:59:10lorenzo92kugel: I was thinking, as a workaround we could write a script that, let's say after 10 reboots, executes a fsck.vfat directly on device
15:59:37lorenzo92well I should check if that really works and doesn't block the device for too much time :)
16:01:50[Saint]That's a bit of a hack...
16:02:29[Saint]Better to find out why the fs is getting trashed, no?
16:03:53lorenzo92indeed...The thing is that behind there's the RFS (for which I found its sources, tough lol)
16:04:20gevaertshm, interesting bug
16:08:42gevaertsscorche: if I create a new topic, and then a spammer replies to that, and akismet intercepts that, and then a mod clears the akismet queue, that topic is no longer listed
16:09:42gevaertsIt still exists, and it's still listed as the most recently modified topic in the board list (well, if it is the most recently modified topic), but it's not listed in the list of topics...
16:09:58*gevaerts is slightly worried by this
16:10:51gevaertsThe most recent post page also still has it
16:26:04 Quit lorenzo92 (Quit: ChatZilla 0.9.88 [Firefox 9.0.1/20111228084940])
16:26:24 Join lorenzo92 [0] (
16:32:39 Quit kevku (Quit: KVIrc 4.0.4 Insomnia
16:33:13gevaertsscorche: I think is worth applying. It's based on clean smf/akismet, and adds (a) hiding spam from recent posts (same as the previous patch), (b) links to user profiles from the akismet spam list (so we can look at sigs or ban easily), (c) deleting spam from individual users from the akismet spam list (which should make dealing with false positives a lot easier), and (d) it doesn't set the last message information on a
16:33:25gevaertstopic if the last message is spam (which should fix various issues with things looking weird)
16:34:21gevaertsI'd still like to add a way to report spam to akismet if it missed it, but I'm not sure where that should be shown
16:50:04 Join CaptainKewl [0] (
17:04:16 Quit advcomp2019 (Read error: Connection reset by peer)
17:05:11 Join advcomp2019 [0] (
17:05:11 Quit advcomp2019 (Changing host)
17:05:11 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
17:07:11***Saving seen data "./dancer.seen"
17:13:29 Quit TheLemonMan (Quit: WeeChat 0.3.6)
17:13:32 Quit lorenzo92 (Quit: ChatZilla 0.9.88 [Firefox 9.0.1/20111228084940])
17:23:44 Quit factor (Ping timeout: 252 seconds)
17:24:10 Join ToxAtec [0] (
17:25:36gevaertsscorche: I also have an sql line that should fix the missing topics that are caused by the bug I mentioned
17:26:02gevaertsIt basically fixes the last message id for topics where the last message doesn't exist
17:26:15gevaertsUPDATE smf_topics set ID_LAST_MSG = (select ID_MSG from smf_messages where smf_messages.ID_TOPIC=smf_topics.ID_TOPIC order by posterTime desc limit 1) WHERE not exists(select * from smf_messages missing_m where smf_topics.ID_LAST_MSG = missing_m.ID_MSG);
17:26:24gevaerts(assuming smf_ is the prefix used)
17:29:29*gevaerts wonders how many topics are lost this way...
17:30:20ToxAtechi, I'm wondering if it's possible to disable charging a sansa fuze v1 on USB connections, or if rockbox has no controll over charging the battery. I saw that there was an option like that on other players (some ipod).
17:30:47Torneif the option isn't there then it's not implemented; whether it's actually physically possible, I don't know
17:30:55Tornebut if so, nobody has written the code for it
17:31:28 Quit SynrG (Ping timeout: 252 seconds)
17:32:19bertrikWe can control charging on the PP, AMS, AMSv2 sansas (enable/disable charger, configure charge voltage and current)
17:32:20 Join TheLemonMan [0] (
17:33:06TorneToxAtec: also note that turning usb charging off doesn't reduce the power used to zero
17:33:07bertrikbut I don't see why you would want to disable charging
17:33:39Tornebertrik: generally, to avoid draining the battery of a laptop that's only going to last a few hours in order to recharge an mp3 player that will last all day :)
17:34:38ToxAtecI thought it might be better for the battery health if you don't start charging it everytime you add a new song; also I found out about the USB HID fuction which would be a perfect scroll wheel for my laptop
17:35:09Tornethe effect on the battery will be probbaly unmeasurable no matter what you do with it
17:36:46bertrikcharge current for a sansa is probably only a fraction of what the laptop itself already uses
17:37:38Tornebertrik: the other reason is to work with broken usb hosts/hubs that can't deliver the power in the first place
17:37:53ToxAtectorne, did you mean the effect on the laptop or fuze battery?
17:37:53bertrikand lithium batteries don't mind if you recharge them often while not empty yet
17:37:59TorneToxAtec: on the fuze battery
17:38:23TorneWell, also the laptop :)
17:39:26Tornebut it depends how you are using the laptop; it'll suck up to 2.5W from the host, which is not a lot compared to a laptop's display, but is possibly measurable if the screen's off :)
17:40:48Tornebertrik: anyway, the option is there because it's easy to do if you already hve the underlying system for charging in a USB-spec-compliant way
17:40:54gevaertsWell, it will only do that if it actually needs to charge
17:41:02Torneif we have the required level of control over the hardware on AMS then we should hook it up
17:41:28Tornein theory the code should be target-neutral after i refactored the beast/ipod support for it
17:41:40bertrikI'm still confused about the need for this. Never heard of anyone having a real problem with the current situation.
17:42:03bertrikyes, it could be done in some generic charging state machine I guess
17:42:10Tornewe already have the generic charging state machine
17:42:19gevaertsbertrik: spec compliance and having all targets behave the same way
17:42:33TorneYou just need a #define to enable it, and implement the underlying control functions in target code
17:42:40Tornethe PP ipods and the gigabeat S have this already
17:42:44Torneand the nano2g i think
17:42:54Tornei wrote this at the devcon in 2010 :)
17:43:14Torneand tested on every PP ipod model with a multimeter, whcih was fun :p
17:43:59 Join lazy [0] (
17:44:04 Nick lazy is now known as lazyguy (
17:45:48lazyguyin the targetstatus, clip+'s usb functionality says yes, but under release 3.1 it's says it's still disabled
17:46:17gevaerts3.1? The clip+ didn't even exist back then...
17:46:35gevaertsIf you mean 3.10, USB was enabled after that
17:46:49lazyguyya 3.10
17:47:06lazyguyso do i need the current build
17:47:12lazyguyor 3.10 is fine
17:47:48gevaertsWell, if you really want USB in rockbox, you need a current build, yes. If you're happy with rebooting to the original firmware, 3.10 is fine
17:48:02lazyguyok thanks
17:48:21gevaertsNote that USB in rockbox might still be a bit unstable. It works well for many people, but it's still relatively new code that hasn't seen very wide testing
17:48:29bertrikTorne, gevaerts, there is a simple state machine to control charging for the as3514/AMS sansas, but it looks like it's not hooked into the new USB/charger detection logic yet
17:49:19lazyguyguess i'll wait then
17:49:27lazyguyi;ve been waiting for years now anyways
17:50:02gevaertsWell, there is of course also the fact that if everybody waits, we'll never know if it works well :)
17:50:06Tornebertrik: well, if you have a go i will take a look at your patch :)
17:50:33bertrikTorne, the stuff you wrote does the 100 mA / 500 mA / "unlimited" detection, based on USB attachment status, charger detection, etc, right?
17:51:23Torneit only ever actually selects 100 or 500, currently
17:51:24lazyguyok, updating now
17:51:36Tornebut if possible the underlying code should handle 0 and unlimited as well
17:51:37lazyguyguess i'll just tell you if it works or not
17:51:39Tornefor future use
17:51:59TorneWe don't currently have any charger detection code
17:52:13Tornei've not gotten around to doing it for ipodvideo yet, and i don't know how to do it on any other devices :/
17:52:21lazyguyalso, i have some superold bootloader, that didnt change or anything right?
17:52:29kugelrockbox on webos seems like a no-go
17:52:48kugelsdl on it only supports sw surface and that's way to slow
17:53:44bertrikI could possibly help with a nano 1g, but IIRC last time I trie I came to the conclusion we can't measure the special resistor networks in the nano 1g
17:54:09kugelCaptainKewl: ping
17:54:11amiconnDetecting classic charger on ipods is simple is?
17:54:31amiconnThe 12V chargers are essentially firewire power
17:54:46TorneOh. That's irrelevant
17:54:50Tornethis is for USB charging only
17:54:53amiconn..which is supported by all PP ipods, even those which don't support firewire data
17:54:56Torneby "classic" there you mean "ancient"
17:54:59Tornei think :)
17:55:05Torneall ipod chargers i have or have seen are 5V
17:55:29CaptainKewlkugel hi
17:55:52kugelCaptainKewl: sdl on webos is too slow to be useful at all
17:58:07gevaertskugel: is sdl the only option?
17:58:51CaptainKewlsoftware, yeah. I got it to run faster by using opengles and rendering the surface to a texture and displaying that instead
17:59:31kugelshow me!
17:59:48kugelI was always interested in us doing opengles :)
17:59:58CaptainKewlhaven't checked the fps but it runs in hybrid without taking 2 seconds to render each frame
18:00:15kugelI can check the fps if you give me the code ;)
18:00:30CaptainKewlI'll see if I can post it up in a bit, will need to reboot
18:00:54 Quit CaptainKewl (Quit: ( :: NoNameScript 4.22 :: ))
18:02:18kugelgevaerts: sdl and opengles, I think. Perhaps fbdev but haven't tried
18:02:37kugelopengles would be interesting for android too methinks
18:03:02gevaertsWorth looking at, anyway
18:03:12gevaertsI suspect fbdev won't be an option on webos
18:03:15kugelthe docs say with opengles you're limited to landscape, though, unless you roate in-app which we can't
18:04:13kugelwell, we actually can as mpegplayer shows
18:04:23amiconnTorne: I thought the charging state machine would handle all potential power sources
18:04:45gevaertskugel: you can rotate within opengl
18:04:52Torneamiconn: no; i'm not aware of a need for it to do so
18:05:02lazyguyare there special drivers for clip+
18:05:05Torneit's specifically about telling the usb interface what to do
18:05:18lazyguyif you load from rockbox
18:05:25kugelgevaerts: the opengles implementation on the touchpad can'T :)
18:05:30Tornecharging direct from a DC port, or from firewire, doesn't come with any special behaviour or rules that i know of :)
18:05:31kugelaccording to anyway
18:06:50funmankugel: not yet
18:07:38gevaertskugel: that sounds like something different. The main viewport will probably always be 1024x768, but what you draw in there is up to you. You can render to a 768x1024 texture, and then rotate that texture before putting it on a surface
18:07:47Torneamiconn: this is not about the actual charging thread/etc, the state machien that runs software controlled charging. this is specifically inside the USB stack, to handle the power requirements in the USB spec
18:07:51kugeli will just commit then. it's mostly an infrastructure change
18:08:03Tornedevices that have software-driven charging still have an independant system to actually run that
18:08:26kugelgevaerts: well sure, that's what I meant with "rotate in-app"
18:08:59gevaertskugel: well, it's two lines of code :)
18:09:10lazyguyokay, i unplugged usb and now rockbox is stuck on the usb screen
18:09:22gevaertsWhich are in the opengl/es specific bit
18:10:02kugelwell yes, but you need to translate touch events also
18:10:07Torneamiconn: the bit that's interesting is the USB stack telling the power managemen system whether it's okay to draw power from the USB +5V pin. if you have some other source of power then that's not the USB stack's problem ;p
18:10:42 Join factor [0] (~factor@
18:11:22gevaertsright. That shouldn't be too hard though
18:11:52gevaertslazyguy: well, at least now you know why that hasn't been enabled in a stable release yet...
18:11:58 Quit linuxguy3 (Ping timeout: 252 seconds)
18:12:03lazyguyno, i got it working now
18:13:38lazyguyis there any way to disable usb functionality while plugged in, i like the option of charging and listening to music
18:15:18gevaertsyes, hold a button while plugging in.
18:15:51gevaertsSelect is probably a good choice. That will bring up the context menu, but that's easy to get back out of, and doesn't do anyhing
18:17:24dfkthow do i update my local patched tree with git? git pull −−rebase refuses to update, it says "refusing to pull with rebase: your working tree is not up-to-date"
18:17:41bertrikwhere (in the source) is the code for determining if it's okay to draw power from USB?
18:18:28bertrikand is it a poll mechanism (asking USB), or does USB just tell someone else what is allowed?
18:18:54Torneusb_charging_maxcurrent tells you how much you can have right now
18:19:05Torneusb_charging_maxcurrent_change gets called when it changes
18:19:16Torneit's in usb_core.c, under HAVE_USB_CHARGING_ENABLE
18:20:32lazyguycool thanks it worked
18:20:46scorchegevaerts: email me if you could and i can get to it when i get back
18:20:57gevaertsscorche: ok
18:21:27lazyguyi got a touchpad you guys need help testing
18:21:33lazyguyif you're still talking about that
18:21:36scorchegevaerts: please roll all the changes into one - i am going to try to make a SMF package from it - which i think means i need the entire file of anything you have modified
18:21:38Tornebertrik: calling the usb stack to ask is cheap, so if the target already has a software charging thread polling for stuff to prod the charge controller, you can just add the call in the polling loop.
18:21:59Tornebertrik: if it's like the ipod where you just poke a hardware register to tell it what the limit is, you can just implement the on-change function
18:22:10Tornedid it this way to accomodate both
18:22:17scorchegevaerts: though, might as well include a diff just so i can double-check =)
18:22:26 Join t0rc [0] (
18:22:31 Quit t0rc (Changing host)
18:22:32 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
18:22:38gevaertsscorche: a complete diff is easiest for me right now :)
18:22:58scorchegevaerts: with any file you have touched =)
18:23:07gevaertsyes, that's the one
18:23:56scorchewell, i mean, i need an email with blah.diff, blah.php, blah1.php, blah2.php, etc
18:24:08gevaertsAh, I see
18:24:27gevaertsok, I'll do my best
18:25:07 Join captainkwel [0] (
18:29:13bertrikTorne, and what if the sansa is connected to a charger, should I detect that myself, and ignore the max current advice from the USB stack?
18:29:38Tornehandle it however you handle it now
18:29:45Torneexcept presumably now you assume USB means 500mA
18:30:04*bertrik isn't actually sure we can distinguish between USB power and charger power
18:30:16Torneif you can't, then assume it's USB
18:30:21bertrikwe're charging with only 300 mA max anyway
18:30:29Torneyes, but 300 is more than 100
18:30:35Tornewhich is all you are allowed to draw without negotiation
18:30:42 Quit lazyguy (Quit: ChatZilla 0.9.88 [Firefox 9.0.1/20111220165912])
18:31:35Torneif the charger provides power on +5VUSB, like most do, and you have no known way of telling the difference, then you are in the same position as the ipods are. if the user picks "force" it will ignore the lack of positive charger detection and allow high current charging after a host doesn't show up within a reasonable time (the code does this for you)
18:31:39bertrikand USB draw = player consumption + charger current, so we better turn the charger off completely when in 100 mA mode
18:32:02Torneif you can power the device without charging it, then sure.
18:32:20Torneon the ipods, the regulation is global, so if you limit the usb power to 0mA and the ipod's battery is dead, it will just die :)
18:32:40Tornewhich is.. not ideal. which is why the logic doesn't do that yet
18:32:51Torneuntil someone decides what a senisble way to handle taht is :0
18:32:58kugelcaptainkwel: :)
18:33:31 Quit dv_ (Remote host closed the connection)
18:34:03captainkwelIt's coming... I haven't looked at this stuff in a week or so. Needs a bit of cleaning.
18:34:49kugelcaptainkwel: great, but don't think you need to clean up just for me
18:36:19amiconnTorne: Do we handle usb suspend?
18:36:36Torneamiconn: no.
18:36:49Tornenot yet, anyway
18:36:51Tornefeel free to add it :)
18:39:33 Join dv_ [0] (
18:49:50 Join leavittx [0] (~leavittx@
18:51:24 Join webguest61 [0] (
18:51:58CIA-81Commit 3c17f28 in rockbox by 03Thomas Martitz: Move pcm_mixer helper routines to firmware/asm.
18:52:00CIA-81Commit 991ae1e in rockbox by 03Thomas Martitz: Create fimrware/asm directory for assembly optimized stuff.
18:52:04CIA-81Commit 9e7b417 in rockbox by 03Thomas Martitz: firmware/target/arm doesn't contain code for hosted anymore.
18:52:04CIA-81Commit 91a3978 in rockbox by 03Thomas Martitz: ypr0: Enable asm optimizations.
18:52:07 Quit t0rc (Quit: WeeChat 0.3.6)
18:52:11CIA-81Commit a5f7f6e in rockbox by 03Thomas Martitz: Use ARCH instead of CPU for selecting optimization levels.
18:52:16kugelheh, the order isn't quite right
18:52:44CIA-81Commit 094cbd5 in rockbox by 03Thomas Martitz: Implement a C version lcd-as-memframe.c and move it and the asm to firmware/asm.
18:52:46webguest61May I ask a question related to ATA error 11 on trying to mod my IRiver H120 to SD card?
18:52:46CIA-81Commit 8e8e978 in rockbox by 03Thomas Martitz: Add framework to let make automatically pick optimized asm implementations over generic C ones to firmware.
18:53:28 Join stoffel [0] (
18:53:56kugelwhat happened to git rename tracking!?
18:54:25kugelman, all the history lost
18:54:36CIA-81Commit a035261 in rockbox by 03Thomas Martitz: Move optimized memcpy and friends and strlen to firmware/asm,
18:54:48kugelI did take care that git has the renames right when doing the local commits
18:54:50gevaertswebguest61: sure. Don't ask to ask, just ask. Keep in mind that people currently online may not know the answer though
18:54:56CIA-81Commit 5e9b62c in rockbox by 03Thomas Martitz: ypr0: Use generic lcd memframe driver.
18:55:17CIA-8191a3978 build result: 292 errors, 28461 warnings (Thomas Martitz committed)
18:55:19CIA-81Commit c94cf3b in rockbox by 03Thomas Martitz: Move optimized ffs to firmware/asm, using the new automatic-asm-picking infrastructure.
18:55:26gevaertskugel: I do see lots of renames when pulling
18:55:38kugelah good
18:55:47gevaerts"firmware/{target/sh/memset-sh.S => asm/sh/memset.S}" and similar
18:55:57kugelthe gitweb ui doesnt show it though
18:56:51kugelsame for git log
18:56:58 Quit Poodlemastah (Ping timeout: 244 seconds)
18:57:06webguest61I have follwed the documents procedure for converting iriver h120 to CF/SD.....Everything works fine with the new bootloader bootloader_h120_7pre4 with the HDD in place. When I put in the CF to SD adapter and try to boot, RAM, ROM, Disk all produce "ATA Error 11"
18:57:09bluebrothertry git log −−follow <file> :)
18:57:13webguest61What am I missing or done wrong?
18:57:22webguest61I've tried 2 different SD crds
18:57:39kugelbluebrother: ah thanks! why is that not default?
18:57:49bluebrotherkugel: don't ask me, ask the git guys :)
18:58:04bluebrotherperhaps because it only works for single files?
18:58:34bluebrother(svn stops on renames by default as well btw)
19:01:43ToxAtecthis USB charging discussion is interesting, although I guess I would not be able to implement anything or so :P but what I think I understood is that controlling USB loading on a sansa AMS is possible but "it's not hooked into the new USB/charger detection logic yet"? So you need some function to tell the.. state machine (?) what has been detected on USB, like 100mA or 500mA?
19:01:57kugelcaptainkwel: sorry, I'm impatient. can't you just show me the unclean code? ;)
19:01:58webguest61I am using the IDE - CF adpater named in posts to work fine with the IRiver, and the same for the CF to SD adapter......Transcend 4GB 133x SD card & PNY 32 GB SD card both wee tested and produce the sme result
19:03:03webguest61from what I've read, this should work, but yet it doesn't
19:03:25 Join Poodlemastah [0] (
19:06:18webguest61any thoughts?
19:06:24 Quit stoffel (Ping timeout: 276 seconds)
19:06:35kugelcaptainkwel: awesome
19:06:47kugelcaptainkwel: can you tell me more about the hybrid mode?
19:07:14***Saving seen data "./dancer.seen"
19:07:53captainkwelUses the HTML-based views for the main application and renders the SDL inside of a window.
19:07:55kugelthere's a sepearate resolution for it?
19:09:20captainkwelYeah... at some point I'd like to figure out some way to maybe get it to stretch to fit get the input to match.
19:09:59captainkwelAlso rescale on the fly when changing screen orientation. Having some trouble getting the video reinitialize to work though.
19:11:56captainkwelMy understanding is if you do a straight SDL-only app, that pretty much just exists by itself and doesn't really interact with the rest of the OS in any meaningful way.
19:14:37*kugel doesnt understand the opengl code there :)
19:16:01captainkwelDon't ask me to explain it, it's mostly copied and pasted. :)
19:16:38kugelwhere did you get it from?
19:20:00captainkwelInternets. I think this was most helpful:
19:23:07webguest61thanks anyway for tolerating my intrusion.....bye
19:23:26 Quit webguest61 (Quit: CGI:IRC)
19:24:09 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
19:26:45CIA-81Commit 3520d7f in rockbox by 03Thomas Martitz: Fix native arm builds. arm_support couldn't build properly.
19:28:02 Join SynrG [0] (
19:29:14CIA-813520d7f build result: 132 errors, 30492 warnings (Thomas Martitz committed)
19:38:01gevaertscaptainkwel: as another opengl newbie, I can't help wondering why you use GL_TRIANGLE_STRIP and not GL_QUADS
19:39:36CIA-81Commit ec36357 in rockbox by 03Thomas Martitz: Fix sdl build warnings
19:41:52jhMikeS?? It still didn't build properly :)
19:42:12CIA-81ec36357 build result: 127 errors, 0 warnings (Thomas Martitz committed)
19:42:47captainkwelgevaerts, looks like the only available primitive in ES is the triangle...?
19:43:03gevaertshm, yes, that would explain it
19:44:22nosa-jgevaerts: about the usb screen i suppose it has to do with the flash memeory which is hard to 'patch' and can ware out that memory bit if flashed too much, thanks for your help :P
19:44:57nosa-jas to why it's not 'theamable'
19:45:28gevaertsnosa-j: no, it's because if you touch the disk during USB, you risk filesystem corruption. Either the host OS has control, or rockbox, but not both
19:46:19gevaertsThe clean way around this would be MTP :)
19:47:06CIA-81Commit 9f4ee79 in rockbox by 03Thomas Martitz: Fix checkwps.
19:48:28gevaertsIt can be themed in theory, but we'd have to be *very* careful to make sure everything needed is in RAM before handing over control to the host OS. We've been bitten by this before, which is why the USB screen now uses the hardcoded system font and not the user/theme specified UI font
19:49:24CIA-819f4ee79 build result: 20 errors, 0 warnings (Thomas Martitz committed)
19:52:48kugelgevaerts: it didn't fix the 5 errors. they just happened on a different build because they only happen on my client!
19:53:00jhMikeSgevaerts: that should be possible if it waits to ack, or is there something with hitting the disk for glyphs in certain cases?
19:53:56gevaertsjhMikeS: the glyph cache is definitely involved. You can change HID mode, which changes some text, or it could be something silly like the clock
19:54:07kugelwhy does gcc fail to inline that
19:56:41kugelhow to fix that?
19:56:43 Quit SynrG (Ping timeout: 240 seconds)
19:56:44kugelupgrade gcc?
19:58:07gevaertsWhich gcc is that?
19:59:09kugelor don't inline for this gcc version?
20:00:00 Quit maraz (Remote host closed the connection)
20:00:53funmansounds hackish
20:01:01funmanwe don't want to test all known versions of gcc
20:01:20funmanwhich -O are you using ?
20:01:38kugelI think the sim uses none
20:01:48kugelI mean no -O in the command line, not -O0
20:02:02funmantry -O or -O2 ?
20:03:34 Quit leavittx (Ping timeout: 240 seconds)
20:03:51kugel-O seems to work
20:04:12kugelbut it creates some warnings (that also happen for some RaaA builds)
20:04:49kugelwhy do we filter -O out anyway?
20:05:45 Join maraz [0] (
20:07:33kugelperhaps someone was musguided a bit and thought you couldnt debug optimized code
20:08:08nosa-ji supose thats the same with the boot screen too gevaerts? :D
20:08:13 Quit PaulJam (Ping timeout: 252 seconds)
20:08:52funmankugel: perhaps to hide the warnings :)
20:09:03gevaertsnosa-j: well, yes and no. The boot screen doesn't have the exclusive access issue, but basically at the point the system is ready to handle themes, the bootscreen isn't needed any more
20:09:31nosa-joh i see
20:10:54nosa-jand mpt would require drivers for the os to use it :(
20:12:20nosa-jthats not too big of a deal though
20:13:02gevaertsWell, other projects are moving to mtp too. Recent android e.g. only does MTP
20:13:23Mirwait wut?
20:13:24gevaertsThe main reason we don't have mtp yet is the usual lack of people working on it :)
20:13:35Mirmtp sucks though
20:13:39Miri hate mtp
20:13:50Mirwhy is droid going mtp
20:14:05gevaertsThat would be off-topic here
20:14:08Mirgevaerts: links to source
20:17:36kugelit's strange that I get strcmp redefined warnings in rockboy when plugin.h does #undef strcmp
20:18:42kugelperhaps plugin.h is already included before string.h, and then not again due to include guards
20:18:58pamaurywhat is the difference between lcd_write_yuv420_lines_odither and lcd_write_yuv420_lines ? I'm confused because of this: lcd_write_yuv420_lines_odither
20:20:34kugel_odither does dithering while transforming
20:21:03pamauryso the fallback C implementation doesn't do dithering ?
20:22:09 Quit captainkwel (Quit: Ex-Chat)
20:24:15 Join captainkwel [0] (
20:24:32kugelpamaury: that's right
20:28:25dfktif git pull −−rebase refuses to update my local patched tree, should i just use git pull to update and keep my patches?
20:29:12Tornewhen you say "local patched", do you mean you have local commits, or uncommitted changes?
20:29:46dfkti applied patches from flyspray to my local tree
20:30:08pamaurycan lcd_update_rect be called with a rectangle that doesn't fit in the screen ? I see that the implementation in lcd-memframe.c does lots of checks
20:30:10dfktand −−rebase complains about my local tree not being up to date
20:30:27funmani use git stash && git pull −−rebase && git stash pop
20:30:35dfktholy hell
20:30:47dfkttalk about intuitive :p
20:31:04funmanwhen rebasing you must remove your lcoal diffs
20:31:10jhMikeSfunman : I've got that same thing set up automated if the basic pull fails :)
20:31:25jhMikeSor commit the changes first
20:31:36gevaertsdfkt: git basically refuses to do anything if you have untracked changes. You need either git stash to put modifications away for a while, or make a local commit
20:31:50n1spamaury: i think it's juat to be safe, it *should* probably never be called with such rectangles
20:31:53Tornecommit early and often :)
20:31:54funmanyou can pull (without rebase)
20:32:06gevaertsIt *is* intuitive if you've studied it for long enough :)
20:32:14dfkti sure wish git was a bit easier to comprehend, like svn...
20:32:17Tornefunman: well, sort of. If you get into the habit of doing that once you actually have local commits you'll be creating merges, though
20:32:27kugeldfkt: most people do :)
20:32:34Torneand even with a merge it won't let you pull changes that touch the same files, even if they don't conflict
20:32:45dfktso i should use funman's line as an equivalent to a good old "svn up"? ;)
20:32:53Torneyes, more or less
20:33:12Torneit basically commits your changes for you, but puts them aside, then pulls from upstream and applies the temporary commit on top again
20:33:19*jhMikeS has under [alias]: up = !git pull −−rebase origin master || (git stash -q && git pull −−rebase origin master && git stash pop -q)
20:33:27Tornewhich will introduce coflict markers if things conflict, which is what svn up does :)
20:33:52Torneyou might want to consider just.. making actual commits more often, thoguh :)
20:34:21*pamaury hopes kugel will soon correct all errors and warnings :)
20:34:22gevaertsAs soon as we have gerrit up and running this particular use case will slowly change
20:34:41Torneyeah, sorry about that
20:34:46Tornei need to write the stuff for that real soon :)
20:34:50*jhMikeS never know what a commit message will be until acutally finalizing the whole shebang
20:34:58TornejhMikeS: you can rewrite it as often as you want
20:35:14gevaertsTorne: I'm not complaining, really just pointing out a benefit :)
20:35:34jhMikeSyeah, there's a −−revise (or whatever it was)
20:36:15jhMikeSaha, that's it
20:36:23Torneyou can also use interactive rebase to flatten multiple commits into one commit (and it'll merge their messags, or throw them away, as you choose)
20:36:28CIA-81Commit 74c8746 in rockbox by 03Thomas Martitz: Don't filter out -O for sims.
20:36:29Torneso yeah. just make commits locally :)
20:36:40Torneonce yo uare done you can squash them together and rewrite the message
20:37:12dfktjhMikeS, where do i find this [alias] section?
20:37:31kugeldfkt: better use jhMikeS' line. if there's nothing to stash than stash pop will harm :)
20:37:43dfktstash pop sounds illegal anyways :D
20:38:15jhMikeSdkft: in .git/config, if not there add it, and define any shortcuts there
20:38:47Torneseriously though, committing is safer
20:39:07Torneonce you have committed the changes you can trivially revert to the earlier state if you make a mistake when resolving conflicts/etc
20:39:19CIA-8174c8746 build result: 0 errors, 43 warnings (Thomas Martitz committed)
20:39:21Torneusing the stash, this is more difficult
20:40:27gevaertskugel: that's why it was filtered :)
20:41:12Tornealso if you stash and pull, this will still not work if you have created a local file that now exists in upstream ;)
20:41:17n1si'd guess it was fro debugability and build speed too
20:41:20Tornesince stash doesn't touch untracked new files
20:41:34Torneso the pull will still fail, because it'll refuse to overwrite the existing file
20:42:13CIA-81Commit e720ad0 in rockbox by 03Thomas Martitz: Fix one more warning.
20:42:57kugeln1s: you can debug code at all levels of -O
20:43:36kugelyou might want to avoid "parameter optmized out" but then you want -O0 instead of nothing
20:43:53n1s-O0 is the default
20:44:45CIA-81e720ad0 build result: 0 errors, 4 warnings (Thomas Martitz committed)
20:46:22kugelwhat the heck, I didn't know there are stubs for storage_*_sectors
20:50:48CIA-81Commit 31a3f76 in rockbox by 03Thomas Martitz: Attempt to fix remaining warnings.
20:53:09CIA-8131a3f76 build result: 4 errors, 6 warnings (Thomas Martitz committed)
20:59:44kugeloh man!
21:00:22CIA-81Commit 38050f4 in rockbox by 03Thomas Martitz: ypr0: Work around warning introduced by buggy alsa headers.
21:00:58 Nick preglow_ is now known as preglow (
21:02:48CIA-8138050f4 build result: 8 errors, 10 warnings (Thomas Martitz committed)
21:02:50kugelthat attempt was unsuccessful
21:03:25 Quit Xerion (Read error: Connection reset by peer)
21:04:59lebelliumkugel: are you trying to fix the warning issues you talked me about yesterday?
21:05:07lebelliumthe annoying warnings*
21:05:25kugelmore annoying ones :)
21:05:35kugelbut it builds wanring-free now
21:05:38kugelat least for me
21:06:40lebelliumI don't really understand how it works. For instance he reads here 8 errors and 10 warnings for your latets build
21:06:56kugeloh ffs
21:06:59kugelthat error is *wrong
21:07:18***Saving seen data "./dancer.seen"
21:07:31 Join TBFOOL [0] (
21:07:43 Join Xerion [0] (
21:07:54 Quit TBFOOL (Client Quit)
21:08:59kugelhelp me :(
21:09:22 Join TBFOOL [0] (
21:10:41 Quit TBCOOL (Ping timeout: 245 seconds)
21:11:19*Mir hands kugel a lifesaver in fruit flavors
21:11:38 Quit y4n (Quit: Today is the perfect day for a perfect day.)
21:12:24n1skugel: looks like the script that parses the logs for errors is buggy
21:12:48n1sit counts the info lines asociated with the pdbox warning as errors
21:13:13kugelthe warning is wrong too
21:13:22kugelit the memcpy params are alright
21:13:51kugelmy gcc doesnt show this error
21:14:17 Quit perrikwp (Read error: Connection reset by peer)
21:15:04 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
21:16:05 Quit bluebrother (Disconnected by services)
21:16:06 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
21:16:22n1sit looks to me like it's copying a 5 byte string into a 4 byte buffer
21:18:03gevaertsn1s: where?
21:18:11n1sor possibly that gcc screws up and things that is what's happening
21:18:32 Quit fs-bluebot (Ping timeout: 244 seconds)
21:18:41n1sgevaerts: i was looking at apps/plugins/pdbox/PDa/src/d_soundfile.c:702:15
21:18:59kugelthats a 4 element char array
21:19:03n1sthe question is if char datachunk_ID[] = {'S', 'S', 'N', 'D'}; is null terminated or not
21:19:04gevaertsn1s: that doesn't look like 5 bytes to me
21:19:09gevaertsIt's not
21:19:16gevaertsWell, it shouldn't be
21:19:26n1swell, then it looks like a gcc bug
21:19:38 Join fs-bluebot [0] (
21:19:44jhMikeSsay what?
21:19:46n1sperhaps sizeof is broken for this case
21:19:56kugelgevaerts: what is storage_*_sectors supposed to return?
21:21:20gevaertskugel: IIRC 0 on success, a (negative, I guess) error code otherwise
21:22:24gevaertsn1s: what's "t_datachunk *aiffdc = (t_datachunk *)headerbuf + sizeof(t_aiff);" supposed to do?
21:23:48CIA-81Commit 7dd65f7 in rockbox by 03Thomas Martitz: Another attempt at fixing storage_*_warnings. Correct return values also.
21:24:01*n1s is no expert on this!
21:24:42gevaertsI'd say it tries to leave toom after headerbuf for one t_aiff object, but I'm not sure it does that. What's the precedence here again?
21:24:51*gevaerts asks Torne to have a look at this
21:24:53n1si'd guess it creates a pointer at offset sizeof(t_aiff)*sizeof(t_datachunk)
21:25:17gevaertsYes, but that sounds weird
21:26:05Tornethe cast binds tighter, so yeah.
21:26:08CIA-817dd65f7 build result: 0 errors, 1 warnings (Thomas Martitz committed)
21:26:38Torneit's equivalent to headerbuf[sizeof(t_aiff)] if headerbuf was a t_datachunk[] :)
21:27:03Tornepretty sure. which as you say, sounds obviously wrong.
21:27:08gevaertsIf that's the case, it might well be overflowing the headerbuf array, which is on the stack and therefore reasonably close to aiffdc, so this could actually be genuinely overlapping data
21:27:39Tornei think they mean it to be sizeof(t_aiff)+sizeof(t_datachunk)
21:27:49Tornebut there's no particularly elegant way to write that using actual pointer arithmetic :)
21:28:01gevaertsNot sure
21:28:18gevaertsI'd say it should be (t_datachunk *)(headerbuf + sizeof(t_aiff))
21:28:23gevaertsheaderbuf is a char *
21:28:37Torneyes, that's plausible also
21:28:42gevaertsWell, a char[] to be precise
21:28:43Torneif t_aiff is the first thing in memory
21:28:54TorneBut in that case, yes, you need the parens
21:29:00Tornebecause the precedence is wrong
21:29:17kugelhow about "t_datachunk data[];" it t_aiff?
21:29:25gevaertsIt's trying to write an AIFF header. We should be able to find out what that's supposed to look like
21:30:35CIA-81Commit 91b52a1 in rockbox by 03Thomas Martitz: Fix another uninitialized var warning.
21:31:39jhMikeSgevaerts: aiff_enc
21:32:01jhMikeSin apps/codecs
21:32:39 Quit TheLemonMan (Quit: WeeChat 0.3.6)
21:32:52CIA-8191b52a1 build result: 4 errors, 2 warnings (Thomas Martitz committed)
21:33:04gevaertsThat writes SSND immediately after the equivalent of the aiff_t bit in d_soundfile.c
21:33:34gevaertsSo I'm pretty sure it's supposed to skip sizeof(t_aiff) bytes after headerbuf
21:33:43gevaertswhich means (t_datachunk *)(headerbuf + sizeof(t_aiff))
21:34:06jhMikeSthe recording codec does in as basic a way as it gets
21:34:16jhMikeSprobably the intent, yes
21:34:25*gevaerts fixes it
21:34:30Torneso yeah, that's just someone forgetting that casts are higher precedence than arithmetic
21:34:40gevaertsAnd then not testing :)
21:34:49Tornewell yes ;)
21:35:57kugelthat datachunk should be part of t_aiff, really
21:37:05n1sgevaerts: seems your guess was correct i think, judging from;a=commitdiff;h=c2a3380c00e6aea9c7505651997937e08d922fdc;hp=de1b2d1defa5a1f7cd48b6adb5bcd3083e6b636a
21:38:19n1sdoesn't seem to match the code imported into rockbox exactly but same thing
21:39:29jhMikeSFS #12249
21:39:30fs-bluebot 3pdbox - Buffer access out-of-bounds: aiffhdr.a_samprate (bugs, closed)
21:39:33kugelshould have taken the upstream fix :P
21:41:19n1sheaderbuf + ((aiffhdr->a_samprate + sizeof(dogdoo))-(unsigned char *)aiffhdr) is really ugly :)
21:41:47CIA-81Commit 7efbd63 in rockbox by 03Frank Gevaerts: Fix wrong pointer arithmetic in the PDbox aiff header writing code
21:42:36gevaertsjhMikeS: ah, nice find
21:42:37 Join swiezak [0] (
21:42:39CIA-81Commit 8a3af26 in rockbox by 03Jonathan Gordon: skin touchregions: fix the 'none' region firing when it shouldnt.
21:43:02kugelgevaerts: hey! posted it early!
21:43:22gevaertskugel: ah, nice find :)
21:44:06*gevaerts just assumes that this would be upstream code so he didn't bother to check its history
21:44:25JdGordonciabot looks good... pity it wont ping us though with full names (unless the red/yellow is still on irc name? :) )
21:44:44CIA-817efbd63 build result: All green
21:44:54*jhMikeS just typed that to get the task description :P
21:47:32CIA-818a3af26 build result: All green
21:50:58bluebrother^JdGordon: just add a hilight for your real name :)
21:51:07 Quit swiezak ()
22:01:29 Quit benedikt93 (Quit: Bye ;))
22:02:38 Join stripwax [0] (
22:04:06 Join SynrG [0] (
22:04:14kugelcaptainkwel: FPS went down, actually
22:04:53captainkweloops :(
22:05:39captainkwelrunning in fullscreen or hybrid?
22:06:24captainkwelI know that if you do software rendering in hybrid it's slow to the point of unusability... hardware it at least runs ok.
22:07:16kugel7fps only
22:09:40captainkwelhmm. I'll play around with it a bit more and see what I can find.
22:11:52 Quit captainkwel (Quit: Ex-Chat)
22:18:37 Join z180 [0] (
22:22:12 Quit SynrG (Ping timeout: 245 seconds)
22:23:53kugel60fps in RaaAoA on the toucpad
22:26:22z180instead of the already mentioned lwIP, Rockbox could use the smaller uIP stack if the license is compatible
22:27:29CIA-81Commit 250a733 in rockbox by 03Dominik Riebeling: Check running processes at startup.
22:28:08 Join MethoS [0] (~clemens@
22:28:34 Nick MethoS is now known as Guest28733 (~clemens@
22:30:10CIA-81250a733 build result: All green
22:34:19 Quit stripwax (Ping timeout: 252 seconds)
22:38:28CIA-81Commit 5ac41a2 in www by 03Thomas Martitz: ypr0: Add normal build to build system.
22:40:20CIA-81Commit aeae94d in rockbox by 03Thomas Martitz: Fix typo in comment.
22:40:59 Join Strife89 [0] (~Strife89@
22:43:00CIA-81aeae94d build result: All green
22:44:56 Join SynrG [0] (
22:45:07 Quit SynrG (Changing host)
22:45:07 Join SynrG [0] (~synrg@debian/developer/synrg)
22:45:47 Quit z180 (Quit: ChatZilla 0.9.88 [Firefox 3.6.25/20111212142243])
22:45:56CIA-81Commit de1bcfa in rockbox by 03Thomas Martitz: Fix incomplete comment.
22:48:19CIA-81de1bcfa build result: All green
22:50:01kugelhmm that didn't work as Zagot promised
22:53:00kugeldid I miss something?
23:05:52kugelZagor too
23:07:19***Saving seen data "./dancer.seen"
23:10:34 Part TomColler
23:12:09 Quit factor (Read error: Connection reset by peer)
23:15:03 Join factor [0] (~factor@
23:15:15 Quit SynrG (Ping timeout: 252 seconds)
23:19:10 Join stripwax [0] (
23:32:46 Join Scromple [0] (~Simon@
23:36:44 Join Scr0mple [0] (~Simon@
23:36:52 Quit Scromple (Client Quit)
23:38:35 Quit n1s (Quit: Ex-Chat)
23:38:43 Quit stripwax (Quit:
23:45:17 Quit ToxAtec (Quit: ChatZilla 0.9.88 [Firefox 7.0.1/20111022212758])
23:55:38 Quit lebellium (Quit: ChatZilla 0.9.88 [Firefox 10.0/20120118081945])
23:58:56 Join JdGord [0] (~AndChat@

Previous day | Next day