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 2010-03-15

00:00:03amiconnAnd I am asking what we would win
00:00:20gevaertsThe common case people are trying to solve is a track with v1 tags where people ran a replaygain tool that added the replaygain tags in v2 only, but didn't copy the other tags
00:00:21rasherThe possibility of reading people's tags
00:00:28amiconnRockbox does that
00:00:31gevaertsat least, IIUC
00:00:41rasherThat was my case, which I doubt is unique
00:01:04rasherI'm sure there are others (people accidently click a checkmark to add a v2 tag but don't fill it, etc)
00:01:08amiconnAnd how likely is it that an id3v1 conatins exactly those pieces of information which are missing in the id3v2 (if the latter is non-empty)?
00:01:14kugelamiconn: fix your tags and you don't get a seek
00:01:17rasheramiconn: non-zero
00:01:43amiconnI guess it's less likely than the percentage of runtime we'd be losing by doing this
00:01:43rasherWhy *not* do it? There's literally no cost, and a non-zero chance of helping people out
00:02:22funmanamiconn: can you measure the percentage of runtime you'd be losing ?
00:02:33 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
00:02:35rasherWe keep getting back to this point, and people keep going "but a seek takes battery runtime!"
00:02:51rasherAnd I keep getting more depressed
00:03:04gevaertsamiconn: the only tracks where you would be losing runtime without gaining useful metadata would be those that have *only* non-human-readable v2 tags like replaygain, but no title or anything like that
00:03:11*funman hugs rasher
00:03:14amiconnfunman: I'm quite sure it's measurable
00:03:26gevaertsunless I'm missing some cases
00:04:03funmanamiconn: well it is (measurable), but according to your detractors it is so small that it can be ignored.
00:04:18amiconnI can't ignore it
00:05:25funmanamiconn: will you measure it anyway?
00:05:37rasheramiconn: if it shaves off more than 30 seconds of battery life for you, I'll buy you a beer. And that offer goes for all Rockbox users who can prove this.
00:05:41funmani can easily ignore 1% less runtime
00:05:52gevaertsamiconn: do you actually have mp3 files that match the criteria I mentioned, or are my criteria wrong?
00:06:46gevaertsi.e. *no* title, artist or album tags at all?
00:06:48 Join stripwax [0] (~Miranda@87-194-34-169.bethere.co.uk)
00:07:10amiconnWell, if it tries id3v1 whenever id3v2 misses fields id3v1 could provide, I do have quite a few such tracks. And it would fail everytime, because I don't have *any* id3v1 tags
00:07:12rasherI have yet to see any argument beyond "well my battery life might be impacted some" against it
00:07:33rasheramiconn: how much battery life do you think you would lose?
00:07:38amiconn(id3v1 fields being title, artist, album, track number, year)
00:08:02rasherThat's not what's being suggested though
00:08:05gevaertsamiconn: that's not what's being proposed. The proposal would only try id3v1 if id3v2 misses *all* of those fields. As soon as one is there, id3v1 will not be tried
00:08:16pixelmahow would you explain why completing tag info from ID3v1 if all of title/artist/album is missing in the ID3v2 is ok, but completing or trying to if one of those is missing is not?
00:08:46pixelmaand what about other tags?
00:08:59JdGordonpixelma: so the only issue with fms now is the peak meter?
00:09:04rasheroh for christ sake
00:09:47gevaertspixelma: one missing tag is reasonably likely to be normal (i.e. some tracks aren't part of an album). Missing all of them is much less likely to be expected
00:09:51 Join hd [0] (~jd@modemcable207.134-202-24.mc.videotron.ca)
00:09:52 Quit hd (Changing host)
00:09:52 Join hd [0] (~jd@Wikipedia/HellDragon)
00:10:37gevaertsAnd I think that the exact list of tags to look at isn't a blocker. If people also want to look at track number, fine
00:11:54amiconnHmm, my estimation says the loss would be a little less than 3 minutes of a total of 15.5 hours with 2100mAh batteries
00:12:08 Quit jd (Ping timeout: 245 seconds)
00:12:14amiconnSo not that much, but it still feels technically wrong
00:12:32Bagderamiconn: assuming all your songs match gevaerts criteria?
00:12:42amiconnThat's with all tracks being played requiring that extra seek
00:13:05Bagderthat's not a very likely scenario though, if I may guess
00:13:48JdGordonjust for the record, I dont ever remember my problems with the id3v1/2 patch being the extra seek
00:13:59amiconnOther values used: average current on recorder 135mA, disk access current 500mA, extra time for the seek: 200ms, 4 minutes per track
00:14:52S_a_i_n_tIf a track is missing all of the relevant id3v1 tags, how likely is it that it will have *any* id3v2 tags?
00:14:55saratogathats twice the seek time you told me before
00:14:56pixelmagevaerts: in theory, but that's a very very weak argument if someone asks about it and why not etc.
00:15:00*S_a_i_n_t is confused by this...
00:15:13saratogayou're becoming more pessimistic as you make your case
00:15:52rasheramiconn: Thanks for injecting some numbers.
00:16:03rasherSo that's 3 minutes at the absolute worst case we could possibly ever imagine
00:16:26Bagderthat's 0.3% btw
00:16:43amiconnsaratoga: Maybe, but then the 500mA are a bit optimistic in turn (of course that depends on the disk model)
00:16:47pixelmaS_a_i_n_t: that's not the question at all if I unders
00:16:55gevaertspixelma: Do we know of people who have a significant number of tracks with mixed v1/v2 useful tags?
00:16:55pixelmatand you correctly
00:17:26funmanamiconn: so is 0.3% small enough to be ignored?
00:18:07gevaertsit's 0.3% for people who either have been missing their title tags all the time and didn't notice, or who have no title tags at all
00:18:17pixelmagevaerts: I saw some very messed up tagging by other people
00:19:47 Quit DerPapst (Quit: Leaving.)
00:20:07 Join Casainho [0] (~chatzilla@87-196-27-159.net.novis.pt)
00:20:16 Quit Horschti (Quit: Verlassend)
00:21:24gevaertspixelma: anyway, I'm not actually against reading v1 in all cases where it could provide one extra string, but I think the numbers are totally different. FS #10016 will not have any effect for people with "properly" tagged files, and a small effect for those who have semi-weird cases like v2-replaygain + v1-main-tags. Fallback for all tags would have a small effect for nearly everyone (because as was pointed out, some tags are very commonly missing
00:21:41rasherpixelma: I thought we didn't care what happened to people with imperfect tags. Now suddenly we do?
00:22:56gevaertsFS #10016 will also have a small effect for people with files with *only* replaygain (or similar) tags, but I'd expect those to be really uncommon
00:22:57pixelmahuh? I just think we are losing arguments against them and then we do extra seeks all of the time
00:23:40rasherNo one is suggesting to seek every time
00:24:08 Quit stripwax (Read error: Connection reset by peer)
00:26:43 Join stripwax [0] (~Miranda@87-194-34-169.bethere.co.uk)
00:26:43 Quit JdGordon (Read error: Connection timed out)
00:27:02***Saving seen data "./dancer.seen"
00:27:10 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
00:32:18 Join Casainho_ [0] (~chatzilla@87.196.27.159)
00:32:45pixelmamy point was just that currently there is a clean line and you know what you can expect, after this change it's not anymore and you lose grounds if someone asks for more
00:33:21 Join Foxx- [0] (~Foxx@pool-68-160-211-89.ny325.east.verizon.net)
00:33:45 Quit ender` (Quit: If at least two walls aren't shored up with balks of timber then it's not a true witch's cottage at all, but merely the home of some daft old bat who reads tea leaves and talks to her cat.-- Terry Pratchett: Lords and Ladies)
00:34:06 Quit Foxx (Ping timeout: 245 seconds)
00:34:11gevaertsI think there's still a reasonably clean line, in that the "more" requires everyone to pay the 3 minutes as soon as they omit one single tag.
00:35:03 Quit Casainho (Ping timeout: 245 seconds)
00:35:03CIA-5New commit by 03funman (r25188): amsinfo: do not assume library blocks always have the same size ...
00:35:12 Nick Casainho_ is now known as Casainho (~chatzilla@87.196.27.159)
00:36:15 Quit bertrik (Quit: sleeo)
00:39:13funmanthe Clip+ likely have a FM si4702 chip (like fuze/e200v2/clipv1/c200v2..), and a "RDA 5802" whatever it is
00:39:39 Quit stripwax (Quit: http://miranda-im.org)
00:48:18 Join planetbeing_ [0] (~planetbei@c-71-236-164-204.hsd1.or.comcast.net)
00:50:13 Quit r0b- (Ping timeout: 260 seconds)
00:51:07 Quit planetbeing (Ping timeout: 252 seconds)
00:51:07 Nick planetbeing_ is now known as planetbeing (~planetbei@c-71-236-164-204.hsd1.or.comcast.net)
00:53:32pixelmabluebroth3r: I even found a bug report about the nopt issue, it was closed because the workaround of rearranging the nopt worked - http://www.rockbox.org/tracker/task/10652
00:57:01 Quit CGL (Ping timeout: 240 seconds)
00:57:52 Quit bmbl (Quit: Bye!)
01:00
01:02:28 Join Llorean1 [0] (~DarkkOne@adsl-99-158-46-229.dsl.hstntx.sbcglobal.net)
01:02:44 Quit Llorean (Quit: Leaving.)
01:02:56 Nick Llorean1 is now known as Llorean (~DarkkOne@adsl-99-158-46-229.dsl.hstntx.sbcglobal.net)
01:03:01 Quit Llorean (Changing host)
01:03:01 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
01:04:47 Join r0b- [0] (~nnscript@adsl-76-235-205-202.dsl.klmzmi.sbcglobal.net)
01:07:23 Join mikroflops_ [0] (~yogurt@90-224-30-68-no112.tbcn.telia.com)
01:11:23 Quit mikroflops (Ping timeout: 258 seconds)
01:15:54 Quit r0b- (Ping timeout: 248 seconds)
01:19:26 Part froggyman
01:21:06 Join r0b- [0] (~nnscript@adsl-76-235-195-8.dsl.klmzmi.sbcglobal.net)
01:30:26 Quit karashata (Quit: The fluffy dragon has left completely!)
01:36:10CIA-5New commit by 03funman (r25189): sd-as3525v2: reduce the delay before reading response
01:36:15CIA-5New commit by 03funman (r25190): Clip+: si4702 FM
01:38:34funman(still no sound)
01:40:51funmanhm boomshine.lua likely crashes (black screen)
01:41:48 Quit r0b- (Ping timeout: 268 seconds)
01:42:21 Nick Foxx- is now known as Foxx (~Foxx@pool-68-160-211-89.ny325.east.verizon.net)
01:43:05 Join r0b- [0] (~nnscript@76.235.195.8)
01:44:29 Quit evilnick (Ping timeout: 276 seconds)
01:46:11S_a_i_n_tAny Nano2g owners present which are fluent in more than one language?
01:46:29S_a_i_n_tOr that know enough of another to tell me if my translations are working?
01:46:42S_a_i_n_terr...Nano1 *or 2g will do.
01:59:52 Join Speedy2 [0] (~johnny@cpe-76-88-67-189.san.res.rr.com)
02:00
02:00:02CIA-5New commit by 03funman (r25191): as3525: do not assume we can record from microphone
02:00:22Speedy2Guys, BTW, thanks for your hardwork. I loaned my Sansa Clip with Rockbox to the GF and she liked the UI and said the sound quality was noticably better. Just wanted to pop in and say thanks!
02:01:03S_a_i_n_tawesome.
02:03:15ConformistSpeedy2: sansa clip has better sound quality than what mp3 player?
02:05:37perfectdrug_S_a_i_n_t: is sim ok?
02:06:44 Quit Casainho (Remote host closed the connection)
02:09:39krazykit`Conformist, the original firmware, presumably
02:09:56Conformistrockbox really improves sq?
02:10:07Conformistthats kewl
02:10:11Conformisthow does it do it
02:10:57LloreanIt's subjective
02:11:03S_a_i_n_tConformist: it has a 5 band EQ...which many players lack.
02:11:08S_a_i_n_tthat helps a LOT.
02:11:12Conformistoh
02:11:20Conformisti really like the eq settings
02:11:31LloreanOur decoders are different from theirs. In some cases ours is more accurate (in others it may be less). As well we may run the hardware a bit differently.
02:11:47Conformistwow
02:11:49krazykit`if the OF applies equalization or other "sound improvements", NOT doing them may improve SQ too
02:11:50LloreanAlso some original firmware may alter sound natively, an internal EQ to "sound better" based on what they think consumers like, etc.
02:12:25LloreanSo there are many reasons why Rockbox may sound different than the original software. "Better" is never something we can promise though, since it also may not sound different, or you may like different things than your neighbor.
02:12:44Conformistwell i've compared like a dozen different spdif sources, and i like rockbox h120 optical output best
02:13:02JdGordonwe go by the "Bag Of Shit" standard here...
02:13:06funmananyone with a clipv2 could test current bootloader/binary ?
02:13:12JdGordonwe strive to not sound like one!
02:13:14Conformisti wonder if its not becuz rockbox has more well written software
02:14:06Conformistif i could rockbox my computer for media playback i would, i have more faith in rockbox than any media player on the computer
02:14:14 Join jd [0] (~jd@modemcable207.134-202-24.mc.videotron.ca)
02:14:15 Quit jd (Changing host)
02:14:15 Join jd [0] (~jd@Wikipedia/HellDragon)
02:15:09Conformisthow much does paypal deduct if i use them to donate to rockbox?
02:16:21funmani think it depends of the country/currency/amount
02:16:35funmanlook up their terms and conditions perhaps
02:16:44Conformistok
02:16:57 Quit hd (Ping timeout: 268 seconds)
02:17:12Conformistif they charge per transaction i'd rather once than multiple smaller donations
02:17:49funmani'd think one transaction is always cheaper than several
02:18:13funmanbtw i'm wondering if i should buy a clipv2
02:18:31Conformistpaypal charges 30 cents per
02:18:37funmani can work with only the Clip+ but there doesn't seem to be a lot of testers with Clipv2
02:19:13Lloreanfunman: Have you tried looking at third-party forums to see if there are users there interested in testing?
02:19:26LloreanI'd bet there's a nice group of people at someplace like anythingbutipod willing to do any crazy thing you ask.
02:19:42funmanLlorean: i was just thinking I could ask on the rockbox forums
02:21:47Conformistdoes rockbox let you donate to individual programmers or should i donate to the link at the bottom of rockbox.org?
02:22:34LloreanYou should use the donate link. You're free to do whatever you want, though.
02:22:42CIA-5New commit by 03funman (r25192): Clipv2: enable FM so testers can confirm if it's detected
02:22:47*JdGordon sends Conformist his paypal details :)
02:23:21Conformistii'll use the donate link
02:23:27funmanConformist: JdGordon is not a developer!! but i am one !!!
02:25:12 Join preglow [0] (thomj@tvilling2.pvv.ntnu.no)
02:25:37*JdGordon doesnt have a typo to fix to prove otherwise :(
02:25:39funmanLlorean: also I assume testers should be able to build from source & edit some files, perhaps if I ask on ABI or another forum people would ask where they can get binaries etc
02:27:06***Saving seen data "./dancer.seen"
02:27:38funmanalso I just see that Torne has a clipv2, and I assume he can "build from source & edit some files" :)
02:29:06 Join Strife1989 [0] (~michael@adsl-154-2-161.mcn.bellsouth.net)
02:29:52Conformistsince im broke rockbox only gets 19$
02:30:00 Part Conformist
02:31:24RadicalREvery little bit helps.
02:38:28 Nick Strife1989 is now known as Strife89] (~michael@adsl-154-2-161.mcn.bellsouth.net)
02:38:36 Nick Strife89] is now known as Strife89 (~michael@adsl-154-2-161.mcn.bellsouth.net)
02:48:23 Join saratoga_ [0] (~9803c20d@gateway/web/freenode/x-stwmzrpmwjdljukg)
02:48:31saratoga_funman: want me to test something
02:48:59funmansaratoga_: just see if you can get to the menu with current build (and current bootloader if you don't have one that works)
02:49:10saratoga_ok will try now
02:49:13 Quit saratoga_ (Client Quit)
02:51:24 Join evilnick [0] (~evilnick@ool-457bccf5.dyn.optonline.net)
02:53:01 Quit r0b- (Ping timeout: 264 seconds)
02:55:59 Join r0b- [0] (~nnscript@adsl-76-253-124-192.dsl.klmzmi.sbcglobal.net)
02:56:47Speedy2Oh, Conformist left
02:56:57Speedy2Rockbox audio quality >> Sandisk original Sansa firmware
02:57:17Speedy2I have a sort of crazy question
02:57:17CIA-5New commit by 03funman (r25193): sd-as3525v2: prepare for multidrive builds
02:57:19saratogafunman: it just freezes after "executing"
02:57:30funmanhm pbxy reported the same thing
02:57:50Speedy2If I retro-fitted a Clip with the FM radio chip that supports RDS, how hard would it be to add the firmware support for it? And if I did so, would others be interested in having their clips modified?
02:58:11Speedy2SiLabs makes the same version of the IC with RDS support.
02:58:29saratogaprobably not hard but i doubt you could modify the clip very easily, the board is tiny and the traces are even smaller
02:58:43Speedy2I have access to the right equipment
02:58:53Speedy2I've already had it opened and fixed its guts
02:59:11S_a_i_n_thttp://pastebin.com/yWPPiBLH looks really wrong to me, but I suspect it *should* work...?
02:59:17Speedy2After heavy use, solder joints on the CPU crack.
03:00
03:01:11 Join bzed_ [0] (~bzed@devel.recluse.de)
03:01:11 Quit bzed (Read error: Connection reset by peer)
03:01:27 Nick bzed_ is now known as bzed (~bzed@devel.recluse.de)
03:01:44saratogafunman: if you need things tested, you can ping me on gtalk
03:02:10funmansaratoga: right now i have no idea, i suppose to get further it would require tinkering with system_init()
03:02:11 Quit krazykit` (Ping timeout: 246 seconds)
03:02:58saratogafunman: is it radio that broke things or was it already not booting?
03:03:06funmanit was already not booting
03:03:13saratogaah ok
03:05:11 Quit perfectdrug_ (Quit: perfectdrug_)
03:05:23 Join krazykit [0] (~kkit@adsl-76-240-216-183.dsl.ipltin.sbcglobal.net)
03:08:06JdGordonS_a_i_n_t: no reason it wouldnt.... maybe we should change the Sx/St tags to use ()'s instead of |'s
03:08:11saratogadoesn't rbutil install the current ams devices?
03:08:34funmanyes it does (not sure for c200v2/m200v4 though)
03:08:52saratogaah i misunderstood the wiki
03:09:41S_a_i_n_tJdGordon: I was thinking it would be better (easier to read) as " %Sx[String] actually
03:09:56S_a_i_n_tthe |'s make the syntax impossible
03:10:30JdGordonor that
03:11:48S_a_i_n_tit only looks weird in a conditional though.
03:15:23 Quit sinthetek (Read error: Connection reset by peer)
03:17:58S_a_i_n_tYear: Unkown would be a lot better than "Unknown Year" when it comes to being translated to another langauge...no?
03:18:19S_a_i_n_tI guess this as the word order may differ in the other language.
03:25:58 Join panni__ [0] (hannes@ip-95-222-52-93.unitymediagroup.de)
03:26:01 Quit panni_ (Read error: Connection reset by peer)
03:26:01 Quit jd (Read error: Connection reset by peer)
03:26:14 Join jd [0] (~jd@modemcable207.134-202-24.mc.videotron.ca)
03:26:15 Quit jd (Changing host)
03:26:15 Join jd [0] (~jd@Wikipedia/HellDragon)
03:32:10*S_a_i_n_t Grrrrrr's at the lang string "Loading..." including the "..."
03:32:20S_a_i_n_tand thus, not being able to use it.
03:32:50S_a_i_n_t*well, not how I'd like to anyway.
03:35:46S_a_i_n_t"Disk" (the best alternative) doesn't seem as appropriate to display upon disk acces as "Loading" to me at least.
03:36:18CIA-5New commit by 03funman (r25194): sd-as3525v2: add card hotswap to sd_thread ...
03:36:46S_a_i_n_tand its being used in a text animation and the "..." in "Loading..." mess that up.
03:38:04 Quit Darkknight512 (Ping timeout: 264 seconds)
03:42:30*S_a_i_n_t guess it would be pretty hard to convince someone to commit a patch to svn just so one word in my WPS can be translated/animated properly...
03:42:32S_a_i_n_t:D
03:49:03funmanFlynDice: (for the logs) perhaps B5 enables power for µSD card, and then we can set the card bits in MCI_COMMAND/other regs
03:50:04 Join planetbeing_ [0] (~planetbei@c-71-236-164-204.hsd1.or.comcast.net)
03:50:10LloreanS_a_i_n_t: Why not use an image or icon instead of text?
03:51:30 Quit planetbeing (Ping timeout: 252 seconds)
03:51:30 Nick planetbeing_ is now known as planetbeing (~planetbei@c-71-236-164-204.hsd1.or.comcast.net)
03:55:04 Quit funman (Quit: free(random());)
04:00
04:05:44S_a_i_n_tLlorean: I have an icon for disk access also, but my wps has "working; working.; working..; working..." (alternated sublines of course) as a text animation in the progress bag. And I'd like to be able to translate this, but apparently can't.
04:06:39S_a_i_n_tI figure it may not be immediately apparent when "working" means to a non english speaker.
04:07:15S_a_i_n_tbut there's no "buffering" or "disk access" in the lang files either.
04:07:40LloreanWhy not just do an animation of some sort. Rotating circle, spinning dots, or whatever, like many PC apps use to indicate "I'm doing something at the moment"
04:08:25S_a_i_n_tI have that at the moment, I just thought the text was a nice tough also.
04:08:37S_a_i_n_ts/that/spinning dot animation/
04:11:56S_a_i_n_tI must say, I did expect "Buffering" to be in the English.lang file.
04:12:10 Quit kugel (Remote host closed the connection)
04:14:05 Join CGL [0] (~CGL@190.207.203.1)
04:19:25 Join SirFunk [0] (~Sir@97-92-38-108.dhcp.aldl.mi.charter.com)
04:20:29 Join CaptainKwel [0] (~jason@207.237.107.203)
04:21:18 Quit CaptainKewl (Write error: Broken pipe)
04:27:10***Saving seen data "./dancer.seen"
04:32:21 Quit Speedy2 (Remote host closed the connection)
04:35:16 Quit TheSeven (Disconnected by services)
04:35:30 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:35:43 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:39:01 Join Barahir [0] (~jonathan@gssn-5f757426.pool.mediaWays.net)
04:40:22 Quit Barahir_ (Read error: Operation timed out)
04:47:25 Quit Strife89 (Read error: Connection reset by peer)
04:47:53 Join Strife89 [0] (~michael@adsl-154-2-161.mcn.bellsouth.net)
04:49:36 Join planetbeing_ [0] (~planetbei@c-71-236-164-204.hsd1.or.comcast.net)
04:52:04 Quit planetbeing (Ping timeout: 256 seconds)
04:52:09 Nick planetbeing_ is now known as planetbeing (~planetbei@c-71-236-164-204.hsd1.or.comcast.net)
04:54:15 Join kylegeek [0] (~kylemagui@c-174-52-134-94.hsd1.ut.comcast.net)
04:54:44 Quit kylegeek (Client Quit)
04:58:32 Join Rob2223 [0] (~Miranda@p4FDCA4FE.dip.t-dialin.net)
05:00
05:01:43 Quit Rob2222 (Ping timeout: 245 seconds)
05:01:55 Quit panni__ (Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ))
05:11:37 Join anewuser [0] (anewuser@unaffiliated/anewuser)
05:13:37 Join kadoban [0] (~mud@cpe-67-247-80-129.rochester.res.rr.com)
05:19:51 Join hd [0] (~jd@modemcable207.134-202-24.mc.videotron.ca)
05:19:51 Quit hd (Changing host)
05:19:51 Join hd [0] (~jd@Wikipedia/HellDragon)
05:19:51 Quit jd (Read error: Operation timed out)
05:41:52 Quit CaptainKwel (Remote host closed the connection)
05:42:12 Join evilnick_ [0] (~evilnick@ool-457bccf5.dyn.optonline.net)
05:43:37 Quit evilnick (Ping timeout: 256 seconds)
05:51:18 Quit evilnick_ (Ping timeout: 245 seconds)
05:52:13 Join Strife1989 [0] (~michael@adsl-154-2-161.mcn.bellsouth.net)
05:52:17 Quit Strife89 (Read error: Connection reset by peer)
06:00
06:01:41 Join evilnick [0] (~evilnick@ool-457bccf5.dyn.optonline.net)
06:08:47RadicalRNGEN
06:17:11 Quit RadicalR (Quit: Leaving)
06:27:13***Saving seen data "./dancer.seen"
07:00
07:07:41 Quit CGL (Quit: me juiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii wenas noches!)
07:26:48 Quit elcan (Remote host closed the connection)
07:26:56 Quit anewuser ()
07:39:11 Quit r0b- (Ping timeout: 260 seconds)
07:58:20 Join _arbingordon [0] (~w@c-71-226-248-30.hsd1.pa.comcast.net)
08:00
08:01:59 Quit arbingordon (Ping timeout: 260 seconds)
08:02:21 Quit krazykit (Ping timeout: 264 seconds)
08:02:21 Quit nimak (Ping timeout: 264 seconds)
08:02:24 Quit stavrob (Ping timeout: 248 seconds)
08:02:31 Join nima [0] (~nima@adsl-75-45-227-30.dsl.sfldmi.sbcglobal.net)
08:02:48 Join krazykit [0] (~kkit@adsl-76-240-216-183.dsl.ipltin.sbcglobal.net)
08:03:24 Join stavrob [0] (~sam@78-105-125-218.zone3.bethere.co.uk)
08:12:11 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:12:35CIA-5New commit by 03jdgordon (r25195): fix the remote font not working when there is no .rsbs loaded
08:17:34 Quit Topy (Quit: Leaving)
08:23:52 Quit planetbeing (Quit: planetbeing)
08:27:16***Saving seen data "./dancer.seen"
08:35:13 Quit JdGordon (Read error: Connection timed out)
08:35:29CIA-5New commit by 03jdgordon (r25196): rearrange the skin loading to remove some code dupe
08:35:51 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
08:40:14 Join ender` [0] (krneki@foo.eternallybored.org)
08:40:40 Quit ender` (Read error: Connection reset by peer)
08:40:58 Join ender` [0] (krneki@foo.eternallybored.org)
08:41:00 Join flydutch [0] (~flydutch@host83-164-dynamic.15-87-r.retail.telecomitalia.it)
09:00
09:00:41 Join B4gder [0] (~daniel@rockbox/developer/bagder)
09:05:16 Join petur [0] (~petur@rockbox/developer/petur)
09:18:47 Quit JdGordon (Read error: Connection timed out)
09:23:01 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
09:23:01 Join DerPapst1 [0] (~DerPapst@p5099d40e.dip0.t-ipconnect.de)
09:30:35 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
09:43:21 Quit Zagor (Ping timeout: 245 seconds)
09:45:15 Join Zagor [0] (~bjst@46.35.227.87.static.tab.siw.siwnet.net)
09:45:15 Quit Zagor (Changing host)
09:45:15 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
09:58:03 Quit TheSeven (Quit: ChatZilla 0.9.86 [Firefox 3.5.8/20100202165920])
09:59:18 Quit 17SAALP2R (Ping timeout: 245 seconds)
10:00
10:06:31 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
10:18:30 Part flyasky ("Bye!")
10:25:02*gevaerts points Zagor to http://www.rockbox.org/irc/log-20100314
10:25:23Zagorthat's a big log. any particular time?
10:25:50gevaertsoops, sorry
10:26:11gevaertshttp://www.rockbox.org/irc/log-20100314#23:15:00
10:27:17***Saving seen data "./dancer.seen"
10:29:37Zagorit seems at least the block was lifted automatically. always look on the bright side! :-)
10:31:45 Join TheSphinX^ [0] (~cold@p54A5D2E1.dip.t-dialin.net)
10:36:36 Join planetbeing [0] (~planetbei@c-71-236-164-204.hsd1.or.comcast.net)
10:40:38CIA-5New commit by 03zagor (r25197): Require new rbclient. ...
10:46:01 Join funman [0] (~fun@rockbox/developer/funman)
10:49:51 Quit JdGordon (Ping timeout: 248 seconds)
10:58:59 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
11:00
11:00:17 Join LinusN [0] (~linus@rockbox/developer/LinusN)
11:01:45 Join m3dlg [0] (~m3dlg@bb-87-81-252-83.ukonline.co.uk)
11:17:58 Join TheSeven|Mobile [0] (~theseven@rockbox/developer/TheSeven)
11:19:14*TheSeven|Mobile wonders what could have caused that nano2g battery life decrease
11:21:44TheSeven|Mobilesome change during the last months seems to have caused a 40% higher battery current, but i can't think of any change that could have caused it
11:22:07 Quit JdGordon (Read error: Connection timed out)
11:23:44 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
11:23:45*TheSeven|Mobile doesn't really like to bisect that
11:28:57 Nick fxb__ is now known as fxb (~felixbrun@h1252615.stratoserver.net)
11:42:55 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:43:38 Join s3t [0] (~4d7ccaf9@giant.haxx.se)
11:43:40 Quit TheSeven|Mobile (Quit: TheSeven|Mobile)
11:48:37 Quit s3t (Quit: CGI:IRC (Ping timeout))
11:49:05 Quit m3dlg (Ping timeout: 256 seconds)
11:53:06 Join s3t [0] (~4d7ccaf9@giant.haxx.se)
11:57:18 Join Genre9mp3 [0] (~yngwiejo@dsl-88-218-63-169.customers.vivodi.gr)
11:57:19 Quit Genre9mp3 (Changing host)
11:57:19 Join Genre9mp3 [0] (~yngwiejo@rockbox/contributor/Genre9mp3)
11:57:38 Quit Genre9mp3 (Client Quit)
11:57:54 Join robin0800 [0] (~quassel@cpc3-brig8-0-0-cust436.brig.cable.ntl.com)
11:58:07 Quit s3t (Quit: CGI:IRC (Ping timeout))
12:00
12:08:37Tornefunman: yeah i will test clipv2 at some point :)
12:10:33funmanTorne: according to pbxy & saratoga the screen stops at "executing"
12:11:20funmanif you want to hack a bit, I would put some calls to _buttonlight_on() in system_init() / main()
12:13:24 Quit JdGordon (Read error: Connection timed out)
12:13:32 Join einhirn [0] (~Miranda@bsod.rz.tu-clausthal.de)
12:14:24 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
12:17:44 Quit Adubb (Read error: Connection reset by peer)
12:19:08 Join m3dlg [0] (~m3dlg@212.183.140.6)
12:19:18funmanor comment the end of system_init() (only executed in rockbox.sansa, not the bootloader)
12:20:24funmanhmm in crt0.S we use bl (when not in the bootloader) but apparently we rely on sp setup made by the bootloader
12:20:51funmanand we don't use the same memory layout between clipv2 and other models in the bootloader
12:21:32funmankugel: ^ perhaps it's why the fuzev2 build fails? what's the size of fuzev2 OF firmware block ?
12:21:57 Join Casainho [0] (~chatzilla@87-196-190-92.net.novis.pt)
12:26:51CIA-5New commit by 03funman (r25198): as3525*: setup stack pointer before we use it
12:27:19***Saving seen data "./dancer.seen"
12:27:22funmanTorne: (or anyone with a clipv2) could you try r25198 current build ? (no need to update bootloader)
12:28:24Tornenot at the moment, i'm at work and haven't even installed hte bootloader yet
12:28:31funmanoops I forgot something
12:29:21CIA-5New commit by 03funman (r25199): as3525* : make sure sp is setup in bootloader
12:40:04 Join watto [0] (~watto@193.203.81.165)
12:47:47CIA-5New commit by 03funman (r25200): lamp: invert backlight for Clipv2/Clip+ too
12:47:50 Quit m3dlg (Ping timeout: 252 seconds)
12:47:52CIA-5New commit by 03funman (r25201): Clip+ : make the bootloader not verbose by default
12:55:50 Quit robin0800 (Ping timeout: 245 seconds)
13:00
13:02:04 Join robin0800 [0] (~quassel@cpc3-brig8-0-0-cust436.brig.cable.ntl.com)
13:31:26CIA-5New commit by 03pamaury (r25202): fat: make fat_open more flexible by accepting the file==&dir->file and make fat_opendir also more flexible by accepting dir==parent_dir
13:32:06CIA-5New commit by 03pamaury (r25203): dir_uncached: fix opendir_uncached to have opened directories keep enough information to update metadata; this is necessary because of the way ...
13:33:06 Join mischasworld [0] (~quassel@193.174.158.81)
13:34:44 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
13:35:19 Join froggymana [0] (~187b533e@giant.haxx.se)
13:44:01 Quit ender` (Quit: The reward for work well-done is more work.)
13:47:42 Quit mt (Read error: Connection reset by peer)
13:47:48 Join mischa [0] (~quassel@193.174.158.84)
13:48:22 Join mt [0] (~mtee@rockbox/developer/mt)
13:48:22 Quit mischasworld (Ping timeout: 246 seconds)
13:50:20 Join Lss [0] (~Lss@cm48.omega219.maxonline.com.sg)
13:51:24pamaurygevaerts: see FS #11108 for usb audio
13:51:54 Join Farthen [0] (~chatzilla@e179237057.adsl.alicedsl.de)
13:51:55CIA-5New commit by 03uchida (r25204): UIsimulator: iAudio M3 remote keys add.
13:52:59 Join evilnick_ [0] (~evilnick@ool-457bccf5.dyn.optonline.net)
13:53:06linuxstbpamaury: Did you manage to decipher the pcm format?
13:54:52 Quit evilnick (Ping timeout: 256 seconds)
13:57:12 Join Genre9mp3 [0] (~yngwiejo@dsl-88-218-63-169.customers.vivodi.gr)
13:57:13 Quit Genre9mp3 (Changing host)
13:57:13 Join Genre9mp3 [0] (~yngwiejo@rockbox/contributor/Genre9mp3)
13:57:22 Quit Genre9mp3 (Client Quit)
13:58:24 Quit parafin (Remote host closed the connection)
13:58:37pamauryI made some tests and a early conclusion is that it is normal signed 16-bit pcm data. I believe the problem has to do with iso transfers: the core can't keep up and misses packets. But as packets are not aligned on samples, the result is...horrible
13:58:48gevaertspamaury: thanks! I'll have a look at it later
13:58:58 Join parafin [0] (parafin@paraf.in)
13:59:12pamaurygevaerts: the code is horrible, close your eyes when reading it ;-)
13:59:18pixelmadoes uchida's commit mean that now only the remote keys are simulated in an M3 sim and not the main target's anymore?
14:00
14:00:20pixelmaah no, I think I see now
14:00:21 Join ender` [0] (krneki@foo.eternallybored.org)
14:01:33 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:01:41*pixelma wonders about other targets with remote though
14:02:43funmanthey would require individual changes, no?
14:02:56 Quit JdGordon (Ping timeout: 276 seconds)
14:05:05pixelmawell, interesting would especially be the X5 and M5 since they use the same remote, the main targets button layout differs from the M3 though
14:10:43 Join adnyxo [0] (~aaron@adsl-065-013-002-216.sip.asm.bellsouth.net)
14:21:54funmanranmachan: there is likely a buffer overflow in recording code
14:23:09 Join evilnick__ [0] (~evilnick@ool-457bccf5.dyn.optonline.net)
14:26:02 Quit evilnick_ (Ping timeout: 256 seconds)
14:27:21***Saving seen data "./dancer.seen"
14:28:04 Join Schmogel [0] (~Miranda@p3EE21A31.dip0.t-ipconnect.de)
14:30:42 Quit kugel (Ping timeout: 246 seconds)
14:39:10funmanany reason why album_art.* is in the recorder/ folder ?
14:39:38linuxstb"recorder" historically also means "bitmapped targets"
14:39:58 Join kugel [0] (~kugel@rockbox/developer/kugel)
14:48:42 Join m3dlg [0] (~m3dlg@212.183.140.32)
14:49:10 Join panni_ [0] (hannes@ip-95-222-52-93.unitymediagroup.de)
14:55:43 Join jgarvey [0] (~jgarvey@cpe-065-190-069-073.nc.res.rr.com)
14:58:19 Part LinusN
15:00
15:03:50 Quit funman (Quit: free(random());)
15:14:32 Quit m3dlg (Ping timeout: 260 seconds)
15:20:13 Quit mischa (Ping timeout: 248 seconds)
15:23:57 Join Kitr88 [0] (~Kitr88@BSN-182-12-189.dial-up.dsl.siol.net)
15:25:00 Join mischasworld [0] (~quassel@193.174.158.84)
15:27:04 Quit mc2739 (Ping timeout: 240 seconds)
15:28:04 Quit Kitar|st (Ping timeout: 240 seconds)
15:28:32 Quit Kitr88 (Ping timeout: 260 seconds)
15:28:41 Quit robin0800 (Remote host closed the connection)
15:29:19 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
15:33:24 Join Kitar|st [0] (Kitr88@BSN-182-47-251.dial-up.dsl.siol.net)
15:37:48 Join robin0800 [0] (~quassel@cpc3-brig8-0-0-cust436.brig.cable.ntl.com)
15:41:34 Join toffe82 [0] (~chatzilla@12.169.218.14)
15:52:56 Join Horschti [0] (~Horscht2@xbmc/user/horscht)
15:54:42 Quit Horscht (Ping timeout: 246 seconds)
16:00
16:12:29 Quit toffe82 (Ping timeout: 248 seconds)
16:14:50 Join funman [0] (~fun@rockbox/developer/funman)
16:18:24funmanranmachan: can you enable logf and still be able to read the logf buffer after starting recordign? (perhaps you need to return early from enc_set_parameters)
16:26:10amiconnfunman: It would be better to check for HAVE_NEGATIVE_LCD in lamp.c
16:26:56funmanright
16:27:25***Saving seen data "./dancer.seen"
16:28:42CIA-5New commit by 03funman (r25205): lamp: use HAVE_NEGATIVE_LCD
16:31:44 Nick Strife1989 is now known as Strife89 (~michael@adsl-154-2-161.mcn.bellsouth.net)
16:32:56 Quit robin0800 (Remote host closed the connection)
16:34:43pixelmawhat does the HAVE_LCD_INVERT do then?
16:35:07funmani wanted to use CONFIG_LCD == LCD_SSD1303 instead of defined(SANSA_CLIP) || defined(SANSA_CLIPV2) || defined(SANSA_CLIPPLUS), but the blue/yellow screen isn't tied to the controller
16:35:17funmanpixelma: reset it to the default state if it was changed in the settings
16:36:09pixelmahmm... probably the OLED display is not really "inverted" in that sense
16:36:23pixelmaalthough it looks like it is
16:36:51funmanwell it just has no backlight
16:38:39amiconnHAVE_LCD_INVERT indicates a feature, HAVE_NEGATIVE_LCD indicates a property
16:39:22pamauryIsn't it known bug that the database build hang when there are no media files on the dap ?
16:39:36 Join captainkewllllll [0] (~2669ecc2@gateway/web/freenode/x-wzelxkwygqadezvu)
16:40:01 Quit captainkewllllll (Client Quit)
16:45:02kugelpamaury: yes
16:46:51pamauryand nobody tried to fix it ? Are they FS report about that ?
16:47:13 Join Adubb [0] (~aldubuc@xplr-ts-t11-208-114-159-122.barrettxplore.com)
16:48:48pixelmapamaury: yes
16:52:13 Quit GodEater (Ping timeout: 256 seconds)
16:53:11 Quit Strife89 (Read error: Connection reset by peer)
16:53:40 Join Strife89 [0] (~michael@adsl-154-2-161.mcn.bellsouth.net)
16:54:36 Join splashote [0] (~fabian@i577BD6B5.versanet.de)
16:55:10splashotehey there, how can i deactivate the "mouse mode" it seems as if I can't reinstall rockbox on my e200 due to that
16:55:47saratogayeah theres an option for it
16:57:23pixelmaUSB HID option in the General Settings > System
16:57:51splashotesaratoga: ok, i found it in the configuration. after all i think it's not responsible for the error in during the reinstallation
16:58:06 Join m3dlg [0] (~m3dlg@212.183.140.18)
16:58:11 Join ender1 [0] (krneki@foo.eternallybored.org)
16:58:27saratogayeah i think its only a problem on some buggy versions of macos
16:59:15splashotethe utitlity says that i has no direct access to the "disk"... why could that be? do i have to start the player with the original firmware in order to reinstall rockbox?
16:59:22 Join GodEater [0] (~bibble@cl-711.lon-02.gb.sixxs.net)
16:59:23 Quit GodEater (Changing host)
16:59:23 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
16:59:36funmansplashote: make sure you run with administrator privileges
17:00
17:00:48splashotehm, the utility doesn't start when i try to run it as an admin (running ubuntu here)
17:01:00 Quit ender` (Ping timeout: 260 seconds)
17:01:00 Nick ender1 is now known as ender` (krneki@foo.eternallybored.org)
17:01:02splashotei don't think i started it as an admin when i installed it..
17:01:40 Join toffe82 [0] (~chatzilla@12.169.218.14)
17:01:41splashotei'll try with the original firmware running
17:01:42 Quit Casainho (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115133306])
17:01:49saratogai doubt it can patch your player without root access
17:01:51funmansudo ./rbutilqt # should do the trick
17:02:24saratogabut if you can mount the device in rockbox USB mode I wonder what exactly you're hoping to accomplish in the OF?
17:05:02splashotei messed up the config and just want to reinstall rockbox. the sudo stuff works, thanks!
17:05:19gevaertsYou can just reset the configuration then
17:05:24Tornereinstalling rockbox won't reset the configuration
17:05:38Tornejust rename or delete the configuration file
17:05:56 Join komputes [0] (~komputes@ubuntu/member/komputes)
17:06:14splashoteok,ill do so. I'm checking out the podcast tagging function. where to run " #! rockbox/tagbrowser/2.0 "?
17:06:31saratogaand if you just want to reinstall rockbox (which won't reset anything) you can just unzip a new build to the player without even needing sudo
17:07:42splashotesaratoga: great!
17:07:51ranmachanfunman: I enabled DEBUG and LOGF using configure
17:08:11ranmachanNow it doesn't crash, instead it says 'The disk is full'
17:08:26ranmachanAnd Debug log is completely empty?!
17:08:36splashoteignore my last question, i figured it out.
17:08:44 Quit lyngaas (Ping timeout: 245 seconds)
17:08:56funmanyou need to add #define LOGF_ENABLE in pcm_record.c before including logf.h
17:10:48 Join CGL [0] (~CGL@190.207.236.167)
17:12:40splashotethere is no tagnavi_custom.config in /.rockbox
17:13:33funmanhum I see some evidence that iram is bigger on Clip+ : OF accessing 0x8106xxxx
17:17:50splashotei want to creat a podcast tag but the /.rockbox/tagnavi_custom.config file is nowhere to find.
17:20:18Torneof course not, it's the *custom* tagnavi configuration file. You create it to customise the navigation..
17:20:43ranmachanthread pcmrec start\000audio_close_recording\000audio_close_recording done\000audio_init_recording\000audio_init_recording done\000audio_close_recording\000audio_close_recording done\000audio_init_recording\000audio_init_recording done\000audio_set_recording_options\000audio_set_recording_options done\000enc_set_parameters\000enc sampr:8000\000chunk size:4112\000resbytes:0\000num chunks:1044054\000enc size:4293150048\000fnq files:20", '\000' <repeats 1599
17:20:45splashoteTorne: ok, did so and it works. sry, new to rockbox configuration.
17:22:04funmanranmachan: can you logf bufsize ? the enc size looks 'a bit' high
17:23:18ranmachanAFAICS "enc size" _is_ bufsize :)
17:23:32ranmachanAh, wait
17:23:55ranmachanIt's a local variable called bufsize
17:27:03 Quit m3dlg (Ping timeout: 240 seconds)
17:27:16ranmachanOk, I added logf for all variables used to calculate bufsize, let's see...
17:31:08 Join Darkknight512 [0] (~Darkknigh@CPE00212968356c-CM00186845dd46.cpe.net.cable.rogers.com)
17:34:55ranmachanfunman: (enc_buffer - pcm_buffer) is too big. enc_buffer=0x30356700, pcm_buffer=0x301518E0
17:36:48 Quit B4gder (Quit: It is time to say moo)
17:38:30 Quit Galois (Quit: Leaving)
17:39:20 Join Galois [0] (djao@efnet.math.uwaterloo.ca)
17:39:39ranmachanfunman: #define PCM_NUM_CHUNKS 256 /* Power of 2 */
17:39:40ranmachan#define PCM_CHUNK_SIZE 8192 /* Power of 2 */
17:39:43ranmachanThats 2MB! :)
17:40:11 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
17:41:55funmanno wonder it's too big then
17:42:35*GodEater aims a hefty kick at his crappy router
17:43:12funmanplayback.c:audio_get_recording_buffer() : buffer_size = end-audiobuf is known at compile time, we should check at compile time if it fits in memory
17:43:24ranmachanNow trying with 8 chunks :)
17:44:19ranmachanIt doesn't crash and I see a volume bar that responds when I hit the mic
17:45:51ranmachanfunman: IMHO it would also be nice to use write-protected mmu pages for code and read-only data :)
17:46:19funmanyeah right
17:46:45funmanapp.lds would be a bit more complex though
17:47:26funmanhm well not necessarily, we'd only need 1 more section
17:47:34TorneYou'd have to 1MB align .data then though
17:47:38Torneto keep it as section mappings.
17:47:53funmanhm not necessarily ?
17:48:27ranmachanarm922 has page sizes 1K, 4K, 64K and 1M AFAIR from the datasheet
17:48:47TorneYes, but at the moment we don't ahve any page tables and use only section mappings.
17:48:52funmanhm i thought it was only 1MB, and wanted to add a translation like is done for uncacheable addresses
17:49:12funmanbut then if the memory is accessed through a writeable section we can't prevent overwrite
17:49:33TorneYah, to do it with sections you'd have to have 1MB of the actual memory read-only
17:49:48funmanranmachan: i'm looking at c200 manual but don't see how to start recording
17:49:49Torneso you'd need to make proper second level pagetables.
17:49:58Tornewell, one.
17:50:54funmani'm not sure how we can add a clean build-time check for recording buffer size in pcm_record.c
17:52:29ranmachanfunman: I have 'Recording' as a menu option directly below 'Settings'
17:53:09funmanon fuze i just press play: 'length' and 'size' will increase
17:54:02 Quit pamaury (Quit: Quitte)
17:54:18ranmachanIf I press the record button it starts recording
17:54:39ranmachanBut the actual writing to FS doesn't seem to be working properly
17:54:45funmanoops
17:55:27funmani can try on clip(v1) to rule out any c200v2 specific problem
17:55:58ranmachanTime is erratically and at size about 100K it says 'The disk is full. Press PREV to continue"
17:56:18ranmachanOf course 8 chunks is a lot less than 256...
17:56:31ranmachanMaybe it just isn't built to handle a buffer that small
17:57:06funmanit's gonna be built to do whatever we ask it to do! show this code who's the boss!
18:00
18:00:45 Part splashote
18:01:15 Quit petur (Quit: work->home)
18:03:10 Quit Tomis (Read error: Connection reset by peer)
18:03:52ranmachanOk, 'The disk is full' really means AUDIO_STATUS_ERROR...
18:04:38*funman remembers seeing something like that, but not the context
18:04:54 Join Tomis [0] (~Tomis@70.134.89.59)
18:06:14 Join perfectdrug [0] (~marko@p5B0EC362.dip.t-dialin.net)
18:12:37 Join lyngaas [0] (~staale@19.81-167-149.customer.lyse.net)
18:17:24ranmachanError seems to be PCMREC_E_CHUNK_OVF
18:19:01funmanwhat if you disable DEBUG ? (one never knows)
18:21:42ranmachanThis is already with DEBUG disabled
18:22:12ranmachanI gather chunk overflow means that it didn't manage to write out in time
18:22:31ranmachanThe watermark calculation is totally not for buffers this small
18:22:59ranmachanlow wmk: 8
18:22:59ranmachanwarning: panic < low
18:22:59ranmachanflood at: 8
18:22:59DBUGEnqueued KICK ranmachan
18:22:59ranmachanwarning: low 'write at'
18:22:59ranmachanwrite at: 8
18:23:08funmanerrors |= PCMREC_E_CHUNK_OVF is only in a #ifdef DEBUG block
18:23:16funmanhm
18:24:11funmantry lowering the watermakr by the same amount you lowered the number of chunks?
18:27:26***Saving seen data "./dancer.seen"
18:28:31 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
18:30:05 Join bertrik [0] (~bertrik@ip117-49-211-87.adsl2.static.versatel.nl)
18:32:43 Join MethoS- [0] (~clemens@134.102.106.250)
18:34:38 Quit TheSeven (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
18:38:01funmanhum there's not much differences between sound code of Clipv1 & Clip+ but I still can't get sound
18:38:14funmanperhaps we need to power explicitely the headphones?
18:38:52 Join domonoky [0] (~Domonoky@g229154033.adsl.alicedsl.de)
18:38:52 Quit domonoky (Changing host)
18:38:52 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:41:37 Quit Tomis (Read error: Connection reset by peer)
18:42:17 Join Tomis [0] (~Tomis@70.134.101.101)
18:42:24bertrikfunman, try it and see what happens :)
18:43:44funmani mean a GPIO pin could control power ^^ i'm not gonna plug 220V on my headphones :)
18:44:56bluebroth3rBagder: can you add rbutil 1.2.5 to FS?
18:46:40 Join r0b- [0] (~nnscript@adsl-76-235-216-125.dsl.klmzmi.sbcglobal.net)
18:46:51r0b-how well does rockbox perform on the e200v2?
18:46:56r0b-just like the v1?
18:48:27 Quit kugel (Ping timeout: 276 seconds)
18:48:32funmanr0b-: yeah, except e200v2 USB mode is handled by the OF
18:48:37 Join anewuser [0] (anewuser@unaffiliated/anewuser)
18:48:50r0b-so doom and mpeg perform about the same?
18:49:23r0b-funman may i pm?
18:49:24funmani was told doom crashes
18:49:27funmanno
18:49:48r0b-unfortunatly i had to order a new e200
18:50:00r0b-i dont know if its gonna be v1 or v2
18:50:05bertrikI guess doom runs much faster on an e200v2 than on an e200v1 because its cpu runs 3 times faster
18:51:29pixelmapamaury: http://www.rockbox.org/tracker/task/9093 (I don't know if the attached patch is the correct solution though)
18:52:50pamaurypixelma: I'll have a look at it, thanks
18:53:24gevaertspamaury: maybe for isochronous we could have an api where the class driver registers a callback with the controller driver, which gets used to notify the class driver about the status of the previous transfer and gets the next buffer to transfer at the same time?
18:53:35saratogadoom probably runs badly on the e200v2 due to lack of RAM
18:53:43 Join m3dlg [0] (~m3dlg@212.183.140.6)
18:53:43saratogai expect it crashes after the first level like on the fuze
18:55:15r0b-well i dont play doom often :P
18:56:22r0b-saratoga care if i pm you?
18:57:00pamaurygevaerts: I was thinking of a similar solution but more precisely I was thinking about a static schedule. Something like [for the receive part at leats] (1) allocate N slots (a slot=a packet=a trasnfer). (2) for each slot, assigna length and a buffer. (3) start the whole thing in a ring buffer way. My idea is that we shouldn't change the buffers once the stream is launched. Just have the usb controller do everything and just call the completio
18:58:31funmanr0b-: well if you have a question just ask it here
19:00
19:02:33pixelmabluebroth3r: I got a working deutsch.voice on my OndioFM now on the 10.4 PPC :)
19:02:59 Quit CGL (Ping timeout: 246 seconds)
19:03:13pixelmashall I test it on Intel too, or with a speex target?
19:04:23r0b-well its not DIRECTLY related to rockbox
19:04:59 Join CGL [0] (~CGL@190.207.236.167)
19:06:00bluebroth3rpixelma: nice :)
19:06:27bluebroth3rpixelma: speex would definitely be interesting as that got endianess fixes too.
19:06:48pixelmafile size is still a bit larger than I would hope for but you'll get something working and bearable wrt quality
19:07:57 Quit ender` (Quit: MCSE = Moron Claiming System Expertise, Must Call Someone Else, Minesweeper Champion, Solitaire Expert, or Must Consult Someone Experienced?)
19:08:55CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
19:08:55*bluebroth3r really hates people putting multiple issues in one tracker task :(
19:09:20 Nick bluebroth3r is now known as bluebrother (~dom@rockbox/developer/bluebrother)
19:09:33 Quit pamaury (Quit: Page closed)
19:10:07 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:10:11r0b-any rough guess on when the AMS v2 models Clipv2 Fuzev2 Clip+ will be rockboxable?
19:10:50Battousair0b-: follow this page: http://www.rockbox.org/wiki/SansaAMS
19:11:04Battousaithey update it regularly as the ports progress
19:11:25r0b-nice
19:11:42*bluebrother wonders who "they" is :P
19:11:50Battousaithey is them
19:11:57 Nick fxb is now known as fxb__ (~felixbrun@h1252615.stratoserver.net)
19:11:59r0b-they is they
19:12:50r0b-the Clips are nice tiny mp3 players
19:15:02CIA-5New commit by 03bluebrother (r25206): Fix manual download links. ...
19:21:35 Join ender` [0] (krneki@foo.eternallybored.org)
19:25:50 Quit adnyxo (Read error: Connection timed out)
19:26:44 Join adnyxo [0] (~aaron@adsl-065-013-002-216.sip.asm.bellsouth.net)
19:27:59ThomasAHfunman: Is disabling the caches still needed for clip+?
19:28:29funmanno
19:28:55 Join stripwax [0] (~Miranda@87-194-34-169.bethere.co.uk)
19:29:00ThomasAHfunman: good, I'll try it as soon as I can :)
19:29:22funmanyou can get the current build at http://build.rockbox.org/data/rockbox-sansaclipplus.zip
19:30:58ThomasAHfunman: should I rather build myself to test if building works with my toolchain or rather the zip so we all test exactly the same binary?
19:33:58funmanlike you prefer
19:35:15ThomasAHfunman: ok, I'll build myself then ... but now I have to leave
19:37:13bluebrotherpixelma: wrt file size: did it decrease with the new version?
19:39:58 Quit CGL (Ping timeout: 245 seconds)
19:44:11pixelmaa bit but not much I think, I still have to resample to the lowest possible to get a reasonable file size on the PPC (haven't tested for differences yet) with the Infovox Klaus and a still understandable speed
19:44:59 Quit antil33t (Ping timeout: 258 seconds)
19:45:16saratogaerror: expected identifier or â(â before â{â token
19:45:25saratogathe line in question: STIN void XPROD32(ogg_int32_t a, ogg_int32_t b,
19:45:35saratogaSTIN isn't the problem, ogg_int32_t seems ok
19:45:53funmansaratoga: use gcc -E to check preprocessor output
19:47:39funmansaratoga: the Clip+ has twice the IRAM of Clipv1 (640kB)
19:47:43 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
19:48:43 Quit DerPapst1 (Quit: Leaving.)
19:49:04 Join antil33t [0] (~Mudkips@203-184-54-232.callplus.net.nz)
19:49:22saratoganice
19:49:30saratogawow there is no mention of that function at all
19:49:38saratogaheh because i defined a macro with the same name before
19:49:39saratogaawesome
19:51:19bertrikfunman, nice!
19:51:32 Quit m3dlg (Ping timeout: 240 seconds)
19:51:59saratogaif its actually fast this time it'll be nice
19:52:09saratogaotherwise i guess we'll just have a larger codec buffer
19:52:14funmani'm goign to try
19:52:29funmansaratoga: 640kB ought to be enough for every codec ?
19:53:33saratogayeah I think so
19:53:46saratogaeventually i want to reduce all targets to have just 512KB of codec ram
19:54:02saratogaonly AAC and a handful of vorbis files ever need more then that now
19:54:12pixelmabluebrother: in other news... I got a working voice file on my M5 now. But with default settings, the file size is twice the size of the slightly older file I still had there
19:54:37pixelmathat's still on the 10.4 PPC box
19:54:40 Quit Farthen (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
19:55:21 Join Farthen [0] (~chatzilla@e179237057.adsl.alicedsl.de)
19:55:32saratogaAMS's intention was probably that you put all the codec and buffer in the IRAM and put the RAM into self refresh mode during much of playback (probably until you needed to buffer from the file system)
19:55:40saratogaat least that'd be my guess
19:55:47bluebrotherpixelma: ok, that's great −− at least the code works correctly now.
19:56:10bluebrotherfor speex you could try narrowband mode. That will downsample the clips to 8kHz instead of 16kHz
19:56:44bluebrothermaybe your old file is using narrowband. At least that would be an explanation for the doubled size
19:59:14funmansaratoga: that's not the case for clipv2 tho
19:59:41funmanhm well it might, now we knwo that the iram is large enough for the OF firmware block to fit
19:59:55saratogawell who knows what sandisk actually did with it :)
20:00
20:00:01saratogathey may be as clueless as us
20:00:34 Join dfkt_ [0] (~dfkt@unaffiliated/dfkt)
20:01:02 Quit mischasworld (Ping timeout: 240 seconds)
20:03:01 Join Luca_S [0] (~57025514@giant.haxx.se)
20:03:29funmanit seems IRAM is slower to read than DRAM (caches disabled)
20:03:55 Quit dfkt (Ping timeout: 248 seconds)
20:04:40funmanand perhaps a bit slower to write
20:05:20funmanreading/writing the same amount of data: 14 ticks to read dram, 14 ticks to write dram, 17 or 18 cycles to read iram, 14 or 15 cycles to write iram
20:05:53*gevaerts is going to be annoying
20:05:58gevaertsHow many cycles per tick?
20:06:13funmans/cycles/ticks/
20:06:32funmananyway the test was with caches enabled (didn't copy the new rockbox.sansa)
20:07:08 Join planetbeing_ [0] (~planetbei@c-71-236-164-204.hsd1.or.comcast.net)
20:09:21 Join Stephen__ [0] (~S@86.45.50.221)
20:09:28funmanwith caches disabled the difference is hardly noticeable (10 times slower)
20:09:40saratogafunman: do you know what the clocks are?
20:09:47funmannope
20:11:10funmanwe should test speed again when we know how memory/pclk are clocked
20:11:24CIA-5New commit by 03funman (r25207): as3525v2 has twice the as3525v1 IRAM (verified on Clip+)
20:12:03 Quit planetbeing_ (Quit: planetbeing_)
20:14:29saratogafunman: fwiw knowing the exact memory timings makes optimizing codecs later on easier
20:14:33saratogaso its very nice to know
20:14:40saratogahttp://duke.edu/~mgg6/rockbox/tremor_ffmpeg_imdctv4.patch
20:14:51saratogaif anyone else wants to look at our tremor upstream merge
20:20:51saratogaa couple lines at the top of one file in that patch are borrowed from ffmpeg
20:21:09saratogathey're absolutely trivial, so if I describe what they do could someone pastebin the c code
20:21:54kugelthe new mdct in C-only is faster than the old tremor one with ASM-only
20:22:08saratogayes, at least one some targets
20:22:25saratogacoldfire and PP i think
20:22:42kugelawesome
20:25:53saratogaseriously someone who hasn't looked at the code pastebin storing an integer that 2 power of another integer
20:26:48funmansaratoga: what's the function prototype ?
20:27:07funmani looked at the pastebin but only browsed, didn't read it
20:27:16 Join Chronon [0] (~quassel@c-67-171-212-229.hsd1.or.comcast.net)
20:27:30***Saving seen data "./dancer.seen"
20:27:35gevaertssaratoga: something like a=2^b;?
20:27:46saratogafunman: void ff_imdct_half(unsigned int nbits, ogg_int32_t *output, const ogg_int32_t *input)
20:27:50funmangevaerts: xor isn't pow() !
20:27:56saratoga2^nbits but preferably faster . . .
20:28:17saratogathen i need nbits-1, nbits-2 nibits-3
20:29:41saratogahuh pbxy got the radio to work
20:30:01 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
20:30:37saratogastripwax: ping
20:32:02funmansaratoga: you can't use the ffmpeg bits in tremor?
20:32:23saratogai want bsd license
20:33:07funmanyou could ask ffmpeg author to relicense this bit
20:34:39 Quit flydutch (Quit: /* empty */)
20:34:46 Join dfkt [0] (dfkt@unaffiliated/dfkt)
20:35:35 Join cfp_ [0] (~cfp@ip-247.net-81-220-136.rev.numericable.fr)
20:35:40pixelmabluebrother: I even got a speex voice file on the 10.4 Intel now, it should be possible to get talk clips too but I'll try :)
20:36:20ChrononTrying to get build client set up. I keep getting "Server message: Fatal build error: Command not found. You have been temporarily disabled." Where should I look to determine what hasn't been set up properly? I seem to have all of the necessary bits installed.
20:36:34*gevaerts isn't sure if he understands what this function should really do
20:37:08pixelmabluebrother: autodetection with two targets connected gave me a weird result though - selected device was one, mount point the other
20:38:26 Quit dfkt_ (Ping timeout: 260 seconds)
20:39:37*gevaerts clarifies: he is sure that he doesn't know what saratoga means
20:39:41domonokypixelma: autodetection could need a improvement. it just takes what if finds first.
20:40:03 Join petur [0] (~peter@78-21-206-95.access.telenet.be)
20:40:03 Quit petur (Changing host)
20:40:04 Join petur [0] (~peter@rockbox/developer/petur)
20:40:09domonokythere is a unfinished patch from me in the tracker which can deal with multiple connected devices.
20:40:32 Quit kugel (Ping timeout: 240 seconds)
20:41:01funmansaratoga: http://pastie.org/870859
20:41:32 Join kugel [0] (~kugel@rockbox/developer/kugel)
20:41:52gevaertsfunman: why not *output<<=nbits?
20:42:28gevaertsor *output=*input<<nbits
20:44:25funmanhm right
20:44:33 Quit r0b- (Ping timeout: 268 seconds)
20:44:52funmansaratoga: void ff_imdct_half(unsigned int nbits, ogg_int32_t *output, const ogg_int32_t *input) { *output = *input << nbits; }
20:45:09funman+ © gevaerts
20:45:31pixelmabluebrother: the speex talk clips work too now on the Intel \o/
20:45:51gevaertsThe function should probably get another name though. That ff_ bit looks suspicious :)
20:51:18kugelgevaerts: is that even copyrightable? that line seems *too* trivial
20:51:42gevaertskugel: I don't know. I also don't really know if it matches saratoga's description
20:52:06kugelwell, I also thought about a simple shift when reading his description
20:52:10 Quit FOAD (Ping timeout: 246 seconds)
20:52:41 Join Lear [0] (chatzilla@rockbox/developer/lear)
20:53:55 Join FOAD [0] (~dok@dinah.blub.net)
20:56:03 Quit Luca_S (Quit: CGI:IRC (EOF))
20:56:25 Join CGL [0] (~CGL@190.207.236.167)
20:56:28 Quit kugel (Remote host closed the connection)
20:57:34bluebrotherpixelma: great! Autodetection is known to be broken when two devices are connected. I was planning to look into this, but there's always this time issue :(
21:00
21:00:52 Quit stripwax (Quit: http://miranda-im.org)
21:00:59 Quit hd (Ping timeout: 248 seconds)
21:01:07 Quit CGL (Client Quit)
21:02:13 Join jd [0] (~jd@modemcable207.134-202-24.mc.videotron.ca)
21:02:13 Quit jd (Changing host)
21:02:13 Join jd [0] (~jd@Wikipedia/HellDragon)
21:03:38 Join punkt [0] (~punkt@dslb-092-077-140-194.pools.arcor-ip.net)
21:05:17 Join r0b- [0] (~nnscript@adsl-76-235-216-125.dsl.klmzmi.sbcglobal.net)
21:09:14cfp_hello everyone! I just noticed that my alarmclock plugin has a serious flaw in the way in handles times (disturbing for an alarmclock, isn't it), which means that it may fail to ring in many situations. I've posted a patch that fixes it on FS #7505. Could somebody have a look at it and possibly commit it? Thanks!
21:10:52 Quit pixelma (Disconnected by services)
21:10:52 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
21:10:58 Quit amiconn (Disconnected by services)
21:11:00 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
21:11:09 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
21:11:25cfp_funman: Hello :) Could you please have a look at the patch? The calculation is wrong when the alarm is set on the next day
21:11:26 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
21:11:42funmancfp_: must be another FS# , this one is about gigabeat flashwriter
21:11:48cfp_wooops
21:12:06cfp_my bad, FS #11110
21:12:40cfp_I was adding 24 hours to the time when there was an overlap, which resulted in the alarm never ringing (ahem :s)
21:13:34gevaertscfp_: let me guess, you planned to report this earlier but you didn't wake up on time ;)
21:13:52funmanis there something wrong: overlap depends on rem_seconds() and rem_seconds() depends on overlap?
21:15:45cfp_gevaerts: exactly :)
21:16:02cfp_funman, not really. The initial value of overlap is determined by rem_seconds
21:16:12cfp_when you launch the plugin, overlap = false
21:16:24cfp_then if its say 10 PM, and you set it to 6 PM
21:16:51funmancfp_: you could just do if (seconds > 0) overlap = false; in rem_seconds()
21:17:00 Join DerPapst [0] (~DerPapst@p4FE8FEC7.dip.t-dialin.net)
21:17:00cfp_rem_seconds = 3600 * (6-22)
21:17:19cfp_sure
21:17:32funmanhm and the value of overlap could be modified in rem_seconds only
21:17:43cfp_funman, how?
21:18:00cfp_for the main loop is not executed when waiting
21:18:28funmanret = XXX; overlap = ret <= 0; return XXX; ?
21:18:55funmansorry, i must leave
21:18:56 Quit funman (Quit: free(random());)
21:19:38cfp_hmmm but then you would return a negative value, therefore trigerring the alarm
21:19:43cfp_*triggering
21:21:01Fartheni didn't look at the code but why don't you set overlap to false on 00:00h ? or do you?
21:22:19cfp_Farthen, I can't be sure that my code will run precisely at 00:00
21:22:29cfp_but the current solution is pretty neat imo
21:23:05cfp_if the next alarm is in the past, then set a flag to add 24 hours to the remaining time
21:23:16cfp_as soon as it isn't in the past anymore, remove the flag
21:23:46cfp_"isn't is the past anymore" meaning that alarm.hour - current_time.hour > 0
21:25:05*Farthen is wondering what happens if it is 23:59 and you set the alarm to 00:00h
21:25:44 Quit Zagor (Quit: Clint excited)
21:26:04Farthenor even if it is any time and you set the alarm to 0:00h
21:26:09cfp_it rings in 59 seconds
21:27:17cfp_the overlap flag is always set in this case
21:27:33cfp_I'll be back in an hour
21:31:00 Join froggyman [0] (~sopgenort@pool-72-69-76-103.chi01.dsl-w.verizon.net)
21:31:15 Part watto
21:36:25 Join FOAD_ [0] (~dok@dinah.blub.net)
21:37:45bertrikWas funman joking about the 640kB IRAM thing? His last commit sets IRAM at 1 MB, not 640 kB, for the as3525v2
21:40:03 Quit FOAD (Ping timeout: 256 seconds)
21:40:04 Nick FOAD_ is now known as FOAD (~dok@dinah.blub.net)
21:40:38 Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
21:46:22TheSevenliar: I have been trying to track down the additional battery consumption, but without success so far
21:46:49TheSeventhe battery current on builds from november 09 and immediately before committing boosting is identical
21:47:02TheSevenso there are 2 possibilities:
21:47:44TheSeven1. something affected some additional battery consumption patches that aren't in svn
21:48:00TheSeven2. it's actually iloader who is at fault here
21:49:01 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
21:52:16pamaurygevaerts: what do you think of what I told you two hours ago ?
21:53:11liari tried a current build from the build.rockbox.org and the battery consumption was ~42mA(backlight on, usb chip off) so i dont think its because of local changes
21:59:05gevaertspamaury: you're probably right. No sense in making this too dynamic
21:59:29gevaertspamaury: were you planning to arrange the TDs in a looping linked list?
21:59:49pamauryyes, that's what I'm thinking of
22:00
22:00:01gevaertsthat should work I guess
22:00:45pamauryThe only tricky point is allocation. We probably don't want to driver to be responsible for allocation but on the other hand, it might need special memory section or alignement, so I don't know
22:01:30pamauryAn idea about that precise point ?
22:02:28gevaertsSomething like USB_DEVBSS_ATTR (as used in usb_storage.c) should work
22:02:55gevaertstargets that need nothing special can just define that as empty
22:03:08gevaertsactually, I'd just use USB_DEVBSS_ATTR
22:05:26pamauryso there would an api like (a) usb_drv_allocate_iso_slots(int nb_slots, <the that hold a transfer descriptor> *tds); (b) usb_drv_fill_iso_slot(int slot, int length, unsigned char *buffer) (c) usb_drv_set_iso_endpoint_mode(int mode) where mode=START|STOP|FLUSH
22:06:26gevaertsAlso a callback that tells you which slot is next and one which slot has just been handled
22:07:02gevaertsI'm not sure if I'd export the tds
22:07:51pamaurywho would allocate them then ? About the callback I'm not sure: each slot should have a different buffer so that the normal callback can be called. The slot are processed in order: 1, 2, ..., N, 1, 2 ...
22:08:40gevaertshm, yes, the normal callback will work, at least for completion
22:08:49pamauryWe don't need to export the tds. We can have something like USB_DRV_ISO_SLOT_SIZE and the users allocate it by unsigned char iso_slots[nb_slots*USB_DRV_ISO_SLOT_SIZE];
22:09:02pamauryThis should work for the receive part
22:09:14gevaertshm, will it? Don't you want the class driver to be told which buffer exactly completed?
22:09:33gevaertsWe could extend it of course
22:09:48pamauryah yes, I forgot that the completion handler doesn't give the buffer pointer !
22:10:15gevaertsthere's no real reason not to give it I guess
22:10:22pamauryindeed
22:11:25gevaertsI'm really a bit reluctant about having TDs outside of the actual driver, even as opaque objects...
22:12:08pamauryI share your opinion but how would the driver allocate them ?
22:12:20gevaertsit can't. That's the problem
22:12:34 Quit Lear (Quit: ChatZilla 0.9.86 [Firefox 3.6/20100115144158])
22:13:10gevaertsBut I'm also not sure if all hardware drivers use a similar concept, i.e. if it's even possible to assume that a fixed size block per ISO slot will work
22:14:37pamauryMost hardware use a concept similar to QHs and TDs no ? I mean, this concept is in all hosts and many devices. What about the other usb devices that rockbox supports ?
22:15:00gevaertsFrom what I understand, this model really isn't common in devices
22:15:46gevaertsA lot of chips use a FIFO based system, but I don't have any idea how exactly those handle isochronous
22:15:58gevaertsThey probably need a driver-defined data structure
22:16:15 Quit cfp_ (Quit: Quitte)
22:16:59gevaertsI think they have enough buffer to store one (or a few) packets per endpoint, and when that's sent out the driver refills the packet buffer
22:17:40pamauryBut can't we assume a fixed size per slot ? I mean, if the hardware manages to do it, there is not reason not to manage to do it in software. Actually, only a few fields are necessary, and it's even simpler if everything is done in software.
22:18:27pamauryHow many different usb drivers are there in rockbox ?
22:18:28gevaertsSo you'd have USB_DRV_ISO_SLOT_SIZE and USB_DRV_ISO_SLOT_ATTR defines?
22:18:34gevaertsfour or five
22:18:39 Quit TheSphinX^ (Quit: XChat)
22:18:45pamauryyes
22:19:13gevaertsThere's ARC, tcc, nano2g, mr500, and onda. The ZVM also has one, but I'm not sure if it still works
22:19:33pamaurydo they support iso transfers ? is it even known ?
22:19:49gevaertsIf we go that way, shouldn't we generalise the driver-allocates-TD concept for all transfer types?
22:19:50 Join pbxy [0] (~pb@2001:5c0:1000:b::59a7)
22:20:06gevaertsI don't know which ones support iso
22:20:55pamauryhum, it's true that the driver allocate td concept is interesting. It would also allow other types of scheduling to be implement nicely, like queueing, if necessary
22:21:08pamauryDo we have datasheets for them ?
22:21:16CIA-5New commit by 03bluebrother (r25208): Fix bootloader files vanishing after installation (FS #11086). ...
22:21:28bluebrotherdomonoky: around?
22:21:40domonokyjup
22:22:07bluebrotherany thoughts on a new rbutil release? The latest ones had various bad behaviour.
22:22:36bluebrotherlike the issue I've just fixed. I assume behaviour of Qt changed in that aspect (or it has always been a bug that was fixed in Qt ...)
22:22:58gevaertsPeople have datasheets, yes, although in some cases they'll be for different SoCs that happen to share the USB bit
22:23:28domonokybluebrother: no problem with a new release.
22:24:18bluebrotherwe really need a way to test at least parts of the functionality. The only big question is: how (without requiring targets)?
22:25:36gevaertsTheSeven: the nano2g uses the s3c6400x datasheet for usb, right?
22:25:55TheSevenyes
22:26:37TheSeven(that's why i called the driver usb-s3c6400x.c, back then i didn't know that this was an OTG from synopsys)
22:26:56pamaurydo you know if it supports iso transfers ? Is the datasheet freely available ?
22:27:32***Saving seen data "./dancer.seen"
22:27:37TheSevenhttp://www.ebv.com/fileadmin/products/Products/Samsung/S3C6400/S3C6400X_UserManual_rev1-0_2008-02_661558um.pdf
22:27:53TheSeven"samsung confidential", but nobody seems to care.
22:28:01bluebrotherdomonoky: oh, and I was thinking about splitting the source part out of rbutilqt.pro into a separate pri file as it's getting quite crowded.
22:28:16TheSevenand yes, it does support iso transfers
22:28:22TheSeveni haven't done anything with them yet though
22:28:24domonokybluebrother: good idea.
22:28:30TheSeventhe driver also doesn't support them yet
22:29:07gevaertsTheSeven: have you followed our discussion a bit? From what you know of the controller, would the model we propose work?
22:29:49 Quit antil33t (Read error: Connection reset by peer)
22:29:50pamauryTheSeven: thanks for the link.
22:30:03 Quit leavittx (Ping timeout: 245 seconds)
22:32:13pamauryTheSeven: does this spec apply to other parts of the nano2g or only the usb part ?
22:35:13gevaertspamaury: changing the API later if it's not workable for some chips shouldn't be too hard, so we can always fix things when needed
22:35:59pamauryindeed
22:36:34TheSevenpamaury: no, i didn't follow
22:36:53TheSevenand the spec does apply to *some* other parts of the nano2g, but not much
22:37:05 Join antil33t [0] (~Mudkips@203-184-54-232.callplus.net.nz)
22:38:09gevaertsTheSeven: is the USB controller one of those FIFO-based things that are mainly software-managed?
22:38:54pamaurylooking at the spec, it seems software managed but TheSeven will probably know better than us
22:39:06TheSeveni'd call it semi-managed
22:39:47TheSevenit has a bunch of fifos of course, and it doesn't have things like QH/TDs, but it will do quite a lot of the groundwork
22:40:32TheSevenyou can basically have one transfer per endpoint at a time, which are dealt with in round-robin order by the hardware
22:40:47TheSevenand it has an AHB master
22:41:48pamauryAHB ?
22:42:04TheSevenadvanced highspeed bus
22:42:09TheSevenit's doing busmaster DMA
22:43:01TheSevenso you basically tell it to receive x packets with a total size of y on endpoint z, and store the result to address a, and it will just do it.
22:43:18TheSevenbut it doesn't have any sophisticated queueing things like the host controllers usually have
22:43:41ChrononBTW: I think I fixed the problem with build client. This is a fairly recently installed system and apparently I didn't have xcolor.sty installed, which was preventing building the manual.
22:46:12gevaertsOK, so it's basically equivalent to a transfer descriptor in the ARC controller, except that that one can actually handle non-contiguous pages (which we don't use)
22:46:33gevaertsAt least for bulk/interrupt. We'll have to see what happens on isochronous
22:46:58gevaertsAlso except that TDs can be chained, but currently we don't do that either
22:48:58 Join mt2 [0] (~chatzilla@41.233.142.159)
22:48:59gevaertsor do we?
22:49:04*gevaerts can't remember
22:49:31 Quit Stephen__ (Ping timeout: 245 seconds)
22:50:00pamaurywe use it I believe, for use transfers
22:50:03pamaury*huge
22:50:11mt2Would a build client be useful on a 256kbps up speed ?
22:50:12pamauryat least arc seems to support it iirc
22:50:35 Join Stephen__ [0] (~S@86.45.50.221)
22:51:04 Part punkt
22:51:05TheSevenmt2: mine with 1mbps seems to be amongst the faster ones, so i'd guess yess
22:51:29TheSeventhere are also quite a bunch of builds that aren't uploaded at all (bootloader, sim, checkwps)
22:52:53TheSevenanother nice-to-have thing (i think i already mentioned that once some time ago): a red/yellow nagging irc bot :-)
22:52:53mt2Great. I'll try adding mine then and see how it performs.
22:55:54 Quit anewuser ()
22:56:15BagderTheSeven: I actually wrote such a "blame script" once that would mention who did the recent commit that caused a yellow/red but I never made it into actually announcing it on the channel
22:56:36Bagderand then we changed the build system so now I think I would have to do it all again ;-)
23:00
23:00:07Stephen__also assuming the person who commits is in irc.
23:00:54gevaertsYes, but lots of us think that that should be the norm anyway
23:01:09bluebrotherthe build system announcing when a build round is done and mentioning the number of targets that are yellow / red would be nice. Now that we have commit announcements.
23:02:10Bagderyeah, they're just a bit separated as things work today
23:02:25mt2Maybe this should be included as a "rule of the thumb" thing when sending out commit-access e-mails.
23:02:36Bagderthat's not a bad idea
23:06:51 Quit pamaury (Quit: Page closed)
23:11:40 Quit saratoga (Quit: Page closed)
23:18:06bluebrotheranyone interested in updating a rbutil translation? -> http://www.alice-dsl.net/dominik.riebeling/weblinguist/ (working on the page, but it should do its job already)
23:24:11domonokybluebrother: regarding the voicefile issue from flyspray. i already fixed the voice download url sometime ago. thats why it works with svn, but not 1.2.5 :-)
23:24:16 Part toffe82
23:25:31bluebrotherdomonoky: ah, ok. Forgot about that. Too much TTS stuff lately :)
23:25:31 Join CGL [0] (~CGL@190.207.236.167)
23:25:39domonoky:-)
23:25:53 Quit liar (Ping timeout: 245 seconds)
23:26:09domonokyif you want todo more tts stuff, you could look at the opensapi thing in flyspray :-)
23:27:41bluebrotherwell, for now I'm planning to do some minor other stuff like looking at the "blurry icon" issue in FS. Currently playing around with replacing the icon with an svn version.
23:28:17bluebrotheroh, and I want to finish my translation website :)
23:28:26bluebrotheror at least improve it
23:28:28domonoky:-)
23:28:55bluebrotherthe translation status of Rockbox Utility has quite some room for improvements
23:30:25domonokytrue. is this translation page php ? maybe it should go into svn to allow other to improve it too ?
23:30:53bluebrotheryes, it's php. Though it's kinda hackish right now.
23:33:42domonokywe should probably combine your .ts handling code with the UI code from rashers translation page (with some modifications as needed).
23:33:44 Join anewuser [0] (anewuser@190.77.16.9)
23:33:48 Quit anewuser (Changing host)
23:33:48 Join anewuser [0] (anewuser@unaffiliated/anewuser)
23:36:48bluebrotherwell, I'm not sure how good this would work. The translations are quite different −− no separate translations for speech in ts files for one.
23:37:19bluebrotherbut separating data and view is something I want to get into that :)
23:37:54 Quit evilnick_B (Ping timeout: 252 seconds)
23:39:07 Quit jgarvey (Quit: Leaving)
23:40:17domonokydo you use any template engine in this page ? (themepage and translate use smarty for templates).
23:40:59 Quit petur (Quit: Zzzzz)
23:43:03 Quit domonoky (Read error: Connection reset by peer)
23:46:06 Quit bertrik (Quit: sleeo)
23:46:34 Join m3dlg [0] (~m3dlg@bb-87-81-252-83.ukonline.co.uk)
23:48:11 Quit Farthen (Quit: sleeping... ;-))
23:49:19 Quit bmbl (Quit: Bye!)
23:52:05 Join dfkt_ [0] (~dfkt@unaffiliated/dfkt)
23:53:41 Quit m3dlg (Ping timeout: 264 seconds)
23:53:54 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
23:54:10 Join stripwax [0] (~Miranda@87-194-34-169.bethere.co.uk)
23:54:59 Join m3dlg [0] (~m3dlg@bb-87-81-252-83.ukonline.co.uk)
23:55:45 Quit dfkt (Ping timeout: 260 seconds)

Previous day | Next day