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

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

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

#rockbox log for 2009-11-29

00:01:49 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
00:04:44 Quit pamaury ("exit(*(int *)0 / 0);")
00:19:12bertrikUnhelpful, as far as I understand the greylib was coded as a replacement for ordered dither because is looks better than ordered dither
00:19:54bertrikfast display will probably flicker regardless of the dither scheme
00:20:45bertrikamiconn is the expert on this
00:23:01 Join Llorean1 [0] (n=DarkkOne@pool-71-114-234-160.austtx.dsl-w.verizon.net)
00:23:09amiconnUnhelpful: When I started developing the greylib (or rather the old graylib), I tried ordered dither first
00:23:16amiconnIt flickered like hell...
00:23:48amiconnAnd the temporal dithering is calculated live now - no need for any matrix
00:24:09amiconnIt works much like the bresenham algorithm
00:25:04amiconnUnhelpful: The diacritics support is fundamentally bugged as-is
00:25:10saratogaif I change the voltage to a GPIO pin externally, should I assume that the debug menu will see that change or do I need to actually put the pin into input mode in rockbox and read it myself?
00:25:20saratogaon AMS that is
00:27:03bertriksaratoga, I don't know, but I think you should at least avoid forcing a pin configured as an output to a different external voltage
00:27:23saratogabertrik: yeah just looking at the code now I realize that
00:28:07bertrikyou could apply a voltage through a 1 kilo-ohm resistor for example, so in case of a mismatch between pin output voltage and externally applied voltage, only a few mA will flow
00:29:37saratogabertrik: I limited my supply to a few mA so hopefully I didn't damage anything
00:29:57bertrikprobably not then :)
00:32:46 Quit bertrik ("De groeten")
00:33:03***Saving seen data "./dancer.seen"
00:44:12Unhelpfulamiconn: ah. i was thinking that it might be best to maximize the number of on-off transitions on the faster displays. but greylib will probably never look *good* on oled.
00:46:21 Quit soap ()
00:50:27 Quit Llorean (Read error: 110 (Connection timed out))
01:00
01:18:23 Join evilnick_ [0] (i=4571af51@rockbox/staff/evilnick)
01:20:22 Quit evilnick (Ping timeout: 180 seconds)
01:20:33 Nick evilnick_ is now known as evilnick (i=4571af51@rockbox/staff/evilnick)
01:26:54amiconnUnhelpful: The number of on-off transitions *is* maximized in the current algorithm
01:30:49amiconnThe dithering itself isn't random at all, it's just bresenham-like. Only the initial "error" value (called "phase" in the greylib) is randomized
01:31:05amiconn...for each pixel
01:31:46Unhelpfulamiconn: and it looks worse with an ordered dither matrix for initial phase?
01:32:06amiconnyes
01:32:48amiconnOr rather, I didn't try an ordered dither matrix as the starting value, I only tried identical starting values, so to say
01:33:24amiconnThe original implementation worked differently, precalculating the pattern instead of calculating it live
01:33:55Unhelpfulit looks to me like a significant part of the problem with diacritics is that it doesn't handle draw modes at all? it seems it's assumed that DRMODE_FG will be "right" for drawing over the existing character, but for some modes this may not be true.
01:34:01amiconnAt first I applied no shift to that pattern, and it flickered like hell. This was on a real lcd btw - the archos recorder
01:34:16amiconnUnhelpful: Yes, that's one problem
01:34:43amiconnAnother problem is that it completely ignores the possibility that a diacritical character might be wider than its base character
01:34:52Unhelpfulamiconn: ordered dither matrix for initial phase *might* be helpful - by making the individual "frames" used for dithering look a bit more ordered-dither-ish
01:35:59Unhelpfulhrm. if it's wider i guess we need to 1) extend the text right by drawing enough columns of "empty" bitmap 2) draw the diacritic centered over the actual base character ?
01:36:27amiconnThe diacritic is centered already
01:37:10amiconnBut if it's wider than the base, it will extend into the previous and next character, meaning it will partially overdraw the previous one, and will be clipped *by* the next one
01:38:01amiconnThe drawmode problem looks like it's even worse than I thought. The hardest one is DRMODE_COMPLEMENT - this will never work right with just overdrawing, it will need pre-combining
01:38:26amiconnBut it looks like DRMODE_INVERSEVID also breaks
01:40:46amiconnI've been thinking about trying precombination for ages - not only for diacritics (in fact I kept forgetting that aspect), but for algorithmic emblodeing and perhaps slanting (pseudo-italic)
01:42:00amiconnI also wonder whether the larger bitmaps would help improving text rendering speed
01:42:15amiconn(my idea was to precombine a whole line)
01:43:01Unhelpfulamiconn: it almost certainly would for the AA case, as that packs pixel horizontally. aren't mono bitmaps packed vertically?
01:43:09amiconnI mean especially on horizontally packed greyscale and on colour displays. For a larger monochrome bitmap it might pay off to transpose the bits
01:43:57amiconnYes they are. Precombining wouldn't change that directly, but one bitmap per line might be a better starting point for optimisation that one bitmap per character
01:44:34amiconnAlso, on coldfire the buffer for that should reside in IRAM, which is probably *very* helpful
01:45:02 Join soap [50] (n=soap@rockbox/staff/soap)
01:45:15Unhelpfulprecombining the string to a buffer and then blitting the buffer to screen *per draw* seems like it would probably *not* be faster?
01:45:39amiconnWhy not?
01:46:33Unhelpfulit seems to me that precombining could potentially cost as much as blitting to buffer
01:46:56Unhelpfulto screen rather. it's *almost* as much work as "really" drawing the string
01:47:14amiconnNo, because precombining can utilize several optimisations direct drawing cannot do
01:47:59 Join Tomis [0] (n=Tomis@70.134.98.13)
01:48:18amiconn(a) The precombination buffer would have the same bit alignment as the source bitmaps (the individual characters). So for the standard case (not extra processing like diacritics and emboldening) it would always use memcpy
01:48:51amiconnThat's optimized quite a bit (e.g. utilizing burst reading on coldfire wherever possible)
01:48:56Unhelpfulhrm, good point. might be worth packing AA fonts vertically for the same reason...
01:50:05amiconn(b) If the intermediate buffer is in IRAM, the following draw-to-framebuffer would profit
01:51:04Unhelpfulprecombine might also be nice in the subpixel-AA case, since we *could* get greater resolution for glyph width/placement by not rounding glyphs to whole pixels...
01:51:33amiconn(c) When dealing with a larger monochrome bitmap, it might pay off to transpose the bits (turn the vertically packed pixels into horizontally packed ones), so that the font bitmaps stay vertically packed on all targets, yet we could switch to horizontally packed mono bitmaps on greyscale ipods and all colour targets
01:52:06amiconnTransposing an 8x8 pixel block of a monochrome bitmap can be done quite efficiently especially on arm
01:56:56Unhelpfulvertical subpixel AA is probably not terribly worth doing tbh
01:56:56 Join Tomis2 [0] (n=Tomis@70.134.83.19)
01:58:36Unhelpfuland horizontal is *less* worth doing without subpixel glyph placement.
02:00
02:02:34 Join drgarrett [0] (n=45b24a7d@giant.haxx.se)
02:03:12drgarretthey - noob here.
02:03:38krazykitdo you have a question, or are you needlessly informing us of that?
02:04:02drgarrettis there a way to search older irc logs to find a previously discussed problem?
02:04:15RavoxX_krazykit, whats about the fuzev2, are there any nightlys?
02:04:59krazykitdrgarrett, yes, via google: search site:rockbox.org/irc searchterms
02:05:13drgarrettbeauty - thanks.
02:05:36krazykitRavoxX_, i know absolutely nothing about the progress of the AMS targets
02:05:43RavoxX_ah, okay
02:06:32RavoxX_Nevertheless, I'll say goodbye now to bed
02:06:39 Quit RavoxX_ ("Verlassend")
02:09:54 Quit geertvdijk (Read error: 110 (Connection timed out))
02:10:41 Quit drgarrett ("CGI:IRC (EOF)")
02:11:09 Quit Tomis (Read error: 110 (Connection timed out))
02:11:09 Nick Tomis2 is now known as Tomis (n=Tomis@70.134.83.19)
02:18:11 Quit liar (Read error: 113 (No route to host))
02:33:05***Saving seen data "./dancer.seen"
02:34:50saratogathe Sandisk connector is actually really interesting
02:35:09saratogait has a PCB on it that lets you easily solder into the power, audio, or USB lines
02:35:23saratogaseems like an obvious way to measure power consumption without even opening the player
02:37:48JdGordonpixelma: thanks for doing the sbs
02:46:06 Join krazykit` [0] (n=kkit@adsl-76-251-250-122.dsl.ipltin.sbcglobal.net)
02:46:06 Quit krazykit (Read error: 104 (Connection reset by peer))
02:53:07 Nick krazykit` is now known as krazykit (n=kkit@adsl-76-251-250-122.dsl.ipltin.sbcglobal.net)
02:55:31 Quit matsl (Read error: 110 (Connection timed out))
02:58:15soapsaratoga, if you want I have some bare connectors for E200s
02:58:58saratogasoap: you mean the sandisk cable?
02:59:14saratogawoot I think I got line out working on my fuze
02:59:35saratogai say "think" because I don't have a cable, but it looks like audio on my oscilliscope
02:59:40soapno - bare connectors and shells
02:59:58saratogasoap: got a link to what they look like?
03:00
03:00:06soaphmm
03:00:52saratogaright now i've just taken apart a spare sandisk cable
03:00:58saratogabut if they're more handy i might want one
03:01:01soaphttp://www.chargeconverter.com/shop/
03:01:14saratogai don't suppose you have a fuze and a line out adapter?
03:01:39soapno :(
03:01:47saratogasoap: wow neat
03:02:12saratogai think I'm going to order one of those breakout boards
03:02:25soapdon't look at SD_K and Sk, I believe I have SD_B
03:04:17saratogawhat did you plan to do with them?
03:07:59 Nick Llorean1 is now known as Llorean (n=DarkkOne@pool-71-114-234-160.austtx.dsl-w.verizon.net)
03:13:59 Join jhulst [0] (n=jhulst@jhulst.com)
03:15:47 Quit jhulst (Client Quit)
03:16:09 Quit Guest35900 (Remote closed the connection)
03:17:01 Quit Zarggg ()
03:21:54 Join Zarggg [0] (n=zarggg@65-78-69-194.c3-0.eas-ubr6.atw-eas.pa.cable.rcn.com)
03:24:55 Part froggyman
03:37:10 Join NoGare [0] (i=NoGare@dyn216-8-173-29.ADSL.mnsi.net)
03:37:17 Quit tchan (Remote closed the connection)
03:37:54 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
03:38:18soapsaratoga, make cables
03:38:21 Quit Sajber^ (Read error: 104 (Connection reset by peer))
03:39:49 Join rikupw [0] (n=rikupw`@c-67-160-68-170.hsd1.wa.comcast.net)
03:40:26saratogasoap: I think i'm just going to order that 35 dollar break out board
03:40:36saratogait'll be easiest if I want to do power measurements on the fuze
03:41:25soapcan't argue with that
03:44:06saratogasoap: any idea where i can buy a line out cable? the part with the actual headphone male plug thing
03:44:46soapno. Those connectors / boards are all I know on this front.
03:51:12 Join froggyman [0] (n=sopgenor@pool-72-69-220-194.chi01.dsl-w.verizon.net)
03:58:30jesusaurusrl
03:58:40jesusaurusthat was weird :/
04:00
04:00:47 Quit rikupw ("_You say hello... but I say goodbye!_")
04:01:34UnhelpfulLlorean: i think what we'd want to do, if possible, is have a window for averaging load to determine clock speed. during playback this is probably easy, we could just have the interval start when the decoder is called to refill the pcm buffer (or alternately when it's done, it doesn't matter).
04:02:03Unhelpfulwhat remains to be seen is whether we can control the window over which load is measured (without just measuring the load ourselves)
04:02:25saratogaUnhelpful: what is this for?
04:02:27Unhelpfulwhen not in playback we probably can't do anything more clever than leaving it in full dynamic mode
04:02:30Unhelpfulsaratoga: beast
04:02:40 Quit bzed (Read error: 110 (Connection timed out))
04:03:06 Join bzed [0] (n=bzed@devel.recluse.de)
04:03:16saratogaUnhelpful: for voltage scaling?
04:03:25Unhelpfulit *looks* per the datasheet as though we can ask it to tune voltage and clock for us. there are a number of tunables to set for how scaling runs.
04:03:35Unhelpfulvoltage and clock
04:04:34saratogaUnhelpful: a simple solution to targets with voltage scaling is to just have a three teer boost system, with normal, boosted, and boosted_high, with the last one being specifically enabled by plugins/codecs if needed
04:04:44saratogawhich would probbaly just be mpegplayer and ape
04:05:41LloreanAnd maybe certain plugins (rockboy) while audio is playing.
04:05:43Unhelpfulsaratoga: since the power consumed has a greater-than-linear relationship to clock in a system where both clock and voltage are scaled together, it's probably best to *try* to keep the clock and voltage as close to constant as possible
04:06:28saratogaUnhelpful: nonlinearity is mostly due to voltage scaling, so if you pick the normal and boost speeds well, you can probably keep the voltage at the lowest setting for everything but APE, etc
04:06:44Unhelpfulif you can cut the time to finish the "work" in half, but it costs more than twice as much to do so, then cycling between boost and idle doesn't make sense.
04:07:09saratogasomething like boosting between 25 and 100MHz should allow for essentially linear boosting unless you have extremely fine grained control over voltage
04:07:23saratogalinear in power boosting I mean
04:08:32Unhelpfulalso changing clock without voltage does not seem to do much at all on beast
04:08:38 Quit bzed (Read error: 60 (Operation timed out))
04:08:50saratogai strongly prefer a simple solution like that FWIW, as I think it will make things like codec optimization much simpler if theres fewer clock speeds to worry about
04:09:12saratogathat way typical codecs still only have two speeds, and thus test_codec results will be easier to make sense of
04:09:46 Quit n1s ("Lämnar")
04:10:36saratogaparticularly on targets like the beast where the main core is so ridiculously powerful even on a per clock basis, it seems unlikely that normal codecs should ever have to run much above 50MHz anyway
04:10:41Unhelpfulwell, like i said, if we can just tell the SoC what window to average load over for self-tuning, then we don't have to do any of that...
04:11:12saratogathat will make test_codec somewhat ackward, since it won't be running at the same clock speed as actual playback
04:12:01saratogathough maybe test_codec could be made to set its own clock speed somehow
04:12:17Unhelpfulheh, only ape and aac-he exceed 50MHz :)
04:12:27saratogaso that its not testing what memory latency in clock cycles is like at 500MHz when codecs run at 50MHz
04:12:28 Join bzed [0] (n=bzed@devel.recluse.de)
04:13:03saratogaUnhelpful: I've been watching all your benchmarking, and I think for 1 quarter the time you put into it you could have just made AAC-HE run 2x as fast :)
04:13:46Unhelpfulyou assume that i comprehend all algorithms as well as i do some. ;)
04:14:08saratogaUnhelpful: i don't think you really need to understand the alogirthms, i usually don't
04:14:11 Quit GeekShado_ ("The cake is a lie !")
04:14:31saratogafor HE it'd probably be enough to just look at teh function names and steal the ASMed versions of those functions from mp3/mpc/atrac
04:14:47saratogaalthough some mucking around with it might be needed
04:15:09 Quit bzed (Read error: 104 (Connection reset by peer))
04:15:26saratogaalso as mt discovered if you bug buschel enough you can also get him to write you ASM optimized filterbanks ;)
04:15:28 Join bzed [0] (n=bzed@devel.recluse.de)
04:15:31Unhelpfuli think it's still worth knowing which gcc settings produce best results. there's no reason not to tune optimization levels per-codec. we already do it, and it only takes three test_codec runs to be able to do it *correctly* ;)
04:16:13saratogatrue, but the differences are very small, and in the case of AAC-HE simply reflect how well GCC is optimizing code that almost certainly will be replaced by ASM eventually
04:17:03Unhelpfulalso this is the sort of thing i work on when too busy or tired to *think* about things :)
04:17:08 Join Rondom_ [0] (n=quassel@dslb-084-057-140-046.pools.arcor-ip.net)
04:17:08saratogaand realistically if you want to make a codec 10% faster its really not too difficult except for MP3/Vorbis/WMA as those are the only ones that are reasonably well optimized
04:17:11saratogayes true
04:17:22 Quit bzed (Success)
04:17:23saratogathats why i'm playing around with line out instead of working on the mdct problem
04:18:17 Quit Rondom (Read error: 60 (Operation timed out))
04:20:26 Quit kugel (Read error: 104 (Connection reset by peer))
04:22:24 Quit Tomis (Read error: 104 (Connection reset by peer))
04:22:25Unhelpfulwhich is very frequently since offspring += 2
04:23:05 Quit fdinel ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
04:23:33Unhelpfulas i see it, if we can play with an averaging window, but let the SoC do the averaging, things are very simple while decoding - each time we start filling the PCM buffer, compare curent_tick to the tick from the last buffer fill. the time passed is a roughly ideal window for load averaging, i'd think...
04:24:22 Join Tomis [0] (n=Tomis@70.134.83.19)
04:27:47Unhelpfulthere are *quite* a few tunables for how the SoC manages scaling, though... and i've not gotten to reading and trying to understand all of it, yet.
04:28:24 Join bzed [0] (n=bzed@devel.recluse.de)
04:30:14 Join krazykit` [0] (n=kkit@adsl-76-251-250-122.dsl.ipltin.sbcglobal.net)
04:33:06***Saving seen data "./dancer.seen"
04:37:00 Quit krazykit (Read error: 110 (Connection timed out))
04:39:19 Quit bzed (Read error: 104 (Connection reset by peer))
04:41:04 Join Sajber^ [0] (n=Sajber@h-65-31.A213.priv.bahnhof.se)
04:58:38 Part froggyman
05:00
05:02:33 Join bzed [0] (n=bzed@devel.recluse.de)
05:15:10 Join bzed_ [0] (n=bzed@devel.recluse.de)
05:18:53 Quit bzed (Read error: 111 (Connection refused))
05:18:54 Nick bzed_ is now known as bzed (n=bzed@devel.recluse.de)
05:35:52 Quit Strife89 ("Bed.")
05:43:32 Join Topy [0] (n=Topy44@f054223067.adsl.alicedsl.de)
05:46:51 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
05:55:39 Join Tomis2 [0] (n=Tomis@70.134.100.9)
06:00
06:01:50 Quit T44 (Read error: 110 (Connection timed out))
06:08:40 Quit Tomis (Read error: 110 (Connection timed out))
06:08:41 Nick Tomis2 is now known as Tomis (n=Tomis@70.134.100.9)
06:09:58 Quit Llorean (Read error: 104 (Connection reset by peer))
06:33:08***Saving seen data "./dancer.seen"
06:33:48 Quit panni_ ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
06:36:55 Quit NoGare ("ChatZilla 0.9.85 [Firefox 3.6b4/20091124213835]")
06:46:52 Quit niekie (Client Quit)
06:47:43 Join niekie [0] (i=quasselc@dreamworld.bergnetworks.com)
06:52:08 Quit shaggy-h (Read error: 113 (No route to host))
07:00
07:00:54 Join bluebrother [0] (n=dom@f053154223.adsl.alicedsl.de)
07:02:40 Quit bluebroth3r (Read error: 60 (Operation timed out))
07:24:50 Quit BHSPitLappy (Remote closed the connection)
07:33:16 Quit evilnick (Ping timeout: 180 seconds)
07:38:35 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
07:56:59 Quit Horschti (Read error: 110 (Connection timed out))
07:58:11 Join AndyIL [0] (n=pasha_in@212.14.205.32)
08:00
08:11:10 Quit AndyI (Read error: 110 (Connection timed out))
08:26:45 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
08:33:12***Saving seen data "./dancer.seen"
08:40:51 Join Kitr88 [0] (n=Kitarist@89.142.36.242)
08:40:56 Quit Kitar|st (Read error: 104 (Connection reset by peer))
08:42:24 Join Kitar|st [0] (i=Kitarist@BSN-142-13-181.dial-up.dsl.siol.net)
08:44:08 Join Rob2223 [0] (n=Miranda@p4FDCD19D.dip.t-dialin.net)
08:58:49 Quit gevaerts (Nick collision from services.)
08:59:01 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
08:59:18 Quit Kitr88 (Connection timed out)
09:00
09:02:11 Quit Rob2222 (Read error: 110 (Connection timed out))
09:12:28 Join antil33t1 [0] (n=Mudkips@203-184-54-232.callplus.net.nz)
09:12:33 Quit antil33t (Connection reset by peer)
09:14:17amiconn[04:17:08] <saratoga> and realistically if you want to make a codec 10% faster its really not too difficult except for MP3/Vorbis/WMA <== Please try ape, then...
09:24:01saratogaamiconn: true theres APE
09:26:25saratogabut my point was that only codecs which aren't already well optimized tend to take advantage of compiler tricks
09:42:42 Part toffe82
09:45:34 Join flydutch [0] (n=flydutch@host205-59-dynamic.5-87-r.retail.telecomitalia.it)
09:51:04 Join n1s [0] (n=n1s@cm-84.215.127.139.getinternet.no)
09:54:19 Join DerPapst [0] (n=DerPapst@p4FE8FAC1.dip.t-dialin.net)
10:00
10:11:23 Join stoffel [0] (n=quassel@p57B4E3CD.dip.t-dialin.net)
10:32:00 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
10:32:27 Join geertvdijk [0] (n=chatzill@cc412026-a.zwoll1.ov.home.nl)
10:33:15***Saving seen data "./dancer.seen"
10:44:17CIA-6New commit by nls (r23780): These macros have been replaced by (UN)LIKELY macros and were not used anywhere.
10:48:10 Join pamaury [0] (n=pamaury@sal63-1-82-243-96-220.fbx.proxad.net)
10:52:53 Join JackNasty [0] (n=d035ff47@giant.haxx.se)
10:53:15JackNastyevening all
10:54:34JackNastyevening all!
10:54:49n1smornin'
10:54:49JackNastyhmm..not voiced? hrm..lol
10:54:55JackNastyoh noes! lag!
10:58:56JackNastylag...may have to reboot router..brb
11:00
11:00:50 Quit JackNasty ("CGI:IRC (EOF)")
11:02:51 Join fyrestorm [0] (n=nnscript@cpe-69-203-150-85.si.res.rr.com)
11:04:38 Join FOAD_ [0] (n=dok@dinah.blub.net)
11:07:35 Join Grahack [0] (n=chri@ip-222.net-82-216-222.rev.numericable.fr)
11:14:02 Join liar [0] (n=liar@83.175.83.185)
11:19:15 Quit fyre^OS (Read error: 110 (Connection timed out))
11:20:47 Quit FOAD (Read error: 110 (Connection timed out))
11:20:47 Nick FOAD_ is now known as FOAD (n=dok@dinah.blub.net)
11:27:54 Quit n1s (Remote closed the connection)
11:50:43 Join Ubuntuxer [0] (n=johannes@dslb-094-221-092-126.pools.arcor-ip.net)
11:53:04 Join bertrik [0] (n=bertrik@ip117-49-211-87.adsl2.static.versatel.nl)
11:54:51 Join polobricolo [0] (n=polobric@AGrenoble-257-1-18-79.w86-194.abo.wanadoo.fr)
11:57:54CIA-6New commit by nls (r23781): Use clz instruction to count leading zeroes on armv5+, gives about 2% speedup with one sample on the Gigabeat S
12:00
12:12:34 Quit polobric1lo (Read error: 110 (Connection timed out))
12:14:29 Quit linuxstb (Read error: 113 (No route to host))
12:21:14 Join petur [0] (n=peter@rockbox/developer/petur)
12:21:24 Join n1s [0] (n=n1s@rockbox/developer/n1s)
12:24:20 Quit liar (Read error: 113 (No route to host))
12:33:16***Saving seen data "./dancer.seen"
12:37:21 Join liar [0] (n=liar@83.175.83.185)
12:51:29 Quit stoffel (Remote closed the connection)
13:00
13:03:41 Part Ubuntuxer
13:11:44 Join bzed_ [0] (n=bzed@devel.recluse.de)
13:18:38 Quit bzed_ (Read error: 104 (Connection reset by peer))
13:22:19 Join lokijasss [0] (n=4dff7132@giant.haxx.se)
13:22:45 Join lokijassso [0] (n=4dff7132@giant.haxx.se)
13:23:30 Quit lokijassso (Client Quit)
13:24:59 Quit flydutch (Read error: 60 (Operation timed out))
13:25:21 Join bzed_ [0] (n=bzed@devel.recluse.de)
13:25:47 Quit bzed (Read error: 110 (Connection timed out))
13:27:20 Quit lokijasss ("CGI:IRC (Ping timeout)")
13:27:34 Nick bzed_ is now known as bzed (n=bzed@devel.recluse.de)
13:27:54 Join flydutch [0] (n=flydutch@host139-200-dynamic.15-87-r.retail.telecomitalia.it)
13:30:13 Join bzed_ [0] (n=bzed@devel.recluse.de)
13:32:11 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
13:33:54 Quit stripwax (Read error: 104 (Connection reset by peer))
13:35:12 Join bzed__ [0] (n=bzed@devel.recluse.de)
13:36:41 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
13:38:01 Quit bzed__ (Read error: 104 (Connection reset by peer))
13:38:26 Join bzed__ [0] (n=bzed@devel.recluse.de)
13:39:59 Quit liar (Read error: 113 (No route to host))
13:43:19 Quit bzed__ (Read error: 104 (Connection reset by peer))
13:44:03 Join liar [0] (n=liar@83.175.83.185)
13:48:55 Quit stripwax ("http://miranda-im.org")
13:49:02 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
13:54:47 Join bzed__ [0] (n=bzed@devel.recluse.de)
13:56:53 Quit bzed_ (Read error: 111 (Connection refused))
13:57:52 Quit bzed (Read error: 111 (Connection refused))
13:57:52 Quit bzed__ (Read error: 104 (Connection reset by peer))
13:57:57 Join bzed [0] (n=bzed@devel.recluse.de)
14:00
14:10:38 Join evilnick [0] (i=4571af51@rockbox/staff/evilnick)
14:15:13 Nick Rondom_ is now known as Rondom (n=quassel@dslb-084-057-140-046.pools.arcor-ip.net)
14:17:56 Quit bzed (Read error: 104 (Connection reset by peer))
14:22:47 Join bzed [0] (n=bzed@devel.recluse.de)
14:26:08 Quit bzed (Remote closed the connection)
14:26:10 Join bzed [0] (n=bzed@devel.recluse.de)
14:27:10 Join polobric1lo [0] (n=polobric@AGrenoble-257-1-29-246.w86-194.abo.wanadoo.fr)
14:27:40 Quit liar (Read error: 104 (Connection reset by peer))
14:29:37 Join bzed_ [0] (n=bzed@devel.recluse.de)
14:29:37 Quit bzed (Read error: 104 (Connection reset by peer))
14:29:55 Nick bzed_ is now known as bzed (n=bzed@devel.recluse.de)
14:33:18***Saving seen data "./dancer.seen"
14:34:30 Quit pamaury ("exit(*(int *)0 / 0);")
14:34:43 Nick fxb__ is now known as fxb (n=felixbru@85.214.97.64)
14:35:07 Join bzed_ [0] (n=bzed@devel.recluse.de)
14:35:07 Quit bzed (Read error: 104 (Connection reset by peer))
14:35:19 Nick bzed_ is now known as bzed (n=bzed@devel.recluse.de)
14:37:28 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
14:40:28 Quit polobricolo (Read error: 110 (Connection timed out))
14:40:58n1si wonder if that clz thing can be used in other places too
14:41:30 Quit antil33t1 (Read error: 104 (Connection reset by peer))
14:41:37 Join antil33t [0] (n=Mudkips@203-184-54-232.callplus.net.nz)
14:41:43 Join bzed_ [0] (n=bzed@devel.recluse.de)
14:42:23 Quit bzed_ (Client Quit)
14:44:23 Quit stripwax ("http://miranda-im.org")
14:44:34 Join Omlet [0] (i=omlet05@238.157-200-80.adsl-dyn.isp.belgacom.be)
14:45:03 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
14:51:01 Join krazykit [0] (n=kkit@adsl-76-251-250-122.dsl.ipltin.sbcglobal.net)
14:51:06 Quit krazykit` ("Lost terminal")
14:55:33 Quit petur ("real life")
14:58:26 Quit bzed (Read error: 111 (Connection refused))
15:00
15:06:05 Quit stripwax (Read error: 110 (Connection timed out))
15:06:48 Join polobricolo [0] (n=polobric@AGrenoble-257-1-129-195.w92-157.abo.wanadoo.fr)
15:19:43 Quit polobric1lo (Read error: 110 (Connection timed out))
15:21:34 Join liar [0] (n=liar@83.175.83.185)
15:24:29 Quit flydutch (Read error: 110 (Connection timed out))
15:25:47gevaertsTorne: I'm not sure about the UDMA things. I seem to remember from my tests that you really need UDMA2 or higher for full performance, although this could be boosting-related
15:26:15 Join fdinel [0] (n=Miranda@modemcable235.127-131-66.mc.videotron.ca)
15:29:21CIA-6New commit by nls (r23782): #if 0 out unused stuff
15:37:19 Join flydutch [0] (n=flydutch@host224-163-dynamic.8-87-r.retail.telecomitalia.it)
15:51:36Rondom~komisch
15:53:04markunRondom: wrong channel?
15:53:26 Join matsl [0] (n=matsl@1-1-4-2a.mal.sth.bostream.se)
15:53:50Rondommarkun: yep :-/
15:54:02markunmacht nichts :)
15:54:13Rondom:-)
15:58:29Unhelpfuln1s: have you tested codecs with the new checksum mode with -fstrict-aliasing? i'd say if they pass then it should be OK to commit.
16:00
16:00:03Unhelpfulspc tracks to test are quite easy to come by. redistributable ones, not so much.
16:01:13n1sUnhelpful: i haven't yet but will do it soon then
16:01:32*n1s is testing a patch that speeds up flac on the beast ;)
16:10:23 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
16:17:00 Join efyx_ [0] (n=efyx@lap34-1-82-225-185-146.fbx.proxad.net)
16:21:02CIA-6New commit by nls (r23783): Use clz instruction to calculate integer logarithms, speeds up decoding of one flac 8 sample by 12%.
16:21:51 Quit stripwax (Read error: 104 (Connection reset by peer))
16:21:59n1sUnhelpful: ugh, the aliasing pach breaks coldfire builds
16:22:05gevaertsn1s: is that measurable for an actual flac file?
16:22:34n1sgevaerts: ?
16:22:54gevaertsn1s: "speeds up decoding of one flac 8 sample by 12%"
16:23:03gevaertsor do you mean one sample file?
16:23:15n1syeah, one sample file
16:23:40n1s(the one from the test_files dir on the download server)
16:23:57*gevaerts was just a bit confused
16:24:10gevaertssorry for waking you up :)
16:24:23n1syeahm maybe a little ambigous
16:24:35 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
16:31:21n1sgcc 3.4 doesn't like sd[raw_voice->waveform]->loop while 4.0 is fine with it
16:31:45n1si get a "error: invalid type argument of `->'"
16:33:21***Saving seen data "./dancer.seen"
16:33:24 Quit Grahack ("Leaving.")
16:35:43n1sO_o it wanted a . operator instead, why doesn't 4.0 care?
16:35:45Unhelpfulgevaerts: the sample files are actual music. ;)
16:36:14Unhelpfuln1s: don't know. you could always leave strict aliasing on for coldfire :/
16:37:07n1sUnhelpful: it is left on for coldfire already, it's the pointer semantics that seems to differ between gcc versions which confuses me
16:37:09Unhelpfulso, if i want to go poking at the beast SoC's power management registers, i assume i should do read, modify, write?
16:44:28n1si don't know jhMikeS is probably the one to ask but he hasn't been around for a while
16:45:14AlexPrasher: Is much needed to add the Clip to the theme site? I've just tried and checkwps builds for it. Screen res is 128x64
16:46:11n1sok, . works on newer gcc too
16:49:28Unhelpfulit looks like there are already some macros, and we use them to read the current clock. i can't make heads or tails out of what it's *doing* with the high two bits of PMCR0, though.
16:55:39 Quit efyx_ (Remote closed the connection)
16:55:51 Join efyx_ [0] (n=efyx@lap34-1-82-225-185-146.fbx.proxad.net)
16:55:59 Quit efyx_ (Remote closed the connection)
16:56:07 Join efyx_ [0] (n=efyx@lap34-1-82-225-185-146.fbx.proxad.net)
16:57:14 Quit polobricolo (Read error: 60 (Operation timed out))
16:57:54 Quit efyx_ (Read error: 113 (No route to host))
16:58:03 Join mirak_ [0] (n=mirak@85-171-108-41.rev.numericable.fr)
17:00
17:02:07 Join robin0800 [0] (n=quassel@cpc3-brig8-0-0-cust436.brig.cable.ntl.com)
17:03:00 Join efyx_ [0] (n=efyx@lap34-1-82-225-185-146.fbx.proxad.net)
17:07:01 Join panni_ [0] (i=hannes@ip-95-222-21-143.unitymediagroup.de)
17:13:20 Quit flydutch (Read error: 110 (Connection timed out))
17:14:53gevaertsAlexP: the nano2 does have USB, only it doesn't seem to work properly for everyone
17:15:03AlexPoh, my mistake
17:15:11AlexPI'll correct
17:16:36AlexPcheers
17:21:57 Quit Omlet (Connection reset by peer)
17:22:58 Join Omlet [0] (i=omlet05@204.96-240-81.adsl-dyn.isp.belgacom.be)
17:23:03 Quit Omlet (Remote closed the connection)
17:23:15 Quit stripwax (Read error: 54 (Connection reset by peer))
17:25:08n1shmm, my test mod hangs test_codec, is that supposed to work?
17:26:29 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
17:26:34*Unhelpful thought you tested the test. ;)
17:27:06n1si did! but only with the official test set which doesn't include a mod
17:31:05Unhelpfulah. does it hang test_codec without -fstrict-aliasing?
17:31:17 Quit stripwax (Read error: 104 (Connection reset by peer))
17:33:39 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
17:37:39 Join Omlet [0] (i=omlet05@204.96-240-81.adsl-dyn.isp.belgacom.be)
17:41:47 Join tomers [0] (n=chatzill@bzq-84-109-85-100.red.bezeqint.net)
17:42:01 Quit stripwax (Read error: 54 (Connection reset by peer))
17:44:20 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
17:47:29 Join jgarvey [0] (n=jgarvey@cpe-174-097-130-131.nc.res.rr.com)
17:50:07 Join Gregg [0] (n=55de5ba5@giant.haxx.se)
17:52:51 Join pamaury [0] (n=pamaury@88-123-145-145.rev.libertysurf.net)
17:53:11 Join petur [0] (n=peter@d54C6F9B2.access.telenet.be)
18:00
18:10:51 Join solexx_ [0] (n=jrschulz@e176125200.adsl.alicedsl.de)
18:11:58n1sUnhelpful: dunno, that's why i asked :) will test that though
18:16:19 Join xxarthur33xx [0] (n=xxarthur@cpe-71-79-163-117.neo.res.rr.com)
18:16:24xxarthur33xxHello
18:18:31n1s'lo
18:19:33xxarthur33xxI have a question
18:20:11xxarthur33xxAlexP said that there is a bug that prevents some nano 2g from using usb while in rockbox
18:20:20xxarthur33xxis there anyway to fix the bug?
18:21:19n1ssure, we just don't know how :/
18:21:30n1salso it seems to work for some people
18:21:50xxarthur33xxIs there anything I can do to help?
18:22:40n1sif you are a programmer/hacker talk to TheSeven, he probably has the best idea
18:23:11xxarthur33xxoh. I can do some java and stuff, but that propably isn't what is needed
18:23:12xxarthur33xx:P
18:24:01n1sjava is certainly not what's needed
18:24:54xxarthur33xxwhy does your nickname sound so familiar..
18:24:57 Join matthew [0] (n=matthew@70.94.217.104)
18:25:25 Nick matthew is now known as Guest97843 (n=matthew@70.94.217.104)
18:26:14n1sxxarthur33xx: dunno :)
18:26:30 Join stoffel [0] (n=quassel@p57B4E3CD.dip.t-dialin.net)
18:27:35 Quit solexx (Read error: 110 (Connection timed out))
18:31:19 Quit Guest97843 ("Leaving")
18:33:23***Saving seen data "./dancer.seen"
18:34:39 Quit robin0800 (Remote closed the connection)
18:34:52n1scrap, battery died mid testround
18:41:53xxarthur33xx>.<
18:42:08xxarthur33xxI might make some themes for the nano 2g
18:43:51 Quit stoffel (Remote closed the connection)
18:46:07 Join YH-820 [0] (n=777aa0b2@giant.haxx.se)
18:48:01YH-820Can anyone help me? I have a Samsung YH-820 player, and I upgraded its firmware to Janus MTP version provided by Samsung. Now I want to replace this firmware with Rockbox, but maybe Janus bootloader is incompatible with rockbox's. What should I do?
18:49:15 Join mt_ [0] (n=mtee@41.233.152.132)
18:50:23 Nick mt_ is now known as mt (n=mtee@41.233.152.132)
18:50:40 Join polobricolo [0] (n=polobric@AGrenoble-257-1-129-195.w92-157.abo.wanadoo.fr)
18:52:32 Quit YH-820 ("CGI:IRC")
18:55:27 Quit stripwax (Read error: 54 (Connection reset by peer))
18:55:59 Join Strife89 [0] (n=Strife89@adsl-154-14-231.mcn.bellsouth.net)
18:57:43 Quit polobricolo ("leaving")
18:58:27 Join stripwax [0] (n=Miranda@87-194-34-169.bethere.co.uk)
19:00
19:03:35xxarthur33xxis it possible to downgrade?
19:06:31krazykitto use older versions of rockbox? sure. just check out the desired revision and build
19:07:07 Join polobricolo [0] (n=polobric@AGrenoble-257-1-129-195.w92-157.abo.wanadoo.fr)
19:07:14xxarthur33xxI was responding to the other person
19:07:30xxarthur33xxIf they can downgrade then they can get to a supprted version
19:07:39krazykitwell, seeing as they left, that's not very helpful :)
19:07:44xxarthur33xxoh
19:07:47xxarthur33xxI didnt notice >.<
19:11:51CIA-6New commit by nls (r23784): Enable strict aliasing optimizations for codecs on gcc versions >= 4.0, fix alising violations that this uncovered, gives small speedups for most ...
19:15:46amiconnn1s: I wonder whether we should just use __builtin_clz() on all archs
19:16:00amiconnDo you know how well it is optimised?
19:17:34n1samiconn: afaik it doesn't exist on archs which miss the instruction...
19:18:04amiconnThat wouldn't make sense...
19:18:35n1sit's gcc after all :) I'll check
19:19:25 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
19:19:25 Quit pixelma (Nick collision from services.)
19:19:32 Quit amiconn (Nick collision from services.)
19:19:37 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
19:19:43 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
19:19:45 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
19:22:45 Join stripwax__ [0] (n=Miranda@87-194-34-169.bethere.co.uk)
19:23:11n1sit seems it does exist
19:23:47n1si have no other targets than the beast to test on though...
19:24:07n1si'd be happy to change it if someone could benchmark
19:25:50 Join Llorean [0] (n=DarkkOne@adsl-76-202-17-167.dsl.hstntx.sbcglobal.net)
19:30:29Unhelpfulamiconn: it certainly does
19:30:32Unhelpfulerm. n1s.
19:30:50n1sUnhelpful: yeah, i had misunderstood
19:33:38Unhelpfulexcept for platform-specific ones those __builtin functions should always "work"
19:35:24 Join polobric1lo [0] (n=polobric@AGrenoble-257-1-129-195.w92-157.abo.wanadoo.fr)
19:35:40 Quit polobric1lo (Client Quit)
19:36:04Unhelpfulit's rather hard to actually benchmark __builtin_clz usefully, though, if we mean "benchmark for all possible inputs" :)
19:36:45xxarthur33xxi have no clue what you guys are even talking about lol
19:38:39 Join toffe82 [0] (n=chatzill@adsl-71-154-234-44.dsl.frs2ca.sbcglobal.net)
19:39:01Unhelpfulamiconn: couldn't we provide our own __clzsi2, linked before libgcc, if we want to be able to use __builtin_clz but aren't happy with the libgcc implementation?
19:40:35 Quit stripwax (Read error: 110 (Connection timed out))
19:41:24 Join webguest22 [0] (n=4cd48795@giant.haxx.se)
19:42:01 Quit stripwax__ (Read error: 110 (Connection timed out))
19:42:01 Quit webguest22 (Client Quit)
19:53:26 Join bzed [0] (n=bzed@devel.recluse.de)
20:00
20:04:39Tornegevaerts: i'll experiment s'more :0
20:08:02 Join dfkt_ [0] (i=dfkt@unaffiliated/dfkt)
20:09:10 Quit dfkt (Nick collision from services.)
20:09:17 Nick dfkt_ is now known as dfkt (i=dfkt@unaffiliated/dfkt)
20:11:55 Nick YPSY is now known as Ypsy (n=ypsy@geekpadawan.de)
20:20:40CIA-6New commit by bluebrother (r23785): Suppress unused warnings for functions that haven't been implemented.
20:31:11 Quit bmbl ("Bye!")
20:33:00 Join JdGordon1 [0] (n=jonno@c-24-22-210-83.hsd1.wa.comcast.net)
20:33:26***Saving seen data "./dancer.seen"
20:33:27 Quit Dhraakellian (Read error: 104 (Connection reset by peer))
20:38:49 Join p3tur [50] (n=petur@rockbox/developer/petur)
20:41:44 Join froggyman [0] (n=sopgenor@pool-72-69-220-194.chi01.dsl-w.verizon.net)
20:42:34 Quit Kitar|st ()
20:43:41 Join solexx [0] (n=jrschulz@e176107119.adsl.alicedsl.de)
20:49:25Unhelpfuloh, also __builtin_clz is undef if passed a zero. ARM's clz instruction produces 32 in this case, i don't know what x86 BSR will do but gcc explicitly ANDs the result with 31 on x86. we might want our own clz for things that *could* be zero - it could easily be a macro calling __builtin_clz on ARM targets that have a clz instruction.
20:52:03 Join Kitar|st [0] (i=Kitarist@89.143.160.166)
20:56:17 Quit solexx_ (Read error: 110 (Connection timed out))
20:57:55 Quit GeekShadow ("The cake is a lie !")
20:59:42 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
20:59:52 Quit dfkt (Read error: 110 (Connection timed out))
21:00
21:04:34 Quit evilnick (Ping timeout: 180 seconds)
21:11:24saratogalibfaad uses a lot of log2, it might be a good candidate for for CLZ
21:13:25CIA-6New commit by Domonoky (r23786): add sansa clip to checkwps.
21:18:30domonokyrasher: can you update checkwps on the theme site so it includes the clip checkwps , or is this done automagically ?
21:21:44 Quit Tomis (Read error: 104 (Connection reset by peer))
21:21:44 Join Tomis [0] (n=Tomis@70.134.102.148)
21:25:05 Join krazykit` [0] (n=kkit@adsl-76-251-250-122.dsl.ipltin.sbcglobal.net)
21:25:34 Quit krazykit ("Lost terminal")
21:26:51 Nick krazykit` is now known as krazykit (n=kkit@adsl-76-251-250-122.dsl.ipltin.sbcglobal.net)
21:41:19 Join MaadMan [0] (n=MaadMan@188-192-221-19-dynip.superkabel.de)
21:44:21n1sUnhelpful: yeah that undef for 0 may be problematic, but if it's undef they should just use the clz instruction on arm but what the implementations do on other targets may differ...
21:47:18 Join FOAD_ [0] (n=dok@dinah.blub.net)
21:47:26CIA-6New commit by tomers (r23787): Reduce bin size of diacritic characters database
21:49:47 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:50:56CIA-6New commit by mc2739 (r23788): Manual updates ...
21:53:31tomersoh goody - it worked :-)
21:53:34tomersgood night
21:54:18bluebrothertomers: I see some red ...
21:54:33tomersbluebrother: It's not from me :-)
21:54:55tomersI don't know why it didn't happened in the previous commit
21:55:02 Join kugel [0] (n=kugel@rockbox/developer/kugel)
21:55:24kugeltomers: did you test whether this affects drawing performance?
21:55:30bluebrothertomers: ah. Sorry, spotted that too late :o
21:55:36tomerskugel: Negative
21:55:58*kugel can imagine those bitfileds have negative impact
21:56:22tomerskugel: I can switch to bit masking
21:56:36*bluebrother wonders if n1s is aware of the red
21:57:33n1sbluebrother: ah, forgot to check the table
21:58:02kugeltomers: you could make some tests too
21:58:21tomerskugel: I'll do it tomorrow. Can I test using the simulator?
21:59:00kugelnot really if you want target representative figures
21:59:18tomersSo I'll test on e200, and d2
21:59:19kugelbut you can hack into test_gfx
21:59:44tomersI'm only interested in the % of change
22:00
22:00:15 Quit FOAD (Read error: 110 (Connection timed out))
22:00:15 Nick FOAD_ is now known as FOAD (n=dok@dinah.blub.net)
22:00:21gevaertsthis is something that could well be CPU dependent I think
22:00:25n1sin fact there are no errors just warnings so the build table is buggy
22:01:09*tomers good night
22:01:09n1sand why are there just speex warnings on *some* sims?
22:01:28gevaertscompiler version?
22:01:32 Quit tomers ("ChatZilla 0.9.85 [Firefox 3.5.5/20091109125225]")
22:02:32n1sgevaerts: maybe but then why does gcc 4.4 not complain while building for targets?
22:02:57*n1s doesn't like this
22:09:23CIA-6New commit by bluebrother (r23789): Add support for OS X. ...
22:10:36n1sugh the spc decoder is #ifdef hell
22:12:09 Quit xordos ("Leaving")
22:16:40n1si could go for cleaner or more ifdef hell...
22:16:58n1sno idea if that pointer thingy is useful for anything
22:18:06 Quit Tomis (Read error: 110 (Connection timed out))
22:20:35 Join Tomis [0] (n=Tomis@70.134.104.79)
22:22:26CIA-6New commit by nls (r23790): Fix yellow by making the ifdef hell slightly worse
22:25:13 Join Tomis2 [0] (n=Tomis@70.134.85.196)
22:28:31CIA-6New commit by mc2739 (r23791): Manual - add missing {} which caused spacing problems
22:31:00 Quit geertvdijk ("ChatZilla 0.9.85 [Firefox 3.5.5/20091102152451]")
22:33:28***Saving seen data "./dancer.seen"
22:34:46 Join Tomis2_ [0] (n=Tomis@70.134.105.111)
22:35:05CIA-6New commit by bluebrother (r23792): cp on OS X doesn't know the option -u. It isn't really needed, so simply remove it.
22:35:46 Join evilnick [0] (i=4571af51@rockbox/staff/evilnick)
22:36:22 Nick Ypsy is now known as YPSY (n=ypsy@geekpadawan.de)
22:37:58xxarthur33xxso
22:37:59CIA-6New commit by nls (r23793): Disable strict aliasing for speex for now to get rid of warnings, some very nasty casting in here
22:38:51n1slet's cast between int* and float* !
22:39:30 Quit Tomis (Read error: 110 (Connection timed out))
22:39:31 Nick Tomis2_ is now known as Tomis (n=Tomis@70.134.105.111)
22:39:32 Join MG_Man [0] (n=MGMan@11.208.101.97.cfl.res.rr.com)
22:42:00CIA-6New commit by nls (r23794): Oops, correct typo that made the previous commit useless
22:43:22 Join Tomis2_ [0] (n=Tomis@70.134.74.179)
22:45:00 Quit Tomis2 (Connection timed out)
22:45:39 Quit DerPapst ("Leaving.")
22:51:14rasherdomonoky: It's done automagically if the clip is built by tools/checkwps/buildall.sh
22:51:55 Quit Omlet ("( www.nnscript.com :: NoNameScript 4.22 :: www.esnation.com )")
22:52:02domonokyrasher: it should, i just added it to targets.txt :-)
22:52:33rasherThen the rest should happen within the next 24 hours (I forget when they're built)
22:53:01domonokyoki, good :-)
22:53:06MG_ManRight, well, I've updated the boxamp theme again, now that I've actually had the chance to use it on a player
22:53:28MG_ManIt would flip out when it went from a song with album art ot a song without, at least whle the HD was looking for the song
22:53:31MG_ManSo I fixed all that
22:53:45 Part froggyman
22:53:49MG_ManI'm going to add custom status bar support to it before I re-upload it though
22:54:19xxarthur33xxI have a question about nano 2g
22:54:21domonokyrasher: could you take a look at FS #10794 (sbs checking for the theme site) but unfornatly i still can really test it.. maybe you can ? :-)
22:54:38xxarthur33xxI was just looking at a picture and I tried to get out of the picture and it wouldnt let me
22:57:12 Join Thundercloud [0] (i=thunderc@persistence.flat.devzero.co.uk)
22:57:24xxarthur33xxHow would I go about getting out of the picture veiwer?
22:57:51n1szz
22:57:56amiconnUnhelpful: __builtin_clz being undef for zero argument is due to how the x86 instruction works (*imo* a silly thing)
22:57:57n1sxxarthur33xx: check the manual
22:58:51n1samiconn: but what result will gcc's implementations on archs without clz instruction give?
22:58:56n1sfor 0
22:59:02amiconnkugel: Your patch doesn't apply. There's a conflict in led-mini2440.h
22:59:05 Quit Tomis (Read error: 110 (Connection timed out))
22:59:06 Nick Tomis2_ is now known as Tomis (n=Tomis@70.134.74.179)
22:59:22amiconnn1s: Could be anything
22:59:40amiconn32, 0 or even random garbage
23:00
23:00:04n1sexactly, so we shouldn't use those if the input can be 0
23:00:24n1sand yes it is silly
23:01:47amiconnHmm, probably
23:01:52n1shmm, actually 31 - __builtin_clz will give -1 for 0
23:02:12amiconn?
23:02:14n1sif clz for 0 is 32 (as it is on arm)
23:02:51xxarthur33xxI just checked the manual and there isn't even anything about viewing images
23:03:21n1samiconn: what i used for the integer log function in flac
23:03:26evilnickxxarthur33xx: There is.
23:04:09xxarthur33xxevilnick: For the nano 2g?
23:04:22 Quit n1s ("Lämnar")
23:04:34evilnickxxarthur33xx: http://download.rockbox.org/daily/manual/rockbox-ipodnano2g/rockbox-buildch10.html#x13-22100010.3.3
23:05:44xxarthur33xxoh wow, I totally missed that.
23:05:49xxarthur33xxThanks for the link.
23:06:08evilnickxxarthur33xx: And also: http://download.rockbox.org/daily/manual/rockbox-ipodnano2g/rockbox-buildch10.html#x13-21500010.3
23:09:42 Join MethoS- [0] (n=clemens@134.102.106.250)
23:15:16Unhelpfuln1s: there's a clz implementation in pictureflow that's *reasonably* fast and gives 32 for 0 input... we also have several of our own elsewhere in rockbox, don't we?
23:16:09Unhelpful"reasonably" meaning i benched it against several other bit-hack methods on arm... and i think i had gevaerts run on coldfire, but i don't remember now.
23:17:17kugelamiconn: strange, that file should be deleted I think
23:17:40kugelamiconn: anyway, the hunk shouldn't matter for the other real-led targets
23:19:44Unhelpfulthe methods i tried were based on the ones here: http://graphics.stanford.edu/~seander/bithacks.html#IntegerLogObvious
23:24:53 Quit p3tur ("Zzzzz")
23:26:10 Quit petur ("Zzzzz too")
23:29:19 Quit bertrik ("De groeten")
23:30:05 Quit Gregg ("CGI:IRC (EOF)")
23:31:53 Quit maraz (Remote closed the connection)
23:39:13 Join robin0800 [0] (n=quassel@cpc3-brig8-0-0-cust436.brig.cable.ntl.com)
23:44:00 Join maraz [0] (i=maraz@xob.kapsi.fi)
23:48:08 Quit robin0800 (Remote closed the connection)
23:48:55 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
23:50:06 Join robin0800 [0] (n=quassel@cpc3-brig8-0-0-cust436.brig.cable.ntl.com)
23:57:17 Join MichaelBerry [0] (n=545c2871@giant.haxx.se)
23:57:21 Join froggyman [0] (n=sopgenor@pool-72-69-220-194.chi01.dsl-w.verizon.net)

Previous day | Next day