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 2010-01-24

00:01:29 Quit intrados1 (Connection timed out)
00:06:56 Part froggyman
00:07:42 Join intrados1 [0] (
00:08:02Mode"#rockbox +v intrados1 " by rb_voicebot (
00:08:50 Part stooo
00:21:50 Join MagusG [0] (
00:22:10Mode"#rockbox +v MagusG " by rb_voicebot (
00:22:14 Quit Bagder_ ("It is time to say moo")
00:23:40 Quit Zagor ("Leaving")
00:23:52 Quit MagusG (Read error: 113 (No route to host))
00:24:10 Join MagusG [0] (
00:24:30Mode"#rockbox +v MagusG " by rb_voicebot (
00:25:31 Quit dantje (Client Quit)
00:26:31 Quit hebz0rl_ (Read error: 110 (Connection timed out))
00:28:32Unhelpfulhrm, could a similar method to the one used for overlay plugins be used to fill the remaining codec iram with a division LUT for APE? or should i just pick a size that leaves some wiggle room and leave it at that?
00:29:30saratogaUnhelpful: did you ever look at that codesourcery gcc branch?
00:30:14Unhelpfulsaratoga: only so far as to realize it was a bigger mess to build than i wanted to deal with at the time... why do you ask? might it help here?
00:30:52saratogasomeone benchmarked it for Cortex and concluded it was quite a bit faster at ffmpeg then plain gcc4
00:33:02Unhelpfulhrm. i wonder how much is down to optimization of the code being compiled and how much is from better library routines... in the case of APE i have dividers now that are faster than libgcc for all ARM versions we use - around 20% faster on arm7 and more like 2-5% on arm9e or arm11
00:33:33Unhelpfulthough i've not actually benched on arm9e, as i don't have a working target that uses it... nor is there one to my knowledge :)
00:33:54Unhelpfulalso, i wonder if gcc's autovectorization is involved
00:34:55saratogai think this was with vectorization disabled
00:36:12Unhelpfuli'd be curious as to how vector-friendly some of our fixed-point stuff might be
00:37:26Unhelpfulthe VFP is much, much slower than libgcc's or my dividers for scalar values, although part of that is due to converting operands and results, which wouldn't happen if the data stayed in float or double
00:37:52 Join hebz0rl [0] (
00:38:12Mode"#rockbox +v hebz0rl " by rb_voicebot (
00:38:18 Quit tchan ("WeeChat 0.3.1-dev")
00:38:36saratogamost of the codecs should be vector friendly, but NEON is only on really high end targets, so that just leaves VFP
00:38:43saratogahigh end ARM that is
00:39:59Unhelpfuladding a reciprocal LUT for small divisors to the arm4 divider that ape uses gets the ape normal sample to 100% realtime... i'm just trying to figure out how to size the LUT appropriately for targets with varying codec iram size
00:40:08 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
00:40:28Mode"#rockbox +v tchan " by rb_voicebot (
00:40:34 Join stooo [0] (
00:40:39 Join S_a_i_n_t [0] (i=S_a_i_n_@
00:40:54Mode"#rockbox +v stooo " by rb_voicebot (
00:40:59Mode"#rockbox +v S_a_i_n_t " by rb_voicebot (
00:56:43 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
00:59:58 Quit robin0800_ (Remote closed the connection)
01:05:05 Quit stooo ("Leaving.")
01:09:36 Quit bertrik ("De groeten")
01:09:38 Quit DataGhost (Read error: 113 (No route to host))
01:26:29 Join moos [0] (
01:26:49Mode"#rockbox +v moos " by rb_voicebot (
01:39:22mc2739saratoga, linuxstb: tags/bootloader_ams_v1/ is indeed the Ver 1.0 bootloaders for the AMS devices. It was created on Oct. 9, 2009. Since the bootloaders for e200v2 and Fuze are dated July 5, 2009, I assume they just never got placed on the download server. I have checked out that version and if you are in agreement, I will create bootloaders for testing and placement on the download server.
01:52:21***Saving seen data "./dancer.seen"
01:55:26 Quit pjm0616 (Client Quit)
02:01:23AlexPmc2739: That makes sense.
02:03:31 Quit geertvdijk ("ChatZilla 0.9.86 [Firefox 3.5.3/20090824101458]")
02:04:18S_a_i_n_tIf there's any WPS savvy devs here, can I please please! *PLEASE!!* get you to look at: and tell me what I've messed up? Alternating sublines for the current/next track info go crazy when returning from the hold screen. There's a recent(-ish) SIM for Nano 1st Gen I've prebuilt here: let a track play for a minute or so with hold on then take hold off & y
02:07:04mc2739S_a_i_n_t: You might want to look at FS #10926 and see if it could be similar to you problem
02:07:36S_a_i_n_tthansk...will do.
02:10:47S_a_i_n_tmc2739: no, wish it was though. when returning from the hold screen all the transitions that the alternating sublines *would* have made (if visible) are backed up and alternate in sequence untill finished.
02:10:52S_a_i_n_tdoes that explain it?
02:12:27S_a_i_n_tif 50 sublines would have alternated while hold is on, when i take hold off they alternate at approx 0.1sec's weird
02:13:31mc2739oh, I understand now
02:13:57Unhelpfulit looks like most targets only define iram size in :/
02:14:57saratogaUnhelpful: usually we just check for PP5022 or the iaudios
02:15:09saratogathey're the only non 96k IRAM targets where IRAM is useful
02:15:44 Join rhodan_ [0] (
02:16:04Mode"#rockbox +v rhodan_ " by rb_voicebot (
02:19:09Unhelpfulso beat f, clip, and nano2g i shouldn't worry about? also this method isn't really applicable to CF as it lacks a high-word multiply - although perhaps it might be worth it even though it's a few extra instructions
02:20:38saratogathe first two yes, i'm not sure about the nano2g actually
02:29:58 Quit rhodan (Read error: 110 (Connection timed out))
02:33:56TheSevennano2g could give up to 128KB of IRAM to codecs
02:34:20TheSevenand IRAM is significantly faster than DRAM, however twice as slow as the cache
02:34:24Unhelpfulfor armv4 without iram i'm not sure how big a win this really is... pretty much it depends on being able to fetch a reciprocal for divisors below 5000 or so without a great delay
02:34:35TheSevencache is 4KB instr and data each IIRC
02:39:47Unhelpful4096 is enough to cover just over half of the divisions done by ape codec on the normal sample file... i bumped it up to 5500 or so on my e200 just because i had the room, but it doesn't make a huge difference. i never actually timed with cutting off reciprocals at 4096, though, i'll try it now.
02:40:57Unhelpfulalso nano2g manages a respectable speed without this rather nasty hack ;)
02:40:58 Quit DerPapst ("Leaving.")
02:42:18TheSevenUnhelpful: could we maybe do ape-4000 realtime with that hack? :-D
02:42:37saratogawhat does the nano 2 use the other IRAM for? IIRC it has a lot right?
02:42:45UnhelpfulTheSeven: gimme a sec, i'll paste my modification...
02:43:11TheSevensaratoga: 176KB total, and ~50 needed by the core
02:43:28TheSeventhe meizus have 256KB total
02:48:11saratogaTheSeven: if you get bored, you should run test_codec on the samples in the wiki and then post the results
02:48:31TheSeveni already did that some months ago
02:48:50*TheSeven wonders where the results are...
02:50:18saratogaif you find them put them on the wiki
02:50:24 Quit hebz0rl ("Ex-Chat")
02:51:45Unhelpfulhere's a patch, change recip_max to change the table size. since 1 can't be handled correctly by this method anyway, the table starts at 3. a subtraction is done, a negative result branches to code that handles divide by 0/1/2.
02:51:56*Unhelpful always forgets :/
02:52:54Unhelpfulhrm, i should probably move the branch for div0 after the 1/2 handling
02:53:22TheSevensome of the wma tests may have been tested without an IMDCT
02:53:53 Quit efyx_ (Remote closed the connection)
02:56:13saratogathats right i've seen these
02:56:20saratogadid you ever figure out why the times were so slow?
02:56:34*TheSeven guesses that it's slow iram
02:56:35saratogawhat about hte WMA tests?
02:57:09saratogaWMA is half as fast as the Nano 1G, and even slower then the gigabeat f which doesn't use IRAM
02:58:23TheSevenI haven't analyzed that in detail
02:58:26saratogadid it really take 2 hours to decode a single APE file?
02:59:17TheSevenc5000 was extremely much slower than c4000
03:00:15saratogatheres no posted battery benches, but I remember you saying it was pretty good even so?
03:01:06TheSeven33h with mp3 192kbit/s, backlight off all the time
03:01:16*Unhelpful tries ape high just to hurt his e200 some
03:01:44saratogathats amazing
03:01:56TheSevenhowever the ipod was connected to a set of pc speakers and not headphones
03:01:58Unhelpfulbut i know this is very unlikely to get that to even half realtime, esepecially as the division count is fixed regardless of bitrate
03:02:05TheSevenmay make a difference of some few mA
03:02:28saratogathe samsung CPU must be many times more power efficient then AMS or PP
03:02:34saratogaper MHz
03:03:21TheSevenliar managed to bring the whole ipod down to ~11mA with the LCD completely powered off but the CPU active (not sleeping, 96MHz most of the time)
03:03:38TheSevenbut that was really aggressive undervolting
03:03:58saratogaPP is about 4 mA for MHz
03:04:14saratogaerr 4MHz per mA
03:04:18TheSevenmy ipod is somewhere at ~12mA average when playing mp3s with backlight off but lcd on
03:04:37TheSeven(assuming the battery is still at its 400mAh)
03:04:48TheSevenit's 3 years old btw, and was used quite a lot
03:04:49saratogathese are 90nm chips?
03:04:55TheSevenno idea
03:05:13 Join JdGordon1 [0] (
03:05:34Mode"#rockbox +v JdGordon1 " by rb_voicebot (
03:05:49TheSevenapple achieves ~15mA average on them when feeding into 32 ohm phones
03:05:52 Quit JdGordon1 (Client Quit)
03:08:38saratogai guess that kind of efficiency is what apple's economy of scale get you
03:10:15 Quit Strife1989 ("If you hold a Unix shell to your ear, you can hear the C.")
03:29:16 Quit scorche (" rawr...that is all...rawr")
03:30:08 Quit TheSeven|Mobile ()
03:30:32Unhelpfulouch, 64%, it's actually *worse*. we don't use iram dynamically for filter buffers, do we? perhaps i need to come up with a way to make this hack selectable at codec startup, or something
03:37:11Unhelpfulvector_add/vector_sub have no optimizations for armv4 other than load/store multiple? the vector_generic implementation seems to suggest that these may be implemented with "normal" integer arithmetic for these sizes, so SWAR might be worth considering.
03:39:04 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
03:39:25Mode"#rockbox +v scorche " by rb_voicebot (
03:41:28 Join JdGordon1 [0] (
03:41:50Mode"#rockbox +v JdGordon1 " by rb_voicebot (
03:42:20 Quit JdGordon (Ping timeout: 180 seconds)
03:48:42 Nick fxb is now known as fxb__ (n=felixbru@
03:52:23***Saving seen data "./dancer.seen"
03:59:11 Quit moos ("ChatZilla 0.9.86 [Firefox 3.6/20100115144158]")
04:04:57 Join Barahir [0] (
04:05:17Mode"#rockbox +v Barahir " by rb_voicebot (
04:07:31 Quit Sajber^ ("Leaving.")
04:08:03 Join Kitr88 [0] (
04:08:23Mode"#rockbox +v Kitr88 " by rb_voicebot (
04:16:45 Quit GeekShadow ("The cake is a lie !")
04:31:30 Quit gevaerts (Remote closed the connection)
04:31:31 Quit FlynDice (Remote closed the connection)
04:31:33 Quit lostlogic (Remote closed the connection)
04:34:46 Quit JdGordon1 (
04:34:46 Quit Spaceghost (
04:34:46 Quit toffe82 (
04:34:46 Quit jfc (
04:34:46 Quit fxb__ (
04:34:46 Quit simabeis (
04:34:46 Quit BlakeJohnson86 (
04:34:46 Quit maraz (
04:34:46 Quit gibbon_ (
04:34:46 Quit jho (
04:34:46 Quit dionoea (
04:34:46 Quit zu (
04:34:46 Quit scorche|sh (
04:34:46 Quit linuxstb (
04:34:46 Quit lyngaas (
04:34:46 Quit Kitr88 (
04:34:46 Quit tchan (
04:34:46 Quit Strife89 (
04:34:46 Quit bluebroth3r (
04:34:46 Quit rb_voicebot (
04:34:46 Quit n17ikh (
04:34:46 Quit alexbobp (
04:34:46 Quit sbhsu (
04:34:46 Quit intrados1 (
04:34:46 Quit lyon8 (
04:34:46 Quit Horscht (
04:34:46 Quit at0m (
04:34:46 Quit Zarggg_ (
04:34:46 Quit Res1 (
04:34:46 Quit crashd (
04:34:46 Quit sinuc (
04:34:46 Quit rvvs89 (
04:34:46 Quit jvd (
04:34:46 Quit Kohlrabi (
04:34:46 Quit Xerion (
04:34:46 Quit ChanServ (
04:34:46 Quit MagusG (
04:34:46 Quit CaptainKewl (
04:34:46 Quit phanboy4 (
04:34:46 Quit yosafbridge (
04:34:46 Quit liar (
04:34:46 Quit TheSeven (
04:34:46 Quit Barahir_ (
04:34:46 Quit elcan (
04:34:47 Quit Zambezi (
04:34:47 Quit FOAD (
04:34:47 Quit chaos (
04:34:47 Quit AlexP (
04:34:47 Quit panni_ (
04:34:47 Quit antil33t (
04:34:47 Quit Kitar|st (
04:34:47 Quit GodEater (
04:34:47 Quit mikroflops_ (
04:34:47 Quit dmb (
04:34:47 Quit avacore (
04:34:47 Quit Tuplanolla (
04:34:47 Quit jordan` (
04:34:47 Quit BeFalou (
04:34:55 Quit crwl (
04:34:55 Quit martian67 (
04:34:55 Quit ps-auxw (
04:34:55 Quit BHSPitMonkey (
04:34:55 Quit pixelma (
04:34:55 Quit amiconn (
04:34:55 Quit Tomis (
04:34:55 Quit AndyI (
04:34:55 Quit fdinel (
04:34:55 Quit Rob2223 (
04:34:55 Quit mc2739 (
04:34:55 Quit Curtman (
04:34:55 Quit linuxguy3 (
04:34:55 Quit nima (
04:34:55 Quit preglow (
04:34:55 Quit tha (
04:34:55 Quit Unhelpful (
04:34:55 Quit beta2k (
04:34:55 Quit tmzt (
04:34:55 Quit Topy44 (
04:34:55 Quit pippijn (
04:34:55 Quit YPSY (
04:34:55 Quit z35 (
04:34:55 Quit tomaw__ (
04:34:55 Quit blithe (
04:34:55 Quit dys (
04:34:55 Quit Naked (
04:34:55 Quit aevin (
04:34:55 Quit rphillips (
04:34:55 Quit B4gder (
04:34:55 Quit parafin (
04:34:55 Quit Barahir (
04:34:55 Quit scorche (
04:34:55 Quit rhodan_ (
04:34:55 Quit S_a_i_n_t (
04:34:55 Quit kloeri (
04:34:55 Quit MethoS- (
04:34:55 Quit killan (
04:34:55 Quit Kopfgeldjaeger (
04:34:55 Quit grndslm (
04:34:55 Quit kloeri_ (
04:34:55 Quit jd (
04:34:55 Quit tarbo (
04:34:55 Quit krazykit (
04:34:55 Quit topik (
04:34:56 Quit Torne (
04:34:56 Quit bzed (
04:34:56 Quit goffa (
04:34:56 Quit ehntoo (
04:34:56 Quit Galois (
04:34:56 Quit rasher (
04:34:56 Quit stacker55 (
04:34:56 Quit CIA-44 (
04:34:56 Quit HBK (
04:34:56 Quit ThomasAH (
04:34:56 Quit saratoga (
04:43:07NJoinrb_voicebot [0] (
04:43:07 Join dys` [0] (
04:43:07NJoinelcan [0] (
04:43:07 Join Kitar|st [0] (
04:43:07Mode"#rockbox +ovvv rb_voicebot dys` elcan Kitar|st " by
04:43:07 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
04:43:07 Join lostlogic [50] (n=lostlogi@rockbox/developer/lostlogic)
04:43:07 Join FlynDice [0] (
04:43:07NJoinBarahir [0] (
04:43:07Mode"#rockbox +vvvv gevaerts lostlogic FlynDice Barahir " by
04:43:07NJoinJdGordon1 [0] (
04:43:07NJoinscorche [50] (n=scorche@rockbox/administrator/scorche)
04:43:07NJoinrhodan_ [0] (
04:43:07NJoinS_a_i_n_t [0] (i=S_a_i_n_@
04:43:07Mode"#rockbox +vvvv JdGordon1 scorche rhodan_ S_a_i_n_t " by
04:43:07NJointchan [0] (n=tchan@lunar-linux/developer/tchan)
04:43:07NJoinMagusG [0] (
04:43:07NJoinpanni_ [0] (
04:43:07NJoinkloeri [0] (i=kloeri@freenode/staff/exherbo.kloeri)
04:43:07Mode"#rockbox +vvvv tchan MagusG panni_ kloeri " by
04:43:07NJoindys [0] (
04:43:07NJoinSpaceghost [0] (i=quassel@unaffiliated/spaceghost)
04:43:07NJoinStrife89 [0] (
04:43:07NJoinBHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
04:43:07Mode"#rockbox +vvvv dys Spaceghost Strife89 BHSPitMonkey " by
04:43:07NJoinpixelma [0] (i=quassel@rockbox/staff/pixelma)
04:43:07NJoinamiconn [0] (i=quassel@rockbox/developer/amiconn)
04:43:07NJoinTomis [0] (n=Tomis@
04:43:07NJoinAndyI [0] (n=pasha_in@
04:43:07Mode"#rockbox +vvvv pixelma amiconn Tomis AndyI " by
04:43:07NJoinCaptainKewl [0] (
04:43:07NJoinfdinel [0] (
04:43:07NJointoffe82 [0] (
04:43:07NJoinMethoS- [0] (n=clemens@
04:43:07Mode"#rockbox +vvvv CaptainKewl fdinel toffe82 MethoS- " by
04:43:07NJoinlyon8 [0] (
04:43:07NJoinphanboy4 [0] (
04:43:07NJoinantil33t [0] (
04:43:07NJoinyosafbridge [0] (
04:43:07Mode"#rockbox +vvvv lyon8 phanboy4 antil33t yosafbridge " by
04:43:07NJoinliar [0] (
04:43:07NJoinkillan [0] (
04:43:07NJoinKopfgeldjaeger [0] (
04:43:07NJoinbluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
04:43:07Mode"#rockbox +vvvv liar killan Kopfgeldjaeger bluebroth3r " by
04:43:07NJoinRob2223 [0] (
04:43:07NJoinHorscht [0] (n=Horscht2@xbmc/user/horscht)
04:43:07NJoinTheSeven [0] (n=theseven@rockbox/developer/TheSeven)
04:43:07 Join GodEater [0] (n=bibble@rockbox/staff/GodEater)
04:43:07Mode"#rockbox +vvvv Rob2223 Horscht TheSeven GodEater " by
04:43:07NJoinjfc [0] (
04:43:07NJoingrndslm [0] (
04:43:07NJoinmikroflops_ [0] (
04:43:07NJoinmc2739 [0] (n=mc2739@rockbox/developer/mc2739)
04:43:07Mode"#rockbox +vvvv jfc grndslm mikroflops_ mc2739 " by
04:43:07NJoinkloeri_ [0] (i=kloeri@freenode/staff/exherbo.kloeri)
04:43:07NJoinmartian67 [0] (n=martian6@about/linux/regular/martian67)
04:43:07 Join jd [0] (n=jd@Wikipedia/HellDragon)
04:43:07NJoinNaked [0] (
04:43:07Mode"#rockbox +vvvv kloeri_ martian67 jd Naked " by
04:43:07NJointarbo [0] (n=me@unaffiliated/tarbo)
04:43:07NJoindmb [0] (n=Dmb@unaffiliated/dmb)
04:43:07 Join Zambezi [0] (i=Zulu@unaffiliated/zambezi)
04:43:07NJoinCurtman [0] (
04:43:07Mode"#rockbox +vvvv tarbo dmb Zambezi Curtman " by
04:43:07NJoinn17ikh [0] (
04:43:07NJoinavacore [0] (
04:43:07NJoinat0m [0] (
04:43:07NJoinkrazykit [0] (n=kkit@
04:43:07Mode"#rockbox +vvvv n17ikh avacore at0m krazykit " by
04:43:07NJoinlinuxguy3 [0] (
04:43:07NJoinTuplanolla [0] (n=jani@unaffiliated/tuplanolla)
04:43:07NJoingibbon_ [0] (
04:43:07NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
04:43:07Mode"#rockbox +vvvo linuxguy3 Tuplanolla gibbon_ scorche|sh " by
04:43:07NJoinzu [0] (
04:43:07NJoinfxb__ [0] (n=felixbru@
04:43:07NJoinsimabeis [0] (
04:43:07NJoinlyngaas [0] (
04:43:07Mode"#rockbox +vvvv zu fxb__ simabeis lyngaas " by
04:43:07NJoinBlakeJohnson86 [0] (
04:43:07NJoinmaraz [0] (i=maraz@
04:43:07NJoinjho [0] (n=jh@unaffiliated/cene)
04:43:07NJoinlinuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
04:43:07Mode"#rockbox +vvvv BlakeJohnson86 maraz jho linuxstb " by
04:43:07NJoindionoea [0] (n=dionoea@videolan/developer/dionoea)
04:43:07NJoinjordan` [0] (n=jordan@
04:43:07NJoinalexbobp [0] (n=alex@
04:43:07NJointopik [0] (
04:43:07Mode"#rockbox +vvvv dionoea jordan` alexbobp topik " by
04:43:07NJoinparafin [0] (
04:43:07NJoinTorne [0] (i=torne@rockbox/developer/Torne)
04:43:07NJoinbzed [0] (
04:43:07NJoinThomasAH [0] (
04:43:07Mode"#rockbox +vvvv parafin Torne bzed ThomasAH " by
04:43:07NJoingoffa [0] (n=goffa@
04:43:07NJoinHBK [0] (
04:43:07NJoinCIA-44 [0] (n=CIA@
04:43:07NJoinehntoo [0] (
04:43:07Mode"#rockbox +vvvv goffa HBK CIA-44 ehntoo " by
04:43:07NJoinGalois [0] (
04:43:07NJoinrasher [50] (n=rasher@rockbox/developer/rasher)
04:43:07NJoinaevin [0] (i=eivindsy@unaffiliated/aevin)
04:43:07NJoinrphillips [0] (
04:43:07Mode"#rockbox +vovv Galois rasher aevin rphillips " by
04:43:07NJoinB4gder [241] (n=daniel@rockbox/developer/bagder)
04:43:07NJoinstacker55 [0] (
04:43:07Mode"#rockbox +vv B4gder stacker55 " by
04:43:07NJoinYPSY [0] (
04:43:07NJoinps-auxw [0] (n=arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
04:43:07NJoinnima [0] (
04:43:07NJoinz35 [0] (
04:43:07Mode"#rockbox +vvvv YPSY ps-auxw nima z35 " by
04:43:07NJoinpreglow [0] (
04:43:07NJointha [0] (
04:43:07NJointmzt [0] (n=tmzt@
04:43:07NJoinpippijn [0] (n=pippijn@unixcorps/staff/pippijn)
04:43:07Mode"#rockbox +vvvv preglow tha tmzt pippijn " by
04:43:07NJoinUnhelpful [0] (n=quassel@rockbox/developer/Unhelpful)
04:43:07NJointomaw__ [0] (i=tom@freenode/staff/tomaw)
04:43:07NJoinTopy44 [0] (
04:43:07NJoinblithe [0] (
04:43:07Mode"#rockbox +vvvv Unhelpful tomaw__ Topy44 blithe " by
04:43:07NJoinBeFalou [0] (n=mamutoi@unaffiliated/befalou)
04:43:07NJoinbeta2k [0] (
04:43:07NJoincrwl [0] (
04:43:07NJoinAlexP [0] (n=alex@rockbox/staff/AlexP)
04:43:07Mode"#rockbox +vvvo BeFalou beta2k crwl AlexP " by
04:43:07NJoinFOAD [0] (
04:43:07NJoinchaos [0] (n=chaos@gentoo/user/ch4os)
04:43:07NJoinZarggg_ [0] (
04:43:07Mode"#rockbox +vvvv AlexP FOAD chaos Zarggg_ " by
04:43:07NJoinXerion [0] (
04:43:07NJoinKohlrabi [0] (n=Kohlrabi@
04:43:07NJoinRes1 [0] (n=Res@
04:43:07NJoincrashd [0] (
04:43:07Mode"#rockbox +vvvv Xerion Kohlrabi Res1 crashd " by
04:43:07NJoinrvvs89 [0] (n=ivo@pdpc/supporter/base/rvvs89)
04:43:07NJoinjvd [0] (n=syscrash@poipu/developer/syscrash)
04:43:07NJoinsinuc [0] (n=sinuc@unaffiliated/sinuc)
04:43:07NJoinsbhsu [0] (
04:43:07Mode"#rockbox +vvvv rvvs89 jvd sinuc sbhsu " by
04:43:07NJoinChanServ [0] (ChanServ@services.)
04:43:07Mode"#rockbox +o ChanServ " by
04:43:37Mode"#rockbox +v soap " by rb_voicebot (
04:43:37Mode"#rockbox +v advcomp2019 " by rb_voicebot (
04:43:37Mode"#rockbox +v xavieran " by rb_voicebot (
04:43:37Mode"#rockbox +v logbot " by rb_voicebot (
04:43:37Mode"#rockbox +v Slasheri " by rb_voicebot (
04:46:59 Quit TheSeven (Nick collision from services.)
04:47:09 Join Space [0] (i=irssi@unaffiliated/spaceghost)
04:47:11 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
04:47:22 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
04:47:29Mode"#rockbox +v Space " by rb_voicebot (
04:47:31Mode"#rockbox +v TheSeven " by rb_voicebot (
04:50:46 Quit dys (No route to host)
04:58:25 Join kadoban [0] (
04:58:46Mode"#rockbox +v kadoban " by rb_voicebot (
05:04:35 Quit Spaceghost (Read error: 110 (Connection timed out))
05:04:36 Quit elcan (Remote closed the connection)
05:05:29 Join elcan [0] (i=user36@
05:05:49Mode"#rockbox +v elcan " by rb_voicebot (
05:11:57 Join rhodan [0] (
05:12:17Mode"#rockbox +v rhodan " by rb_voicebot (
05:12:48 Quit rhodan_ (Read error: 60 (Operation timed out))
05:16:36 Quit lyon8 ("Bis dann..")
05:17:28 Quit Strife89 ("Bed.")
05:23:04 Nick Space is now known as Spaceghost (i=irssi@unaffiliated/spaceghost)
05:24:56 Part toffe82
05:33:06 Quit MethoS- (Remote closed the connection)
05:46:03 Join ball [0] (
05:46:23Mode"#rockbox +v ball " by rb_voicebot (
05:47:11 Join webguest20 [0] (
05:47:32Mode"#rockbox +v webguest20 " by rb_voicebot (
05:49:14 Quit webguest20 (Client Quit)
05:52:24***Saving seen data "./dancer.seen"
05:54:44 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
05:55:04Mode"#rockbox +v togetic " by rb_voicebot (
06:01:44 Join openstandards [0] (n=openstan@
06:02:05Mode"#rockbox +v openstandards " by rb_voicebot (
06:08:07 Join intrados1 [0] (
06:08:28Mode"#rockbox +v intrados1 " by rb_voicebot (
06:10:21 Quit ball ("leaving")
06:11:36 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
06:11:58Mode"#rockbox +v bmbl " by rb_voicebot (
06:13:40 Quit intrados1 (Remote closed the connection)
06:14:14openstandardshi there is it possible to install onto the sansa c200v2 via the rockbox utility?
06:15:50 Quit togetic ("WeeChat 0.3.0")
06:16:00 Join intrados1 [0] (
06:16:20Mode"#rockbox +v intrados1 " by rb_voicebot (
06:20:22 Quit intrados1 (SendQ exceeded)
06:21:13 Join saratoga [0] (i=9803c6dd@gateway/web/freenode/x-knkmpqqrdrdnpmpk)
06:21:33Mode"#rockbox +v saratoga " by rb_voicebot (
06:21:34saratogaopenstandards: i believe it is disabled until the c200v2 port is usable
06:21:53 Join intrados1 [0] (
06:22:13Mode"#rockbox +v intrados1 " by rb_voicebot (
06:27:30 Quit intrados1 (Remote closed the connection)
06:34:17 Quit goffa (Read error: 60 (Operation timed out))
06:35:08 Quit saratoga ("Page closed")
06:36:56 Join AndyIL [0] (n=pasha_in@
06:37:16Mode"#rockbox +v AndyIL " by rb_voicebot (
06:48:13 Quit AndyI (Read error: 110 (Connection timed out))
06:56:15 Quit Tomis (Read error: 104 (Connection reset by peer))
06:56:27 Join Tomis2 [0] (n=Tomis@
06:56:35 Nick Tomis2 is now known as Tomis (n=Tomis@
06:56:47Mode"#rockbox +v Tomis " by rb_voicebot (
06:59:49 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
07:16:53 Join pjm0616 [0] (n=user@
07:17:13Mode"#rockbox +v pjm0616 " by rb_voicebot (
07:19:02 Join ball [0] (n=ball@
07:19:23Mode"#rockbox +v ball " by rb_voicebot (
07:19:27ballAh nice... playing music on my Rockboxed iPod mini.
07:33:01 Quit JdGordon1 ("Leaving.")
07:33:25 Join Llorean [0] (n=DarkkOne@
07:33:29 Part Llorean
07:36:54ballI'm off to bed now then. Goodnight all.
07:37:08 Quit ball ("leaving")
07:45:01 Quit amiconn (Excess Flood)
07:45:02 Quit pixelma (Excess Flood)
07:45:04 Join pixelma [0] (i=quassel@rockbox/staff/pixelma)
07:45:04 Join amiconn [0] (i=quassel@rockbox/developer/amiconn)
07:45:24Mode"#rockbox +v pixelma " by rb_voicebot (
07:45:24Mode"#rockbox +v amiconn " by rb_voicebot (
07:48:33 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
07:48:54Mode"#rockbox +v Horschti " by rb_voicebot (
07:52:25***Saving seen data "./dancer.seen"
07:55:46 Quit Horscht (Read error: 60 (Operation timed out))
08:14:18 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
08:14:39Mode"#rockbox +v togetic " by rb_voicebot (
08:19:52 Join JdGordon [0] (
08:20:12Mode"#rockbox +v JdGordon " by rb_voicebot (
08:30:40 Quit Spaceghost (Client Quit)
08:33:45 Join Rob2222 [0] (
08:34:06Mode"#rockbox +v Rob2222 " by rb_voicebot (
08:52:03 Quit Rob2223 (Read error: 110 (Connection timed out))
09:15:04 Quit pjm0616 (Remote closed the connection)
09:29:19 Quit phanboy4 (Read error: 110 (Connection timed out))
09:32:06 Quit togetic (Read error: 110 (Connection timed out))
09:32:30 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
09:32:51Mode"#rockbox +v togetic " by rb_voicebot (
09:34:18 Quit elcan (Client Quit)
09:34:25 Join elcan [0] (
09:34:45Mode"#rockbox +v elcan " by rb_voicebot (
09:40:02 Quit panni_ ("( :: NoNameScript 3.81 :: )")
09:43:58 Quit CaptainKewl ("( :: NoNameScript 4.22 :: )")
09:52:28***Saving seen data "./dancer.seen"
09:55:20 Join flydutch [0] (
09:55:41Mode"#rockbox +v flydutch " by rb_voicebot (
09:57:52 Join asper [0] (n=asper@
09:58:13Mode"#rockbox +v asper " by rb_voicebot (
09:58:18asperhi guys
09:58:33aspercan I ask a few things about rockbox and archos ?^
09:58:50asper(wow what a big community !)
09:59:06 Quit asper (Client Quit)
09:59:42 Join asper [0] (n=asper@
10:00:03Mode"#rockbox +v asper " by rb_voicebot (
10:00:35asperis there anyone who is developing rockbox for archos AV series ?
10:01:06advcomp2019asper, i do not know but you might have to wait
10:01:47asperI have an AV500 to "offer" as a develop unit
10:01:50pixelmanot that There once was the start of a port to the AV300 but stalled in early stages and for a long time (years) already
10:02:07pixelmanot that I know of, I wanted to say
10:02:13asperis there a reason why AV series is not supported ?
10:02:37pixelmaporting is hard work and no-one has done it
10:03:30asperI heard about "no unit to test with"... I can solve this problem if this is the real one
10:05:14pixelmayou still have to find a developer (or more than one) who is capable and interested in a port enough to do the work. Rockbox is a volunteer effort and you can't get people to work on something they are not interested in
10:05:24 Join stooo [0] (
10:05:34asperyou are right pixelma
10:05:44Mode"#rockbox +v stooo " by rb_voicebot (
10:05:54pixelmais there a AV* users community?
10:06:01asperis there any deeloper here to ask ?
10:06:12pixelmamaybe you can gather interest there...
10:06:34aspermy av500 is already "archopen" tested
10:07:12aspernow that new generation of archos devices had been "opned" with all those info I think it should be not so hard to implement rockbox for that device
10:09:10asperor is htere a mail and/or a form to compile to say taht I have a av500 to send ?
10:13:18asperanyway av300 has a MAS chip while AV500 has a new TMS320
10:13:43asperif someone is interested here I am
10:39:28 Join mt [0] (n=mtee@rockbox/developer/mt)
10:39:48Mode"#rockbox +v mt " by rb_voicebot (
10:41:52 Quit bmbl (Read error: 110 (Connection timed out))
10:42:12 Join moos [0] (
10:42:32Mode"#rockbox +v moos " by rb_voicebot (
10:44:02 Join stoffel [0] (
10:44:22Mode"#rockbox +v stoffel " by rb_voicebot (
10:45:05 Join hannesd [0] (
10:45:26Mode"#rockbox +v hannesd " by rb_voicebot (
10:46:53mtIs there a way for the delete_handler to know whether an audio file is playing ( like a status or an attribute for the file) or should I go on and think of something ?
10:49:12linuxstbmt: What do you want to do?
10:52:55mtlinuxstb: As a first step in a solution for FS #10874, the delete handler should be able to take actions based on whether the file is playing or not.
10:55:42 Join FOAD_ [0] (
10:56:03Mode"#rockbox +v FOAD_ " by rb_voicebot (
10:56:48mtlinuxstb: I'm thinking of adding a file attribute to filetypes.h, FILE_ATTR_AUDIO_PLAYING, or something like that, that would be (FILE_ATTR_AUDIO | (value) ). When a file played, it's attribute is set to that, and then it would be possible to determine the playback status from the delete_handler for example.
10:58:49linuxstbI would expect you to be able to get the filename of the currently playing track easily.
11:07:57 Quit asper ("CGI:IRC (Ping timeout)")
11:08:01 Join bluebrother [0] (
11:08:21Mode"#rockbox +v bluebrother " by rb_voicebot (
11:09:55 Quit moos (Remote closed the connection)
11:11:13 Join pamaury [0] (
11:11:34Mode"#rockbox +v pamaury " by rb_voicebot (
11:11:51 Quit FOAD (Read error: 110 (Connection timed out))
11:11:51 Nick FOAD_ is now known as FOAD (
11:11:53 Join moos [0] (
11:12:13Mode"#rockbox +v moos " by rb_voicebot (
11:14:14 Join rhodan_ [0] (
11:14:35Mode"#rockbox +v rhodan_ " by rb_voicebot (
11:14:55 Join _zic [0] (
11:15:16Mode"#rockbox +v _zic " by rb_voicebot (
11:21:06 Quit rhodan (Read error: 60 (Operation timed out))
11:25:21 Quit bluebroth3r (Read error: 110 (Connection timed out))
11:26:52 Join Highlander [0] (
11:27:12Mode"#rockbox +v Highlander " by rb_voicebot (
11:28:00 Join DerPapst [0] (
11:28:21Mode"#rockbox +v DerPapst " by rb_voicebot (
11:30:27 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
11:30:48Mode"#rockbox +v DataGhost " by rb_voicebot (
11:36:10 Join bertrik [0] (
11:36:31Mode"#rockbox +v bertrik " by rb_voicebot (
11:50:51 Part moos
11:52:29***Saving seen data "./dancer.seen"
11:53:38 Quit stooo ("Leaving.")
12:02:55 Join pjm0616 [0] (n=user@
12:03:06 Join stooo [0] (
12:03:16Mode"#rockbox +v pjm0616 " by rb_voicebot (
12:03:27Mode"#rockbox +v stooo " by rb_voicebot (
12:06:54 Quit flydutch (Read error: 104 (Connection reset by peer))
12:10:11 Join flydutch [0] (
12:10:31Mode"#rockbox +v flydutch " by rb_voicebot (
12:12:31 Join MethoS- [0] (n=clemens@
12:12:52Mode"#rockbox +v MethoS- " by rb_voicebot (
12:14:54 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
12:15:15Mode"#rockbox +v bmbl " by rb_voicebot (
12:18:41 Quit togetic (Read error: 110 (Connection timed out))
12:26:06mtShould the user be notified that the file to be deleted is the currently playing one ?
12:32:06 Quit stooo (Read error: 104 (Connection reset by peer))
12:32:29 Join stooo [0] (
12:32:50Mode"#rockbox +v stooo " by rb_voicebot (
12:33:27GodEatermt: seems reasonable to me
12:35:17pixelmais this about the "delete playing" option in the WPS context menu?
12:35:35pixelmaisn't it obvious then?
12:36:09mtCurrently, what I'm thinking of is that, when a file is to be deleted, the yes-no message "Delete ?" will be preceded by a warning that the playback will skip to the next track in playlist once the file is deleted.
12:36:35AlexPBut it won't
12:36:40AlexPIt carries on playing
12:36:45AlexPAs long as it is buffered
12:36:55mtAlexP : That's what I intend to work on next.
12:37:04AlexPYou want to stop it playing?
12:37:16mtpixelma: FS #10874
12:37:41mtAlexP: because of that bug ^ .
12:38:08AlexPAnd that bug can't be fixed in any other way?
12:39:16rasherThat seems to be hiding the symptoms rather than fixing the bug
12:40:50pixelmathe bug says there is no problem if the song to be deleted is already fully bufferd and playback "just" gets confused if it's only halfway buffered
12:40:58 Quit BHSPitMonkey (Remote closed the connection)
12:40:58 Nick fxb__ is now known as fxb (n=felixbru@
12:41:04mtI'm still just testing, haven't settled on a solution yet. And won't settle on one until it's been discussed.
12:43:09AlexPI'd prefer to let it keep playing, then when get to the end of the buffered section skip to the next (instead of crashing on attempting to rebuffer the non-existing file)
12:43:41mtHow about this : User selects file for deletion -> User receives a warning that if the file is not completely buffered playback will skip to the next track after it consumes the current buffer.
12:44:18mtThen playback engine handles the skipping if it couldn't get more buffer for the said track.
12:44:40AlexPThe bigger issue here is stopping the crash, we can think about messages afterwards :)
12:45:12 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
12:45:16mtum .. good point :)
12:45:33Mode"#rockbox +v GeekShadow " by rb_voicebot (
12:48:45*pixelma wonders if this bugs also happens on hwcodec and if finding this out would help pinning doen the actual problem
12:48:55pixelmadown too
12:51:28mtI'm not 100% sure, but I don't think this bug depends on whether the codec is a sw or hw one.
12:53:35 Nick Naked is now known as Hadaka (
12:53:39pixelmawell, playback and buffering is entirely different
12:55:48amiconnUnless someone broke it recently, hwcodec playback has no problem with the currently playing file going awol
12:56:24amiconnIt will just skip to the next file (and repeat as necessary if the next file(s) are also missing)
12:56:46rasheramiconn: does it skip when the file is deleted, or when it reaches the end of what's been buffered?
12:56:53amiconnIf the swcodec engine doesn't manage that, it's a bug
12:57:09amiconnIt skips when it tries to buffer more
12:57:23rasherI don't think anyone's arguing that it's not a bug
12:57:50amiconnA bug in the swcodec playback engine to be precise
12:58:06pixelmamaybe it once worked on swcodec too...
12:58:33pixelmaeven with half-buffered tracks (or just that this case wasn't tested)
13:03:02amiconnTheSeven: The higher APE compression levels don't change the necessary processing power for the range decoding (where division is involved).
13:03:24amiconnOnly the amount of filtering is what changes
13:04:21amiconnAnd the filtering doesn't involve division, and is optimized a lot already
13:07:22amiconnUnhelpful: All filter buffers which fit there are in iram. And armv4 vector* operations have no special optimisations because we use 32 bit there
13:07:33amiconnIt's faster than using 16 bit and swar
13:08:11amiconnThe reason is the scalarproduct stuff; vector_add/sub would be a little faster but not enough to make up for the way slower scalarproduct
13:09:04 Join dantje [0] (
13:09:25Mode"#rockbox +v dantje " by rb_voicebot (
13:09:52 Join stooo1 [0] (
13:10:13Mode"#rockbox +v stooo1 " by rb_voicebot (
13:20:14 Quit hannesd ()
13:22:08 Quit stooo (Read error: 110 (Connection timed out))
13:26:30 Join maruk [0] (
13:26:35 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
13:26:51Mode"#rockbox +v maruk " by rb_voicebot (
13:26:56Mode"#rockbox +v domonoky " by rb_voicebot (
13:27:12 Join teru [0] (
13:27:33Mode"#rockbox +v teru " by rb_voicebot (
13:33:48 Join S_a_i_n_t_ [0] (i=S_a_i_n_@
13:34:08Mode"#rockbox +v S_a_i_n_t_ " by rb_voicebot (
13:38:17 Join Casainho [0] (
13:38:38Mode"#rockbox +v Casainho " by rb_voicebot (
13:38:41 Join petur [0] (
13:39:02Mode"#rockbox +v petur " by rb_voicebot (
13:41:06CIA-44New commit by alex (r24320): Correct the instructions for returning to Rockbox from the OF on the Gigabeat S.
13:44:49 Quit maruk (Read error: 60 (Operation timed out))
13:45:51 Quit S_a_i_n_t (Read error: 60 (Operation timed out))
13:46:18 Join S_a_i_n_t [0] (i=S_a_i_n_@
13:46:38Mode"#rockbox +v S_a_i_n_t " by rb_voicebot (
13:49:52 Nick stooo1 is now known as stooo (
13:52:30***Saving seen data "./dancer.seen"
13:52:41pixelmameh, something in my settings crashes a recent built on my Ondio (default settings seem to be ok)
13:57:55 Join kaniini [0] (
13:58:15Mode"#rockbox +v kaniini " by rb_voicebot (
14:00:02 Join kugel [0] (n=kugel@rockbox/developer/kugel)
14:00:23Mode"#rockbox +v kugel " by rb_voicebot (
14:01:00 Quit S_a_i_n_t_ (Read error: 113 (No route to host))
14:01:07kugelUnhelpful: shouldn't we use your divider for the core too?
14:02:02pixelmanow it's ok (after I flashed it), weird
14:03:55pixelmaI got a blank screen and no playback with the new build only RoLo'd in and chosing a file to play
14:04:08kugelspeaking of it, there's a bug which causes a dataabort when skipping back from the end of a song (maybe it's limited to the ams devices)
14:04:37kugelit definitely doesn't happen on default settings but with the settings I use on my fuze, I still need to investigate that
14:07:14linuxstbkugel: Do you know what's going on with the fuze and e200v2 bootloaders? "v1" is tagged in svn, but the versions on the download server are "1.0RC" mc2739 said he is planning on building the versions from that tag and releasing them.
14:10:05kugelwell, it seems they never landed on the download server
14:10:31kugelbut the only difference between 1.0RC and 1.0 is the version string, and this version still works as of now
14:10:37linuxstbWhy were "RC" bootloaders made available anyway?
14:10:50kugelwe had test builds in the forum
14:15:44Unhelpfulamiconn: i saw the scalarproduct stuff after that... yeah, it would be hard to make up for the cost there... also i notice that my swar code for jpeg reuses some of the values from the addition to compelete the subtraction, so it would be 6 ops for either by itself.
14:19:18Unhelpfulkugel: i'm working on that. the divider for APE handles certain cases in ways that are only optimal for APE - for example it handles large numerators by branching to code that calls itself after shifting left. this makes sense for APE, where this case happens around 1/10^7 of the time, but is handled inline (with slightly more delay for the fast case) by the divider i'm working on for core. also arm9e and arm11 cores will not see any
14:19:18Unhelpful improvement over gcc with this divider
14:19:24 Quit parafin (Read error: 60 (Operation timed out))
14:19:57 Join parafin [0] (
14:20:11 Quit mt (Read error: 113 (No route to host))
14:20:18Mode"#rockbox +v parafin " by rb_voicebot (
14:20:24kugelUnhelpful: getting rid of the libgcc dependency would be good enough
14:22:55Unhelpfuli don't see why we'd want to make a big deal of getting rid of libgcc? i've only coded replacements for its div and mod functions, and clz, and only for one operand size. libgcc also provides support functions for 64-bit multiply and divide, as well as for variable shifts on processors that don't offer variable-shift instructions. this is not getting rid of libgcc, this is replacing a few small bits of it with more optimized ones. :)
14:23:51kugelhm, ok
14:24:40kugelbut it bugs me that eabi gcc pulls in that exception handling in its division routines
14:24:47Unhelpfullibgcc is *huge*. i'm not going to be able to replace all of it :)
14:25:02Unhelpfulunless you build libgcc with exception-handling disabled?
14:26:34kugelyes, but then it bugs me that we need to patch gcc so much :)
14:27:07 Join FOAD_ [0] (
14:27:28Mode"#rockbox +v FOAD_ " by rb_voicebot (
14:28:01Unhelpfulit's a very small makefile patch. perhaps you could convince the gcc developers to offer a configure option along the lines of −−disable-libgcc-execptions-no-really-i-mean-it-do-it-everywhere
14:28:04kugelUnhelpful: did you ever try a 4.3.x eabi gcc? someone linked a blog post which concludes that 4.4.x introduced performance regressions on arm
14:30:20 Join funman [0] (n=fun@rockbox/developer/funman)
14:30:29kugelsadly I'm not able to find any mention of the changes from 4.4.2 to 4.4.3 only
14:30:41Mode"#rockbox +v funman " by rb_voicebot (
14:34:08 Join fdinel [0] (
14:34:29Mode"#rockbox +v fdinel " by rb_voicebot (
14:34:37 Join maruk [0] (
14:34:58Mode"#rockbox +v maruk " by rb_voicebot (
14:35:12funmanfdinel: hi
14:36:26kugelviewvc apparently can't handle showing the log of 150k revisions
14:37:04funmanfdinel: nice to see you ! you're catching up with development again?
14:38:17amiconnUnhelpful: This is already verified. Go back a bit in rockbox' libdemac history, and you'll find 16 bit code for armv4 filtering
14:38:28 Nick dys` is now known as dys (
14:38:30amiconn...which is slower than what we have now
14:38:44fdinelfunman: hey hi mate nice to see you... nope sadly, mostly idling in here ;)
14:39:11funmanok :)
14:41:43funmanbuilding apps/*.c but {eq_arm,fixedpoint,eq,dsp_arm,codecs,main,metadata,pcmbuf,dsp}.c makes binsize 39kB smaller
14:41:55funman+ with -mthumb
14:42:42 Quit FOAD (Read error: 110 (Connection timed out))
14:42:42 Nick FOAD_ is now known as FOAD (
14:44:01kugelfunman: subdirs also?
14:44:34funmanalso, not firmware/
14:45:00funmanbut well the binary doesn't run
14:45:35kugelthat doesn't surprise me
14:46:01funmanwhy? -mthumb-interwork should handle switching between the 2
14:46:30kugelah, you didn't note that you added that too
14:47:20kugeldo the files you excluded contain inline asm?
14:47:45funmani guess so, i just tried them all and kept the version without -mthumb if they didn't build
14:48:12funman"Undefined instruction at X", and at X there's "ldr r3, =something"
14:49:14kugelit seems cutting down the plugin buffer is more promising. how big is it currently?
14:49:40kugel0x4b0e seems unaligned
14:49:41funman0x80000 for fuze
14:49:50kugelI mean c200v2
14:49:59funmankugel: 0x4b0e is the instruction
14:50:02 Join efyx_ [0] (
14:50:10kugeloh ok
14:50:22Mode"#rockbox +v efyx_ " by rb_voicebot (
14:50:24funmansame size for c200v2
14:51:13funmanbiggest .rock on fuze is doom with 391KBi
14:51:26funmannext one is clock at 236
14:51:54funmanafair doom isn't built on c200v2, so we could tiny the buffer immediately
14:52:22kugelno wonder it has no buffer ;) wasting 512KiB on plugins
14:52:44funmanbtw the gcc post on shows interesting variations in code size for the different compilers
14:52:51funman(in the comments)
14:53:27Unhelpfulamiconn: then i'll leave it. :) i don't think i have anything to add for the arm5te version either, unless there's a case where you will want the sum *and* difference of values.
14:53:31kugelso it looks like the plugin buffer could easily be halved without loosing functionality?
14:54:56funmanwe need to not make it too close to the smallest plugin size to avoid breakages when plugins are updated
14:55:01 Quit pamaury (Read error: 113 (No route to host))
14:55:50amiconnUnhelpful: It's always either-or, as you can see in the C function calling this stuff
14:56:54Unhelpfulhrm... so the next place to look for savings after the divider is probably COP use?
14:56:58funman0x4b0e decodes to ldr r3, [pc, #0xe] but that's not what i see in ida
14:58:29 Join robin0800 [0] (
14:58:32funmanobjdump shows ldr r3, [pc, #56] Oo
14:58:36kugelwell, it's an undefined instruction :)
14:58:50Mode"#rockbox +v robin0800 " by rb_voicebot (
14:58:51funmanno it's well encoded
14:59:21funmanthe undefined instruction is not necessary what i see in the .elf file though, memory could have been corrupted
14:59:27amiconnUnhelpful: COP use is planned, but don't expect -c3000 or higher to become usable on PP
15:01:23amiconnIt might be possible to speed up the filters a bit more by converting a larger part to asm, which probably allows to cut a bit of load/store away
15:01:47Unhelpfulamiconn: i don't, but it might be worth trying, as the table hack gets that to 64% on e200
15:04:03funmani don't see the undefined instruction trap defined in the arm922t technical reference manual (to check what lr refers to exactly)
15:11:16funmanah 56 = 4 * 0xe (ldr is 32 bits load) so that makes sense
15:14:40 Quit Tomis ()
15:16:18funmanand printing the instruction in UIE shows that it's correct ..
15:18:01 Join Buschel [0] (
15:18:22Mode"#rockbox +v Buschel " by rb_voicebot (
15:20:06kugelfunman: does UIE print the instruction at PC?
15:20:38kugelor at pc-8 (IIRC pc is always 2 instr ahead)
15:20:51funmanit only prints the failing address (the value of lr)
15:21:11funmanin thumb, lr is set to address of undefined instruction + 2 so i was reading the wrong instruction
15:21:40BuschelI have got some question here: during investigating fs#10637 I am quite stuck now. It seems like trying to play back any file with the codec_type=AFMT_UNKNOWN (=0) will lead to playback engine stopping until the next restart of the target. Is this wanted behaviour?
15:21:47 Join phanboy4 [0] (
15:21:58funmani don't see what lr's set to in ARM
15:22:07Mode"#rockbox +v phanboy4 " by rb_voicebot (
15:22:13funmansince we don't offset it i assume it's the address of undefined instruction?
15:22:33kugelisn't lr the link register used to store return addresses?
15:22:56funmanARM Architecture Reference Manual for v5 (DDI0100E) shows the THumb Exceptions in A6.1.2 (page 285)
15:23:03kugel(r14 IIRC)
15:23:44funmankugel: yes, it's the return address after the exception has been handled (we don't return because we don't know what to do)
15:23:51funmanso in ARM it might be pc+4
15:25:27funmanthat's it
15:26:56kugelpc+4 is correct?
15:27:20funmanwhen entering the undef instruction vector, lr = address of fault instruction + 4
15:27:25funman(in arm)
15:31:36 Quit petur ("real life")
15:32:57Unhelpfuli think i can squeeze a *tiny* bit more out of the table divider...
15:36:03funmanso the faulty instruction is 0x310c = add r1, #0xc ..
15:37:43Unhelpfulby adding 1 to the reciprocals, i can guarantee that qa=(r*n)>>32 is no *less* than n/d. then calculate n-d*qa, if positive qa is correct and we can return immediately, else subtract 1 from qa and return
15:39:50funmanin ARM mode, the instruction at lr - 4 is undefined (for this v4 cpu) though
15:40:05funmanperhaps i can read psr and extract the T bit
15:40:33 Quit teru ("Quit")
15:40:52TheSevencan anybody tell me quickly how a .rock file should look like?
15:41:59pippijnlike binary
15:42:41pippijnlook at (or something named similarly)
15:42:46pippijnit's model dependent
15:42:55TheSevenhm, it probably won't be possible to trigger a storage_flush from a plugin anyways
15:43:18pippijnwell, you can always add storage_flush to the syscall interface
15:43:34*TheSeven doesn't want to change the core because of this
15:43:51funmanfunctions available in plugin interface can be conditional
15:44:35TheSeventhen i'll rather look for a way to integrate a debugger stub into rockbox itself, to be able to do post-mortem dumps
15:45:41S_a_i_n_tis a post mortem dump like when a dead body loses bowel control? :P
15:50:53 Join evilnick [0] (n=evilnick@rockbox/staff/evilnick)
15:51:13Mode"#rockbox +v evilnick " by rb_voicebot (
15:51:47*kugel wouldn't call the plugin api "syscall interface" :)
15:52:03domonokybluebrother: i just seperated the serverInfo from rbsettings into its own class (, should i do the same for the StaticInfo from rbutil.ini ?
15:52:32***Saving seen data "./dancer.seen"
15:56:25funmanT bit is unset in spsr, now i'm not sure how/if the psr is saved before entering the exception vector
15:58:23funmanit is saved : each exception mode also has a "saved program status register (SPSR)"
16:02:35funmanwell now i'm sure: the code is built in thumb but being run as ARM
16:04:23CIA-44New commit by funman (r24321): ARM: lr addresses the next instruction after the failing one in undefined instruction vector ...
16:06:03 Join mitk [0] (
16:06:24Mode"#rockbox +v mitk " by rb_voicebot (
16:11:57 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
16:12:18Mode"#rockbox +v dfkt " by rb_voicebot (
16:13:10funmanrebuilding only some .o files now works but loading codecs crashes
16:15:15funmani give up for now
16:19:09funmanah doom is built for c200v2
16:19:31funmani was confused with mpegplayer probably, which needs at least 2MB of memory
16:20:34funmancan someone checks if doom runs on c200v2 ?
16:21:48Unhelpfulfurther improvement, this takes 74.22MHz for ape normal on e200:
16:23:17Unhelpfulthe assumptions it makes about reciprocals are tested for divisors up ot 16384... which is much more table than we can afford anywhere
16:23:47 Quit kaniini ("BRB! IRL!")
16:28:21 Quit maruk ("WeeChat 0.3.0")
16:29:23 Join kaniini [0] (
16:29:44Mode"#rockbox +v kaniini " by rb_voicebot (
16:31:49 Quit Highlander ("Quitte")
16:32:25CIA-44New commit by funman (r24322): *.link: depends on config file where plugin buffer size is set
16:33:46 Quit kaniini (Client Quit)
16:34:55 Join hebz0rl [0] (
16:35:16Mode"#rockbox +v hebz0rl " by rb_voicebot (
16:36:13domonokybluebrother: if you have time, please look at FS #10928
16:36:46funmandoom almost fills all the c200v2 plugin buffer
16:37:23*bertrik tries to find the wads
16:39:02funmanbertrik: ah nice, i was going to post on the f orum
16:41:00 Quit _zic ("Ex-Chat")
16:44:23bertrikfunman, the doom plugin starts, but when trying to start the actual game, I get "Z_Malloc: Failure tryi", "Please wait: Tranmap b" and "Computing: 8..7..6.....1"
16:44:39bertrikthen nothing, the game doesn't start
16:45:10bertrik(this is with svn 24322)
16:45:49 Quit MethoS- (Remote closed the connection)
16:47:10funmani'll disable it for memory size <= 2
16:47:23 Join MethoS- [0] (n=clemens@
16:47:44Mode"#rockbox +v MethoS- " by rb_voicebot (
16:47:46 Quit MethoS- (Remote closed the connection)
16:49:32 Join pamaury [0] (
16:49:52Mode"#rockbox +v pamaury " by rb_voicebot (
16:50:14 Quit phanboy4 (Read error: 113 (No route to host))
16:51:01funmandoom.tex manual is conditional on swcodec
16:51:28funmanbut it's compilation is conditional on LCD_COLOR || LCD_DEPTH== 2 || MROBE_100 || !AV300 (&& MEMORYSIZE > 2)
16:52:21AlexPprobably an old manual opt
16:52:36 Join MethoS- [0] (n=clemens@
16:52:57Mode"#rockbox +v MethoS- " by rb_voicebot (
16:53:15funmanlooks like i didn't disable mpegplayer for c200v2 either
16:53:19funman(in the manual)
16:53:38funmanthose aren't released so i'll ignore it atm
16:54:40funmanwe could use a "lowmemory" feature?
16:57:31funmanbertrik: can you try rockboy ? (you need a rom)
16:57:58bertriksure, but I don't have ROMs
16:58:22funmanyou can find some on emulation websites, it's legal if you have the original game
16:59:29CIA-44New commit by funman (r24323): Disable DOOM for c200v2, it won't run ...
17:01:26funmanAlexP: hm the c200v2 hasn't even a manual, so ..
17:06:25 Join stripwax [0] (
17:06:35 Join kugel_ [0] (
17:06:41 Quit stripwax (Client Quit)
17:06:55Mode"#rockbox +v kugel_ " by rb_voicebot (
17:07:47 Join froggyman [0] (
17:08:08Mode"#rockbox +v froggyman " by rb_voicebot (
17:09:12bertrikfunman, the game does start and I'm hearing some music and seeing a game demo, I can't actually start a game but I guess that's due to an unfinished keymap or something similar
17:09:38funmanok, thanks!
17:09:50funmanwe already won a lot of audio buffer with removing doom
17:10:03bertrikI can switch between 1 player and 2 player by means of the volume buttons
17:14:07funmanassuming rockboy is the biggest plugin now on c200v2 and needs to be fast, we can't use thumb on it
17:14:15funmanso no need to study the use of thumb on plugisn
17:16:36 Quit kugel (Read error: 60 (Operation timed out))
17:16:51 Join fyrestorm [0] (
17:17:11Mode"#rockbox +v fyrestorm " by rb_voicebot (
17:18:09 Join bimbel [0] (n=Miranda@unaffiliated/bmbl)
17:18:30Mode"#rockbox +v bimbel " by rb_voicebot (
17:27:07amiconnDoom won't work on the e200v2. Iirc it needs at least ~4MB RAM
17:27:22amiconnRockboy should, at least with small ROMs
17:27:30amiconn(Smallest are 64KB afaik)
17:29:28funmane200v2 has 8MB
17:29:44funmanand i can play it on fuze which is almost identical to e200v2 (save for screen orientation)
17:30:39amiconnEh, I meant c200v2
17:30:57*amiconn thinks that there are way too many Sansa models
17:31:09funmanand way too many revisions of each model :(
17:31:30gevaertslet's drop support for half of them!
17:32:45 Quit bmbl (Read error: 113 (No route to host))
17:32:51 Nick bimbel is now known as bmbl (n=Miranda@unaffiliated/bmbl)
17:34:24*amiconn would actually try doom on archos (if only to completely invalidate that old april fool's joke) if he had an 8MB-modded archos
17:38:33 Quit kloeri_ (Read error: 54 (Connection reset by peer))
17:38:57 Join kloeri__ [0] (
17:39:05 Quit GeekShadow (Read error: 113 (No route to host))
17:39:18Mode"#rockbox +v kloeri__ " by rb_voicebot (
17:43:41 Join toffe82 [0] (
17:44:02Mode"#rockbox +v toffe82 " by rb_voicebot (
17:45:42 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
17:52:36***Saving seen data "./dancer.seen"
18:01:30 Quit mitk ("Leaving")
18:07:17 Join kugel__ [0] (
18:07:38Mode"#rockbox +v kugel__ " by rb_voicebot (
18:08:18 Join moos [0] (
18:08:38Mode"#rockbox +v moos " by rb_voicebot (
18:13:29 Quit kugel_ (Read error: 60 (Operation timed out))
18:16:32 Quit rhodan_ (Remote closed the connection)
18:16:54 Nick kugel__ is now known as kugel (
18:18:38 Join rhodan [0] (
18:18:59Mode"#rockbox +v rhodan " by rb_voicebot (
18:21:10 Quit robin0800 (Remote closed the connection)
18:25:43 Join panni_ [0] (
18:26:04Mode"#rockbox +v panni_ " by rb_voicebot (
18:26:21 Quit fyrestorm ("Ur skills' fireproof like a wooden panel -- U got feds talking leet on your IRC channel!")
18:29:29 Join saratoga [0] (i=463f90ed@gateway/web/freenode/x-ieuwkknkstygzlsa)
18:29:50Mode"#rockbox +v saratoga " by rb_voicebot (
18:29:52 Quit Casainho (Remote closed the connection)
18:30:01saratogafunman: for what its worth, doom will run on 8MB players, but will quickly run out of memory on nearly all WADs
18:30:22saratogait needs slightly more then 8 MB, atleast as implemented in rockbox
18:31:29gevaertssaratoga: have people tried changing things like the max playlist size and friends? That might help
18:31:47saratogagevaerts: no I just tried with default settings
18:32:30saratogaif you put a print in the sim's doom malloc you could pretty easily check how much it uses while playing
18:33:42saratogaBuschel: for the playback problem you reported, I have played unsupported WMA flavors before without seeing that effect, though it was some time ago
18:33:47saratogamaybe this is a more recent bug?
18:34:06gevaertssaratoga: true. If I ever played doom, I'd do it
18:34:23saratogaso does playback work on the c200v2 now?
18:34:43funmanit "worked" already, don't know if it crashes less
18:35:04Buschelsaratoga: it was reported for mpc on Oct, 3rd. So, at least 3 months old.
18:35:26saratogathis was years ago when i added wma
18:35:50saratogafunman: in theory if you get the buffer up to ~150KB it should work all the time
18:35:53funmanit randomly works
18:36:04funmansaratoga: is there anything special with that limit ?.
18:36:04Buschelsaratoga: i also remember that the disabled sv4-6 mpc-files did not lead to stuck playback (maybe 2 years ago)
18:36:43saratogaits roughly the size of the largest .codec files that can be buffered
18:36:58saratogabut i'm not sure if thats the problem
18:37:18saratogaits just the size on the clip where if we went smaller thigns became buggy
18:38:01funmanmc2739: do you want to rebuild the 1.0RC bootloaders from the svn tag with the version changed to 1.0 ?
18:38:21saratogamastch said this: "MP3-playback (192kbit) is fine with 140kB ringbuffer size."
18:38:24saratogareferring to his clip
18:38:51saratogaso if this is the same as the clip, then you probably fixed playback
18:39:48funmani posted a note on the forum thread
18:39:49kugelwouldnt it be useful to disable codec buffering if the buffer is low anyway?
18:40:13funmankugel: i think you should ask matsch :)
18:40:27kugelthe idea of codec buffering is that you don't need to wake up the disk if the playlist contains mixed content
18:40:51saratogamatsch posted a patch that checks if the file buffer is too small to buffer codecs
18:40:55kugelbut with a buffer of <=400k the problem doesn't exist since that's not enough for a single file anyway
18:40:57soaphow costly is waking up the disc on flash players?
18:41:05saratogabut i never looked into it since we seemed to have fixed clip playback
18:41:24funmansoap: there's no disc to wake up so i believe it's free :)
18:41:27kugelit's really only useful if you are able to buffer multiple files
18:41:33 Quit gevaerts (Read error: 60 (Operation timed out))
18:42:02saratogaits not exactly free since theres some power increase when you energize the nand chip, but it is very small and probably irreelevent for .codec files
18:42:11funmansoap: i think i read here that buffering on flash devices has no consequences on battery life
18:42:16saratogai suggested disabling it a while ago for different reasons and people were opposed
18:42:34saratogathe last two patches in that task
18:42:35soapjust curious, if the cost is so low, why flash targets bother to buffer codecs, then.
18:43:10saratogaNicoP posted a patch that disabled buffering on flash targets
18:43:16kugelsaratoga: well it would be nice to fix the root cause of course, but it doesn't change the fact that codec buffering is useless if you can't buffer mutliple files
18:43:45saratogabut it seemed to decrease battery life on the e200v1 so most people concluded having a small buffer was preferable
18:44:05saratogathe first of those two patches will disable buffering on the c200v2
18:44:28saratogaits probably a good idea since it will mean more buffer available for actual audio too
18:44:46funmansaratoga: remember the FS# for this patch
18:45:05 Quit panni_ (Read error: 104 (Connection reset by peer))
18:45:28 Join panni__ [0] (
18:45:37saratogaoh that task
18:45:44funmansaratoga: (for NicoP patch) i see no mention of battery life
18:45:48Mode"#rockbox +v panni__ " by rb_voicebot (
18:45:51funmanexcept him asking for tests nobody made: )
18:45:59kugelI made a test
18:46:11saratogathey were mentioned in IRC, but for what its worth I was never 100% sure what to make of them
18:46:18kugelI got <14h on my e200v1 while svn was >20h
18:46:30saratogasvn with what audio buffer size?
18:46:42saratogathe whole buffer or 300KB?
18:47:17kugelI'm refering to testing Nico_P's patch
18:47:50saratogayeah but you compared it to buffering with a full 32MB of memory right?
18:48:48saratogai wonder what the difference is on a player like the clip where the memory buffer is so tiny it has to rebuffer every couple seconds anywya
18:49:57 Join kugel_ [0] (
18:50:18Mode"#rockbox +v kugel_ " by rb_voicebot (
18:51:08kugel_saratoga: I didn't test a svn version specifically
18:51:31saratogai just mean did you test stock svn for svn with a smaller amount of memory to simulate the ams targets?
18:51:40saratogafor -> or
18:53:24funmansaratoga: i can do that
18:54:06saratogado you think we need direct from flash playback?
18:54:47TheSevenyou can't really do that
18:54:53TheSeventhere will always be minimal buffering
18:55:25TheSevenand it will probably be beneficial to only read from the flash once a second or something
18:55:40saratogain this case we mean it basically just loads whatever portion of the "audio buffer" directly from flash when the codec requests it
18:55:46saratogaso you'll be loading 16KB or so at a time
18:55:51TheSevenhaving the flash powered up all the time will indeed eat some mA
18:56:39TheSevenanother question is how much latency is added by first needing to power up the flash
18:56:40saratogaits not clear to me why theres a big difference between 1 X millisecond long access and 10 X/10 millisecond longs accesses
18:57:20TheSevenat least on nano2g, I need to wait several milliseconds after powering on the flash before i can actually access it, which means some wasted current for every powerup/down
18:57:41saratogado you explicitly power up/down the flash?
18:57:53saratogafunman: is that needed on AMS too?
18:58:10*kugel_ also wouldn't underestimate the code overhead for microtransfers
18:58:18TheSevenshutting off the flash in the PMU will save ~7mA IIRC
18:58:53funmansaratoga: we disable the clocks, not sure what power we gain / how much time it needs
18:59:02TheSevenbut powering it on means waiting for voltages to stabilize, resetting it, waiting for the reset to complete and the flash to get ready again to accept commands
18:59:46saratogathats a pretty good argument in favor trying to buffer at least 140KB worth at a time
19:00:11saratogaalthough i like to think that when sandisk designed these 2MB players, and the NAND chips to go with them, they optimized for this case
19:00:33TheSeventhey are probably doing some buffering as well
19:00:55TheSevenI don't think it really matters if that buffer is 100 or 200K, but it will matter if it's 10 or 20K
19:01:01funmanafaict the 2MB ram is only used for usb mode
19:01:35saratogathe vorbis and wma decoders probably use it as well
19:01:36funmanand the 384kB ram for everything else (codecs, firmware, buffer?)
19:01:49saratogaat least i think it would be challanging to fit those codecs in so little memory
19:01:50 Join baptiste [0] (
19:02:11Mode"#rockbox +v baptiste " by rb_voicebot (
19:02:43kugel_funman: that sounds unlikely to be honest
19:03:03kugel_I mean it would be pretty stupid
19:03:08funmansaratoga: the "ogg" decoder binsize is 66kB
19:03:19funmankugel_: unless you show me some proof i won't believe you
19:03:43kugel_you don't need to believe me, I haven't claimed anything ;)
19:03:45TheSevenkugel_: don't ever think that DAP vendors aren't stupid ;-)
19:03:46saratogahuh theres someone on google with a 100KB WMA decoder, so maybe these thigns can be made smaller then I think
19:03:48funmanperhaps not having to power the 2MB SDRAM is where the OF has more battery life than rockbox
19:04:12saratogathats something i should test with a DMM
19:04:25funmanwma decoder is split across multiple "library blocks"
19:04:44saratogawell 60KB for the vorbis binary isn't too bad, but the codec has malloc's everywhere
19:05:01saratogaso they dump parts of the decoder to disk when they're not using it?
19:05:11kugel_maybe they use the entire 2MB for the heap?
19:05:32funmansaratoga: afaict the wma dec is 93kB on clipv1
19:05:44saratogai've seen Tremor malloc 800KB on some files, although I suppose sandisk has little concern about compatibility with all vorbis files
19:06:20funmankugel_: you can find several locations in the OF where the memory is set to self refresh and can't be accessed anymore
19:06:54saratogahow quickly can that be done?
19:07:05funmani didn't time
19:07:13 Quit kugel (Read error: 113 (No route to host))
19:07:20saratogais it worth having the codec thread entirely in IRAM, and disabling DRAM while it runs?
19:07:26 Nick kugel_ is now known as kugel (
19:07:36funmanit's a matter of cleaning dcache, setting a bit in MPMC registers and waiting for a status bit
19:07:45saratogawe'd have to move some of buffering in there as well
19:07:57funmansaratoga: the whole binary is in DRAM already including the vectors, and codec api
19:08:13funman+ buffering, etc..
19:08:41saratogayeah but if we can do it quickly, we could disable while in the actual codec, and just have the buffering functions exposed to the codec in IRAM
19:08:50saratogaand then enable whent he codec thread yields
19:09:06funmanwhat happens on interrupts ?
19:09:18saratogahmm yeah thats a problem
19:10:15funmanperhaps with jtag we could check if DRAM is enabled when playing in the OF
19:10:59saratogawith jtag we could dump this whole stupid ams clocking scheme
19:11:19*kugel can't imagine they manage the squeeze the entire system, audio data and decoders into 384K ram
19:11:27kugelespecially not on the c200v2
19:11:46funmanhmm sd_reload__ and (some) drm modules are linked at 0x3xxx (DRAM) in the OF
19:11:47mc2739funman: I have already built Fuze and e200v2 bootloaders from the 1.0 tagged branch. I have tested the e200 bootloader, but the Fuze bootloader needs to be tested.
19:12:12funmankugel: don't forget they use thumb
19:12:19funmanand do not have a lot of features
19:12:49funmanmc2739: ok, you want me to test it?
19:13:49*kugel would be surprised if it didn't work, the only change was to remove the "RC" from the string
19:14:08mc2739funman: That would be good. I'll email it to you/
19:15:06kugeldidn't we release mkamsboot 1.0 at some time?
19:15:46kugelwhy didn't we make 1.0 bootloaders at the same time?
19:16:07 Join darkham [0] (
19:16:18funmanthere's no link between mkamsboot & the bootloaders
19:16:28Mode"#rockbox +v darkham " by rb_voicebot (
19:16:41funmanno requirement
19:17:15 Join stooo1 [0] (
19:17:19 Quit stooo (Read error: 104 (Connection reset by peer))
19:17:36Mode"#rockbox +v stooo1 " by rb_voicebot (
19:17:43 Nick stooo1 is now known as stooo (
19:18:21 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
19:18:23kugelsure but i'm still curious
19:18:42Mode"#rockbox +v BHSPitMonkey " by rb_voicebot (
19:19:30mc2739funman: email sent
19:23:50 Join phanboy4 [0] (
19:24:11Mode"#rockbox +v phanboy4 " by rb_voicebot (
19:25:16funmanmc2739: works fine
19:26:18mc2739funman: Thanks for testing. Should I send these to bagder?
19:29:05funmanyes, i don't thing anything else is needed
19:29:29 Quit moos (Read error: 60 (Operation timed out))
19:31:42darkhamwhen developper almost think as 3.5 release date?
19:32:02funman2nd february
19:32:04AlexPdarkham: see the dev mailing list
19:33:56 Join Ebert [0] (
19:34:17Mode"#rockbox +v Ebert " by rb_voicebot (
19:45:30 Quit saratoga ("Page closed")
19:49:16 Part stooo
19:52:26 Join utchybann [0] (
19:52:40***Saving seen data "./dancer.seen"
19:52:47Mode"#rockbox +v utchybann " by rb_voicebot (
19:56:56 Quit at0m (Client Quit)
19:59:38CIA-44New commit by funman (r24324): Italian translation for rbutil ...
20:03:20 Join Grahack [0] (
20:03:40Mode"#rockbox +v Grahack " by rb_voicebot (
20:06:27bluebrotherfunman: rbutil has its own CREDITS file (for historical reasons). Is there a reason you add people only to the main CREDITS?
20:06:43bluebrotherbut this also brings up the question if it makes sense to have two CREDITS files at all.
20:06:55bluebrotherdomonoky: will check the FS task later. In general I like the idea.
20:07:04mc2739three including the manual
20:07:24bluebrothermc2739: the manual one was recently merged AFAIK
20:07:42bluebrotherdomonoky: btw, please try to not introduce trailing spaces in diffs, it makes them harder to read ;-)
20:07:44funmanbluebrother: didn't know about this other CREDITS file, i guess it could be merged
20:07:46*domonoky thinks it would be good to merge all CREDITS files to on.
20:08:25funmanbluebrother: btw you can use diff -w to ignore whitespace changes in diff (including indentation changes)
20:10:05mc2739funman: you have red from r24322
20:10:34 Join stripwax [0] (
20:10:55Mode"#rockbox +v stripwax " by rb_voicebot (
20:11:32 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
20:11:49 Join stooo1 [0] (
20:11:53Mode"#rockbox +v gevaerts " by rb_voicebot (
20:12:10Mode"#rockbox +v stooo1 " by rb_voicebot (
20:12:10domonokybluebrother: i just checked again, but cant find any changed trailing spaces in this patch. Can you point me to one ?
20:12:31kugelit looks like an error from the target renaming
20:12:46 Nick stooo1 is now known as stooo (
20:13:38 Quit Rob2222 ()
20:14:11kugelfunman: Zagor made some .make files depend on errno.o (teru changed it to ctype.o recently) to depend on <target>.h indirectly before the target renaming happened
20:14:49bluebrotherdomonoky: seems I mixed that up with the pastie :o Sorry.
20:14:54 Join stripwax_ [0] (
20:15:14Mode"#rockbox +v stripwax_ " by rb_voicebot (
20:15:19bluebrotherwell, the main CREDITS file is rather long. Does it make sense to have 400+ entries in the rbutil one?
20:15:48funmanwhy not ? rbutil installs work by these 400+ people/projects
20:17:06rasherWe include it with Rockbox itself, surely there's room in rbutil
20:17:25AlexPmc2739: The manual one was removed
20:17:49AlexPI think the same should apply here - one cvredits file for Rockbox, be it the firmware, rbutil, manual, etc etc
20:19:21funmanwhat is the difference between ondavx747 and ondavx777 ?
20:19:31 Quit darkham (Read error: 104 (Connection reset by peer))
20:19:40bluebrotherhmm. I'm just somewhat unsure, but you've got a good point (well, had that in mind myself before too ;-)
20:19:50 Join Jaykay [0] (
20:20:01pixelmafunman: one digit? Sorry, couldn't resist ;)
20:20:05bluebrotherok, so unless there is someone against merging those I could do it.
20:20:10Mode"#rockbox +v Jaykay " by rb_voicebot (
20:20:47 Join intrados1 [0] (
20:21:07Mode"#rockbox +v intrados1 " by rb_voicebot (
20:21:52CIA-44New commit by funman (r24325): zenvision.h -> zenvision30gb.h
20:22:12funmancould someone svn cp ondavx747.h ondavx777.h for me ? git has a mv but not a cp command
20:23:22bluebrotherfunman: I _think_ git-svn should handle that by its rename detection
20:23:48bluebrotherat least it works for moves.
20:23:55funmaneven if i modify the file after that ?
20:24:13bluebrotherok, that makes the question harder ;-)
20:24:26funmanhm i'll just do it and mention vx747, so one browsing the history will be able to look at the vx747 file
20:24:35bluebrotherNo idea, but if git gui shows it as copy I'd expect git-svn to create a copy in the history
20:24:46JdGordonjust looking back through the logs... regarding buffering crashing when you delete the current file. is read() sane if the fd is open but removed?
20:25:04JdGordonbuffering.c seems to be able to handle negative return values from it just fine
20:25:25CIA-44New commit by bluebrother (r24326): Respect default values for ServerSettingsList as well. No idea why I missed that.
20:26:57 Quit pixelma (Nick collision from services.)
20:26:59 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
20:27:16 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
20:27:16 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
20:27:16 Quit amiconn (Nick collision from services.)
20:27:20Mode"#rockbox +v pixelma " by rb_voicebot (
20:27:36 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
20:27:37Mode"#rockbox +v amiconn " by rb_voicebot (
20:28:09JdGordonS_a_i_n_t: if that wps acts how you say it acts (I havnt looked at it yet) then you found a bug... put it on flyspriay
20:28:54CIA-44New commit by funman (r24327): ondavx777 : use own config file
20:29:41S_a_i_n_tcould you? I'm pretty much guaranteed to be fucking something up...I know I could do the viewports better..but the sublines seem sweet. I haven't put it on flyspray yet as I have 0 ability to confirm it.
20:29:50S_a_i_n_tno rush though...
20:30:19JdGordonif its in flysrapy there is less chance of it being forgottn
20:30:26JdGordonand even fixed
20:30:36funmanhum i did something bad for zen vision
20:30:54S_a_i_n_tyeah, I understand that...I just don't know if the error is mine or not.
20:31:19*domonoky sees bluebrothers commit, and checks his patch..
20:32:36*domonoky sees that his patch is fine. :-)
20:32:49bluebrotherdomonoky: thanks :)
20:33:11pixelmaJdGordon, mt (for the logs): for what it's worth, deleting the currently playing track does not confuse playback on hwcodec
20:33:17*bluebrother missed yet another ebay auction :(
20:33:30JaykayAlexP: FS #9794 :D can i help you with it?
20:33:40JdGordonpixelma: thanks, but yeah, I woudlnt expected it to
20:33:48 Quit stripwax_ (Read error: 110 (Connection timed out))
20:33:48AlexPJaykay: I'd forgotten about that
20:33:53CIA-44New commit by funman (r24328): Add a config file for zenvision (which includes zvm30gb config file)
20:34:16AlexPJaykay: Yeah, just need to check that the release bootloaders al say file not found not error -1 for all targets bar h100,h300
20:34:37pixelmaJdGordon: yeah, not unexpected but I thought it's worth checking in case something else than everyone thinks is broken
20:35:00 Join petur [0] (
20:35:21Mode"#rockbox +v petur " by rb_voicebot (
20:35:50JaykayAlexP: if i should do that you need to tell me how... i guess that this will be as much work as if you would check it :)
20:40:23funmanmc2739: thanks for showing me the red... i should have looked myself especially after this kind of change
20:42:12 Quit stripwax (Read error: 110 (Connection timed out))
20:49:47 Join robin0800 [0] (
20:50:08Mode"#rockbox +v robin0800 " by rb_voicebot (
20:50:28bluebrotherdomonoky: merging the CREDITS brings up the problem of the header in docs/CREDITS. My idea is to simply display the credits in a fixed-width font. Thoughts?
20:51:16funmanbluebrother: you could add a marker and remove the header from the gui
20:51:55domonokysounds fine as long as the header doesnt wrap. else do it like funman suggested :-)
20:52:17bluebrotherhmm, we could of course do the same processing as done when adding it to the manual
21:06:03 Quit DerPapst ("Leaving.")
21:08:05 Part froggyman
21:12:48 Quit pamaury ("abort();")
21:13:17 Quit toffe82 (Read error: 104 (Connection reset by peer))
21:18:08 Quit flydutch ("/* empty */")
21:20:11 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
21:20:31Mode"#rockbox +v togetic " by rb_voicebot (
21:23:19 Quit S_a_i_n_t (Read error: 110 (Connection timed out))
21:24:36 Join jake__ [0] (
21:24:56Mode"#rockbox +v jake__ " by rb_voicebot (
21:29:19 Join S_a_i_n_t [0] (i=S_a_i_n_@
21:29:41Mode"#rockbox +v S_a_i_n_t " by rb_voicebot (
21:29:46 Part Ebert
21:31:49 Quit Buschel ()
21:34:16 Quit robin0800 ("No Ping reply in 180 seconds.")
21:34:37 Join robin0800 [0] (
21:34:58Mode"#rockbox +v robin0800 " by rb_voicebot (
21:36:50 Quit jake__ (Client Quit)
21:39:09 Nick AlexP is now known as BigBambi (n=alex@rockbox/staff/AlexP)
21:39:18 Nick BigBambi is now known as AlexP (n=alex@rockbox/staff/AlexP)
21:41:09 Join robin0800_ [0] (
21:41:30Mode"#rockbox +v robin0800_ " by rb_voicebot (
21:42:19bluebrotherdomonoky: do you think it's a good idea to keep the "People that have contributed to the project ..." line in the credits output? I'm tending to keep that.
21:43:01JdGordonor change it to "who to blame if this rapes your grandma and kills your cat!"
21:43:20funmani'd reverse the two verbs position
21:43:38domonokyif you already strip the header, you could strip that line too.
21:43:49*bluebrother looks if there are children around
21:44:04bluebrotherdomonoky: sure, but it kinda fits. IMO :)
21:45:11domonokytrue, but better add it as a rbutil text. then it can be translated.
21:45:20*bluebrother removes the line for now
21:46:29domonokyany comments to my patch ? i want to commit it fast, because it touches so many files that it gets conflicts on nearly any rbutil commit :-)
21:48:30bluebrotherdomonoky: I'll look at it after finishing the credits stuff. Which is about now :)
21:48:41domonokygood :-)
21:48:45CIA-44New commit by bluebrother (r24329): Merge rbutil CREDITS file with main one and use that. ...
21:49:27 Quit funman ("free(random());")
21:51:33 Join Casainho [0] (n=chatzill@
21:51:54Mode"#rockbox +v Casainho " by rb_voicebot (
21:52:44***Saving seen data "./dancer.seen"
21:54:45bluebrotherdomonoky: that "class QSettings;" in serverinfo.h looks strange. including QtCore should be sufficient.
21:54:47 Join stripwax [0] (
21:55:14 Quit robin0800 (Read error: 110 (Connection timed out))
21:55:14Mode"#rockbox +v stripwax " by rb_voicebot (
21:55:30bluebrothersame for systeminfo.h
21:56:36domonokybluebrother: true, its a left over thing, i will remove it. but systeminfo needs it, its QSettings based.
21:57:55bluebrotherdomonoky: but you include QtCore, which does contain QSettings so that declaration shouldn't be needed
21:58:01 Quit baptiste (Read error: 104 (Connection reset by peer))
21:58:02*domonoky is wrong, its not needed in all three settings/info classes.
21:59:10bluebrotheroh, and you change some configure_modelname entries. Is there a reason for them to be in this commit? AFAICS those should go separate (and there is still some renaming needed −− some sections should get renamed too)
21:59:43 Quit Jaykay (Read error: 110 (Connection timed out))
21:59:59bluebrotherapart from that I haven't seen anything obvious. It's a lot code shuffled around ;-)
22:00:02domonokyi renamed them, because you suddenly notice the wrong ones, with the new Stable/unstable handling from the server.
22:00:48domonokybut there it surely more renaming work needed. But we have to be carefull with that.
22:00:57bluebrotherwe should check all those names again. For example there is still a section gigabeatf, but if we want to unify names it should be gigabeatfx
22:01:14bluebrotheror some samsung sections. But we indeed need to be careful ;-)
22:02:25 Quit bmbl ("Bye!")
22:04:06domonokytheoretically we should now be able to combine buildserver_modelname and configure_modelname. But still cant use the rbutil internal name, because rbutil knows more targets then configure/buildserver
22:05:23bluebrotheryeah, that's kinda bad. Haven't checked lately, but maybe we can use the configure name + some rbutil extension for that.
22:05:38domonokythat might be possible..
22:05:56bluebrotherlike iriverh10_5gb.mtp
22:06:10domonokyarchosfmrecoder.8mb ?
22:06:23bluebrotherlike that, yeah.
22:06:47bluebrotheror use any other character for separation that won't appear in configure names.
22:06:52domonokysounds nice..
22:07:37bluebrotherI was thinking of something like that since the beginning of playing around with those names in my head
22:08:15 Quit stoffel (Remote closed the connection)
22:08:26kugeldoesn't configure still produce archosfmrecorder8mb if it was selected in the advanced options?
22:08:32bluebrotherbut anyway, IMO you should go for FS#1092. Maybe committing the rbutil.ini changes separately first.
22:08:52domonokyhm, the 8Mb archos are the only ones with buildserver/configurename differences.. any idea why ?
22:10:31bluebrotherbecause it's a configure target while all others are only rbutil internal differences. Like MTP vs UMS installation on the h10
22:10:52bluebrotherhmm, there is something broken now. I get two configuration dialogs after startup :o
22:11:18CIA-44New commit by Domonoky (r24330): rbutil: rename a few more targets.
22:11:51 Quit Grahack ("Tu m'as vu ?")
22:12:40CIA-44New commit by Domonoky (r24331): rbutil: split RbSettings. use Stable/unstable status from server.
22:13:39 Join CaptainKewl [0] (
22:13:59Mode"#rockbox +v CaptainKewl " by rb_voicebot (
22:16:16CIA-44New commit by Domonoky (r24332): rbutil: fix doubled configure dialog.
22:16:55 Join Oxymis [0] (
22:17:16Mode"#rockbox +v Oxymis " by rb_voicebot (
22:18:30bluebrotherdomonoky: beat me ...
22:19:14*domonoky beats bluebrother :-)
22:19:37*bluebrother feals beaten :o
22:20:03 Quit HBK ()
22:22:07 Quit kloeri__ (Read error: 104 (Connection reset by peer))
22:26:48 Quit stripwax (Read error: 110 (Connection timed out))
22:27:25 Join kloeri__ [0] (
22:27:46Mode"#rockbox +v kloeri__ " by rb_voicebot (
22:28:08 Quit robin0800_ ("No Ping reply in 180 seconds.")
22:28:28 Join robin0800 [0] (
22:28:48Mode"#rockbox +v robin0800 " by rb_voicebot (
22:28:49 Quit gevaerts (Nick collision from services.)
22:29:01 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
22:29:21Mode"#rockbox +v gevaerts " by rb_voicebot (
22:30:21 Quit tomaw__ (K-lined)
22:32:18 Join tomaw_- [0] (i=tom@freenode/staff/tomaw)
22:32:38Mode"#rockbox +v tomaw_- " by rb_voicebot (
22:33:06 Nick tomaw_- is now known as tomaw__ (i=tom@freenode/staff/tomaw)
22:40:19 Quit hebz0rl ("Ex-Chat")
22:48:02 Join HBK [0] (
22:48:22Mode"#rockbox +v HBK " by rb_voicebot (
22:50:35 Quit Casainho (Read error: 113 (No route to host))
22:52:08 Join Casainho [0] (n=chatzill@
22:52:29Mode"#rockbox +v Casainho " by rb_voicebot (
22:55:04 Join Strife89 [0] (
22:55:24Mode"#rockbox +v Strife89 " by rb_voicebot (
22:55:44 Join Rob2222 [0] (
22:56:04Mode"#rockbox +v Rob2222 " by rb_voicebot (
22:58:25 Join JdGordon| [0] (n=jd@rockbox/developer/JdGordon)
22:58:46Mode"#rockbox +v JdGordon| " by rb_voicebot (
23:02:56 Quit stooo (Read error: 104 (Connection reset by peer))
23:03:13 Join stooo [0] (
23:03:33Mode"#rockbox +v stooo " by rb_voicebot (
23:04:27*bluebrother uploaded a new set of rbutil svn binaries; in case someone wants to try:
23:04:35 Quit stooo (Read error: 104 (Connection reset by peer))
23:04:58 Join stooo [0] (
23:05:18Mode"#rockbox +v stooo " by rb_voicebot (
23:05:21 Join stripwax [0] (
23:05:42Mode"#rockbox +v stripwax " by rb_voicebot (
23:07:37 Quit JdGordon| ("Bye")
23:09:10Unhelpfulthis uses roughly the maximum available space for the APE reciprocal LUT on PP targets. i've only an e200, so if somebody would care to give the ape normal sample a go on pp5020 or pp5002, i'd be interested in the results. if anybody wants to try some *other* armv4 target that uses codec iram and tell me that the LUT is set too large or too small for them, that's fine, too. :)
23:12:41 Quit stripwax (Read error: 60 (Operation timed out))
23:36:55 Quit grndslm (Read error: 60 (Operation timed out))
23:49:41Strife89Unhelpful: Sorry, but I'm not sure which targets use which hardware.
23:49:54Strife89Unhelpful: I have a c200, a Fuze, and an iPod Color.
23:51:31 Quit efyx_ (Remote closed the connection)
23:52:48***Saving seen data "./dancer.seen"
23:53:21 Quit phanboy4 (Read error: 110 (Connection timed out))

Previous day | Next day