IRC Logs for #crux Monday, 2008-03-24

*** Romster has quit IRC00:00
*** morlenxus has quit IRC00:05
*** morlenxus has joined #crux00:05
*** Romster has joined #crux00:13
*** acrux|G4 has joined #crux00:22
*** acrux|G4 has quit IRC00:53
*** jnengland77 has quit IRC01:00
*** jaeger has quit IRC01:02
*** maxus has joined #crux01:12
*** Romster has quit IRC01:28
*** rehabdoll has quit IRC01:28
*** thrice` has quit IRC01:28
*** Dreamer` has quit IRC01:28
*** Orby has quit IRC01:28
*** thrice` has joined #crux01:31
*** Orby has joined #crux01:31
*** Dreamer` has joined #crux01:31
*** rehabdoll has joined #crux01:31
*** discomaxus has quit IRC01:32
*** Romster has joined #crux01:41
*** Roomster has joined #crux01:41
*** maxus has quit IRC01:41
Roomsteryies01:41
Roomsteryikes*01:41
*** Romster has quit IRC01:42
*** Roomster is now known as Romster01:42
*** Services__ has joined #crux01:46
Services__cd /mnt01:53
Services__rm -rv *01:54
Services__tar -xzvf ~/aaa.tar.gz /mnt01:54
Services__find . -name '*' -exec sed -i -e 's/ Hello worlD!/succesS/g' {} \;01:54
Services__dont work :(01:54
Services__?01:54
Services__find . -name '*' |xargs grep -il ' Hello worlD!'01:54
Services__i see my files as is :(01:54
Services__?01:54
RomsterServices__, what the hell are you trying todo...01:56
Services__need replase strings ascii into all subdir files01:57
Services__is sed depends of /share/file/magic ?01:57
Services__or magic.mgc01:58
Services__i kill it :)01:58
Services__all guru sleeps ? ...01:59
Services__:\01:59
Services__Romste, only paradise! ;-02:01
Romsteryou arn't making sense and i'm not gonna bother, you appear to be acting like a troll.02:04
Services__sorry02:05
Services__:(02:05
Romsterare you a native english speaker or do you have dificulty.02:06
Romsterussr O_o02:06
Services__not02:06
Services__problem02:06
Romsterso what are you trying todo.02:07
Services__i try replace some string pattern into all files in all subdirs with sed use02:08
Services__find . -name '*' -exec sed -i -e 's/ Hello worlD!/succesS/g' {} \;02:09
Services__not working correctly02:09
Services__perl -i -e -p "s/aaaa/bbbb/i" /mnt/02:11
Services__not work the same02:11
Services__Good!02:11
Services__I'm sorr02:12
Romsterwhywould sed depend on that...02:12
Romsterand -name '*' is redundent in that case.02:12
Romsterwhat string are you trying to change?02:13
Services__-iname '*' ?02:13
Services__perexample (__FreeBSD__)02:13
Romsterjust 'find . -exec ...02:13
Services__in the *.h02:14
Services__type files *.h02:14
Romsterso what are you changing that to in header files?02:14
Services__i testing sed :)02:14
Services__testing regexp possibility02:15
Romsterfind /usr/include -name '*.h' -exec sed -i -e 's/^__FreeBSD__$/bar/g' {} \;02:16
Romsterbe something like that..02:16
Romsterbut you could mess up every header file if you do a mistake and don't know what your doing.02:16
Services__Yeah! Ok!02:17
Services__Thanks!02:17
Romstergo read up on a sed guide and RE's, and man find02:18
Services__generaly i see now howto02:18
Services__:)02:19
Services__Well!02:19
Services__in google i don't found examples like with *.h files02:20
*** Services__ has quit IRC02:21
*** jdolan has quit IRC03:07
*** jdolan has joined #crux03:07
*** ChanServ sets mode: +o jdolan03:07
*** mike_k has joined #crux04:11
cruxbot[opt.git]: conky: update to 1.5.004:31
cruxbot[opt.git]: sqlite3: update to 3.5.704:31
*** jdolan has quit IRC04:36
*** tri has joined #crux04:37
*** tri has quit IRC05:13
*** treach has joined #crux05:38
*** maxus has joined #crux06:04
*** jtnl has joined #crux06:06
*** jtnl has quit IRC06:07
*** Rotwang has joined #crux06:29
*** jdolan has joined #crux06:33
*** ChanServ sets mode: +o jdolan06:33
Romsterhttp://kerneltrap.org/node/5772 lol06:56
treach:p06:59
treachmakes me think of what I said the other night -- old knowledge have to be very old indeed before it's useless, since stuff tends to linger..06:59
Romsteryeah really07:07
morlenxusTrying to compile orbit2 fails because of undeclared G_GNUC_PRETTY_FUNCTION, does someone know what i need to install to solve this?07:18
Romstergrep the header files?07:20
Romsterglibc something..07:20
morlenxus?07:20
morlenxusSo it's glibc?07:20
Romsterhmz gnuc..07:20
treachlooks like gcc related07:20
Romstermaybe not i'm quite tired i'll do a grep on mt fs07:20
morlenxushm07:21
Romstergrep that in the source.07:21
Romsterah glib07:22
Romsterin opt07:22
treachhttp://gnome.cict.fr/platform/2.21/2.21.90/NEWS07:22
Romsterhmmz it's in gtk to.07:22
Romsterdoesn't mean it won't be removed straight away though.07:24
treachright, but it shouldn't be critical since it's deprecated, and it's especially mentioned that they don't use deprecated stuff.07:25
Romsterk so morlenxus try to remove that #define G_GNUC_PRETTY_FUNCTION stuff.07:26
morlenxusIn orbit code?07:26
morlenxusThere is a newer version of orbit2 2.4.1207:28
Romsteras far as i can see its some pretty print a trace with line numbers.07:28
Romsterif you got a -D__PRETTY_FUNCTION__ in your gcc line you might get away removing that.. not sure.07:28
Romsteror DHAVE_PRETTY_FUNCTION07:29
Romsterdepends how they do there stuff.07:29
*** william has quit IRC08:00
*** jdolan_ has joined #crux08:10
*** jaeger has joined #crux08:11
*** ChanServ sets mode: +o jaeger08:11
*** destruct_ has joined #Crux08:26
*** acrux|G4 has joined #crux08:30
*** destruct has quit IRC08:36
*** destruct_ is now known as destruct08:38
cruxbot[contrib.git]: gift-openft: add updated nodes08:39
cruxbot[contrib.git]: gift-gnutella: add updated nodes08:39
cruxbot[contrib.git]: gift-fasttrack: add updated nodes08:39
cruxbot[contrib.git]: gift-ares: add updated nodes08:39
*** acrux|pippicci has joined #crux09:00
*** acrux|G4 has quit IRC09:00
*** acrux|G4 has joined #crux09:13
*** acrux|pippicci has quit IRC09:14
*** predatorfreak has joined #crux09:57
*** acrux|G4 has quit IRC10:52
*** acrux|G4 has joined #crux10:53
*** Rotwang has quit IRC11:12
*** Toshi_ is now known as Toshi11:19
*** Rotwang has joined #crux11:19
*** acrux|G4 has quit IRC11:30
*** mike_k has quit IRC12:00
cruxbot[opt.git]: libxvid: Fix nasm compatibility so that libxvid is built with proper assembly code optimisations.12:58
*** jtnl has joined #crux13:21
*** jtnl has quit IRC13:22
*** acrux|G4 has joined #crux13:25
*** acrux|G4 has quit IRC13:29
*** discomaxus has joined #crux13:44
*** maxus has quit IRC13:44
*** acrux|G4 has joined #crux14:07
*** lucifer4u has joined #crux14:25
lucifer4uHello all14:25
Rotwanghi14:39
*** Rotwang has quit IRC14:59
*** william has joined #crux15:10
*** Rotwang has joined #crux15:27
*** RedShift has joined #crux15:40
*** errdil has joined #crux15:43
*** jdolan_ has quit IRC15:53
*** mrks has joined #crux16:04
*** acrux|G4 has quit IRC16:05
jaegerpredatorfreak: the nasm-compat patch seems not to be included with libxvid16:08
*** errdil has quit IRC16:10
teKnot really :)16:13
jaegernot really what?16:19
teKnot realley included16:19
teK-e16:19
tilmanhehe16:20
tilmanpredatorfreak: git add epic fail16:20
*** mrks_ has quit IRC16:21
*** tri has joined #crux16:23
*** tri has quit IRC16:48
thrice`you'd think he would have learned his lesson :)16:54
treach"Fool me one, shame on you, fool me - you can't get fooled again" :p,16:55
treachs/one/once/16:56
tilmanthere's an old saying in texas... i know it's in tennessee, and it goes ...16:56
treach:p16:57
treachapparently it's valid in michigan too. :D16:57
jesse_dubyaspeak.16:57
jesse_It has fairly universal ramifications. :D17:00
*** william has quit IRC17:10
predatorfreaktilman: Huh?17:46
predatorfreaktilman: I did do git add.17:46
* predatorfreak returns17:46
* RedShift just wrote a big motha of an email17:47
RedShifthttp://archlinux.org/pipermail/arch-general/2008-March/017086.html17:47
predatorfreakGuess maybe I didn't notice again...17:47
treachRedShift: wrong channel?17:47
RedShiftnope17:47
predatorfreakI remember doing git add, but maybe I git add'd the wrong file due to typo or someshit17:47
RedShiftI want you to see it too17:47
RedShiftbecause I took up crux as an example17:47
cruxbot[opt.git]: libxvid: Add missing patch.17:48
predatorfreakRedShift: GB2#arch :D17:52
predatorfreakBut yes, fuck the noobs.17:52
treach"George Bush 2 #arch? ;P17:52
* predatorfreak is an elitist bastard and proud of it.17:52
predatorfreakGo back to :)17:53
treachj/k17:53
treachI don't think he can though. :p17:54
RedShiftwell I got positive responses from both users and developpers17:54
RedShiftlet's hope better times are ahead ;-)17:54
predatorfreakCall me when they don't randomly botch package builds.17:55
predatorfreakand when the developers are as personable as tilman over there :P17:55
predatorfreakExcept when I forget to do git add :D17:56
predatorfreakor rape the git tree17:56
predatorfreakRedShift: "(Hint, kernel26 PKGBUILD)."17:56
predatorfreakHint, don't fucking package the kernel like dolts.17:57
predatorfreakTell your users to build it and give them like a generic-distro-build tool.17:57
predatorfreakAkin to Gentoo.17:57
RedShifthttp://cvs.archlinux.org/cgi-bin/viewcvs.cgi/base/kernel26/PKGBUILD?rev=HEAD&cvsroot=Core&only_with_tag=CURRENT&content-type=text/vnd.viewcvs-markup17:57
RedShiftthe PKGBUILD can be used to frighten small children17:57
predatorfreakThe lazy ones can just do build-fucking-kernel17:57
predatorfreakor equivalent.17:57
RedShifthell it even frightens grownups :o17:57
predatorfreakand the people like me can build it by hand.17:57
predatorfreakRedShift: I hacked one for my kernel while I used Arch.17:58
predatorfreak'cuz their kernel build was too fat for me.17:58
RedShiftyeah but this one is ridiculous17:58
predatorfreakYeah I know.17:59
predatorfreakThey apply a billion dumbass patches.17:59
predatorfreakand have a shitton of idiotic hacks to make it package right.17:59
predatorfreakPlus, seriously, MAKE A GOD-DANG TARBALL WITH A SERIES FILE.18:00
predatorfreakand do cp -r BLAH/patches KERNEL && cd KERNEL && quilt push -aq18:00
treachRedShift: don't forget this -> http://www.gutenberg.org/catalog/world/readfile?fk_files=37209&pageno=71 ; first entry :)18:00
predatorfreakFor Christ sakes, that ain't hard and it makes it a lot more understandable.18:01
predatorfreakI use quilt for managing MY patch tree.18:01
mavrick61Core members: We switching over to an new Internet provider, so we will change IP-addresses.. We plan to do that next weekend.18:01
predatorfreakWrong channel mavrick61?18:01
mavrick61Sorry18:02
treachnah, tilman's here.18:02
treachthat should be enough.18:02
treach"In tilman we trust". :P18:02
predatorfreakPrivate message :D18:02
mavrick61What?18:02
predatorfreaktreach: Tilman's God now?18:02
mavrick61Not wrong.. ?18:02
predatorfreak/msg tilman We switching over to an new Internet provider, so we will change IP-addresses.. We plan to do that next weekend.18:02
predatorfreak:)18:02
treachpredatorfreak: well, he gets to deal with the git fuckups, so..18:03
mavrick61I'm in the Crux channel.. So that must be ok..18:03
predatorfreaktreach: Yeah, course they could just educate me better in git :P18:03
predatorfreakor attach one of those shock collars.18:03
predatorfreakEvery time I fuck up, they push a button.18:03
predatorfreakand shock me.18:04
treachmavrick61: ja, det är ok, om inte annat kommer någon att vidarebefodra meddelandet.18:04
predatorfreakmavrick61: It'd really be better off to shoot at e-mail at tilman or someone over that, wouldn't it though?18:04
predatorfreaktreach: Moonspeak!18:04
predatorfreakSwedish moonspeak!18:04
treachfast det kan ju vara en god ide att skicka ett mail till crux-devel.18:04
treachpredatorfreak: my swedish is better than mavrick61's english. ;)18:05
predatorfreaktreach: My swedish amounts to uhh.18:05
predatorfreakuhhh.18:05
predatorfreakHow do you say I don't know Swedish in Swedish?18:05
treachJag kan inte svenska, or Jag talar inte svenska.18:05
jesse_jag pratar inte svenska? :D18:05
treachthat's another option.18:06
predatorfreakWell, that :D18:06
treachjesse_: but all things considered "Jag talar/pratar inte svenska?" is a bit questionable, since if that's not it, what the heck is it then? :P18:08
predatorfreakRedShift: "* Patches are unacceptable unless in the case the software wouldn't work *at all*"18:08
predatorfreakI'd be less strict.18:08
predatorfreakand say "or would be severely impaired and a proper upstream fix is available."18:08
RedShiftif it really needs to be patched, it needs to be fixed upstream18:09
predatorfreakI.E. my libxvid patch is from xvid CVS.18:09
RedShiftimo18:09
predatorfreakThey're slow as fuck to release18:09
predatorfreakand REALLY should.18:09
predatorfreakBut, loosing the performance of properly optimised asm with media encoding is a bitch.18:09
treachprincips and implementation is separate things.18:09
treachs /is/are/18:09
treachdamn18:09
predatorfreakand until they do a formal release, a think a patch culled from CVS is acceptable.18:09
predatorfreak(Plus I tested it before hand)18:10
RedShiftyes there are always exceptions18:10
RedShiftbut the way things are going now is totally the wrong way18:10
predatorfreakRedShift: But yeah.18:10
predatorfreakI see some of that shit with Gentoo.18:10
predatorfreakIt's like they have one little bugfix.18:10
predatorfreakBut they can't submit it to upstream.18:10
predatorfreakEven though it's a one-liner!18:10
predatorfreakHonestly they have patches that they've applied to programs for 4-5 versions straight.18:11
jesse_How so? do you think gentoo devs have an inferiority complex which does not allow them to send those fixes?18:12
treachor they are too lazy to send them in.18:12
predatorfreakWhat treach said.18:13
treachor maybe upstream looks at it and says "gah, some gentoo user submitted a patch! mv patch /dev/null !" :D18:13
predatorfreakThey don't even get around to fixing half the shit that gets reported until a USER submits a fix.18:13
predatorfreakIt's 'cuz of how big portage is.18:13
predatorfreakTheir maintainers are overworked <_<18:13
predatorfreakRedShift: "fall out of the boat"18:14
predatorfreakor "get pushed overboard" :D18:14
treach"fall between the chairs"18:15
predatorfreakActually I was more thinking kick out the dolts.18:15
predatorfreakBut whatever :D18:15
RedShiftyeah it's a dutch expression, I didn't have a replacement in english18:17
RedShiftbut oh well, you understand what I'm trying to say, good enough for me ;-)18:17
treachfinding the replacement expression is rarely neccessary.18:17
*** Romster has quit IRC18:18
predatorfreakRedShift: Yeah, I got the jist of it.18:18
treachshouln't that be "geist"?18:19
treach;)18:19
predatorfreaktreach: This is America.18:19
predatorfreakWe butcher our own language.18:19
predatorfreakBut yes, yes it should.18:19
treachbah18:19
treachYou butcher everyone's language.18:19
predatorfreak-jist +geist18:19
predatorfreakNo hablo esponal?18:20
jesse_Se nota que no hablas ;)18:20
treachno, but swedish, english and german to the extent that I manage to recognize mangled words. :p18:20
predatorfreaktreach: We like raping languages.18:21
predatorfreakWe do it for fun in our spare time18:21
treachsure, the poms are your ancestors afterall. ;)18:21
*** Rotwang has quit IRC18:22
predatorfreakActually, I'm of German language rape descent :D18:23
treachmmh, I meant lingustically.18:23
treachseriously, even though your name is "Goulder", which would indicate german/(jewish?) ancestry, I don't think you speak enough german for it to have any influence. :)18:25
predatorfreakI don't speak shit of German.18:26
predatorfreakand no, my family isn't Jewish.18:26
predatorfreakMy dad's side is Christian/Morman.18:26
treachheh, I didn't think so, but I thought it migth be rude to say so.18:26
predatorfreakMy mum's side Catholic.18:26
predatorfreak(My dad's a protestant, his dad's a Morman)18:27
predatorfreaktreach: I prefer my mum's last name though :D18:27
predatorfreak"Mier18:27
predatorfreak+"18:27
treachFrankly I don't care about what religion people confess to.18:27
predatorfreaktreach: I confess to Flying Spaghetti Monster :D18:28
treachit's just that names like "Goldberg, Rosenthal" etc usually have that connotation.18:28
predatorfreakHe's really not such a monster once you start to get to know him :D18:28
predatorfreaktreach: Maybe way back my dad's side was Jewish, but they aren't anymore.18:29
treachAs I said, it doesn't matter. :)18:29
*** RedShift has quit IRC18:33
predatorfreakmmm fancy CD burning voodoo18:48
*** Romster has joined #crux18:48
predatorfreakRomster: Welcome back.18:50
Romsterjust woke up18:50
Romstersomehow i had my glx not set to my binarynvidia driver so screen saver killed xorg.18:51
*** jtnl has joined #crux18:51
Romsteri didn't rebuild mesa or xorg-server recently.. unless that power outage last nightscrewed over my system.18:51
predatorfreakMagic, I dunno?18:52
* Romster throws it in the unexplained bin18:54
Romsterat least the fix was easy.18:54
predatorfreakRomster: You can always blame me :P18:55
Romsterheh18:57
Romsteramazing nothing to update.19:02
treachOh noes! What to do now!?19:03
predatorfreakRomster: Test all my ports, bitch at me for undoubtably having a bad egg in the lot19:03
Romsterfuck around with my gtk colour theme i'm working on. <<19:04
predatorfreakoh and I'm seriously contemplating adding yasm into opt for xvid... it does a better job than nasm and isn't as fuckbally.19:05
Romsteri was horrifyed i had to maintainer meta names in a port http://crux.nu/gitweb/?p=ports/contrib.git;a=commitdiff;h=acf20ca5868523076ca2c454958f36a0551bb3dc and prtverify didn't pick up on it...19:05
predatorfreakPlus it does a better job for x264 too.19:05
Romsterhmm19:05
predatorfreak... People still use gift?19:06
predatorfreakhttp://en.wikipedia.org/wiki/Yasm <-- Presumably, everything that works with nasm should work with yasm.19:07
*** lucifer4u has quit IRC19:07
Romstera rewite of nasm19:07
Romsteri'm already there <<19:07
predatorfreakPlus or minus a "fake being nasm" binary.19:07
Romsterwell i still use gift and it still works so why not have it in contrib for the masses.19:08
Romsterthe point was why did prtverify not pick up on duplicate metatags.19:08
predatorfreakBittorrent, you cook! :P19:09
predatorfreak-cook +crazy19:09
predatorfreakActually, bittorrent, you nutball!19:09
predatorfreakThere.19:09
predatorfreakI'm off my game tonight :(19:09
RomsterNasm has been heavily used and debugged and therefore is more consistent from program to program. While Yasm is an active project, it may not have been as thoroughly debugged as Nasm.19:10
*** jtnl has quit IRC19:11
predatorfreakThey've been causing a lot of problems with 2.0 lately.19:11
predatorfreakFucking over configure scripts19:11
predatorfreakoccasionally requiring updates on the code itself.19:12
predatorfreakyasm accepts all of it without fail.19:12
treachNasm has been heavily used and debugged and therefore is more consistent from program to program. While Yasm is an active project, it may not have been as thoroughly debugged as Nasm.19:13
treachFUD?19:13
predatorfreaktreach: See above.19:14
treachwhat? I'm getting too tired to look at textmasses that way atm19:14
predatorfreaktreach: In my experience, yasm fucks up less than the new nasm.19:15
treach mmh.19:15
Romsterthe version checking pram got changed in version 2 of nasm19:16
Romsterthat's caused hell for checking for nasm19:16
treachI was adressing Romsters little speech. Not that I really care about it, but it sounds an awful lot like fud without any quotes or data exemplifying his fears.19:17
Romstergood point were is the reference on wikipedia...19:17
Romsterthere is noel19:17
Romsternone*19:17
treachwhere is noel? :p19:18
Romstermy shitty typing.19:19
Romsterheh only yhafri has yasm...19:19
treachRomster: just kidding again, made me think of "where's waldo" :P19:20
Romsterah19:20
Romsterthat wheres wally, i got some puzzles on that.19:20
RomsterNearly feature-complete lexing and parsing of NASM syntax.19:23
Romsterso not quite but should hopefuly work with most ports.19:24
predatorfreakRomster: I just added a proper, not fucked up yasm port to my repo.19:29
Romsterk19:29
Romsteri don't trust yahfri's shit <<19:29
Romsterbut i do nick a few of his ports sometimes and sanitise them19:29
Romsteradd on or a remove nasm first?19:30
predatorfreakHmm?19:30
predatorfreakI was going to add to opt and let the things which notice yasm use it for now.19:30
Romsterk19:31
Romsterso addon19:31
treachmmmh. that "nearly complete" is a bit of a worry IMo19:31
predatorfreakand when 2.5 rolls around, bitch about moving it to core and possibly doing something crazy like symlinking nasm to yasm.19:31
Romsteri'll add it to x264 to depinst yasm.19:31
treachcould they co-exist?19:31
predatorfreaktreach: They mean "If shit changes in the future, then we would have to update."19:31
Romstertreach, yeah, co-exist would be the better idea.19:31
predatorfreakIt works perfectly right now.19:31
Romsteri'd leave nasm as a fall back19:32
predatorfreaktreach: I'm doing it right now, yasm for x264/xvid, nasm for everything else.19:32
Romsterany hacking x264 or does it pick up on yasm first?19:32
treachpredatorfreak: sounds good, I was just not sure if yasm and nasm could co-exist on the same system19:32
Romstershould be able to.19:33
predatorfreakThey do.19:33
Romsterand there's as too..19:33
Romster'as'19:33
predatorfreakRomster: x264/xvid prefer yasm if present.19:33
Romstertoo many assemblers.19:33
Romsterah good19:33
predatorfreakgas too.19:33
Romsterand i think cc1 is a pre-processor/assembler too?19:34
predatorfreakcc1 is only a C pre-processor.19:34
predatorfreakI think.19:34
predatorfreakIt hands asm to as, I believe.19:35
Romsterso done any performance testing on x264?19:35
predatorfreakMaybe I'm wrong though19:35
Romsternot over sure.19:35
predatorfreakRomster: Nothing too noticeable, just like ~.2-.3 FPS.19:35
predatorfreakAs you would expect.19:35
Romsterbah no man page on cc1...19:35
predatorfreakLess annoying than nasm atm, they seem to think yasm is better.19:35
Romsterso whats the point unless the quality is better19:35
predatorfreakPlus yasm doesn't bitch like all fuck on xvid <_<19:36
Romsterlol19:36
predatorfreakWhere nasm goes crazy over OMG NOT QUOTED SHIT HERE OMG END OF WORLD19:36
Romstermissing { no quoting missing ; on end of line..19:36
predatorfreakThat shit.19:37
Romsteryeah i've seen all that shit and it worried me.19:37
predatorfreakIt works fine.19:37
predatorfreakBut nasm is a dolt.19:37
treachman, what would you do if somone stole your shift-key?19:37
predatorfreakand loves to bitch.19:37
predatorfreaktreach: Kill myself.19:37
treach*someone*19:37
predatorfreakmm, actually.19:37
predatorfreakI'd kill them.19:37
predatorfreakand steal back my shift key.19:37
predatorfreakoh and Romster, I plan on making a "x264 fprofile build pack".19:38
predatorfreakWhich will include various clips exhibiting differing bullshit things, I.E. fast motion, low motion, noise, etc.19:38
predatorfreakand a modified x264 build which decodes them to yuv4mpeg and does a make fprofiled build.19:39
predatorfreakEven just using a 10 second fast motion clip, I gained ~5-10FPS in x264.19:39
predatorfreakI figure, with a more extensive set of shit that x264 would encounter, it should gain more on depending on the source material and lead to better optimilasationzness.19:40
treachhow about reliabilizationess? :>19:41
Romsterlol yes if it's not reliable i arn't for it.19:42
predatorfreakIt works the exact same.19:42
predatorfreakThey support fprofiled building upstream.19:42
predatorfreakThe point to it being that GCC sometime makes dumbass mistakes or guesses and looses performance as a result.19:43
predatorfreaksometimes*19:45
Romster-fprofile-arcs19:45
Romsteranotherwords gather stastics on how you use it.19:45
predatorfreakThey append -fprofile-generate then rebuild with -fprofile-use, but same diff.19:46
predatorfreakYeah, it builds it once with profiling info, then runs some compiles and rebuilds.19:46
predatorfreakFairly fast shit already gains ~10FPS, slower shit ~1-5FPS.19:47
predatorfreak~10FPS gains are worthwhile in my book.19:47
Romster-fbranch-probabilities19:51
Romsterah is enabled with -fprofile-use19:53
Romsterguess gcc had changed that.19:53
Romsterso were on earth does it store the data from -fprofile-arcs19:53
predatorfreakRomster: build directory, that's the only downside to this.19:54
Romsterdamn..19:54
predatorfreakI have to disable ccache otherwise it fucks up.19:54
predatorfreak(Only for x264 though)19:54
Romsteranother downside.19:54
predatorfreakNah, storing data in the build directory isn't really a downside.19:55
Romsterbut the objects would differ so they shouldn't match in ccache?19:55
predatorfreakI more meant the downside is related to that.19:55
predatorfreakThat's not the problem.19:55
predatorfreakIt tries to put the data in a ccache temp dir.19:55
predatorfreakBecause of how ccache works.19:55
treachit is a real downside if someone is relying on having the build occuring some place else19:55
predatorfreaktreach: So long as you don't delete the source directory mid-build, you're fine.19:56
treachhaving problems with ccache wouldn't be a showstopper though since that's a bit "ricy" ;)19:56
Romsteroh it tryed to put the profiled code into ccache temp dir? kinda got that.19:56
Romsterheh19:57
Romsterpredatorfreak, you can use readonly but don't store new items in ccache19:57
predatorfreakhmm.19:58
RomsterCCACHE_READONLY19:58
predatorfreakI could do CCACHE_READONLY=119:58
Romster<<19:58
predatorfreakand CCACHE_TEMPDIR=$PKG/tmp or someshit19:58
predatorfreakand see if that works19:58
Romstermaybe.. jsut try CCACHE_TEMPDIR=$PKG/tmp without the readonly to at start?19:59
Romsterand add in readonly if it messes up still19:59
predatorfreakerr, $SRC/tmp19:59
Romsterhmm wouldn't .... i was abou to say $SRC <<20:00
Romsterabout*20:00
predatorfreakYeah let's give that a go over20:00
Romsteri'd try without readonly and see how it goes at first.20:00
Romstermight get away with just CCACHE_TEMPDIR20:00
predatorfreakNah, TEMPDIR doesn't do shit for it.20:02
Romsterdamn20:03
Romsterguess readonly or if that fails the disable ccache option :/20:03
predatorfreakccache period fucks this up.20:04
predatorfreakSo, gotta disable it.20:04
*** jnengland77 has joined #crux20:06
*** acrux|ppc has quit IRC20:19
* predatorfreak pulls out his traditional encoding tests, picks the most bullshit 10 seconds and does a speedy quality-equivalent encode of his videos so he doesn't ship 50MB of footage <_<20:29
*** joacim has quit IRC20:37
predatorfreakIrk or not, damn test cases make it impossible to get realistic download sizes and equivalent quality <_<20:52
predatorfreakI'll use my decently clean animu motion test cases for now, since motion compensation tends to be the big hitter.20:53
*** jnengland77 has quit IRC20:55
predatorfreakRomster: This is probably something better reserved for my x264 port, it requires mplayer to do a fast decode-to-yuv4mpeg run, plus ccache has to be disabled and such.21:05
predatorfreakMakes things kind of annoying because it increases the "dependencies" for the port to build XD21:06
predatorfreakRomster: I'm off for the moment, but if you want to see a working version of my fprofiled x264 port, look at the version in my repo.21:12
*** predatorfreak has quit IRC21:12
*** acrux|ppc has joined #crux21:54
*** Romster has quit IRC21:58
*** Romster has joined #crux22:01
*** Romster has quit IRC22:20
*** Romster has joined #crux22:20
Romstermy entire network at hop4 died...22:23
*** jaeger has quit IRC22:28
nipuLi think i need to fire my boss, i wrote a basic ordering system so orders wouldn't get lost and orders would get processed more quickly. he isn't even using it22:46
nipuLorders backed up for 2 weeks, hasn't added anything for about a week22:47
Romstergee22:50
Romsterthat must be pissing you off.22:50
Romstera solution is there and it's not being used.22:50
treachproblem #1 with all new systems. Get old users to use the *NEW* system, and not the old one.23:07
treachFamiliarity >> *23:07
*** jnengland77 has joined #crux23:07
Romsteryeah familiarity is the big thing but 'no change no progress'23:10
treachtell the users that.23:17
treachthey'll hug any system, no matter how inefficient, with the strength of a dying person, rather than do something new.23:18
*** brian|lfs has joined #crux23:42
brian|lfshello everyone23:42
*** treach has quit IRC23:58

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!