IRC Logs for #crux-devel Sunday, 2007-08-05

*** jaeger has quit IRC01:34
*** maro has quit IRC01:55
*** deus_ex is now known as pedja05:16
*** treach has joined #crux-devel05:19
*** treach has quit IRC05:33
*** jaeger has joined #crux-devel08:42
*** m94mn1 has joined #crux-devel08:56
*** m94mn1 has left #crux-devel08:56
tilmanjaeger: do you know about poppler-data?09:10
tilmanjaeger: we might want to ship that.09:10
tilmanthings to ponder:09:10
tilmanseparate port or not (it might only be useful to a subset of users)09:11
jaegerit's for CJK and Cyrillic, right?09:11
tilmanopt or secret repo, since the data isn't free, it seems09:11
tilmanit can be freely copied09:12
tilmanso that's not an issue09:12
tilmanand btw, glib 2.14 is out09:13
jaegerI'm sure someone would complain a lot if it's in the same port as poppler with 2 different licenses, free or no09:13
jaegerok, thanks09:13
tilmanbut i could imagine our ukrainian users would appreciate a port of poppler-data09:15
tilmanRomster: does glibmm 2.12 build against glib 2.14?09:17
tilmani guess it should09:18
Romsteri've made a update diff of sudo but of course i /still/ can not emil to anyhting with :/09:18
jaegeryeah, just thinking a separate port would be better for those two reasons: 1) maybe not useful to all, and 2) adobe license09:18
Romstertilman, i haven't tryed09:19
jaegeroh, if anyone would like to test the 2.3 netinst iso, find it at
jaegerif you DO try it, please use as the package URL, that's the only place I've uploaded 2.3 packages09:19
Romsterwould a version bump alone be suficent for glib?09:21
tilmani guess so09:21
jaegerI've built it on this machine, testing it now... so far it hasn't broken09:21
tilmanjaeger: are the package names hardcoded in the setup script?09:22
tilmanor could i try it with the packages, too?09:22
tilman(i'd set up a local ftpd)09:22
jaegerI generate a list and md5sums09:22
tilmani see09:22
tilmanvery cool09:22
tilmanyep, just found it09:22
tilmani'll try to test it in qemu later09:23
Romsterjaeger, while your at it change the /2.12/ to /${version%.*}/ inthe source line.09:27
jaegerwhat, no please? :)09:28
Romsteraww why not i hate version bumps and finding the source array has another version number to alter too.09:28
Romsterplease jaeger :)09:28
tilmanit's unreadable :P09:29
Romsterpfft lol.09:29
jaegerit's a tad cryptic, true09:29
Romsternot really, hasn't been to me *shrugs*09:30
Romsteryour call as its your port.09:30
jaegerI think I'll stick with 2.14 for now, might change it later09:31
jaegerdon't have to change that too often, fortunately09:32
Romstertrue, not in this case but some packages ya do, like in xorg i think i found a few.09:36
*** treach has joined #crux-devel09:43
Romsterhmm might of been kde i found.09:44
jaegerAFK, gotta run an errand09:45
clbUpdate from opt: 5 Aug 14:44 - glib: updated to version 2.14.009:46
Romsterfigure i gotta rebuild all of the ports that depend on glib... better check to revdep09:48
tilmanyou figure wrong09:48
tilmanglib releases are always source and binary compatible to previous ones09:48
tilmansince 2.0 obviously09:48
Romsterwell if any of the API's have altered ya do.09:48
tilmanatm, pkgutils2 is a speed devil10:00
tilmanpkginfo -l:10:00
tilmanpkgutils 5.30.0: .12s10:00
tilmanpkgutils2: 0.02s10:01
treachprt-cache listinst is just as fast. :)10:14
Romsteris that using -O3 ?10:56
tilmanwell, i just wanted to say that the new thing is going to be faster than the current one10:57
tilmanat least it looks like this atm10:57
tilmanRomster: O210:57
Romsterah standerd then.10:59
treachas if that would make much difference...11:03
tilmani don't even know where O3 wins over O211:04
treachwhen you gain from inlining and unswitching loops, probably ;)11:05
tilmanif i understand -funswitch-loops correctly, it's probably boring11:05
tilmancause if you have stuff that can be optimized with this, your code is really silly11:05
treachmmh, if you're using a lot of gotos it might make a difference as well.11:07
tilmanmmh, maybe11:08
treachbut all in all it sounds like a bunch of cornercases, more or less11:08
tilmani think in general goto-ladden code is hard to optimize11:08
tilmancause it's harder to make assumptions about the control flow11:08
tilmani think i'm almost being cptn obvious ;)11:08
treachgotos are good, but you should use them with care..11:09
tilmandoes it make sense to omit the leading slash in eg "pkginfo -l"?11:15
tilmanand pkginfo -o etc11:15
tilmanor: is it bad to show the leading /?11:16
jaegerI don't think it's a problem either way11:20
tilmanwonder whether there's a market for 'christian crux'11:52
tilman1.) do a survey on this11:52
treachhaha, I doubt it.11:52
tilman2.) produce christian crux11:52
tilman3.) ???11:52
tilman4.) profit!11:52
treachdamned religious nutjobs.11:53
treachI wonder what's so unholy about ubuntu that they need a "christian" edition.. Is it "angels" instead of daemons and "ghosts" instead of zombies..?11:57
treach*shakes head*11:57
tilmanmaybe they didn't like the default wallpaper with the three guys and 2 dekolletes12:00
treachor maybe they needed something to go with their christian computer, bought in a christian shop, standing on a christan table, etc. :p12:03
tilmanchristian ubuntu *is* real, right?12:04
treachafaik, yes.12:04
tilmani think the spoof was satanic ubuntu12:04
tilmani'm asking because forwards to, LOL12:05
treachthere are definitely enough religious wackos out there to make it possible.12:05
tilmani'd like to see a christian ubuntu user map12:05
tilmangnihihi, they have a bible box firefox theme12:06
treachgnucash is part of it. :D12:06
tilmanerr, bible fox12:07
tilman about half way down12:07
tilmangreat icons >12:07
treach"I really wish those who are opposed to the Christian faith could take their guards down for a moment and look at the bigger picture of what I am trying to accomplish."12:09
treachsounds like a nice representative.12:09
tilmanjaeger: do you know whether gtk 2.12 is due "soonish", too? i can't find a release plan12:10
treacha through and through hypocrite.12:10
tilmantreach: source?12:10
treachthe wikipedia page above12:10
tilmanyeah, that's what i had in mind12:12
treachI got that, nice little bible quote. :)12:12
treachdamn, it almost makes me sorry I don't use ubuntu. ;D12:13
tilman# There is no abort() support in Ubuntu Christian Edition's glibc12:13
jaegertilman: no idea there12:14
treachHAHA ;D12:14
tilmanIn Ubuntu Christian Edition, there can be only root and 12 more users, one of which will help crackers to acces the system.12:14
tilmanthat's a really good one ;>12:15
treachsudo is blasphemy. :P12:15
jaegerif I had to guess, I'd guess the christian version is made solely so the author can laugh at people who use it12:17
tilman "Six Six Sixty Four Bit"12:18
treachdon't forget there's Ichthux, and Ubuntu Muslim Edition (rather convenient to short as "Ubuntu ME"..)12:19
tilmanthere's an "Ubuntu Ultimate Edition" according to a german wikipedia page12:20
tilmani just find it stupid to poke fun at the vista naming ;)12:21
treachputting religious labels on stuff isn't too bright either.12:23
treachmakes me think of some medival hawker trying to sell miracle water and "genuine relics".12:24
treach"Sire, buy this genuine toe of St Thomas" *holds up skeletal detail from a dog*12:24
tilman> I want to release Pango 1.18 early next week.13:05
tilmanmaybe gtk 2.12 is around the corner13:05
tilmanthat would be so sweet13:05
tilman is at 2.12 already :O13:07
tilmanactually, "for GTK+ 2.11.6" ;)13:08
*** jjpk has joined #crux-devel13:47
tilmanthe state of pdf rendering on a free desktop is a sad one14:49
tilmanand/or the developers of acrobat reader aren't very capable14:50
tilmans/and\/or /14:50
tilmanthey definitely aren't14:50
tilmanthey don't seem to have *any* caching14:51
treachHate to say this, but afaik there aren't any *developers* on acrobat for linux. :/14:55
treachafaik it's *one* guy, or so.14:55
tilmanit's kind of nice that they list all the people involved in Help->about->Credits14:57
treachIt must be adobes variant of putting someone in the freezer.14:57
treachwhat happened to rakt, btw. I saw you sent sip a tarball. Does that mean there's still hope? :)14:58
treachepdfview feels rather... buggy. :>14:58
tilmangunzip optimtut1.ps14:58
tilmanps2pdf optimtut1.ps14:58
tilmanepdfview optimtut2.pdf14:59
tilmannotice it's completely unreadable14:59
tilman-> poppler sucks14:59
tilmantreach: but epdfview annoys me, too. so yeah, there's hope :D15:00
treachwell, I think it's because of the font used. Some looks pretty ok to me, others, like this one, is hopeless. :/15:01
treachs is/are15:01
tilmanmaybe i'll try poppler 0.5.9 sometime15:02
treachalso, I'm getting fairly tired of that "LOADING" crap. :/15:02
jjpkI don't know what adobe is doing with acroread in linux, but it sucks.15:04
treach..and seems to be down, again. Are they hosted by sourceforge..?15:05
tilmanMIT is having connectivity problems15:05
tilmanthat's why is down15:05
tilmanfdo is at portland ;)15:05
tilmanpeople.fdo works for me15:05
treachMIT having "connectivity problems" is hardly news. :>15:05
tilmanas does www.15:06
treach$ doesn't15:07
tilmanwith $name being what?15:07
treachpoppler, for instance15:08
tilmanjust tried that one, too :)15:08
tilman"suggested fontconfig font substitutions"15:10
tilmanmmh, maybe i should try those15:10
tilmanalthough i don't think those fonts are very popular15:10
treach"dingbats".. mmh, that sounds like something drepper would use.15:11
treachZapfDingbats -> Dingbats15:12
tilmanwell yes15:12
tilmanbut that's some kidn of symbol font, isn't it?15:12
tilmandrepper's paper uses helvetica afaict15:12
treachof course15:12
treach.. sorry. forget about it. :)15:13
tilmanforget about what? ;]15:13
treachthat thing you didn't see passing you. :p15:14
* tilman goes through poppler's bug list in bugz15:14
treachwhat I don't get is why evince/kpdf manages to look decent, while everything else looks like crap.15:15
tilmanreally? evince = good whereas epdfview = bad?15:16
treachwell, at least I don't have any recollection of it behaving nearly as badly.15:16
treachsame goes for kpdf15:16
tilmanthat's weird, since both use the same poppler backend (the glib one)15:17
tilmanor do they15:17
treachand "good" is probably an exaggeration15:17
tilmanmaybe evince uses the cairo backend15:17
tilman"pdffonts optimtut1.pdf" yields itneresting output15:20
tilmanmaybe acroread just has a better fallback font15:20
treachhmm, something's still strange.15:22
treachI just tested with "SumatraPDF", which is a windows app, but also uses poppler apparently.15:22
treachThe result on the original pdf was way better than epdfview15:23
jjpkThat bug is still floating unfixed?15:25
tilmancomment #2 suggests it's not a bug15:25
tilmanas i understand it, cairo uses poppler to test its own pdf system15:25
jaegerargh, can't get terminus sources right now15:26
tilmanwant me to upload them?15:26
jaegerthat would be great :)15:26
jjpkWouldn't be the first time terminus is unavailable.15:27
jaegerjjpk: aye, was thinking about mirroring it myself but I never remember that when it's up :)15:27
jaegertilman: thanks much15:27
treachscary, running that silly program through wine works a million times better than epdfview..15:28
jjpkkdpf was nice, but it unfortunately depends on kde.15:30
treachyeah, kind of a downer.. :>15:30
jjpkWhat else did I find annoying with kpdf...15:31
jjpkIt was unable to open multiple files iirc.15:31
tilmanwith the current version of cairo, i would not get the impression that it depends on poppler again15:32
tilman    AC_MSG_WARN([PDF backend will not be tested since poppler >= $POPPLER_VERSION_REQUIRED is not available])15:32
jjpkAt least it confirms what was said in poppler's bugzilla.15:34
tilmantreach: does sumatra render optimtut1.pdf correctly, too?15:35
treachwell, "reasonably readable" I'd say15:36
tilmani should install 0.5.9 *now*15:36
*** Romster has quit IRC15:36
*** Romster has joined #crux-devel15:37
treachcurrently building it.. :>15:42
tilmanx just died for me15:42
treachmh, wonder if I need to rebuild epdf..15:44
tilmandoesn't help anyway15:45
treachsame crap, but faster, I'd say. :/15:46
tilmanevince uses poppler_page_render(), which renders to a cairo surface15:59
tilmanraktpdf uses _render_to_pixbuf()15:59
tilmanditto for epdfview15:59
tilmanbut i don't think it's likely that this is causing the problems16:00
treachhaha, just played a bit with rakt, too bad it doesn't have zoom.16:09
treachbecause it absolutely kills epdf on pdfs that makes you wait forever on "Loading" :>16:10
treachlike the fhs-2.3.pdf and "lshort.pdf" ("short" guide to latex)16:10
tilmanif you want to play, feel free to make raktpdf use cairo instead of gdk_pixbuf16:10
tilman0.5.9 really sseems faster16:11
treachhm, I'm too tired right now for such exercises, but I might have a look at it tomorrow.16:11
treachyeah, quite a bit16:12
tilmanlshort.pdf doesn't show much "loading" at all here16:12
treachI bet my system is slower than yours. :(16:12
tilmanbut this is a 1.5 ghz athlon :D16:12
treachthat should be ":)"16:12
treachwell, this one is clocked at 700Mhz. :P16:13
tilmando modern desktop cpus support dynamically changing the clock speed?16:13
tilman"speedstep" etc16:14
treachseriously, it's fast enough to watch mkv's etc, so why shouldn't it be enough to render a few lousy pdfs..?16:14
treachathlons do, "cool'n quiet"16:14
treachthe more modern ones that is.16:15
tilmanthat's good to see16:15
treachthis is only possible because I have a nforce2 mb.16:15
treachother than just setting it in bios, obviously.16:15
tilmanhehe, just found a dumb memory leak in raktpdf16:21
treachone of those discoveries you get both happy and annoyed by. :)16:23
tilmani'll do the cairo move now i think16:25
jjpk"yay found bug. :) nay, it's a mem leak. :("16:26
tilmanjjpk: more like "omg, i can't believe i didn't notice that :("16:27
tilmani usually often run valgrind to find memory leaks etc16:27
tilmanwhen i started on raktpdf, i found a gtk bug16:28
tilmani reported it, i fixed it, but they rejected it: "the fix looks good, but i can't reproduce the bug. sucks to be you."16:28
treachjjpk: I meant like "something you'd rather not see, but non the less are happy you found"16:28
treachsounds really sane..16:29
jjpkThe mind is a terrible thing to taste. :p16:29
tilmani can see why they hesitated16:29
tilmanbut some more attempts and work and they would have been able to realize that there really was a bug16:30
tilmanoh well16:30
jjpkA shit attitude for a dev. :s16:32
tilmanyes, i was pretty annoyed about him >:16:33
tilmanespecially because i spent 2 hours or so diving into that horrible gtk-file-chooser code16:33
tilmanto understand and fix the bug :x16:33
jjpk"gee, thanks"16:35
treachpass the patch to jaeger. :p16:38
tilmanrendering is *much* faster now16:50
tilmanusing cairo16:50
tilmanif you want to test the *rendering speed*16:52
tilmanyou'll have to resize the window until you see all the contents16:52
tilmannot quite finished16:52
tilmani don't know if that patch will work with poppler 0.5.4 bw16:53
treachit doesn't matter. :)16:53
treachhm, that patch didn't apply cleanly.16:55
treachI guess I'm behind the times. :/16:55
tilmanhead is e8ff9ad0234828d58e8b4cf81680b7650d6b43b316:56
tilmanbut don't fear!16:56
treachmh, did you intend to send the .git dir?16:59
tilmancause you can do everything with it16:59
tilmancd /tmp16:59
tilmanmkdir raktpdf16:59
tilmancd raktpdf16:59
tilmanbsdtar xf raktpdfgit.blah16:59
tilmanmv newdirectory .git16:59
tilmangit checkout master16:59
jaegerwhere should the poppler-data files go? /usr/share/poppler?17:00
jaegerlooks like it17:01
tilmanthat's the default in Makefile17:01
jaegersince it wants to put them in /usr/local/share/poppler by default17:01
jaegerI'll go ahead and put poppler-data into opt17:07
treachmh, the patch still partly fails. I must be doing something wrong.17:09
tilmanwhat's the first line of 'git log'?17:10
treachI should probably not play with this when I barely can keep my eyes open. :P17:10
treach Fixed an off-by-one error that enabled the "next" button on the last page.17:10
tilmanare you misreading the output of patch perhaps?17:11
tilmani don't see how the patch could fail o_O17:11
treachHunk #1 succeeded at 40 with fuzz 1 (offset 2 lines).17:11
treachHunk #2 succeeded at 143 (offset -7 lines).17:11
treachHunk #3 FAILED at 150.17:11
treachHunk #4 succeeded at 207 (offset -10 lines).17:11
treachHunk #5 FAILED at 230.17:11
treachpardon the spam.17:11
tilmancan you try:17:12
tilman$ git reset --hard17:12
tilman$ patch again17:12
treachthat seems to work.17:13
tilmanvery strange17:13
tilmanworks as well as before the switch to the cairo stuff afaict17:13
treachdo I need to reset before I apply that one as well? :)17:13
tilmanit's not an incremental patch17:14
tilmani think zooming should be very easy to implement btw17:14
tilmanmaybe ~5 lines to request the zooming, and 200 to add the widgets17:14
treachahem, that wasn17:15
treach't nice.17:15
treachapparently something went wrong. :/17:15
clbUpdate from opt: 5 Aug 22:10 - poppler-data: initial import, version 0.117:17
treachthat's supposed to be lshort.pdf, btw.17:19
tilmanthat's idiot2.diff?17:19
jaegerthat's some nice rendering, there17:19
tilmanthis stupid raktpdf thing can't even read file names from the command line17:20
tilmantreach: does it magically fix itself when you switch pages around a bit?17:20
tilmanlshort.pdf works for me17:21
treachno, it'd doesn't fix itself.17:21
treachs it'd/it17:21
tilmanthe size of the window looks about correct17:21
tilmanso i wonder if this really is related to the changes between idiot.diff and idiot2.diff17:22
treachhmm, I haven't tried "idiot" yet, I was busy, so I went straight for nr217:24
treachbut if I hard reset to the original that file works, if I use the #2 patch it fucks up.17:24
tilmanyour copy of cairo is up to date?17:25
treachthe first one screws up too.17:25
tilmanpoppler 0.5.9, right?17:25
tilmanand the latest gtk?17:25
treachafaik yes.17:26
treach(at least prt-get diff doesn't tell me there are any updates)17:26
tilmanokay, it wasn't updated recently i think17:27
*** jjpk has quit IRC17:39
*** pedja has quit IRC17:56
*** maro has joined #crux-devel19:21
jaegerwow, no ndiswrapper port anymore?21:09
*** jaeger has quit IRC22:36
*** jaeger has joined #crux-devel22:43
Romsterany reason why wxgtk is not at current?23:12
Romsterand if it's not being maintained could it be droped so i can maintain it in contrib?23:14
*** jaeger has quit IRC23:55

Generated by 2.11.0 by Marius Gedminas - find it at!