IRC Logs for #crux Thursday, 2011-06-30

*** dru1d has quit IRC00:13
pitillogood morning00:54
enteso many failed buildlogs today :)01:03
*** lasso|qt has joined #crux01:37
*** mike_k has joined #crux01:40
Ovim|afk_nAMorning everyone :)02:11
*** Ovim|afk_nA is now known as Ovim-Obscurum02:11
*** j^2 has quit IRC03:25
Romsterente, better be none from mine -_-03:39
Romstergood thing we use xulrunner in firefox now as it now links to that and not the system supplied one.04:06
enteRomster: I already submitted one ;)04:09
entebut it's almost done with contrib at least04:10
enteand that's everything, including footprint mismatches04:10
entexmms is taking it's time...04:11
enteI might pick up some of nipul's old p5-ports04:13
entep5-berkeleydb has a footprint mismatch and is ages old04:13
Romsterhow can the link be outdated already i checked for 404 links on all my ports a couple of weeks ago.04:26
*** ente has quit IRC04:28
cruxbot[contrib.git/2.7]: p5-berkeleydb: adopted; 0.36 -> 0.4304:30
*** aubic has joined #crux04:31
*** ente has joined #crux04:33
*** ente has quit IRC04:33
*** ente has joined #crux04:33
Romsteralso what happed to telling me on irc i fix my stuff pretty fast.04:40
teK_have you had time to add libgcrypt to chromium's deps? :)04:41
enteI think it's easier to keep track via bugtracker04:42
entefrinnst disconnected yesterday and I didn't even notice and kept talking :P04:42
enteRomster: tor is outdated, I guess that's all04:43
RomsterteK_, not yet getting to it soon, i've made a note of it when is aw it at my lunch time.04:44
Romsteri fought i updated all perl footprint stuff ages ago for new perl04:45
Romsterit's hard to keep track of everything.04:46
Romsterand work at the same time.04:46
* ente nods04:46
entethat's why I procrastinate work :P04:46
RomsterteK_, the only reason i dind't bump chromium is because it wanted gconf or pam i can't remember either is not desirable to me.04:47
entebut it builds fine with a fake-gconf04:47
entei.e. echo | gcc -xc -o libgconf.so04:47
entewhich is plain dumb04:48
entemissing a - at the end I guess04:48
enteRomster: it "depends" on gconf, but works fine with an empty library that's just called gconf04:48
enteI guess patching it is the correct way to go04:49
Romsterah ok...04:53
Romsterwish the dumb thing had a --disable-gconf04:53
entebut the build system is rather special04:54
enteso I suggest digging through it and patching the need for gconf out in the buildsystem04:54
enteRomster: I read it doesn't need gconf anymore in newer versions04:55
enteI didn't try building it for some time04:55
Romsterfirefox deps line should read # Depends on: alsa-lib curl desktop-file-utils gtk lcms libidl mesa3d nss yasm zip04:56
cruxbot[contrib.git/2.7]: spamassassin: 3.3.1 -> 3.3.205:05
cruxbot[contrib.git/2.7]: skipfish: 1.94b -> 2.00b05:05
cruxbot[contrib.git/2.7]: tornadoweb: 1.2.1 -> 2.005:05
cruxbot[contrib.git/2.7]: tftp-hpa: 5.0 -> 5.105:05
cruxbot[opt.git/2.7]: libgcrypt: 1.4.6 -> 1.5.005:06
cruxbot[opt.git/2.7]: unrar: 4.0.3 -> 4.0.705:06
Romsterdownlaoding chromium 14.0.806.0 -_- will take a little while05:06
teK_yep :P05:07
teK_unpacking, too05:07
*** Evil_Bob has joined #crux05:18
enteRomster: I reported a bug for ff yesterday05:18
Romsteryeah for missing zip only.05:19
cruxbot[contrib.git/2.7]: tor: ->
Romsteri built it with that depends on line just before05:19
Romsterin a chroot05:19
enteRomster: and another one for mesa3d and something else05:20
RomsterDependencies: alsa-lib,curl,desktop-file-utils,gtk,lcms,libidl,mesa3d,nss,yasm,zip05:20
Romsteryasm i'd say05:20
cruxbot[opt.git/2.7]: Revert "firefox: added zip as a dep. Closes FS#730."05:20
Romsterrevert O_O05:20
Romstercruxbot> [opt.git/2.7]: Revert "firefox: added zip as a dep. Closes FS#730."05:21
enteyeah, seen that05:21
entezip is in core I guess05:22
Romsteris it?05:22
Romstersince when was that in core... ah wait for pkgutils05:23
entewell, the other two are not in core05:24
Romster-# Depends on: libidl, nss, gtk, perl, python, alsa-lib, yasm05:24
Romster+# Depends on: alsa-lib curl desktop-file-utils gtk lcms libidl mesa3d nss yasm05:24
Romsterelse it wont compile.05:24
enteyeah, mesa3d is still missing05:25
Ovim-Obscurumeh weird.05:25
entecurl.. hm.. curl is installed anyway05:25
enteI guess I should finally install the whole set of core ports05:26
entebut meh05:26
Romsterhmm it is indeed remove curl off it.05:26
Romsterit's not linking to it. my bad05:26
Romster+# Depends on: alsa-lib desktop-file-utils gtk lcms libidl mesa3d nss yasm05:27
entegcc -lidl -aldi05:27
Ovim-Obscurumhrhrhrh ente05:27
Ovim-ObscurumI thought about it when I read the lib name for the first time.05:28
entegcc -library05:28
enteIMO newlib is a stupid name05:28
Romsterprt-get info libidl05:28
RomsterName:         libidl05:28
RomsterPath:         /usr/ports/opt05:28
entewhy not call it libibrary :D05:28
Ovim-Obscurumlibsexy is a stupid name05:28
RomsterA library for creating trees of CORBA Interface Definition Language (IDL) files05:29
Ovim-ObscurumBut it really exists - was a dependency for compiz05:29
Romsteruh is that even needed... i got the deps line off another distro.05:29
Romstertime to try rebuilding it again without it grr05:29
Romsteri have ways to find missing deps but none to prune unneeded ones. other than ones not linked to port if i scripted that.05:31
Romstersome ar e build time deps.05:31
ente <- shorter link to the same task ;/05:32
Romsterchecking for libIDL-2.0 >= 0.8.0... Package libIDL-2.0 was not found in the pkg-config search path.05:33
Romsteryes it needs libidl05:33
Romsterbuild time.05:34
enteit depends on perl *and* python05:35
entebinary of babel :P05:35
enteperl is in core05:35
entedoes it link against it or is it a buildtime dep?05:35
Romsterchecking it now05:36
Romster3.4GHz quad core still isn't fast enough for me it's not even using 100% of my cpu :/05:36
enteI'm building with -j16 on some i705:36
entemodel name      : Intel(R) Core(TM) i7 CPU         860  @ 2.80GHz05:37
entethat machine is pretty fast05:37
Romster-j4 with BFS scheduler05:37
enteit's better to start more jobs than cores, I think05:37
entethat way I/O gets less in the way05:37
enteI built all the packages for now :)05:38
enteand it threw > 300 errors05:38
Romster For years we've been doing our workloads on linux to have more work than we05:38
Romsterhad CPUs because we thought that the "jobservers" were limited in their05:38
Romsterability to utilise the CPUs effectively (so we did make -j6 or more on a05:38
Romsterquad core machine for example). This scheduler proves that the jobservers05:38
Romsterweren't at fault at all, because make -j4 on a quad core machine with BFS is faster than *any* choice of job numbers on CFS.05:38
entethe mbox containing the buildlogs is 21M of size05:38
Romsteruh thought i removed all the CRLFS on that sorry..05:39
entecan't patch the kernel on that machine :)05:40
enteand I don't really want to anyway05:40
Romster includes it, it is faster.05:40
Romsterand it's stable.05:40
Romsterjust some compiles seem to have some other bottleneck.05:41
Romsterand i can't see what when i compile in ramfs too.05:41
Romsterreading disks for header files,libs?05:41
Romstershould cache it.05:41
Romsterfinddeps says libidl is linked into firefox.05:43
RomsterDependencies: alsa-lib,desktop-file-utils,gtk,lcms,libidl,mesa3d,nss,yasm05:44
Romsterwant me to try and prune some more? or are you happy with that.05:44
Romstereven chromium taxes all my cores to 100%05:49
Romsterso i dunno what firefox is doing during compile that makes it so underutilized at compile time.05:50
Romsteri read a while ago that google let MS buy skype because google things they can do a better chat program in there browser :D05:51
Romstererr s/things/thinks05:52
Romsterwho the heck wants chatting in a web browser anyways.... stupid idea.05:52
Romsterother than a irc gateway in javascript.05:53
Romsterwhy is everyone let me monologue...05:56
Romsterok i think chromium takes longer than a bootstrap of gcc...06:07
entebecause that way I'm not the one this time who does the monologue ;)06:07
enteyeah, browsers are bloatware06:07
RomsterSQLiteFileSystemChromiumPosix.cpp:(.text._ZN12_GLOBAL__N_112chromiumOpenEP11sqlite3_vfsPKcP12sqlite3_fileiPi+0x22): undefined reference to `chromium_sqlite3_initialize_unix_sqlite3_file'06:11
Romster -Duse_system_sqlite=106:18
Romstermy sqlite3 probably needs a bump06:18
teK_had the same error, too (reported it here iirc)06:23
Romsternewer sqlite3 than in crux opt sqlite-autoconf-3070701.tar.gz06:24
Romsterwho's stupid idea was it o use 0 as separator...06:27
Romsterlucky i got ccache installed.06:28
ente0 as separator for what?06:33
*** aubic has quit IRC06:40
*** bandito85 has joined #crux06:50
teK_well, it's sponsored by adobe and oracle :>06:58
Romsterstill no excuse for such stupid thing as that.07:01
teK_of course not.07:01
Romstersome how i think chromium is bypassing my ccache...07:04
Romsterstill building.07:04
Romstershould of been done by now07:04
Romsteri set a shadow directory to shadow gcc g++ names to /usr/bin/g++ by appending it to my PATH.07:05
Romsterunless this thing is using absolute paths... don't wanna stop it.07:05
teK_you better shouldnt ;)07:05
enteyou could mv g++ to /usr/local/bin :)07:09
Romstergod for sake i think using out of tree sqlite3 si a bad idea even with bleeding edge sqlite3 on my system...07:10
teK_i just defined use_system_sqlite=0..07:10
Romsteryeah it's fucked up using the system one :/07:11
teK_fun fact: i compiled it on my 1,6GHz Dual Core System and the above error occured while linking the browser binary i.e. after quite SOME time ;)07:12
Romsterwhy did i not notice the stupid prick of a thing prefixes the names with chromium_sqlite3_07:13
Romsterwhile it's stopped i'm gonna see why it bypasses my ccache.07:14
enteis it still about chromium?07:15
enteis it in contrib?07:16
ente... because imo it should be moved there07:16
entegiven it's importance07:17
teK_but I'd like to see/put it there, too07:17
teK_i'm switching from FF, as it seems.07:17
enteI recently switched to opera07:18
enteI get a lot less reports now07:18
entemostly because I cache broken deps :D07:18
Romsteri wont move it there until the bugs are ironed out.07:34
Romsterah i see i missed symlinks for gcc 4.5 when i use gcc 4.6 on my system07:43
teK_which bugs?07:44
Romstercompiling for starters :D07:44
Romsterand being reliable to compile.07:44
Romsteri don't fully understand the buld system yet.07:45
Romsteri'm not throwing half baked barely know how it works, and is it gonna break later on builds in contrib until i know it's going to be stable enough, else i'll get the whole why haven't you bumped it yet it's out of date. every time a new release won't compile.07:46
teK_so I'm your first satisfied customer ;)07:47
Romsteri guess. taking a risk using my personal repo :D07:48
Romsterit's starting to get some cobwebs in there on other ports.07:48
Romsterexcept a few i keep bumping.07:48
Romsterthat reminds me who's doing wine isn't that up for adoption? I got a wine dev build in my personal repo.07:49
teK_so yes it is.07:50
Romsterthough i must admit mine needs a little work on the deps line. and i eed to include the geck stuff too.07:50
Romsterbut i'm quite familiar with it as i hang out in the wine dev and other channels.07:51
Romsterthough i doubt anyone wants a dev release of wine in here. probably want so called stable.07:51
Romstermuch better i got ccache working on g++-4.507:52
teK_dunno, I'm on 64bit07:53
teK_so idc07:53
Romsterhmm someone else would have to do the 64bit wine it needs 32bit libs. i do plan to move to 64bit sometime too but not /just/ yet.07:55
teK_I'm glad I don't need it07:56
Romsteri can do it but it's just a little more work...07:56
Romsterwine 64bit only will rul 64bit PE executable.07:57
Romstererr rul run...07:57
ente zomg07:57
entefirst time I had that on crux :P07:57
ente... except when I tried to package something of course07:57
Romsteryeah everyone should be building in a clean chroot with out core and what depinst pulls in.07:58
Romsteri do that for contrib ports i've slacked off on that on my personal repo though.07:58
entebuilding in a chroot in a chroot sounds tedious07:59
Romsterfor the time being, until other projects are done it'll be all self automated.07:59
enteand without core... I don't know07:59
enteudev won't even build without having the whole core installed07:59
enteor at least some packages from core I didn't have installed07:59
entespeaking of build-time deps from core here07:59
Romsteri have SBPREFIX=$HOME/.safe-build-foo and each chroot of that has core only...07:59
enteand I don't see why zip is essential enough to be in core, tbqh07:59
entebsdtar does zip just fine07:59
Romsterhmm bsdtar has zip capability... that's a good argument ente08:00
Romsterperhaps something else in core needs it.08:00
entecrux ~# for x in /usr/ports/core/*; do prt-get quickdep asename $xawk '/\<zip\>/ {print $NF}';done08:02
enteseems not08:02
entecrux ~# for x in /usr/ports/core/*; do prt-get quickdep `basename $x`|awk '/\<zip\>/ {print $NF}';done08:02
entehaving ` as tmux prefix always makes me have paste problems :P08:03
Romstermore than likely it's not listed as a build time dep...08:04
Romsterrebuild all of core without zip then report back.08:05
Romsteri'd try but right now chromium is still going..08:06
Romsterand to think i thought webkit and boost were a pain.08:06
Romsteractually chromium includes webit...08:06
enteI did build all of core without zip, and it worked08:06
Romsterdo i even dare see if it'll build with system webkit after this compile08:07
enteit was only through firefox that I even noticed zip is in core08:07
Romsterok so report to flyspray about zip being in core?08:07
Romsterit probably got overlooked08:07
Romsterit was needed for pkgutils before bsdtar took over.08:08
entemaybe I could post on crux-devel08:09
Romstermy gawd it finally finished compiling... so committing this right now.08:09
enteI feel bad for reporting so many fs tickets :P08:09
teK_don't (feel bad)08:09
entelet's pretend it's not a bug08:09
Romsterheh i used to tons before i got a full time job.08:10
teK_product improvemnet08:10
teK_still compiling *sigh*08:10
Romstercould imagine me updating all my ports every single day and reporting others that are out of date.08:10
Romsteri must of got on some peoples nerves.08:10
enteI could automate sending mails to the persons reported in the pkgfile :P08:12
Romsterthat's why i got a fastish computer for the compiling mostly. how sad is that.08:12
entethat would *really* go on people's nerves08:12
Romsterget on...08:12
pitilloI think some people who reported many "bugs" got that feeling :)08:12
enteget, sorry08:12
teK_I'm happy about every single user reporting bugs to me..08:12
enteI'm discovering all bugs at once with crux-bin :P08:13
enteI also find out which packages are rarely used08:13
RomsterHEAD is now at 384b15d chromium: 10.0.616.0 -> 14.0.806.008:14
pitillotrue teK_, IMHO that means that things can be done a bit better08:15
jaegerfuck you, people who don't give enough information!08:15
Romsterwhats up jaeger?08:15
jaeger"here's our public IP address, we need to set up a VPN connection"08:15
jaegerOpenVPN, IPSec, PPTP, L2TP?08:15
jaegeretc. etc.08:15
Romsterto what local ip etc08:15
jaegerlocal subnet? endpoint hardware?08:16
Romsterand yeah which VPN program.08:16
jaegergive me something useful08:16
Romsteri get that crap all the time at work jaeger08:16
jaegerMe too and I hate it08:16
Romsterthough it's rarely about a VPN08:16
jaegerHow the hell do people like that get ANYTHING done?08:16
Romsteri have no clue.08:16
* teK_ is reading an article about Foreskin-Man. Crazy Americans :P08:16
Romsterso yeah anyone after chromium it's bumped in romster REPO.08:17
Romsterlike more testers.08:17
Romsteralso here is a tip if your not already using -fvisibility-inlines-hidden on your CXXFLAGS in pkgmk.conf do so it makes C++ code so much quicker but it can break on some C++ compiles but not on chromium that worked ok without missing symbols.08:19
Romstercome to think of it only codeblocks breaks on that flag afaik...08:20
Romsterand it's easy to add -fno-visibility-inlines-hidden08:20
Romsterjaeger, it was so damn busy at work that i was losing track of what i was doing and who's computer was who's, though i did remember what i was doing to each computer.08:21
Romsterand then the dc socket for one notebook i unsoldered the broken one and the new one isn't the correct pin out so i have to wit on that one08:22
Romsterand the other notebook i gave the exact model number and number off the broken LCD screen only to get a 16" screen sent to me when the thing uses a 15.6" screen argh...08:23
Romsteri was about ready to murder someone -_-08:23
jaegermost of our users are on windows or macosx, which isn't my job, thankfully08:23
jaegerI mostly do infrastructure and networking, though we have around 5 full-time linux users08:24
Romstersadly i deal with windows and the mac stuff i pass onto the other guy as i really don't want to learn that.08:24
Romsterleast time i played on the mac i was what no lspci....08:24
jaegersystem profiler gives a ton of info, it's pretty nice08:25
Romsterand other commands and damn menu and x on wrong side of GUI.... give me back mu Linux desktop :D08:25
Romsteryeah using that i was just surprised it's lacking lots of stuff that i'd prefer to use.08:26
jaegerhrmm... "tracking servo failure" on my cdrw drive... awesome08:26
Romsterwe had one guy that does media editing with a camera and firewire... turns out he had nothing but hell on a mac, and went to windows all worked like a charm08:26
jaegerGonna be one of those days, I think08:26
Romsteri had a thought on a optical drive testing if it's good or dodgy and needs replacing by writting a bunch of sessions at once and re-reading it back from all over the place. to see if i can find a fault.08:28
Romsteri haven't done much searching if such a program exists yet.08:28
Romsteras i seem to be in procession of lots of optical drives and i dunno which ones are flakey or not.08:28
jaegerThey all are :)08:29
jaegerJust to different degress08:29
Romsteryeah that's the thing i want to try and measure...08:29
Romstergrade them by how flakey they are.08:29
Romsterthe really flakies get thrown out and the not so bad ok for reading jobs.08:30
Romsterwhere the really good ones for writing.08:30
* Romster plays with chromium08:31
RomsterteK_, your on the latest version already or some versions behind the one i just made?08:31
*** pitillo has quit IRC08:31
Romsterhaha rainbow dash theme tilma_n  would love that one.. it's got a pony on it.08:32
Romsterhmm this might take over firefox if it's good for my use.08:33
Romsterthough i'm more hoping midori matures more.08:34
Romsteri want my web developer plugin :D that FF has.08:35
Romster[2944:2944:1466635213189:ERROR:CONSOLE(6509)] "Uncaught TypeError: Cannot read property 'can_uninstall' of undefined", source: chrome://newtab/ (6509)08:35
entethese scripts feel dirty somehow :P08:39
slashbeastente: ` is deprecented and use more quotes08:42
enteno, ` is perfectly fine08:43
Romsterflash-player-plugin is now at
slashbeast` is like $[] for math08:43
ente` is bourne shell, while $() is not08:43
slashbeast$() works in posix sh.08:44
Romsteri thought it was (( )) for math08:44
slashbeastand anyway, you script in bash, not bourne shell08:44
ente(( )) is for meth :P08:44
slashbeastRomster: it is, older syntax was $[math]08:44
entewell, as long as I don't nest ``?08:44
Romsteri tend to like $() over ``08:45
ente$() is longer :P08:45
slashbeasteasier to read08:45
Romsterit can get fun with ` and '08:45
jaegerthis is a dumb argument :P do what you want, as long as it works08:45
slashbeast15:45 |    slashbeast | !`08:45
slashbeast15:45 |       greybot | The backquote (`) is used in the old-style command substitution, e.g. foo=`command`. The foo=$(command) syntax is recommended instead. Backslash  handling inside $() is less surprising, and $() is easier to nest. See
enteRomster: yeah, if you use a font that's not suited for terminals08:45
slashbeastvia #bash @ freenode08:46
enteI don't see any backslashes in my code ;/08:46
Romster[ -n "$(eval "echo \$PKGMK_${UPPER_NAME}_CFLAGS")" ] && export CFLAGS+=" $(eval "echo \$PKGMK_${UPPER_NAME}_CFLAGS")"08:46
Romsterfor example.08:46
*** mike_k has quit IRC08:46
enteand I only use bash because I'm not sure whether shopt -s nullglob is posix :P08:47
slashbeastRomster: instead of eval you can use $(!PKGMK etc.08:47
slashbeastin bash.08:47
slashbeast$ foo=bar; bar=tak; echo ${!foo}08:48
slashbeastente: about your 2nd script, you could use bulitin test instead of putting test in ` ;p08:49
enteslashbeast: hm?08:49
slashbeastente: WC_BUILT=`test -f $BUILT && wc -l <$BUILT 2>/dev/null`08:50
slashbeastand wc can read just file, dunno if using redirect is faster.08:50
enteyou mean  WC_BUILT=`[[ -f $BUILT ]] && wc -l <$BUILT 2>/dev/null` is better?08:51
slashbeast[ -f "$BUILT" ] && WC_BUILT="$(wc -l $BUILT 2>/dev/null)"08:51
slashbeastI am preffer[ [ over [[ unless I need to put regex there.08:52
slashbeast[[ have some adventage over old [, like 'auto-quoting' variables.08:52
ente...and they only work in bash, whether that's an advantage or not...08:53
enteoptimising shellscripts for speed is futile08:53
slashbeastyou decide what is in header, if it is /bin/bash then you can use everything bash offer.08:54
slashbeastand optimizing for speed or not, just clean code.08:54
enteexpansion doesn't work in :=08:55
slashbeasthm? Gimme example08:56
enteI had to change /usr/ports/{core,opt,xorg,contrib}/* to ${PACKLIST:=/usr/ports/core/* /usr/ports/opt/* /usr/ports/xorg/* /usr/ports/contrib/*} to make the globbing work08:56
Romsterslashbeast, damn wonder why i didn't pick up on that one !08:56
slashbeastRomster: reffering to?08:57
ente:+ too, I don't actually need to assign PACKLIST08:57
enteI tried escaping the {}, but that didn't work either08:57
Romsterfoo=bar; bar=tak; echo ${!foo}08:57
slashbeastRomster: ah08:57
Romsterthan my ugly eval08:58
enteslashbeast: is there a bash trick for that?08:58
slashbeastente: if $PACKLIST is empty use /usr/ports/core/* /usr/ports/opt/*?08:58
entebut expand the *08:58
slashbeastPACKLIST is variable or array?08:58
entean environment variable08:59
slashbeastarray would be a way better08:59
slashbeastfor what you need08:59
enteit wouldn't.08:59
slashbeastyou want space separated list of ports?08:59
enteI can't pass an array via environment variables08:59
slashbeastente: I do not follow, let me show you example08:59
slashbeastente: bashcode with arrays09:00
entefor x in /usr/ports/{core,opt,xorg,contrib}/*; do ... worked fine, for x in ${PACKLIST...} didn't, so I had to expand the {core,opt,...} by hand09:00
Romster what the hell...09:00
slashbeastente: then for x in ${array[@]}; do09:00
Romsteri know the song but on taht video O_O09:01
entewhat I want to do is being able to do PACKLIST=/usr/ports/core/{zip,barf,bleargh} sh BUILD_PACKAGES09:01
enteand environment variables don't work with arrays09:01
slashbeastif you do so, the PACKLIST will be expend before it is moved to script09:01
enteand for-loops over space-separated lists work just fine09:01
slashbeastand stop running sh for bash script09:01
Romsterjsut had to see how flash worked on chromium, and had nothing better to search for than tism..09:01
entebash* then, that doesn't make a difference right now09:01
Romsteri always use /bin/bash for my shebang line when i use bash features.09:02
enteand yes, I want it to be expanded in *THIS* case09:02
enteRomster: it's not +x yet09:02
entestop shedding bikes09:02
ente#!/bin/bash -ex09:02
entehappy now? :P09:02
enteslashbeast: I want to default to compiling all of core, opt, xorg and contrib, except when the environment variable PACKLIST is set09:03
slashbeastente: can you show me your code?09:03
entein which case I want it to compile only what's in PACKLIST09:03
entehah, z8009:04
slashbeastente: it have to be 'PACKLIST=foo bar bash' or it may be 'bash foo bar'?09:04
enteit may be the latter, yeah09:04
entemaybe that's a better idea09:04
slashbeastgimme sec to write it.09:05
enteslashbeast: [ "$#" -eq 0 ] && set -- /usr/ports/{core,opt,xorg,contrib}/*09:06
entefor x; do ... done ;)09:07
entethat works, yep09:08
slashbeastthis is how I would do it.09:08
entethat's long-ish and ugly09:09
Romsteranyone ever seen the movie brainstorm?09:09
enteI just love for loops over lists without the word "in" :)09:09
slashbeastente: you can replace `basename $x` with ${x##*/}09:09
slashbeastand you really should quote variables what you pass to rm.09:10
Romsteroh good lord... bash needs to be bashed for some of it's replacements.09:11
slashbeastRomster: simple string manipulation09:11
Romsteryeah look out if it has a space...09:11
slashbeastRomster: FYI posix sh also support it09:11
Romsterah k09:11
enteslashbeast: why?09:12
enteor what exactly?09:12
enterm -f /var/log/pkgbuild/*.log <- expands to '' if there are no files in pkgbuild/09:12
enterm -f $FAILED $BUILT <- never contain spaces09:13
slashbeastente: one day you will do fuckup like it
slashbeastand then it will be like 'OMGWTF all my porn was removed!'09:14
entewell, luckily at least / is preserved ;P09:15
slashbeastalso that is why I suggested you to use array intead of space separated list in variable09:16
slashbeastas long as IFS=[[:blank:]] something like it may happen09:16
entehappen how?09:24
enteI don't rm any of the ports directories09:25
entebesides, cosmic rays could happen09:25
slashbeastnot strict related to your current code, just to your habbits I presume.09:25
slashbeastif you have two arguments 'foo bar' and 'qwe' and you put it into variable, it will be 3 arguments09:25
slashbeastfor i in $args, for example09:25
enteyes, but about every shellscript in the world doesn't work right with spaces09:26
slashbeastif you use array=( 'foo bar' 'qwe' ) and do for i in "${array[@]}";09:26
entewe don't have any port names containing spaces09:26
slashbeastit will be two arguments09:26
slashbeastente: there is IFS variable to set whatever space or something other is separator09:26
enteand normally, I write posix sh, which luckily doesn't have arrays09:26
slashbeastIFS=:; for i in $PATH; do echo $i; done09:26
Romsternice git commit must refer to that in future.09:27
entebtw, $PATH doesn't work with paths containing colons...09:27
entemkdir -p /usr:/bin and good luck appending it to $PATH09:28
slashbeast$PATH is by default : separated list, shells need it to be like it.09:29
enteI guess about every unix program in the world expects $PATH to be a colon seperated list09:30
ente(those that ever getenv("PATH"), at least)09:31
Romsterexport PATH="$PATH:~/bin"09:39
Romstereasy as09:39
Romsteryou could just make a for loop too append to PATH09:42
enteno, I meant that you can create directories containing colons09:43
slashbeastif ! [[ "${PATH}" =~ "^${HOME}/bin" ]]; then export PATH="${HOME}/bin:${PATH}"09:43
entebut you can't append them to $PATH without breakage09:43
slashbeastI have something like it in zshrc09:43
slashbeastente: about what breakage you are talking?09:44
Romsteri thought : is illegal in file/directory names.09:45
teK_. is09:46
teK_"illegal" ;)09:46
teK_I like confusing people by telling them to create a directory or file named - and deleting it again :->09:46
enteif [ -d ~/bin ] && echo "$PATH" | grep -v ~/bin >/dev/null 2>&1; then09:46
ente        export PATH=~/bin:"$PATH"09:46
enteI'm not sure if ash etc read /etc/profile so I try to keep it sh-compatible09:47
enteusually my ~/bin is ~/.bin though because I have too much crud in ~09:47
slashbeastgrep -v bin?09:47
enteteK_: . is illegal?09:47
slashbeastwill strip /bin09:47
enteslashbeast: hm?09:48
teK_ente: well it already exists09:48
teK_so it's kinda illegal09:48
enteteK_: ah, I thought you meant files containing .09:48
entethe only two characters that are illegal inside filenames on UNIX are \0 and /09:49
slashbeastente: instead of forking another bash (pipe) and using grep, you can use string manipulation to remove it09:49
ente. and .. are perfectly legal, although special09:49
slashbeastbash is so dumb, try rm -rf .*09:49
enteyou can make ... if you want to :P09:49
slashbeastand then find that .* is also parent dir09:49
slashbeastso ../*09:49
enteI think rm excludes . and .. per default09:50
slashbeastrm is bulitin09:50
slashbeastas root, go to /home/joe and do rm -rf .*09:50
slashbeastit will wipe your /home09:50
entecrux ~# type rm09:50
enterm is /bin/rm09:50
entecrux ~# echo $SHELL09:50
enteand no, rm checks for . and ..09:51
slashbeastrm suppose to be bulitin09:51
enteotherwise rm -rf .* would recursively remove your entire filesystem09:51
entebecause .. contains a ../..09:51
entewhich contains a ../../..09:51
slashbeast$ rm .* -rf09:52
slashbeastrm: cannot remove directory: `.'09:52
slashbeastrm: cannot remove directory: `..'09:52
slashbeastwhen it was added?09:52
slashbeastI've remember when it worked, proapbly the when rm -rf / was hardened09:52
enteI think your memory is wrong09:52
enterm must never touch neither . nor ..09:53
slashbeastlet me google then.09:53
entebecause there's always a ./.. and ../..09:53
enteso it must exclude these, at least for the second level09:54
Romster. is current directory inode and .. is parent directory that holes the inode for it's sibling... or osme thing along those lines09:54
slashbeastbefore rm was hardened .* was expended to .. and it worked, I will just find when it was changed09:54
slashbeastlike rm -rf / no longer works, you need rm -rf /*09:54
entei.e. . and .. *not* passed via argv must be left out in any case09:54
enteand rm -rf / is that --preserve-root thing, which is iirc entirely unrelated09:55
enteslashbeast: do you have a source?09:55
entewe're just guessing around on this09:55
slashbeastente: searching.09:55
enteprohibiting rm -rf / is a dumb thing to do09:55
enteyou can still rm -rf /usr09:55
enteor rm -rf ~ /.mozilla09:56
slashbeastor rm /*09:56
teK_find / -delete09:56
enteand having special cases is something a java programmer would do09:56
Romsterwasn't there a unlink command too09:56
entenot somebody who's halfway sane in their head09:56
enteRomster: yes09:56
entethe difference is that unlink only works on files09:57
enteoh, and character devices09:57
enteapparently just not on directories09:58
entethat's due to the behaviour of unlink(2)09:58
entethe syscall behind unlink(1)09:58
entethere's also a rmdir(1) wrapper around the rmdir(2) syscall09:59
enteRomster: gcc34 fails09:59
Romsterdamn it09:59
Romsterit used to work who broke it10:00
ente403, I know10:00
entereload :)10:00
enteI'll get you that config.log10:01
Romsterwonder if anything even needs gcc34 anymore10:01
entegcc 3.4 and gcc 2.95 are portable10:02
entethey miss a bit c99, but some architectures are better supported with older gccs than current ones10:02
entealso, GNU decided the gcc project may use C++ a few releases ago10:02
entewhich again limits portability10:03
Romsteri've never used 2.95 in ages but it's been the past few years i tink that i've used 3.410:03
entebesides, 3.4 and 2.95 are also faster10:03
entesabotage uses them for bootstrapping a musl-based system10:03
Romstercompiler is faster as it don't optimize the built objects as much10:04
enteyet 2.95 sometimes produces smaller code than 4.x :)10:04
enteor even faster, but that's hard to measure10:04
*** lasso|qt has quit IRC10:05
entealso, some people need older gcc releases to test if their software compiles with them, for backwards compatibility reasons10:06
enteiirc windows ships gcc 3 or 2 in their "MS Services for UNIX" package10:06
enteiirc, the same goes for MINGW, but I might be wrong10:07
Romsterbuilding gcc34 now in chroot i'll check it later, gonna go sleep now another madness day left before end of financial year then the start of sped tax money mania...10:07
entesolaris used to use some older gcc release for some time, but that might be wrong too, since I never used solaris10:07
Romstermingw can use far more recent gcc builds.10:07
Romsteri sued to package mingw but haven't lately.10:07
enteoh, damn, I wanted to make a remark ;P10:08
Romsterhehe i don't know why but i type used wrong too damn often. i'm glad i'm not a lawyer10:08
Romsterah that runtest thing is breaking it..10:09
Romsterlook into it alter g'night.10:10
entesleep well :)10:10
*** j^2 has joined #crux10:46
*** j^2 has quit IRC10:47
*** j^2 has joined #crux10:48
*** j^2 has quit IRC10:49
*** j^2 has joined #crux10:49
*** lunarvalleys has joined #crux11:09
enteRomster: seems all gcc34 is missing is unset MAKEFLAGS before configure11:21
enteRomster: but it's not done yet11:21
*** pitillo has joined #crux11:26
*** jse has joined #crux11:40
*** Rotwang has quit IRC11:50
enteRomster: argh11:50
enteRomster: <- it fails in libstdc++11:51
*** Rotwang has joined #crux11:51
*** deus_ex has quit IRC11:57
*** deus_ex has joined #crux12:11
*** lowe_ has joined #crux12:23
ente[19:26] < aep> i love how nm disables internet when you connect an ubuntu machine to our intranet, for security reasons12:27
ente[19:51] <ente> main(a,b)char**b;{kill(atoi(b[2]),atoi(b[1]));}12:58
Rotwanggawk 4.0.013:02
Rotwangente: is this c code?13:03
teK_Rotwang:  pidgin 2.9.0 is out, fixing a security vuln13:03
RotwangDifferences from gawk 3.1.8 are not available; they would be too large. -__-13:03
*** SiFuh has quit IRC13:04
Rotwangente: doesn't look like13:04
enteK&R C13:04
Rotwangbut I'm not a c expert13:04
RotwangteK_: thanks, will have to look into it13:05
teK_version bump, I guess13:05
*** SiFuh has joined #crux13:07
*** SiFuh has quit IRC13:18
Rotwangwhat was the best way to credit someone?13:27
Rotwangin a git commit13:27
teK_commit message I guess13:27
ente(thanks Rotwang)?13:27
Ovim-ObscurumteK_: Already made a Port for the ncurses-only version finch13:27
Rotwangno, no no13:28
* ente throws a bucket of rocks at Ovim-Obscurum13:28
Ovim-ObscurumHuh oO13:28
Ovim-ObscurumWhat's wrong, ente?13:28
entefinch horribru.13:28
RotwangteK_: tilman was very cross when using commit message that way ;f13:28
Ovim-ObscurumI like it, using it for many years.13:29
Ovim-ObscurumWorks pretty fine.13:29
teK_Rotwang: I do that all the time13:29
entemaybe finch works13:29
teK_Ovim-Obscurum: I'm using bitlbee + irssi13:29
*** SiFuh_ has joined #crux13:29
entebut it certainly does not work fine...13:29
Ovim-Obscurumente: Why exactly? I can't understand you.13:29
enteI used to be on bitlbee, I have irssi-xmpp nowadays13:29
teK_i use its icq and twitter transport, too13:30
Ovim-ObscurumteK_: teK_ Bitlbee? That's a gateway using irc, isn'tit?13:30
Rotwangfinch is nice imo13:30
entebecause it is a totally horrible curses frontend to a totally horrible IM client13:30
entedoes it still do this weird windowmanagement?13:30
Ovim-Obscurumente: Come on, that can't be your earnest.13:30
tilmanRotwang: "cross"?13:31
tilmanRotwang: do you mean gross? (though that doesn't seem to fit the context either :P)13:31
Ovim-Obscurumente: Shame on me.13:31
cruxbot[contrib.git/2.7]: lxpanel: fix for #72613:31
enteOvim-Obscurum: das war der denglischste satz den ich seit langem gehört habe13:32
teK_Ovim-Obscurum: si.13:33
Ovim-ObscurumJa, ich hätte Vokabeln lernen sollen :(13:33
teK_Wieso, Deutsch ist doch auch eine schoene Sprache13:33
Ovim-ObscurumteK_: I tried several IRC servers with transports and had many problems with connectivity - that's why I had chosen a "real" IM. :)13:33
Ovim-ObscurumAh noch ein Deutscher! :D13:33
enteau?er man hat keine Xmodmap oder ein schei? encoding13:33
Ovim-Obscurumente: I have to fix that, I know.13:33
teK_ente: das kommt darauf an, ob man tatsächlich kein setmap hat oder halt doch.13:34
Rotwangtilman: nr. 413:34
entemeine freundin hält mir immer vor wie grässlich deutsch ist13:34
teK_wo kommt die her13:35
Ovim-ObscurumteK_: Bestimmt Ostdeutschland.13:35
entealso.. muttersprache ist russisch, aber im moment lernt sie norwegisch13:35
* Rotwang has almost finished watching whole allo allo tv series13:35
tilmanRotwang: finch ftw, btw!13:37
Ovim-Obscurumtilman: +113:37
cruxbot[contrib.git/2.7]: lxpanel: fix for #726 and fixed URL13:40
cruxbot[contrib.git/2.7]: pidgin: update to 2.9.014:18
entelibpurple-minimal 2.8.0-114:22
enteguess I could update libpurple :p14:22
cruxbot[contrib.git/2.7]: filezilla: update to 3.5.014:30
*** lasso has joined #crux14:39
cruxbot[contrib.git/2.7]: medit: updated to 1.0.314:53
cruxbot[contrib.git/2.7]: mtpaint: fixed sourceforge source url15:00
Rotwangis it only me that there is dd-mm-yy on the flyspray where date should be?15:07
RotwangOpened by Ian Denhardt (zenhack) - dd-mm-yy15:10
teK_Opened by Ian Denhardt (zenhack) - Sunday, 07 November 2010, 22:48 GMT+215:11
teK_this line?15:11
Rotwangok, my setup is fucked then15:11
teK_what locale does your browser send15:12
Rotwangindeed it was15:12
teK_less work for me ;)15:12
enteteK_: I think the font used for the report should be a monospaced one though ;P15:14
teK_which report?15:14
entethe "Details" part of any bug reported on the bugtracker15:14
teK_not getting the joke15:14
teK_is there one at all?15:14
enteteK_: yeah, the actual text of the report15:16
enteyou know, the actual bug report, not the subject line or the dropdown stuff on the left15:16
entesorry, I suck at describing things15:16
teK_ich verstehs nicht ;)15:16
teK_Moritz hat heute ueber deinen Nick gelacht uebrigens15:17
RotwangI understood15:17
Rotwangbut I don't cate what kind of font it is15:17
teK_er kennt naemlich ente15:17
entewelcher Moritz?15:18
teK_mein Sohn.15:18
teK_hab ich dir damals erzaehlt, dass der heisst wie du ;)15:18
enteah :)15:18
enteich vergesse auch ziemlich schnell sachen wieder :(15:18
entewie alt ist er?15:19
teK_I think it's impolite to speak non-english in here :]15:19
enteyeah, me too :)15:19
* Rotwang doesn't mind15:20
enteteK_: of this report this part should be monospace rather than sans-serif I think15:21
entebut it's not really that important15:21
teK_yeah, but why? :)15:21
entebecause sometimes the text comes from a terminal and it looks all wrong15:21
entestop making me think I have OCD :(15:22
teK_ok but is there a <code> Tag?15:22
entethanks :)15:22
enteI guess I just spent so much time in terminals, all kind of non-monospaced text looks wrong15:23
ente*browser settings*15:23
enteunfortunately I can't configure what "sans-serif" expands to :S15:26
entestupid browser15:26
Rotwangsome bugs at the flyspray are really old15:29
enteyeah, and the new ones get fixed immediately15:31
entemostly those are assigned to someone who doesn't care about crux anymore15:32
enteor some nontrivial stuff15:32
enteseems nobody tried my packages yet :(15:32
Rotwangente: I'd play around with them, but I have not much time on my hand15:35
enteit's ok :)15:35
enteI'd play around with them too, but my only crux box is at my parents' place15:35
enteI wanted to post something about pkg-get, but I'm too lazy15:37
entethe dependency that's most often missing seems to be mesa3d btw15:38
enteall kinds of stuff depending on GL/gl.h without saying so in autoconf15:38
entewhat's the point of autoconf then?15:39
*** j^2 has quit IRC16:06
*** lasso has quit IRC16:12
*** aubic has joined #crux16:45
entedillo depends on xorg-makedepend16:47
*** lunarvalleys has quit IRC17:00
*** Rotwang has quit IRC17:09
*** Dudde has quit IRC17:26
*** cruxbot has quit IRC17:27
*** cruxbot has joined #crux17:30
*** Evil_Bob has quit IRC17:51
*** Ovim-Obscurum has quit IRC18:25
*** bandito85 has quit IRC18:32
enteis simon glossner still active?18:38
*** Ovim-Obscurum has joined #crux18:39
enteseems his last commits to contrib were in '0918:39
entecan I take care of his ports without adopting them?18:39
entep5-class-std is a bit outdated and the footprint mismatches18:40
enteRomster: I get a 403 from p5-date-manip18:41
teK_so you'd bump the version without picking it up?18:42
enteteK_: I'd at least fix the footprint18:42
entebut anyway, I might just pick it up18:42
enteit's just that I don't use it18:42
teK_best solution ;)18:42
enteprobably :)18:43
enteuh.. weird18:43
enteprt-get info p5-io-multiplex18:43
entesmells like copypaste :D18:43
enteis a new perl something that's supposed to wait until the next release?18:44
teK_perl6? :>18:44
entebecause 5.14 was released a few days ago18:44
entereleased 2011-06-1718:46
entemore than just a few days, but still.. :)18:46
teK_dunno tbh18:46
teK_but I'd say no there's no need to wait until the next release18:46
cruxbot[contrib.git/2.7]: p5-io-multiplex: adopted; 1.10 -> 1.1319:24
enteok :)19:27
enteI wonder if people actually use them, or if they get them straight from cpan19:27
cruxbot[contrib.git/2.7]: p5-class-std: adopted; 0.0.8 -> 0.01119:35
cruxbot[contrib.git/2.7]: p5-class-std: mixed up Packager and Maintainer...19:38
enteI guess that's enough for today :P19:39
teK_brave ente19:39
enteich sollte besser mein usb_modeswitch fixen anstatt anderer leute zeug das noch funktioniert :)19:43
enteda geht der download link nicht mehr weil es outdated ist19:43
entesowas gehört verboten :<19:43
teK_ja mach das Mal19:43
teK_und ich geh schlafen19:43
teK_habe die Ehre19:43
entehab keine möglichkeit es zu testen19:43
enteich gehe wohl eher spülen :P19:43
enteteK_: gute nacht19:43
teK_gn8 ;)19:43
*** RyoS has quit IRC20:00
*** RyoS has joined #crux20:00
cruxbot[contrib.git/2.7]: p5-date-manip: 6.23 -> 6.2421:08
Romsterente, fixed21:09
enteyou're fast :)21:10
Romsteri try to be.21:12
Romstergcc34 will have to wait for the weekend.21:12
Romsterdoing another run of it's crude but works mostly.21:15
enteI wonder if pkgadd restores the file owner from packages21:18
entethat would simplify binary packages21:18
entei.e. my user fcron has the uid 1234 and yours has 23, so the tarball contains "fcron" and "1234" as uid, and pkgadd would first try resolving "fcron"21:19
entefind out it's 23 and apply that21:19
enteapparently, that's what bsdtar and gnutar do, but I don't know if pkgadd does it too21:19
enteRomster: what does it do? search for dead http links?21:20
enteanyway, off to bad21:21
*** Dudde has joined #crux21:25
Romsterit lists the return http code including 40421:35
cruxbot[contrib.git/2.7]: taglib: 1.6.3 -> 1.721:56
Romstersadly it don't work on google or sourceforge yet22:07
Romsterback to work later22:07

Generated by 2.11.0 by Marius Gedminas - find it at!