IRC Logs for #crux Thursday, 2008-05-08

Romsterand the channel is dead...00:05
*** pitillo has joined #crux00:08
*** Rotwang has quit IRC00:19
Romsterhi pitillo00:32
*** mxs has quit IRC00:43
pitillogood morning01:08
pitilloyo Romster :)01:08
Romster:)01:09
Romster4pm here but good morning to you :)01:09
pitilloand good afternoon for you :P01:09
Romsterslowly working on contrib improvements to my ports if anyone is wondering why i haven't pushed anything for a few days.01:10
pitillooh nice, sounds interesting01:10
Romstermoving symlinks for ccache/distcc to post-{install,remove} and a couple of port updates.01:11
Romstersince jue bitched about that.01:11
Romsteri didn't see a problem but i'm doing it anyways.01:11
tilmanwhat symlinks?01:12
Romsterthe ones for /usr/lib/{distcc,ccache}01:12
cptncreating symlinks in build() depending on whether ccache/distcc is installed01:12
Romsterjue says there anoying and should be in a post-install.01:12
Romsterso that's what i'm doing.01:13
tilmani see01:13
cptnactually, predatorfreak said they should be a post-install01:13
Romstermade a nice function that does it all and a post-remove too.01:13
cptnjue just said they're annoying01:13
Romsterhmm right.01:13
Romsteryeah and sepen and i think others agreed on that.01:13
Romsteri'm nearly done with it then i can just copy it to the other directories of my ports and edit a few lines.01:14
Romsteron another note should 'locate *.egg-info' files be added as a file to bad files to prtverify and removed from Pkgfiles?01:16
cptnlocate works on a cache01:16
Romstersince it's setuptools that will use them to update them but we have pkgutils todo that for them.01:16
cptnso that would probably not help a lot, would it?01:16
Romsteryes i know.01:16
Romsteri was just posting the command to run to list them.01:17
cptnokay01:17
Romsterhttp://rafb.net/p/UWUhKP50.html01:17
cptnsounds definitely reasonable01:17
Romsteroutput01:17
cptnalthough this might break "egg" installs01:18
cptni.e. if a user uses CRUX' python port, then wants to install something via egg01:18
Romsteronly exception maybe setuputils itself.01:18
Romsterwell if it's installed by pkgutils pkgutils should be maintaining the updates, if there manually installed as python eggs, then the *.egg-info files will be installed on the system and not managed by pkgutils.01:19
Romstersetuputils would still work the same as it does now.01:19
tilmanan egg contains little pythons?01:20
cptnbut don't additional egg stuff need to know what's provided by the system?01:20
Romsterjust python ports that are packaged with pkgutils shouldn't be allowed to be updated by setuptools.01:20
tilmanis it something like gems for ruby? :p01:20
tilmann/, the page loaded now01:20
Romsterhmm not sure i noticed nipul removed it from one of his ports and that raised my idea.01:21
cptntilman: yeah01:21
Romsteryes it's like gems01:21
Romsterpython eggs and ruby gems01:21
Romsterit was just something on my mind i wanted to point out and get a opinion on.01:22
cptnerr01:22
cptnnot sure how this happened :-)01:22
Romstera lot of non-python people didn't see it i guess.01:23
cptnI'm not that much of a python person so I wouldn't know whether anyone would ever want to install eggs instead of a crux packaged version01:23
cptnI know that for some ruby stuff, it's a lot more effort to port it as non-gem01:23
*** namenlos has joined #crux01:23
cptnat least it was back then :-)01:23
Romsteri was aware but i didn't think about it much until i saw nipal removed  the egg file from one of his ports.01:24
cptnso assuming I wanted to install a pytho egg now01:24
cptnwhich would depend on setuputils01:24
cptncould I install it if my setuputils had its egg removed01:25
Romsteri know prologic swearers by python eggs but i prefer to have pkgutils handle it all.01:25
cptni.e. does removing egg-info files break egg?01:25
Romstercptn, haven't tested that, that could be white listed and be excepted from the removing of egg file if that is the case, i suspect it's to update itself from a egg though. but that's not been tested.01:26
cptnI suspect it's for dependency management01:26
Romsteri do know removing egg files from python ports that pkgutils manages shouldn't break anything01:26
Romsterthat's possible too. not sure how it tracks that.01:27
cptnwell, that was my question01:27
Romsteri didn't look that far into it.01:27
cptnso if you _know_ that it doesn't break anything, they can certainly go away01:27
Romsterbut i suppose i should, but i'm busy right now.01:27
cptnwell, you just said that01:28
cptn <Romster> i do know removing egg files from python ports that pkgutils01:28
cptn                  manages shouldn't break anything01:28
Romsteri'll do some testing on my theory.01:28
Romsterthat part is correct, but not sure on the removal of the egg file from setuputils itself would break setuputils.01:28
cptndon't think it would01:29
cptnbut it would break installing eggs depending on setuputils01:29
Romsteryou think it's to track i think it's to update setuputils from itself.01:29
cptnhttp://peak.telecommunity.com/DevCenter/PythonEggs01:29
cptnsetuptools allows you to declare dependencies in your project's setup script, so that they will be bundled inside the egg's metadata directory, and both the runtime and EasyInstall can then automatically find the additional eggs needed, adding them to sys.path when your project is installed or requested at runtime via require()01:30
Romstermy idea was to leave setuputils as is and remove egg files from all .footprints01:30
Romsterhmm so removing them would break other stuff...01:30
Romsterthere goes my nice idea..01:31
cptnshould be simple to verify though01:31
Romstermy theory was setuputils shouldn't manage ports that pkgutils is managing.01:32
cptnfully agreed01:32
Romsterthat's as far as i got.01:32
Romsterbut eggs could still be installed if the user so wished to do so.01:33
cptnI think one can argue that CRUX users should not use egg to install python stuff01:33
Romsteri prefer to use everything with pkgutils personally, but others have there own opinion of course.01:33
tilmanunless egg installs stuff to /usr/local/lib/python*01:33
cptnI think if you remove for eaxmple pygame's egg-info, and then try to install a game via egg, egg will not know pygame  is already there01:34
Romsternot sure what saturation is with ruby gems.01:34
cptnand try to reinstall the pygame egg01:34
cptnmaybe they could be ignored via pkgadd.conf by default01:35
Romstercptn, yeah it would break setuputils from installing eggs..01:35
cptnRomster: why don't you bring it up on the mailing list?01:35
Romsteryeah i will post it. i just remembered about it now.01:36
cptnmaybe there's someone who has worked with python eggs already01:36
Romsterwas a semi-bring to attention thing.01:36
Romsterbefore ML posting.01:36
Romstermaybe we could remove egg files and if a user wants to use python eggs they can avoid the pkgutils versions and use eggs in /usr/local/lib/python*01:39
Romstertilman, although we have a /opt directory wouldn't that be better than a local directory?01:39
cptnthat said, my setuputils egg-info is 32k01:40
tilmanRomster: dunno, imo /opt sucks ;)01:40
Romsterthen why do we have a /opt at all <<01:41
tilmanmaybe because i haven't grabbed control over everything in crux yet?01:42
tilmanor maybe because i think others might think otherwise :p01:42
Romster<_<01:43
cptnhttp://fedoraproject.org/wiki/Packaging/Python/Eggs01:43
* tilman runs to teh lecture01:43
cptn"Since eggs establish a base of functionality that upstream authors can expect, we need to be sure to include the egg files if a package builds them. Starting with Fedora 9 any package that uses setuptools or distutils will build egg-info."01:44
cptnthat's what they say01:44
Romsterlocal was meant to be local files and the rest for system wide like from a server, seems a lot of programs default to this but we force --prefix=/usr01:45
Romsterhmm01:46
cptnso maybe there could be a policy to not remove them01:46
Romsterpycrypto and twisted remove them.01:47
Romsterwhich made me think about that.01:47
Romstermaybe, i just think to me at least they are a waste since i prefer to use pkgutils to handle it all.01:48
Romsterand having more than one package manager fight over a program is silly IMO.01:48
Romsterthen again maybe we all should be using python eggs and ruby gems and not pkgutils for them.01:49
cptnI agree with that01:49
Romsterbut i don't like that idea.01:49
Romsteri'll throw it on the ML soon and let others have there say.01:50
cptnmmmh, okay01:50
Romsterjust discussing it here has raised some other points about it.01:51
cptnIn any case, it would be nice to have a recommendation01:51
Romsteryeah that's what i would like so we all know what is the best action to take.01:52
Romsteri'm all for consistency.01:52
Romstercptn, glad you are back too.01:53
cptnhey, what else should I do with my time? :-)01:53
Romsteroften wondered what kept you away from crux :)01:53
Romstertrue.02:06
pitillocan someone take a look to this and tell me if sounds reasonable to have a build like it? (I mean to provide a port wich builds some stuff inside instead of using 1 port for each) http://lokalix.dyndns.org/crux/e17/e17-snapshot/Pkgfile02:15
*** sepen has joined #crux02:19
sepenmorning02:20
Romsterpitillo, not sure i agree with wvstreams having both programs together. but that i dunno.02:24
Romsterprivate repo i guess it's fine but i wouldn't do that for contrib or anything else.02:24
Romsterhi sepen02:25
sepenhey Romster02:25
pitilloRomster, what does wvstrams mean? (the fact of having libs and the program itself in the same pkgfile?)02:27
pitillocan be better split it in some snapshots-ports?02:28
Romsterpitillo, was wvstreams and wvdial in the same Pkgfile since both come from the same project and nothing else uses wvstreams02:28
Romstersplit into snapshot ports would be my prefered method.02:29
pitillooki, then split it is the way to go02:29
pitillooki, thank you Romster02:29
Romsterdon't think it's much more work other than editing the version tags and pkgmk'ing a new version.02:29
Romsteryou don't need to have them all installed to work?02:30
pitillowell, I like the idea of provide both ways (snapshots and cvs builds)02:30
pitilloRomster, all must be installed in that order to work02:31
Romstersnapshots are builds of cvs/svn/git/...02:31
pitilloyes02:31
Romsterlike i have done with x26402:31
Romsterso it requires all to work.02:31
pitillobut they are specified versions/dates provided by the packager/maintainer, and I like to let the user check the last version too, if he wants02:32
Romsterhmm *thinking*02:32
Romsterhaving them seperate is easier to patch one if you have too or are they locked to the same date/revision.02:32
pitilloI had a little discussion with sepe_n about this, and the first problem I can see is the way of providing the snapshots02:33
Romsteri run a script todo that.02:33
Romster./generate-<portname>-snapshot02:33
pitilloRomster, they are locked to the same version. I made a little script to make that snapshots when I find a good version working here02:33
pitilloRomster, yean, same here02:33
Romsterthat fetches and tarballs the source.02:33
pitilloand uploads them to a target02:33
Romsteryou don't have a stable set of ports as well?02:34
pitilloummm well, I think ths snapshots must be the stable version of them.02:34
pitillo(at least stable because here are working fine when they were tested)02:35
pitillobut this case may be can be different between users, and updating to the last cvs version can solve problems for them (in the case they have problems)02:35
Romsteronly reason i think separate ports is use is because of existing ports are separate.02:36
Romsterbut have e17-<port>-snapshot, maybe.02:36
Romsterso you see e17-<port> and e17-<port>-snapshot02:37
Romsterif there was no existing e17-<port> i'd be very inclined to keep the one Pkgfile.02:38
pitillowell, I will think about this and posibly I will rename all ports, letting the normal name (e17 for example) with snapshots and the cvs ports with a little tag in their names (e17-cvs)02:38
Romsterthese are my opinions though :)02:38
Romsteror keep e17-<port> stuff in a seperate repo for snapshots.02:39
pitillothank you for comment this wth me02:39
Romsteris another option so they keep the same names.02:39
Romstersometimes it's hard to choose and decide.02:39
pitilloummm Iprefer use the same repo for both versions, snapshots and cvs02:39
Romsteri can share with you my generate snapshot script02:39
pitillowell, mine is working perfect here, I can take a look to it to see if I can get ideas for mine02:40
Romsterah right so e17 calls them all snapshots, that's confusing as cvs are snapshots...02:40
Romsterusually it's a release and a snapshot.02:40
pitilloRomster, no no, cvs versions are really a checkout of cvs repository (last revision there)02:41
Romsterah cvs head then.02:41
Romsterwhich could be broken.02:41
pitillocheck e17 and e17-snapshots ports  in the same url I posted before if you can02:41
pitilloRomster, yeah and more on E, but sometimes it's good to keep up-to-date with head's cvs02:41
pitillo(cvs' head)02:42
Romsterah i see... they don't provide tarballs yet.02:42
Romsteri much prefer the snapshots.02:42
pitilloyeah, the only port which provides the tarball is the e17-snapshot02:43
pitillo(also provides tarballs for all deps)02:43
Romsterand test head from updating the snapshots and if it fails just do a revert02:43
Romstere17 and e17-snapshot02:43
Romsteri'd drop both and do e17 as a  'e17' snapshot port.02:44
pitilloyeah, that is what I mean. And make another e17-cvs port02:44
Romsteryeah keep e17 as a snapshot and e17-cvs02:45
pitilloand keep e17 port with snapshots02:45
pitilloyes yes02:45
Romstere17-testing seems a better name though.02:45
pitilloperfect. Thank you for your comments02:45
pitilloummm I prefer cvs over testing, more easy to understand from my point of view (may be testing can confuse people)02:46
Romsterk02:46
pitilloI will try along this morning to make some changes there02:46
Romsterone of my generate snapshot scripts http://rafb.net/p/beScZ473.html02:46
Romsteri run one for each snapshot port i have.02:46
Romstergenerate-libdvdnav-snapshot.sh generate-tidy-snapshot.sh etc..02:47
Romsterin some cases i redo autoconf and stuff on the snapshot to keep the Pkgfile cleaner.02:48
Romsterso it's to the normal ./configure make routine.02:48
Romsterhope all this helps02:48
pitillowell, I am using this http://lokalix.dyndns.org/tmp/e17-snap02:49
Romsterah nearly what i'm doing now.02:50
pitilloyep02:50
Romsterexcept you made it handle all the modules.02:50
pitilloa bit more ugly...02:50
pitilloyes, and I try to update (wich has not much sense using /tmp atm, but I am thinking in change this to another location)02:51
Romstermight pay to compress it as one archive?02:51
Romsteri would edit to save to PKGMK_SOURCE_DIR02:52
Romster. /etc/pkgmk.conf02:52
Romsterecho $PKGMK_SOURCE_DIR02:52
Romsterthen just edit the Pkgfile and rebuild.02:53
Romsterof course upload the archive still as you do.02:53
Romsternow for me to attach more shell. nearly got this done but it's giving me a good fight <<02:54
pitillowell, time to take a bite here... my stomach is talking.... or better said, roaring...02:55
Romsteri'd choose grumbling <<02:56
Romstereither works02:56
*** namenlos has quit IRC03:01
*** namenlos has joined #crux03:11
pitillomine doesn't grumb, directly roars xD03:21
Romsterhehe03:21
Romsterand sweet one file done one to go.03:21
Romsterfew hours i can git push all this.03:22
Romstermaintaining ports can be hard work.03:22
*** lennart has joined #crux03:23
pitillowell, depends in how many you maintain... in your case, enought to go crazy... (like others maintainers)03:23
Romsterheh03:27
Romstersomeone has to. :P03:27
pitillotrue03:29
*** mike_k has joined #crux03:31
*** namenlos has quit IRC03:33
*** namenlos has joined #crux03:36
*** schniggie has quit IRC03:42
*** RedShift has joined #crux03:48
Romsteractually a stats of how many ports aperson has be interesting and the percentage of total for core opt contrib, no point for xorg tilman has that covered <<03:59
pitilloumm for core hasn't much sense while maintainers are moving to core-team, but sounds interesting for opt/contrib04:00
Romsteroh yes i forgot about that...04:00
pitillolike you said... about xorg... the master rules04:00
Romsterwell opt and contrib would be nice.04:00
Romsteri might do one sometime.04:00
Romsterwhen i have nothing else better todo <<04:00
pitillowell,, a  little script to grep maintainers line and wc -l them can be a good start point (I think someone can do it better)04:01
Romsteryeah something like that is what i had in mind but present it on a html page.04:01
Romsterin the end i want to be able to get what ever stats from Pkgfiles directly with a server side web page04:02
Romsterdirectly without having to regenerate it.04:02
Romsterone day i'll do that.04:03
pitillowell, I hope sepen can read this and if he has time (which is quite hard too) he can make a firsst aproach for the script and the webpage04:03
sepenhmm reading now, Im too much work today at office04:04
pitillos/Im/I have04:04
sepenohh thanks04:04
sepen*I've04:05
pitilloyeah04:05
pitilloI have too much work here too, but I haven't the desire to work...04:05
pitilloxD04:05
*** rxi has joined #crux04:08
Romsterpitillo, that happens with me, plenty todo jsut don't always want to do it04:12
Romsterjust*04:12
pitilloxD04:13
pitilloI mean here at work.... :P Making time while waiting a cisco employer which will help me a bit with a cisco catalyst switch stack04:13
Romsterah04:17
Romster'help me a bit' = assist :)04:18
pitilloyeah :)04:19
cruxbot[contrib.git]: wesnoth-server: initial commit04:20
cruxbot[contrib.git]: distcc: moved symlinks out of Pkgfile04:20
cruxbot[contrib.git]: ntfs-3g: 1.2412 -> 1.250604:20
cruxbot[contrib.git]: gcc34: add post scripts for ccache and distcc symlinks04:20
cruxbot[contrib.git]: wesnoth: 1.4.1 -> 1.4.204:20
cruxbot[contrib.git]: wesnoth-server: 1.4.1 -> 1.4.204:20
cruxbot[contrib.git]: vobcopy: fix source url04:20
Romsterstill more to do04:21
*** Rotwang has joined #crux04:39
*** onestep has joined #crux04:54
nipuLis ck4up going nuts for anyone else?05:06
nipuLerror: Hostname not known: ftp.gnome.org05:06
nipuLi get 404's or domain not found all the time05:06
*** mike_ has joined #crux05:09
pitillonipuL, have you tried to ping it?05:09
RomsternipuL, nope..05:10
Romsteri used too but my isp dns servers have gone to shit so i switched to OpenDNS in my maradns config.05:10
*** mike_k has quit IRC05:11
Romsteryou may want to lower your number of threads in the ck4up file.05:11
Romsteror setup QoS or both.05:11
*** jtnl has joined #crux05:14
cruxbot[contrib.git]: gpm: 1.20.1 -> 1.20.305:16
rxinooo .. westnet is selling out to iinet05:56
nipuLs/is selling/has sold/06:03
nipuLjust as i switched to WN too06:04
rxilol06:04
rxilike them?06:04
nipuLthe good news is they are ditching the optus links and going back to telstra06:04
nipuLmight finally get some decent OS bandwidth06:04
rxioverseas bandwidth?06:05
nipuLyeah06:05
nipuLi can't get over 70KB/s06:05
nipuLeven to singapore06:05
*** Rotwang has quit IRC06:05
nipuLwhich is why i ditched comcen, only to find westnet is having the same problem06:05
rxiadsl2+?06:06
nipuLthat said, westnet's customer service is outstanding06:06
nipuL8mb adsl106:06
rxii never had any problems with my 8mb at nambucca06:07
nipuLhow long ago?06:08
nipuLi used to get great speeds until about 6 months ago06:08
rxitil i moved about 3 months ago06:08
rxion adsl2+ on an optus dslam and getting over 70k06:09
rxithey use telstra dslams for adsl106:10
nipuL 4  vlan497.52gdc76f02.optus.net.au (59.154.21.89)  814.159 ms  637.828 ms  608.982 ms06:11
nipuLfail06:11
rxii hit that router at 42ms06:12
nipuLwestnet was with telstra, then they jumped over to optus to get cheaper bandwidth06:12
nipuLi'm doing other stuff too06:12
nipuLbut it's become obvious that optus doesn't have enough bandwidth, so it's back to telstra they go06:13
nipuLthe switch is due for completetion tomorrow, or so i've been told by a tech from westnet06:13
rxilol06:13
Romster204 ms for me to vlan497.52gdc76f02.optus.net.au06:16
mike_tilman: do you think "php-extension-foo" should depend on "php"? strictly speaking it is not dependent, but otherwise useless.06:17
Romsterwhat annoys me the most is i get about 27.873 ms to  4  VLAN309.o3mlc76f05.optus.net.au, then i get a slower 5  203.208.145.154 (203.208.145.154)  187.356 ms06:18
tilmanmike_: i think it should depend on php06:19
Romster30ms is about average i get to my isp then anything up to 600ms or so overseas for the really slow locations06:19
Romstermike_, i'd go along with tilman on that one it should depend on php so it's functional.06:20
mike_tilman: thanks. I'll adjust few ports I am doing now. jue just probably don't care =)06:20
Romsterdunno jue seems to care to bitch at my ports. so i figure jue does care.06:21
Romstermaybe jue likes to be a perfectionist, more so than me.06:22
*** jtnl has quit IRC06:22
*** jdolan_ has quit IRC06:23
mike_anyway we can force him to add that three letters to php-mysql/Pkgfile06:23
*** onestep has quit IRC06:23
mike_hmm, I guess updating php port without updating extensions might give unexpected results somewhen06:26
* mike_ doesn't want to be punished06:27
mike_I hope, someone will update opt/pear to current version. It looks somewhat tricky. There are few usefull pear extensions around.06:37
*** joacim has joined #crux06:40
nipuLrxi: can you tell me how fast this file goes for you? http://static.die.net.au/crux/snapshots/emacs.tar.bz206:43
rxi~315KB/s06:44
nipuLthat's not fair06:44
rxilol06:44
nipuLis that through optus?06:44
rxiyeah06:45
nipuLwtf06:45
nipuL 68.9K/s for me06:45
*** destruct_ is now known as destruct06:46
rxithats not on your westnet ip06:47
rxi?06:47
nipuLno06:47
nipuLthat's my vps06:47
mwansawohooo 234K/s06:47
mwansathrough optus06:47
nipuLit's on san jose i think06:47
nipuLin06:47
rxiahh06:47
nipuLone of the he data centers anyway06:47
rxido you go out through singtel on that?06:48
nipuLyeah06:48
cruxbot[contrib.git]: distcc: minor improvements to post-install and post-remove06:52
rehabdoll~330kb/sec06:57
cruxbot[contrib.git]: distcc: minor improvements to post-install07:06
cruxbot[contrib.git]: gcc34: minor improvements to post-install07:06
cruxbot[contrib.git]: gcj: add post-install and post-remove for ccache and distcc07:06
*** SiFuh_ has quit IRC07:07
cruxbot[contrib.git]: gcc-fortran: add post-install and post-remove for ccache and distcc07:10
*** jue has joined #crux07:11
*** ChanServ sets mode: +o jue07:11
juehello07:11
mike_jue: hello. would you mind adding "php" as a dep for php-mysql. Rationale: is useless otherwise.07:15
jueno, works with php-fcgi or mod_php as well07:16
cruxbot[contrib.git]: gdc: add post-install and post-remove for ccache and distcc07:18
mike_argh... stupid me07:19
Romstera README saying that would be handy.07:20
Romsterthere i think i got them all.07:21
mike_...and playing with prt-get aliases isn't worth that07:22
Romsteri'd of listed a Optional: .. of them ports but seems it'sa bad thing even if it's for informal only and like others would say best to put it in a README.07:23
* mike_ wonders who have invented that "Optional:" header07:26
jueit's not a optional dep, because the mysql.so extension needs one of the three interpreters07:26
juestrictly I'd say the typical php user has enough knowledge to master that problem ;-)07:27
cptnjue: is there one port people will typically want?07:28
cptnI mean, one that would work with most httpds07:29
juedon't think so, mod_php is for sure widely used07:29
Romsterand jue i've moved them symlinks for distcc and ccache, tilman has yet todo the ccache port itself if he will alter that.07:30
juethe cgi/fgci version can be used with nearly any http server07:30
jueRomster: thanks07:31
cptnwhere do they go for ccache/distcc?07:33
cptnRomster: I don't think you'd have to move those of distcc itself07:35
cptnor ccache, for that matter07:35
Romsterhmm07:36
Romstershould the ones in distcc be where it was?07:36
cptnyeah, I think so07:36
Romsteri was wondering about that.07:36
cptne.g. for gcj the symlinks only make sense if distcc is installed07:37
cptnfor distcc, they make sense if distcc is installed07:37
cptnso that should always happen07:37
Romsterbut one might want to not use a shady directory with distcc or ccache.07:38
Romsteronce can do CC="ccache gcc" or CC="distcc gcc"07:38
Romsters/once/one/07:38
Romstershadow not shady...07:38
cptnthen, you'd have to change the README though07:39
cptn"b) prepend /usr/lib/distcc to your $PATH environment variable."07:39
Romsteryeah i better edit the readme i thought i have that in there.07:39
Romsterb) prepend /usr/lib/distcc to your $PATH environment variable.   (e.g. in pkgmk.conf)07:40
cptnyeah, that's what I pasted07:40
cptnthis implied that /usr/lib/distcc is there07:40
Romsterbut it's listed hmm i should list most modes of operation and about the post-install file that does the shadow method.07:41
juenipuL: just read in the logs that you have problems with ck4up. Looks like a resolve problem. You can try to comment out the line "require 'resolv-replace'" (line 30), than dns requests are serialized.07:41
Romsterright now it arn't very well explained.07:41
cptnI didn't get a lot of complaints07:42
cptnRomster: did you?07:42
Romstercptn, no but i get a lot of how do i do this or that.07:42
Romsterand i can make note of a few other variables i am using related to distcc too.07:43
juebbl07:43
cptnRomster: how about a nice Public wiki page07:44
cptnincluding notes how to combine ccache and distcc?07:44
Romsterhmm i can do that. an keep the readme too i guess and make note of the wiki page.07:45
cptnsounds good07:45
Romstercptn, i notice you origionally packaged that stuff07:45
Romsternoticed*07:45
*** jdolan_ has joined #crux07:46
cruxbot[contrib.git]: php-bcmath: initial import07:48
cruxbot[contrib.git]: php-ftp: initial import07:48
cruxbot[contrib.git]: php-gd: initial import07:48
cruxbot[contrib.git]: php-gettext: initial import07:48
cruxbot[contrib.git]: php-mbstring: initial import07:48
cruxbot[contrib.git]: php-postgresql: initial import07:48
mike_here you go07:48
cptn:-)07:48
Romsterwhoo more contrib git push flooding.07:52
*** f1y has joined #crux07:54
thrice`yikes07:54
*** f1y has quit IRC07:56
*** f1y has joined #crux08:00
Romsterthrice`, yeah contrib and opt is being busy.08:09
Romstercptn, initial page created http://crux.nu/Public/Distcc08:10
Romsterjust a quick rundown of what i have now. now i got more wiki code to learn and add in sections on what does what. and i'll have that part i have now as a how i use it at the bottom.08:11
cptnnice08:11
Romsterah i should of made the page as SetingUpDistcc maybe..08:12
Romsteri'm not so good at wiki.08:12
cptnI think you can easily rename it08:12
cptnerr, no08:12
Romsteroh well i'll keep adding to it.08:13
Romsterand get to that later by making a enw page or geting someone to move it.08:13
Romsternew*08:13
Romsterok now i'm making a how to compile in a tmpfs and move that part out of my distcc and link to it.08:21
Romsterhow a wiki should be.08:21
tilmanRomster: what's wrong with opt/ccache?08:25
Romsterthe symlinks are created in the Pkgfile and i  biased all my other ports on that concept and Jue said it was annoying so i moved all that to post-install files now.08:30
cptntilman: nothing, it was only for those ports that check in build() whether distcc or ccache is installed08:31
cptntilman: doesn't affect ccache itself08:31
tilmangood08:31
*** jaeger has joined #crux08:32
*** ChanServ sets mode: +o jaeger08:32
Romsterhttp://crux.nu/Public/HowToSpeedUpBuildingPackages messy but it's a start others can improve on it too.08:35
Romsterjust moves the symlinks to a post-install file. and i've done that with distcc aswell since some may not want to use a shadow directory and set it with CC="distcc gcc" or ccache.08:36
Romsterso i was wondering if you would follow suite with the change tilman.08:37
cptnRomster: why would they not want to do that?08:37
jaegerRomster: under "Mount the tmpfs" the mount isn't specified08:38
Romsteroh sugar.. i'll fix that.08:39
tilmanand building in tmpfs isn't related to having a separate build user08:39
jaegeralso, after it's in the fstab, "mount /usr/ports/work" is sufficient08:39
tilmanthe latter is already explained in vektori's howto, no? :)08:39
cptntilman: does that still exist? :-)08:39
tilmanmmh08:39
tilmandidn't check, but i think so08:39
cptnah yeah, it is08:40
cptnor does08:40
Romsterah i need to make a note on the uid too, and i've made a group myself but isn't needed on that.08:41
tilmanhttp://crux.nu/Main/FakerootPorts08:41
Romsterdamn it i'll just reference that wiki page then and edit again...08:42
*** Rotwang has joined #crux08:42
cptnlooks like a good start for building tips08:42
cptnI think there's lots of knowledge also in READMEs that could be in the Public wiki08:42
cptnallows to make so READMEs shorter, and extend the wiki pages with links and such08:43
cptns/so/some/g08:45
Romsterthere.08:49
*** Rotwan1 has joined #crux08:50
Romsterthat wiki engine isn't like the ones that links are created automatically from CamelCase08:50
Romsterwish there was one standard wiki format.08:51
Romsteri think http://crux.nu/Public/HowToSpeedUpBuildingPackages is passable now.08:52
Romsteri know i've spent a lot of time hunting for information, and having it in one assessable location would be handy.09:01
*** Rotwang has quit IRC09:10
mike_An easy way to fill hard disk*09:21
Rotwan1cat /dev/urandom ?09:21
mike_Rotwan1: I am reffering to the first lines of http://crux.nu/Public/HowToSpeedUpBuildingPackages09:22
Rotwan1k ;]09:23
cptnmike_: I think he refers to avoiding disk access09:24
cptnmike_: and thus prolonging the life of the disk09:24
mike_I see, but it sounds freeky. 2GB of ram is not wortless too.09:25
mike_*worthless09:25
tilmanit's not like it's reserved for the tmpfs :)09:25
cptnmike_: it's dynamically assigned09:25
cptnso it's at most 2GB09:26
mike_yes, I hardly can afford that amount on a laptop with 1GB ram & integrated video card =). And the page sounds like it is a "very good practice", while my experience shows that having it on slow hdd still gives a huge boost.09:27
tilmanhuh?09:28
tilmanare you saying that your hdd is slower than your ram? :)09:28
mike_no. my slow hdd is faster than my CPU =)09:29
mike_900MB ram,  4500RPM 2.5" drive vs 1.73Gh/2MB Pentium M. And it still makes sense.09:31
* jaeger gets out and pushes dbca09:39
Romstersave your hard disk, = use it less09:39
Romstermike_, i only have 1,280MB ram on this pc anymore and it goes to swap.09:41
tilmanmike_: how can memcpy be slower than read&write on the hdd?09:41
tilmanon a machine with limited memory you can still use a little ram disc (maybe 500M)09:42
tilmanit won't be enough to build firefox, but it will suffice for lots of programs ;)09:42
pitillodoes someone use e17?09:43
Romsterbut will fail on bigger ports saying 'out of space', but extra space from swap is slow but only the very large packages will ever hit that.09:44
jaegerpitillo: haven't recently but I used to09:46
pitillojaeger, I want to know if it's used by people and can be interesting make a repo (to upload some snapshot which will be tested here before) and svn ports too to build it09:49
jaegerno idea how many use it, sorry09:49
pitilloif this sounds a good idea I hope someone can contat me, if it isn't used, I'm not sure if it's a good idea wasting space for it09:50
pitillojaeger, well, no problem, thank you for your answer btw :)09:50
jaegerno problem09:51
Romsteri always say 'build it and they will come'09:51
tilmanaon: popcon would be helpful here ^^^^09:51
Romsterpopcorn?09:52
pitilloRomster, that comment was for me? (sorry if it wasn't)09:53
thrice`pitillo: there is an e17 repo09:53
tilmanthat's pitillo's repo i think ;)09:53
thrice`ooh, sorry...heh09:54
pitillothrice`, yes, I am the maintainer.... but I am making snapshots and I uploaded them to my router, which has a low bandwith, that is why I asked if it's used09:54
Romsterthat was fun.09:54
pitilloI think it hasn't too much sense to upload some tarball to the server if I am the only one using it09:55
pitillo*tarballs09:55
Romsterpitillo, i'll mirror the archive if that will help, but on low bandwidth too09:55
pitilloRomster, then is the same, but thank you anyways (here I have a upload conection around 36KB)09:56
Romstertrue but adds redundency.09:56
pitilloummm that's true09:56
mike_tilman: of course memcpy is faster. I mean it sometimes still makes sense using hdd for ccache on a machine with limited ram.09:57
thrice`pitillo: building from svn isn't too time consuming, though :)  your ports work well09:58
Romstermike_, compileing in ram is faster than hdd.09:58
Romstermainly seek time and lower bandwidth of hdd compared to ram.09:59
pitillothrice`, yeah, here I am working with cvs versions, but may be people can think it's a bit dangerous to update it (I think it sometimes)09:59
mike_brrr, that ram might be used for other stuff too09:59
mike_say, running kde* during sysup09:59
Romsterccache on hdd would be insane in ram unless you happen to to afford 4GB or ram for a ccache that can be mirrored to hdd before reboots.09:59
Romstermike_, that ram is dynamicly assigned for port building.10:00
Romsteranother words it's free to the system when you arn't building a port...10:00
Romsterpkgmk                 2.0G     0  2.0G   0% /usr/ports/work10:00
Romstersee 0% used at the moment...10:01
pitillothrice`, and the fact to install cvs in it's deps (can be a bit hard like sepen told me a few days ago if you are behind a fw for example)10:01
mike_ah, my point is lost between the lines10:01
tilmanmike_: i got your point ;)10:01
Romsterr dn't want to installl git/subversion/cvs/bzr or another other cms justto be able to build a port.10:02
cruxbot[xorg.git]: xorg-xtrans: updated to 1.2.10:02
Romsterapparantly i don't..10:02
jaegerhrmm... I can't get virtualbox to go to my full resolution (in X) of 1440x900, only 1024x76810:03
jaegerI remember having this problem but not how to fix it10:03
Romstermike_, what ram isn't being used goes to swap and oh forget it pointless discussion you should know enough to make your own decision.10:04
mike_Romster: =)10:04
Romsterjaeger, dunno i run at 1152x864 75Hz10:04
Romsterfull screen seems to work.10:05
Romsterfor those that want to is why i made the wiki page.10:06
jaegerit used to work for me at 1440x900 on this machine but I can't remember how I fixed it10:06
jaegerfullscreen mode just displays the 1024x768 window in the center with black borders10:06
Romsterjaeger, wiki page it when you figure it out again.10:06
jaegermaybe I just set a modeline10:06
pitilloummmm I don't remember (first thoughts are addons or driver used by X)10:07
*** Rotwan1 has quit IRC10:09
*** predatorfreak has joined #crux10:16
jaegerboth the virtualbox guest additions and the x driver are installed (and the x driver is selected in xorg.conf)10:19
jaegerhrmm, not really improving, now I can get it to do 720x400 :P10:29
*** maxus has joined #crux10:31
Romstermike_, "An easy way to reduce disk access and prolonging the life of the disk, by using faster ram for compilation time." there a better description.10:44
Romsterjaeger, sounds like fun :P10:44
*** RedShift has quit IRC10:50
mike_Romster: nice10:59
*** mike_ has quit IRC10:59
Romsterslowly getting there editing the distcc page a bit more then i'm gonna go sleep and i'll do more later.11:01
*** sepen has quit IRC11:04
*** jtnl has joined #crux11:11
Romsteredited http://crux.nu/Public/Distcc some more.11:18
Romsterthat'll do for now i need sleep before i do anymore.11:18
*** jtnl has left #crux11:18
*** RedShift has joined #crux11:18
thrice`export MAKEFLAGS='-j 6'11:19
Romsterwhat about it?11:20
thrice`oh, I thought -j6 was right11:20
Romsteryou can have spaces.11:20
Romsterbut on a variable it needs to be quoted.11:20
thrice`i've never used jobs before, though.  didn't know you could do spaces :)11:20
thrice`neat11:20
Romstermake -j 1 and make -j1 both work. same as patch -p1 and -p 111:21
thrice`looks good :)11:21
Romsterjust with variables you need to quote when there are spaces.11:21
Romstergot more too add it's too specific for my setup.11:21
Romsterbut ti's stuff others would benefit from knowing.11:22
Romsterit's*11:22
Romsterwhen it's done and someone comes in and asks i can point them to that page.11:24
*** Rotwang has joined #crux11:27
jaegerRomster: mind testing a libmp4v2 build for me? just take the libtoolize line out of the current contrib port11:35
Romsterjaeger, k11:37
Romsterjaeger, hmm that worked11:46
jaegerI know very little about libtool, I wonder if it would cause any problems to leave it out11:47
Romsterwhy was it in there in the first place.11:49
jaegerno idea. perhaps it was needed at some point in the past11:49
Romsterhmm might be wise to add -fno-strict-aliasing to CXXFLAGS11:50
Romsteralso touch bootstrapped shouldn't be needed i'll test without that11:51
Romsterhmm it's some hack to avoid there bootstraping.11:53
*** namenlos has quit IRC11:54
Romsteri don't think leaving it out would be wise, might create some library incompatibility.12:05
jaegersome other dists don't bother with it, gonna build it without and test it for a while12:11
Romsterk12:11
Romsternow i'm off to bed be back in 8 hours or so.12:12
jaegertake care12:12
Romsteri will night.12:13
*** jnengland77 has joined #crux12:18
*** Rotwang has quit IRC12:38
*** jue_ has joined #crux13:01
*** teK has quit IRC13:06
*** jue has quit IRC13:14
*** RedShift has quit IRC13:21
*** jue_ has quit IRC13:23
*** predatorfreak has quit IRC13:40
*** lennart has quit IRC13:53
*** jnengland77 has quit IRC13:55
*** Rotwang has joined #crux14:07
*** RedShift has joined #crux14:12
*** Rotwan1 has joined #crux14:18
*** predatorfreak has joined #crux14:31
*** Rotwang has quit IRC14:34
*** Rotwan1 is now known as Rotwang14:34
*** sepen has joined #crux14:54
*** teK has joined #crux14:57
Rotwanghttp://www.jwz.org/doc/emacs-timeline.html14:58
*** RedShift has quit IRC15:16
*** mike_k has joined #crux15:29
mike_kat least http://crux.nu/portdb/?a=repo&q=contrib seems to be outdated15:31
*** jdolan_ has quit IRC16:14
mike_kRotwang: fox has two "Packager" headers16:25
Rotwangmike_k: which one should i leave?16:25
Rotwangfirst one? or second?16:25
mike_kRotwang: maybe who was the last one(Maintainer) before you've took it16:26
Rotwangk16:27
Rotwangpitillo said its better to leave first one16:37
Rotwangare there any strict rules about it?16:38
RyoSPackager is the guy who originally made the port16:38
RyoSMaintainer keeps it up to date16:39
pitilloRotwang, , well, I don't know if it's the correct, but if I were me, I think I keep the original packager16:39
Rotwangok so the first one16:39
pitilloif someone can clarify this, better :)16:39
RyoSpitillo: thats right ;)16:39
Rotwangmaybe there should be additional file: CREDITS :P16:40
Rotwang(joke)16:40
mike_kRotwang: few minor notes: please check the formatting of your contrib-candidate ports. For example, phun has one ugly leading space inside build() and strange offset in "URL" header; phun, asciijump have world writable directories; pidgin-osd has no .md5sum16:45
mike_kare those writables essential?16:46
Rotwangmike_k: pidgin is downloaded inside build so md5sum takes no part16:49
Rotwang?!?16:49
mike_khuh, teeworlds is fun16:49
Rotwangusr/var/games/asciijump/16:50
Rotwangis writable16:50
mike_kRotwang: hmm, it should be empty. just do pkgmk -um inside.16:50
Rotwangprobably it should be16:50
Rotwangbut asciijump isnt candidate for now ;]16:50
Rotwangok ill make md5sum16:51
Rotwangand yes, teewars are fun :D16:51
mike_knot a candidate but anyway...16:51
Rotwangusr/var/games/asciijump/ is prolly for records and stuff16:52
Rotwangphun also has world writables16:52
Rotwangmaybe i should add sticky bit to that dir in asciijump16:53
mike_kmaybe asciijump's ./configure has some options to point it to another (more common place), say /var/lib/$name. yes, sticky can save if it will work16:53
Rotwangk16:54
mike_khmm, do you have any port which depends on fox toolkit?16:55
mike_kI can't find any to test it16:55
Rotwangi had :<16:55
Rotwangnao but it doesnt compile16:56
mike_kok, nevermind16:56
Rotwangso fox is not necessary for now16:56
mike_kit still could be used by other contributors or by private repos. not really useless, but potentionally untested.16:58
*** f1y has quit IRC16:58
Rotwangi wouldnt put that, and bashburn for now17:01
Rotwangi need to test them first17:01
mike_kyep17:09
mike_kRotwang: good luck17:11
*** mike_k has quit IRC17:11
*** pitillo has quit IRC17:12
*** predatorfreak has quit IRC17:20
Rotwang+ from mike_k, great :D17:58
*** Rotwang has quit IRC18:09
*** bjohan_ has joined #crux18:27
*** bjohan__ has quit IRC18:44
*** errdil has quit IRC19:51
*** jdolan_ has joined #crux20:05
*** predatorfreak has joined #crux20:19
*** errdil has joined #crux20:34
Romstermike_k> at least http://crux.nu/portdb/?a=repo&q=contrib seems to be outdated <- huh? when i spend time updating my ports at least. and i see others update too.20:46
*** destruct1 has joined #Crux20:50
cruxbot[contrib.git]: gcc34: post-install and post-remove fix path for checking for Pkgfile20:55
*** mwansa has quit IRC21:00
*** destruct has quit IRC21:01
*** mwansa has joined #crux21:08
*** mwansa has quit IRC21:12
*** mwansa has joined #crux21:12
*** mavrick61 has quit IRC21:15
*** Dudde has quit IRC21:15
*** rx1 has joined #crux21:16
*** mavrick61 has joined #crux21:16
*** Dudde has joined #crux21:16
*** mrks_ has quit IRC21:24
*** rxi has quit IRC21:31
*** mrks has joined #crux21:35
*** joacim has quit IRC21:43
*** mwansa has quit IRC21:44
*** mwansa has joined #crux21:46
*** predatorfreak has quit IRC22:09
*** jaeger has quit IRC22:21
*** jdolan has quit IRC22:43
*** jdolan has joined #crux22:44
*** ChanServ sets mode: +o jdolan22:44
Romsterhi jdolan23:03
*** morlenxus has quit IRC23:05
*** morlenxus has joined #crux23:05
*** jnengland77 has joined #crux23:15
*** jnengland77 has quit IRC23:39

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