IRC Logs for #crux-devel Friday, 2008-06-27

*** pitillo has joined #crux-devel00:00
*** mike_k has joined #crux-devel01:33
*** mike_k has quit IRC02:14
*** mike_k has joined #crux-devel02:14
cptnoh my, XML02:56
pitilloXML for Pkgfiles?03:22
cptnyes :-)03:22
pitilloxD03:22
cptnpitillo: how was the party last night? :-)03:23
pitillojajajaa there were no party03:23
pitilloI was watching the match at home03:23
pitilloI think this sunday will be better03:23
pitilloI will watch it with all people. I can't have fun and wake up soon in the morning...03:24
pitillobut this sunday will be different, I will try to make an effort :)03:24
cptnheh, nice03:24
cptnI was watching the game in the city last night03:25
cptnlots of happy people03:25
pitilloyeah, here there are some public sites to watch the match too03:25
pitilloyeah, and was a very good match in my opinion03:25
pitillothey made great football03:25
cptnyeah, it was intense03:26
pitilloI hope we can see goals this sunday too (not like vs italy which was a bit hard)03:26
cptnyeah, I hope so too03:26
pitillobut in a final, both teams will be a bit conservative03:26
pitillowe will see how it goes03:27
pitillochanging talk, one Q about prt-get. I readed the TODO list, are you working on it?03:27
cptnhaven't gone through the TODO yet, no03:28
pitilloor do you have an important (or what do you think it can be interesting) item there?03:28
pitilloah oki03:28
cptnthere's a couple of things going on right now, but I hope to get back to programming soon03:29
pitillowell, I was thinking in look for an excuse to start reading code, and of course, try to do something03:29
pitillothese next weeks I will be offline on afternoon and I want to pick something todo offline03:30
pitilloand about the next tuesday, will be another meeting?03:31
cptnheh, being offline sometimes helps to get things done03:31
cptnsince there's so many ways to waste time online03:31
cptnpitillo: yes, there will be a meeting03:31
cptnit would actually be the first tuesday of the month03:32
pitillothat's true, I have some things todo (I have a little server which I let without doing anything on it since some time ago)03:32
pitillobut I want to focus a bit more on coding (and if it can be, on C coding...)03:32
pitillosounds great, I will try to make an antenna to see if I can get a wifi signal, but I can't I will try to move to a friend's house to be here too03:33
pitillobut I / but if03:33
cptnheh, a self-made wifi antenna?03:35
pitilloyeah03:36
cptnsounds cool :-)03:36
pitilloI have now a "cd based" antenna which gives me a bit more of strenght (it's in my friends house)03:36
aonthe reflectors from car headlights make rather good "dishes" too, i've heard03:37
pitilloI am thinking in make another one based on the chinesse "frying pan" (not sure if it's well said)03:37
pitilloaon, ummmm never saw one of that. Taking a look to see if I can get some info related03:37
cptnokay, have to get some work done, ttyl :-)03:38
pitillohave fun :P03:38
mike_kI still have no N-type connectors nearby  to proceed with double bi-quad antenna =(03:43
aonhttp://www.usbwifi.orconhosting.net.nz/03:44
aonheh03:44
pitillojajajaaa03:45
pitillothat is what I meant by chinesse "frying pan" xD03:45
aonhttp://koti.mbnet.fi/t20aaa/WLAN.JPG03:46
aoni guess the headlight reflectors are a bit expensive to buy for just that purpose, though :)03:46
aonthe VAZ-2101{,1,3} headlight could be good as it's round and rather cheap :)03:47
pitilloummm may be looking for them in a second hand cars stuff store03:48
pitilloit's a good idea too03:48
*** sepen has quit IRC03:58
*** sepen has joined #crux-devel03:58
*** clb has joined #crux-devel04:21
pitilloif someone see jue can give him a link to this? http://lokalix.dyndns.org/crux/logs/prtverify/06:04
pitillothere are 2 files, 1 with the prtverify.wl config file (I added one report from 1 of my ports) and the other is a diff  with the output of sepen's repoverify based on prtverify showing that the prtverify WL works very well here06:05
pitilloI don't know if can be better to send him a mail explaining this instead of boring you here06:06
sepenohh nice pitillo06:10
sepeni'm glad to known about that06:10
pitilloI am sending jue a mail to see if he can take a look to it and verify the report06:10
sepenpitillo, sure06:12
sepenor maybe this afternoon he will connect here06:12
pitillowell, I thought can be good to make a note here, but I will send him a mail too06:15
pitillothank you both for your work06:15
sepenI'll modify my repoverify script this afternoon06:32
tilmannipuL: does my patch for #311 work for or not?07:04
nipuLyeah07:15
tilmanokay, i'll push that07:16
nipuLwhy not [ -n "$CC" ] && cat "CC=$CC" >> config/linux-dri07:17
tilmanbecause i'm useless with shell scripting07:17
nipuLheh, have you read my fumbling in #315?07:19
tilmannot yet :D07:19
tilmanjust saw that it took you 3 attempts or so ;p07:20
nipuLyeah, and it's probably still not a complete patch07:20
nipuLi went all "angry german kid" on it07:21
nipuLwell almost, i managed to not smash the keyboard07:21
tilman;]07:21
tilmanwhat about #270?07:25
tilmancptn: i see you already looked at the patch once07:25
nipuLbah, can't the default mime type just be plain/text?07:26
cptnnipuL: heh, the same just annoyed me too07:27
cptntilman: yeah I think it's okay07:27
tilmannipuL: don't think so :x07:27
cptnhttp://flyspray.org/manual:attachments07:29
cptnseems doable07:29
nipuLwouldn't #270 be a downstream patch?07:29
nipuLor do we already have crux specific udev rules?07:29
cptnwell, our udev configuration is quite modified already07:30
cptnbut maybe we could see what other distros do07:30
cptnhaven't looked at them for a long time07:30
nipuLit might be just me, but i've always preferred sudo over using setuid07:32
aonsame07:32
cptnfine with me too07:34
cptnit's easy enough to adjust07:34
*** jue has joined #crux-devel07:35
nipuL-1 for me, it's just a convenience, doesn't really add any extra functionality07:35
cptndo we need cdrom/floppy/tape groups?07:36
tilmani think a cdrom group would be a good thing07:37
cptnI mean, we have them all already07:37
tilmans/would/is/07:37
tilman;p07:37
cptnfair enough :-)07:37
tilmancptn: i want to be able to read/rip CDs with my user07:37
tilmanso chown root:group /dev/hda yada yada07:38
cptntilman: so #270 would be good for you :-)07:38
tilmanare you trying to say that i should merge it?07:38
tilman;p07:38
tilmanheads up: new git's gc command got noticably better in 1.5.607:41
tilmanmy xorg.git shrunk by more than 50% \o07:41
tilmanbut it didn't do anything on my xserver clone, mmh07:42
juecptn: I added that groups and some more to /etc/groups some times ago because our udev-rules used them, dunno if that fit perfect nowadays07:42
juenipuL: comments on the '[crux64] multilib vs. pure 64' thread?07:57
nipuLwell my thoughts on multilib are well known, but to keep the purists happy maybe there should be 2 core repos07:59
nipuLthere's only a handful of ports that differ, so maintainence shouldn't be a problem07:59
juehmm, sorry I mean the open question07:59
nipuLif the 64bit crux is base on multilib, the purists are going to have to have glibc-compat32 on their system08:01
nipuLotherwise i don't think gcc will build08:01
nipuLbut tbh, i haven't tried to build a nultilib gcc without it08:01
nipuLit's also a bit of chicken and the egg08:02
nipuLi can have a play with it tomorrow night08:03
nipuLoh no i can't, have to go to a cocktail party08:03
cptnheh08:03
nipuLsat night's are usually my hardcore geek out until 4am night08:04
nipuLthe joys of married life, it used to be my drink lots of bourbon and try to pull a root night08:04
tilman;D08:04
juenipuL: sorry, still not clear for me, will gcc work without glibc-compat3208:08
nipuLprobably not with --enable-multilib, but as i said, i haven't actually tried08:08
jueI mean not building gcc, but using gcc to compile other stuff08:09
nipuLi think so08:12
nipuLthere was a time when i had a broken glibc-compat32 but could still build x86_64 stuff without any problems08:12
jueok, thanks08:13
nipuLactually, i should be able to test it now08:15
nipuLi'll remove glibc-compat32 and try to build gcc08:16
nipuLif it does build there'll probably be a footprint mismatch though08:18
cptnI think it would be nice if one wouldn't have to exchange the toolchain to install the compat32 layer08:31
nipuLnah, won't build08:46
cptnso we'd either have to install a 32-bit glibc as well08:51
cptnor build it in the gcc build08:52
cptnin the later case, those that want a multilib system would then build it twice in the worst case08:53
cptn(64-bit install; gcc update; glibc-compat32 update)08:53
nipuLbuilding glibc inside the gcc port seems like a bade idea08:55
juehaving glibc-compat32 as a build-time-only dep for gcc isn't a problem IMO. We never will update glibc within a release cycle, so a gcc update means: install glibc-compat32 binary from the CD, rebuild gcc, remove glibc-compat09:00
cptnmmmh09:02
cptnI was hoping we could allow the purists to run their system without any restrictions compared to pure 6409:02
juenot very smart, I know09:02
juemaybe we should have a deeper look at that gcc build problem09:03
cptnor maybe duplicating the toolchain really is the cleanest solution09:12
cptnwe could have both toolchains on the iso09:12
cptnso if the need arises, switching would be very fast09:13
cptnmaybe we should write down the scenarios09:16
*** pitillo has quit IRC09:32
cptnmmmh09:33
cptnwhat if we had both gcc-pure and gcc-multilib in core?09:34
cptnokay, that would be a pain to switch09:34
cptnit's just that if we have two toolchains, the user will have to add another prtdir to prt-get.conf to prefer the multilib toolchain09:35
nipuLis that such a bad thing?09:39
nipuLimo, being able to overlay prtdirs is one of the nice features of prt-get09:40
sepenjue, ping09:42
nipuLanyhoo, off to lose a fight with my pillow09:43
tilmannight09:44
sepengnight nipuL09:44
cptnnipuL: good night09:46
*** sepen has quit IRC10:22
*** sepen has joined #crux-devel10:22
juesepen: ?10:36
sepenjue, I write a tiny patch for adding whitelist path support as a switch >> http://mikeux.dyndns.org/crux/files/prtverify.in-whitelist_path.patch10:36
juehmm, not sure if I want that runtime configurable10:40
sepenok10:42
sepenI was thinking on unprivilege users, for who sounds reasonable10:44
sepenwell, whitelists runs fine in both cases, many thanks, but who decided what is a candidate for adding to whitelist and what not?10:46
juethe supplied one covers core ports only10:49
jueprtverify is only a tool to find bugs in ports, finally the maintainer has to decide whats ok and whats not. If the whitelist-feature becomes more important it makes no seens to supply it with the prtverify sources.10:57
sepenhmm10:58
sepenjust rocks!10:59
tilmanjue: this seems to work http://263ee167538d621d.paste.se/11:05
juetilman: yes, sure. But the configure.in patch still test the full sqlite-version.11:11
jueBut as our version is 3.5.9 so the simple patch should be ok as well11:12
jueas of https://bugzilla.mozilla.org/show_bug.cgi?id=424063 they have problems with 3.5.7, but well ...11:14
tilmani'm looking at configure.in, and i don't see why my naive patch should fail11:15
tilman    PKG_CHECK_MODULES(SQLITE, sqlite3 >= $SQLITE_VERSION)11:15
tilmanthat's what they are doing11:15
jueno, it dosn't fail11:16
juebit it dosn't detect versions <= 3.5.711:16
jues/bit/but11:16
tilmanthat's correct11:16
tilmanbut imo it's a non issue11:16
tilmanwe can expect users to have an up to date sqlite port IMO11:16
jueyes, correct my opinion as well, others will see taht differently ;-)11:17
tilmaninstead of patching configure.in with the Right Way[tm] to check for the version i'd rather put a small C program in the port :D11:17
tilmanand run that ;p11:18
tilmanit's insane, but it's better than shipping a 200K patch imo ;D11:18
jueyep, right11:18
juebtw, its easy to check the exact version, just run 'sqlite3 --version'11:19
tilmanoops11:19
tilmani thought you needed to check the values of SQLITE_VERSION_FOO from sqlite3.h :)11:19
jueif you don't have the commandline interface, yes11:20
*** sepen has quit IRC11:38
tilmanjue: can you please paste your current nss/Pkgfile?11:45
tilmani built xulrunner against system-sqlite, but firefox still claims it cannot initialize the security component11:45
juehttp://sh.nu/p/2457311:47
juecomplete port here -> http://jue.ods.org/tmp/nss-3.12.0-1.prt.tar.gz11:50
tilmanthanks11:52
tilmanjue: haha, missing /usr/lib/xulrunner in ld.so.conf11:57
mike_kyep, I've spotted this today12:03
jueoops, do you need it, works for me without12:03
mike_khowever it does not make my rssowl work12:03
tilmanjue: it used to work without it -- but with those nss changes it's needed apparently12:04
mike_ktilman: can you  echo "discuss /etc/ld.so.conf.d stuff before 2.5 release" > ~/TODO ?12:06
tilmanmike_k: let's see if we find a spot for that in the wiki12:07
juemike_k: as it is part of glibc, we hardly find a senseful way to do that before12:10
tilmanmaybe i misunderstood..12:11
juetilman: FS #25312:11
tilmanwell, yes12:11
tilmani thought mike_k wanted to make sure that we get this in 2.512:11
mike_kyep12:12
jueok, if so I misunderstood ...12:12
tilmanhttp://crux.nu/Wiki/MeetingAgenda12:17
juemike_k: just added my vote ;-)12:17
* tilman voted, too ;p12:17
mike_knice, friday evening should not be that long....12:41
*** mike_k has quit IRC12:41
cptnmmmh13:33
* cptn like #253 too13:33
juevotes are gladly accepted ;-)13:48
jueeven though #253 is the most wanted task already13:50
cptncan't hurt to vote anyway13:51
jueyep13:51
*** jue has quit IRC15:14
*** jmvr has joined #crux-devel16:56
nipuLpkgadd.conf.d would be nice too19:12
nipuLre: x86_64 filesystem port, i used lib64->lib so ports didn't need to have the libdir changed, technically it would be more correct to have lib->lib64, opinions?19:23

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