| ice9 | i just built the rhythmbox and gedit sources using make but after the build, i can't find the binary exec file to start the application | 00:02 |
|---|---|---|
| pvl1 | make install | 00:31 |
| pvl1 | ice9 | 00:31 |
| ice9 | pvl1, no i don't want to install it on my system, just want to make a fix and test it | 00:32 |
| pvl1 | is there a bin dir | 00:37 |
| pvl1 | in the project directory | 00:37 |
| ice9 | no | 00:38 |
| Noskcaj | Can someone please rebuild dx? It should allow it to pass the libtiff5 transition | 01:11 |
| === freeflying_away is now known as freeflying | ||
| Noskcaj | And feel++ needs a rebuild to pick up openmpi1.6 | 01:24 |
| Noskcaj | Actually just about everything at http://people.canonical.com/~ubuntu-archive/transitions/html/openmpi1.6.html needs a rebuild | 01:26 |
| xnox | Laney: thanks. | 01:28 |
| Noskcaj | to rebuild a package should i just submit a branch with a build1 changelog? | 01:34 |
| xnox | Noskcaj: "everything" is a vague definition, given it's 78% complete, and many things are not accounted for (e.g. fixed packages in proposed) | 01:45 |
| xnox | Noskcaj: rebuilding feel++ is jumping the ship, since gmsh is not yet fixed across all architectures. | 01:45 |
| xnox | Laney: not sure about the tracker. E.g. openmpi1.6 didn't get an ppc64el column =/ | 01:47 |
| Noskcaj | xnox, ok. So i do just submit a branch with a build1 revision for the relevant packages? | 01:53 |
| xnox | ? | 02:00 |
| xnox | Noskcaj: which package? | 02:00 |
| xnox | Noskcaj: LoganCloud uploaded all those that didn't FTBFS, the rest seemed to need fixes. | 02:01 |
| Noskcaj | ok | 02:01 |
| xnox | Noskcaj: and has the packages deps migrated to openmpi1.6 already? | 02:01 |
| Noskcaj | idk | 02:01 |
| xnox | Noskcaj: across all arches... it's easy to get into arch skew =/ | 02:02 |
| xnox | Noskcaj: no change rebuilds is such a low level task, that sponsoring those is, well, a lot of overhead. people with upload rights motu/core-dev usually do those. | 02:02 |
| Noskcaj | ok, makes sense | 02:03 |
| xnox | Noskcaj: patches to fix FTBFS for packages in-current transitions is always nice to know =) | 02:03 |
| xnox | Noskcaj: so again which "the relevant package" that can just be no-change rebuild that you have identified? | 02:03 |
| Noskcaj | I was going to manually try them, was just checking the process first | 02:05 |
| xnox | Noskcaj: locally? why a branch merge proposal? | 02:13 |
| LoganCloud | Did I do something wrong | 02:13 |
| LoganCloud | I was mentioned in this channel. Not a good sign | 02:13 |
| Noskcaj | that one i just changed libtiff4-dev to libtiff-dev. Was there something other than building i needed to check? | 02:14 |
| xnox | Noskcaj: one checks current depends in -proposed/-archive, does a local no-change rebuild to check that new library name is picked up, and if all checks out good prepares no-change upload. | 02:14 |
| Noskcaj | LoganCloud, you're fine | 02:14 |
| Noskcaj | ok | 02:14 |
| xnox | LoganCloud: nah, all good =) just pointing out the good work you did on pushing openmpi transition. | 02:14 |
| xnox | Noskcaj: as far as I can tell tiff transition is complete, and it has been requested to remove tiff from the archive. | 02:15 |
| xnox | Noskcaj: unless it's creeping back in? | 02:15 |
| xnox | Noskcaj: so again, can you tell me which _src_ package you are looking at? | 02:15 |
| Noskcaj | xnox, Should be done, just a package that still needed libtiff4 | 02:15 |
| LoganCloud | Phew | 02:15 |
| Noskcaj | And none | 02:15 |
| xnox | Noskcaj: not the transition name, e.g. tiff4. but the reverse dep in question... | 02:16 |
| xnox | Noskcaj: i'm confused. =( | 02:16 |
| Noskcaj | I have no idea what you're talking about | 02:16 |
| * Noskcaj thinks he needs to find something else to do with his holidays | 02:18 | |
| xnox | Noskcaj: "<Noskcaj> [01:11:51] Can someone please rebuild dx? It should allow it to pass the lib" & "<Noskcaj> And feel++ needs a rebuild to pick up openmpi1.6" & "<Noskcaj> to rebuild a package should i just submit a branch with a build1 changelog?" | 02:18 |
| xnox | Noskcaj: I got the impression, that you want to help with transitions and believe you have something ready to be uploaded. | 02:19 |
| xnox | Noskcaj: yet, I still don't know which packages you want to be rebuild. | 02:19 |
| Noskcaj | 1 and 2: basic guesses, probably wrong. 3. Moure general info | 02:19 |
| Noskcaj | I want to help, don't really know what to do | 02:19 |
| xnox | Noskcaj: looking at dx -> whilst it build-depends on libtiff-dev, it doesn't actually result in binary depends on tiff. | 02:20 |
| xnox | Noskcaj: in the transition tracker feel++ is at the very top of the stack, and typically things feel++ depends on should transition before feel++ itself will. | 02:20 |
| xnox | Noskcaj: otherwise one may end up in an unfortunate situation of trying to load both ABIs into a single binary =/ which is not good. | 02:21 |
| Noskcaj | ok, so that's what the dependancy level thing means | 02:21 |
| xnox | Noskcaj: yeah, roll the mouse over the package name to see it. | 02:21 |
| Noskcaj | oh | 02:21 |
| xnox | Not very discovarable, unless one reads the source code of the page ;-) | 02:21 |
| LoganCloud | xnox: thanks :) it was my first real transition | 02:23 |
| LoganCloud | The fact that Debian did it already helped, for sure :P | 02:23 |
| xnox | LoganCloud: there are plenty of FTBFS to fix... left =) | 02:23 |
| xnox | still though. | 02:24 |
| xnox | yeah, but it's now different, cause they did it at different timing. | 02:24 |
| LoganCloud | Yes but they're also FTBFSing in Debian | 02:24 |
| xnox | LoganCloud: did they remove them from testing? | 02:24 |
| LoganCloud | At least most of the ones with issues | 02:24 |
| LoganCloud | I think so | 02:24 |
| xnox | lazy =) | 02:24 |
| LoganCloud | Because the ones with problems say "Sid only" I think | 02:25 |
| LoganCloud | I'm not by my computer right now. On my phone | 02:25 |
| xnox | fixed openmx, cctools and others now. | 02:25 |
| LoganCloud | Cool, I'll take a look in like 10 min | 02:25 |
| Noskcaj | speaking of that, do you guys have any tips for finding what is causing an ftbfs? It's very rare for me to find one i know how to fix | 02:25 |
| xnox | Noskcaj: read the build-log, that's the only.... | 02:27 |
| Noskcaj | As in working out how to fix what the buildlog says | 02:28 |
| xnox | one solves a sudoku by solving it..... | 02:28 |
| xnox | read what it says, and depening on what it says, try to understand why, and resolve the root cause behind it..... | 02:29 |
| xnox | it's not any different from any other puzzles =) | 02:29 |
| LoganCloud | Noskcaj: knowing basic stuff about coding and stack traces helps | 02:31 |
| LoganCloud | Also spotting patterns, and taking note of fixes to common problems | 02:31 |
| Noskcaj | LoganCloud, There's my issue. I only know basic python, and all the python ftbfses are from test errors, or architectures i don't have. | 02:32 |
| xnox | Noskcaj: learn C, it's easy. | 02:32 |
| Noskcaj | sigh. I really should | 02:32 |
| xnox | http://www.cprogramming.com/ | 02:32 |
| xnox | Laney: looks like my last config change did break the trackers. Revert that commit. | 02:49 |
| xnox | Laney: can you take a look at how to enable the new arch? | 02:49 |
| Logan_ | xnox: Tsk tsk. You modified hdf5 right after I fixed the default mpi for ppc64el and arm64. :P | 03:33 |
| xnox | Logan_: really? | 03:36 |
| xnox | Logan_: where? | 03:36 |
| Logan_ | mpi-defaults | 03:36 |
| Logan_ | I reverse-traced the logic it was using to pick up the default MPI and found it was coming from the depends for that package | 03:37 |
| Logan_ | so I added arm64 and ppc64el as arches for openmpi in that package's build-depends | 03:37 |
| xnox | Logan_: right. and ppc64el is finished now. | 03:38 |
| xnox | Logan_: do we need to rebuild everything on top on mpi-defaults on arm64 & ppc64el ? | 03:38 |
| xnox | Logan_: i'll revert hdf5 | 03:39 |
| xnox | Logan_: actually hdf5 needs a no-change rebuild once, mpi-defaults passes. | 03:39 |
| xnox | Logan_: and it's still wrong for hdf5 to pick a non-existant package as a depends. | 03:39 |
| xnox | Logan_: will you submit mpi-defaults patch to debian? | 03:40 |
| Logan_ | xnox: yes | 03:45 |
| xnox | cool, thx | 03:45 |
| Logan_ | xnox: Ugh. This hdf5 thing is a mess. | 05:16 |
| Logan_ | xnox: just sent you an e-mail | 07:18 |
| alkisg | Laney: could you please have a look at https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/474392/comments/11 ? | 07:40 |
| ubottu | Ubuntu bug 474392 in indicator-session (Ubuntu) "indicator-session menus are not policykit aware" [Low,Confirmed] | 07:40 |
| alkisg | /dev/sdb1 on /media/alkisg/usb-stick type vfat (rw,nosuid,nodev,uid=1010,gid=1010,shortname=mixed,dmask=0077,utf8=1,showexec,flush,uhelper=udisks2) | 08:29 |
| alkisg | ...that means that when I copy a file from that stick to my home folder, other users don't have read access to it | 08:29 |
| alkisg | That was not the case in the past... shouldn't dmask provide read access to others by default, or at least some other way could be found so that I don't need to manually update the rights of the files I copy from usb sticks? | 08:30 |
| alkisg | Hmm completely hardcoded... :( http://cgit.freedesktop.org/udisks/tree/src/udiskslinuxfilesystem.c?id=aa02e5fc53efdeaf66047d2ad437ed543178965b | 08:38 |
| alkisg | For example, if they were mounted under /run/user/uid/mounts, they wouldn't need strict permissions because /run/user/uid would block other users | 08:43 |
| === doko_ is now known as doko | ||
| Laney | xnox: what happened? | 10:01 |
| Laney | alkisg: I could maybe take a look after the holidays | 10:07 |
| Laney | If you're able, a patch would be welcome | 10:07 |
| alkisg | Laney: I could try a patch for https://bugs.launchpad.net/indicator-session/+bug/1263438, but not for https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/474392/comments/11, that's a bit beyond my current knowledge of the libraries involved :-/ | 10:10 |
| ubottu | Ubuntu bug 1263438 in indicator-session (Ubuntu) "Indicator-session assumes there's only one active user" [Undecided,New] | 10:10 |
| ubottu | Ubuntu bug 474392 in indicator-session (Ubuntu) "indicator-session menus are not policykit aware" [Low,Confirmed] | 10:10 |
| === ValicekB_ is now known as ValicekB | ||
| alkisg | (12:10:15 μμ) alkisg: Thanks, I wish you happy holidays! (to Laney... pidgin was too lazy in notifying me that I was disconnected :)) | 11:09 |
| === freeflying is now known as freeflying_away | ||
| doko | $ cat test.c | 16:18 |
| doko | int main(){return(0);} | 16:18 |
| doko | $ gcc -fPIE -pie -static test.c | 16:18 |
| doko | *** Error in `/usr/bin/ld': corrupted double-linked list: 0x0a329260 *** | 16:18 |
| doko | <and ld hangs ...> | 16:19 |
| doko | infinity, ^^^ seen with glibc-2.18, not 2.17 | 16:19 |
| doko | kees, is the use of hardening-includes in graphicsmagick correct? the above example calls with -pie -static | 16:21 |
| doko | gold correctly complains about it: | 16:21 |
| doko | $ gcc -fuse-ld=gold -fPIE -pie -static test.c | 16:21 |
| doko | /usr/bin/ld.gold: fatal error: -pie and -static are incompatible | 16:21 |
| doko | collect2: error: ld returned 1 exit status | 16:21 |
| xnox | LoganCloud: yeah, don't worry about it =) easily to sort out. | 16:33 |
| xnox | Laney: so all trackers stopped to update after I've commited ;"ppc64el" in the global config + some trackers that list arches explicitely (ghc?) see commit + revert in the config/ branch. | 16:33 |
| xnox | Laney: try updating config branch to -r-2 and rerun tracker so check where it blows up? | 16:34 |
| cjwatson | xnox: Based on the previous syntax my guess would be that ";" is only supposed to separate items in a list, rather than terminating items | 18:23 |
| cjwatson | xnox: That is, you need to not have the semicolons after ppc64el | 18:23 |
| cjwatson | (global.conf and monitor/permanent/ghc.ben) | 18:23 |
| === jono is now known as Guest49214 | ||
| ice9 | does ubuntu plan to continue using GTK or will move to Qt at sometime? | 22:01 |
| Noskcaj | ice9, Unity is moving to Qt5 with unity 8 | 22:13 |
| ice9 | Noskcaj, so no more GTK? when unity 8 is coming? | 22:27 |
| mdeslaur | ice9: gtk will still be there | 22:29 |
| ice9 | but why Unity to move to Qt instead of GTK? | 22:29 |
| Noskcaj | ice9, We'l still use gtk, but unity 8 (which should be out for 14.10), should be Qt | 22:29 |
| mdeslaur | ice9: unity isn't using gtk at the moment, it's using a toolkit called nux | 22:30 |
| mdeslaur | ice9: it's moving from nux to qt | 22:30 |
| xnox | cjwatson: ta. | 23:16 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!