| === sergiusens_ is now known as sergiusens | ||
| === pbek_ is now known as pbek | ||
| mup | Bug #1766079 opened: ripgrep installed via snap fail to work in /lib/systemd/system folder <Snappy:New> <https://launchpad.net/bugs/1766079> | 12:39 |
|---|---|---|
| === sergiusens_ is now known as sergiusens | ||
| === mup_ is now known as mup | ||
| === molum is now known as ybden | ||
| === jero is now known as Guest41378 | ||
| cjwatson | nsg: I'm far from an expert (I help operate the build service, but snapcraft isn't really my thing), but one thing that's often helpful in general is to grab the build log of the last version known to work and diff it against the first version known not to work | 16:13 |
| cjwatson | nsg: Let me know if you need help with tracking down the files | 16:13 |
| nsg | Yup, I grabbed the build logs last weekend. I have diffed them, both manually and programmatically but no luck. Only thing I noticed is that the output from snapcraft has changed a little, so I guess a new version was released between these builds. | 16:34 |
| nsg | snapcraft_2.39.3+really2.35_all.deb vs. snapcraft_2.40_all.deb | 16:36 |
| nsg | But it's also possible that it has nothing to do with snapcraft, it's related to a upstream dep. But the thing is, all the versions from the build logs looks the same. | 16:37 |
| nsg | The error I'm getting when I launch the snapped application is "malloc(): memory corruption: 0x00000000026afe30" ... lovely error. Feels like I'm dynamic linked the incorrect libs? Then, I'm not that experienced to debug C/C++ code :) | 16:40 |
| nsg | The annoying thing is, I have change 0 lines of code. It broke it self :\ | 16:40 |
| nsg | I wonder if the "Improved elf mangling" feature does something amazing ... "This functionality is triggered when the build environment is newer than a given base (experimental)" I'm on 17.10 so...." | 16:44 |
| nsg | (yes I do lxd based builds) | 16:44 |
| nsg | But that would not explain why the build services creates the same broken packages :) | 16:45 |
| ogra_ | nsg, perhaps related to https://forum.snapcraft.io/t/patchelf-broke-my-binary/4928 | 16:56 |
| nsg | oh, good find. I will test with "build-attributes: [no-patchelf]" | 16:58 |
| nsg | and while I wait for that, I will create a 16.04 VM for a reference build :) | 16:58 |
| ogra_ | you said above yu already use lxd ... so just use "snapcraft cleanbuild" that should automatically build in an lxd container | 17:01 |
| nsg | yes, but, I'm not sure if that's identical with a "real" 16.04 install ... I'm a little out of ideas here to test (the cleanbuld is already running in the background) :) | 17:03 |
| nsg | I have also have a lot of problems with "SNAPCRAFT_CONTAINER_BUILDS=1" builds, never works the 2nd time, not sure why, and a cleanbuld takes like a hour. So a VM will be useful for me anyway. | 17:05 |
| thresh | а куда wayland-egl делся из портов? | 17:52 |
| === hayden is now known as matlock | ||
| matlock | I am trying to build my first snap from a github project, here is the snap file https://github.com/sirredbeard/FeedReader/blob/master/snap/snapcraft.yaml | 19:24 |
| matlock | I can't use the cmake plugin in snapcraft because it breaks under nexted cmake files | 19:24 |
| matlock | so I tried using a build script to call cmake, make, and make install but none of the built libraries and binaries come over into staging or prime, even though other parts built by autotools plugin do | 19:25 |
| matlock | the github project also has meson build features, so I tried using snapcraft's meson plugin for a yaml here | 19:25 |
| matlock | https://github.com/sirredbeard/FeedReader/blob/master/snap/snapcraft-meson-version.yaml | 19:25 |
| matlock | but it fails with gcc, glibc, and zlib issues | 19:26 |
| matlock | can anyone point me in the right direction? thank you! | 19:26 |
| matlock | I also have filed a bug with snapcraft here https://bugs.launchpad.net/snapcraft/+bug/1766101 | 19:27 |
| mup | Bug #1766101: Snapcraft not handling cmake nested builds <Snapcraft:New> <https://launchpad.net/bugs/1766101> | 19:27 |
| matlock | I posted it here too I also have filed a bug with snapcraft here https://bugs.launchpad.net/snapcraft/+bug/1766101 | 19:37 |
| mup | Bug #1766101: Snapcraft not handling cmake nested builds <Snapcraft:New> <https://launchpad.net/bugs/1766101> | 19:37 |
| matlock | I mean on the forum here https://forum.snapcraft.io/t/snapcraft-breaking-nested-cmake-not-moving-build-script-files-over-and/5084 | 19:37 |
| nsg | FYI ogra_, no-patchelf did nothing for me. I tracked it down to desktop-gtk3 and the desktop-launch script. If I launched it w/o it, it started. I did a test build with desktop-gtk2... works. | 20:00 |
| nsg | The snap in question is somewhat old, I have learned a few things since then. I will clean it up a little and try to call desktop-launch directly (it's at the moment called by a wrapper of my own). | 20:02 |
| nsg | I will then give desktop-gtk3 a 2nd chance, plan B, just ignore desktop-launch or use gtk2 :) | 20:03 |
| matlock | I got some initial feedback, snapcraft still doesn't seem to be bundling one part's binaries and libraries into the snap though | 23:28 |
| matlock | so it's really three issues (1) snapcraft breaks on nested cmake files (2) snapcraft doesn't bring over binaries/libraries when using build-override and (3) something is broken with the meson plugin for snapcraft leaving snaps asking for glibc7 and zlib; if I could solve any one of these three that would be fantastic | 23:29 |
| matlock | if someone could look at my yaml and (1) help get the cmake nested issue sorted out (2) help get the binaries/libraries over into the snap or (3) solve the glibc and zlib dependency when building snap using meson plugin that would be a win for snaps | 23:31 |
| matlock | thanks for your help in advance | 23:39 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!