| Mister_Magister | who in the world is building freebasic for sfos | 01:19 |
|---|---|---|
| Mister_Magister | i know you aint programming in that | 01:19 |
| *** ggabriel is now known as Guest9005 | 02:02 | |
| nephros | Mister_Magister: I know, sorry. It does work though :) | 05:33 |
| nephros | Mister_Magister: https://build.sailfishos.org/package/show/home:nephros:release/c-intercal | 05:34 |
| mia | https://git.0x0.st/mia/c2y-stuff/src/branch/main/defer.h here’s a C2y defer polyfill that works with every GCC version since 2004. have fun cleaning up your code | 08:52 |
| nephros | piggz[m], rinigus: what is your opinion on adding compat library packages to Chum? | 12:36 |
| nephros | like these: https://build.sailfishos.org/project/show/home:nephros:sailfishos:backports:5.1 | 12:36 |
| piggz[m] | nephros: purpose? do they conflict with anything in 5.1? | 12:42 |
| mal | is it not possible to detect the version in the code using those? | 12:43 |
| nephros | purpose: Making apps work in 5.1 which link against these libs, but have not (yet) been recompiled. | 12:44 |
| mal | so 5.0 and 5.1 builds would both work | 12:44 |
| mal | isn't that quite temporary? until someone fixes the builds in chum | 12:44 |
| nephros | of course, yea. | 12:44 |
| nephros | But for e.g. Openrepos apps that's not that certain. | 12:45 |
| mal | true | 12:45 |
| nephros | and in those packages I have removed the -devel parts so they're really just runtime. | 12:47 |
| nephros | could put them in /usr/local too. | 12:47 |
| nephros | btw, mal: who's the person to talk to about Python? Apart from Voldemort. | 12:48 |
| mal | nephros: what about python? | 12:55 |
| nephros | Getting these failures a lot: https://build.sailfishos.org/package/live_build_log/sailfishos:chum:testing/python3-flit/5.1_aarch64/aarch64 | 12:56 |
| nephros | possibly broken python-rpm-macros? | 12:56 |
| mal | nephros: same worked before? | 13:01 |
| nephros | According to https://build.sailfishos.org/package/show/sailfishos:chum:testing/python3-flit, yes. | 13:02 |
| nephros | and it's not just that package. | 13:03 |
| mal | I'll have a look | 13:06 |
| nephros | thanks. didn't want to impact your weekend tho :) | 13:10 |
| mal | trying to figure out where that came from in 5.0 | 13:10 |
| rinigus | nephros: I would have avoided backports (that's probably not a common way this term is used if I understand correctly) - better keeping pressure on apps to be recompiled against 5.1 | 13:16 |
| nephros | rinigus: understood, makes sense. backports is just a name I chose for that repo (I have others which have actual backports) | 13:18 |
| mal | I can't see any sign of that pathfix.py in 5.0, maybe it never did anything, the way that is called is diffenent in 5.1, but anyway I'll fix the macro package | 13:36 |
| piggz[m] | rinigus: any view about removing qt6 etc from chum:testing? i think its a lot of noise, and wont work until 5.2, so in the mean time, im happy to use it in my home repo with sb2 | 13:36 |
| mal | nephros: want to do a test build in some project with this https://github.com/sailfishos/python-rpm-macros/tree/jb64238 | 13:41 |
| piggz[m] | qt6-qtbase seems to have hit this error https://forum.qt.io/topic/163096/qt-6.8.3-build-fails-for-arm64-when-compiled-using-gcc-cross-compiler-x86_64-aarch64-toolchain. | 13:48 |
| rinigus | piggz: I don't know if anyone is using it. I guess not many as the one in your repo makes more sense to use | 13:48 |
| rinigus | and as it is testing ... well ... test didn't work out :( | 13:49 |
| nephros | mal: sure, give me a minute :) | 13:50 |
| mal | nephros: seems to work | 13:57 |
| nephros | mal: yep, confirmed: https://build.sailfishos.org/project/monitor/home:nephros:branches:sailfishos:chum:testing | 13:59 |
| mal | PR created | 14:05 |
| mal | I'll try to get that into 5.1 still if possible | 14:06 |
| nephros | ^^^ :thumbs_up: | 14:09 |
| nephros | next thing: Chum has "lz4", and 5.1 has "lz4-libs" now. A case of https://github.com/sailfishos-chum/main/issues/118 | 14:15 |
| nephros | ... and https://github.com/sailfishos/issue-tracker/issues/7 | 14:16 |
| nephros | rinigus: https://github.com/sailfishos-chum/lz4/ is yours. If you give me access I can try to fix/retire it. | 14:23 |
| nephros | Oh, there is already an issue about it. | 14:25 |
| rinigus | .. and you have access to it | 14:26 |
| nephros | right. sorry, I lose track of things. Will try to update :) | 14:31 |
| nephros | rinigus: OK I have pushed to master. If that's ok, please accept OBS request 4207 | 15:01 |
| nephros | The change drops the -tools package without replacement, hope that's ok. | 15:03 |
| rinigus | nephros: done by piggz :) | 15:29 |
| nephros | piggz[m]: you touched it last, therefore any breakage is on you! :D | 16:02 |
| nephros | no but it does seem to have done tthe trick, at least for rsnapshot. | 16:04 |
| mal | nephros: https://build.sailfishos.org/package/show/sailfishos:chum:testing/python3-setuptools_scm fails badly even with your latest changes | 16:17 |
| mal | nephros: maybe try a fewer test builds of that since obs is busy anyway so those just take a lot of workers | 16:28 |
| nephros | mal: on it! selected 7.1 as the newest feasible version. | 16:32 |
| nephros | ah, sorry, misunderstood. Will do branches for other changes on python stuff, sorry, | 16:33 |
| nephros | I shall also play Die Internationale in the background to support the workers. | 16:34 |
| nephros | (Kansainvälinen I guess) | 16:36 |
| mal | hehe | 16:36 |
| piggz[m] | nephros: im tempted to make you a chum maintainer and just let to go at it until everything is fixed | 16:40 |
| nephros | piggz[m]: dont know about everything, but I can try looking at the obvious stuff. | 16:42 |
| nephros | mal: this: https://build.sailfishos.org/package/live_build_log/sailfishos:chum:testing/python3-setuptools_scm/5.1_i486/i586 | 16:43 |
| nephros | Are specs supposed to be defining %__python? Seems wrong. | 16:43 |
| mia | i’m not familiar with chum spec guidelines/py packaging macros, but for the opensuse ones that would not be okay | 16:55 |
| mia | those target multiple py versions at once though so in general you’d request incompatible ones to be disabled instead of selecting the ones you want | 16:57 |
| piggz[m] | nephros: you have the power in :testing | 16:58 |
| * nephros bows. thank you for the trust | 16:59 | |
| nephros | Also https://decider.com/wp-content/uploads/2018/08/he-man-3.jpg?quality=90&strip=all&w=1284&h=856&crop=1 | 16:59 |
| nephros | mia: agreed, seems like a bug in the macros/build env. | 17:03 |
| mia | are we gonna get rpm 4.20 at some point btw? there’s some fun stuff in there with declarative build system macros | 17:09 |
| mia | i have a PoC for meson where the entire spec is reduced to the preamble with the buildrequires and a BuildSystem: meson-autospec. the rest is inferred via build system introspection and adapted to distro guidelines | 17:11 |
| mia | could handle things like libs, -devel, -doc -data etc. automatically for probably the majority of meson projects | 17:12 |
| mia | and i can see similar things working for python | 17:13 |
| piggz[m] | mal: what linux headers are used in 5.1? STATX_MNT_ID was added in 5.8, maybe the glibc update expects that, as it seems to be defined for the qt6-qtbase build, but doesnt actually exist | 17:14 |
| mal | 4.4 | 17:15 |
| mia | 💀 | 17:15 |
| piggz[m] | mal: i guess ill have to patch that out then | 17:15 |
| mia | thanks qualcomm? | 17:15 |
| mal | nephros: I wonder what is trying to use unversioned python | 17:16 |
| *** Mister_Magister_ is now known as Mister_Magister | 17:20 | |
| mal | mia: rpm 4.19 is next waiting to be merged (couldn't be added to 5.1 because we needed to do rpm database migration first in 5.1), but related to generated spec files we can't use those because those are generated too late for OBS to understand them | 17:20 |
| mal | same goes for build dependencies generated by macros etc | 17:21 |
| mia | mal: no no, what those do works with OBS | 17:21 |
| mia | yeah build deps won’t work | 17:21 |
| mia | but that’s better anyway | 17:21 |
| mia | would rather have those be explicit | 17:22 |
| mal | ok, misunderstood the earlier comment then | 17:23 |
| mia | 4.19 adds this for example https://rpm.org/docs/4.20.x/manual/dynamic_specs.html | 17:23 |
| mal | piggz[m]: probably should update kernel headers again | 17:25 |
| mia | i forgot which version but the main sections also got -a and -p options so you can use macros to append/prepend stuff to %prep %build etc. | 17:26 |
| mal | need to make a test project on obs to see what breaks, I know at least some packages will fail with newer headers because new things get build when build detects support for newer things | 17:26 |
| mia | realistically the new rpm features probably have limited practical use outside of some packaging macros to reduce boilerplate, mostly because upstream build systems are often not that nice and need manual work to bring things into compliance… but it’s nice to have ways to make packaging easier and more consistent in a few more cases than before | 17:34 |
| nephros | mal: what is using unversioned python? only thing I can find is /opt/testing/etc/xdg/rpmlint/sailfish.toml | 17:38 |
| nephros | which has a variable ValidShells | 17:38 |
| mal | nephros: I was wondering what that "python_module" uses | 17:38 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
