*** frinring_ is now known as frinring | 02:26 | |
NeoChapay | abranson: https://git.merproject.org/mer-core/libzypp/merge_requests/17/diffs i create pr for fix build master branch | 06:41 |
---|---|---|
NeoChapay | Who can help me with mic error ? https://pastebin.com/P4XbTzQU | 08:31 |
r0kk3rz | ValueError: invalid literal for int() with base 10: '1+git2' | 08:32 |
r0kk3rz | sounds straight forward enough | 08:32 |
abranson | NeoChapay, r0kk3rz: mic didn't work with the latest rpm, but a fix has been merged now I think. maybe that needs an upgrade? lbt? | 08:34 |
NeoChapay | abranson: version 0.14+git5 work fine http://repo.merproject.org/obs/home:/neochapay:/mer:/tools/latest_i486/noarch/ | 08:37 |
r0kk3rz | abranson: you had to delve into the arcane innards of mic? i hope you had a bottle of whisky handy | 08:43 |
abranson | r0kk3rz: wasn't me, I haven't got my wizard robe and hat yet | 08:57 |
abranson | NeoChapay: about the libzypp - thanks for the patch, but I've got a newer upgrade of that in progress. you really shouldn't use the master atm - that version won't be used. | 08:58 |
abranson | https://git.merproject.org/abranson/libzypp | 08:59 |
abranson | *it's not ready yet though* | 08:59 |
r0kk3rz | jusa: https://git.merproject.org/mer-core/libogg/merge_requests/2 | 10:21 |
jusa | r0kk3rz: thank you! merged | 10:53 |
r0kk3rz | that was quick :) | 10:56 |
*** Son_Goku is now known as Conan_Kudo | 11:33 | |
*** Conan_Kudo is now known as Son_Goku | 11:34 | |
r0kk3rz | lbt: another one for the mirror please - https://git.xiph.org/vorbis.git/ | 11:52 |
mal | why do we keep mirrors of so many repos? | 11:54 |
r0kk3rz | because thats TheWayTM | 11:55 |
r0kk3rz | it works nicely with gitlab at least | 11:55 |
mal | I don't really understand why using the original git as the submodule would be any different | 12:00 |
r0kk3rz | all i know is that lbt requested it be this way, im sure he has his reasons | 12:01 |
lbt | some of the packages in mer:core use a git submodules. That means they don't actually have the source in them. By using a local mirror we ensure that we can build even if gitlab/hub/orious/... goes down for a bit. | 12:09 |
lbt | we have had issues in the past where github was blocked for our users too | 12:09 |
lbt | so like any distro we have a local copy of the src we build ... that's all there really is to it | 12:10 |
Son_Goku | it's generally not a bad idea to have a local copy in some form | 12:10 |
Son_Goku | usually having the tarball stored in something like dist-git is usually what's done, but having mirrors of the code tree works too | 12:11 |
r0kk3rz | i like that i can click the module commit ref and view the source in a familiar interface, instead of whatever upstream happens to use | 12:13 |
mal | lbt: btw, how should we handle projects with using mercurial? for example SDL uses that and the current way is a bit annoying for updating, could we use mercurial to git converter and push te generated git repo as a mirror? | 12:13 |
w00t | lbt: or worse, if it vanishes altogether - you're then left holding a license-compliance bag in some cases :) | 12:13 |
lbt | w00t: *nod* :) | 12:13 |
lbt | mal: drop SDL ? | 12:14 |
mal | lbt: ? | 12:14 |
lbt | but yeah - it's the Maintainer's call really. I'm not sure about using a convertor automatically but it seems fine if they do that | 12:15 |
mal | lbt: I was trying to update SDL and I was thinking how that should be done | 12:15 |
r0kk3rz | src dump miror? | 12:15 |
mal | lbt: there is the SDL mirror on github but afaik that is unofficial | 12:15 |
lbt | essentially get the source into git - either by converting or by doing a mega-commit | 12:16 |
lbt | the mega-commit needs to be super well documented in the commit so that it can be trusted - that's really hard | 12:16 |
mal | there seems to be at least two tools for making git mirror of a mercurial repo, hg-git and git-remote-hg | 13:12 |
pvuorela | mal: if sdl git mirror is not official enough, i'd maybe go with extracted sources + patch files. | 13:46 |
mal | pvuorela: would that extracted sources method be more preferred than using hg-git or something for creating a git mirror in mer-core? | 13:51 |
pvuorela | mal: guess it depends on how it goes. if the git mirror is automated, sure that would be nice, but if's updated with a PR it would be hard to verify. | 13:55 |
lbt | I would prefer to see a reproducable mirror - ie if hg-git produces an identical git mirror on 2 runs then that's a good solution to me | 13:57 |
pvuorela | if that mirror is updated by a contributor, i expect more work to me doing the same thing with hg-git compared to extracting upstream .tar.gz and running diff. | 13:59 |
mal | lbt: I need to test how those two tools I mentioned work | 14:00 |
r0kk3rz | im not sure hg-git is what we want | 14:00 |
r0kk3rz | its a plugin for mercurial, not the other way around | 14:00 |
mal | r0kk3rz: yes, but it can be used to mirror also, saw instructions how to do that | 14:01 |
r0kk3rz | ah nice | 14:01 |
mal | r0kk3rz: so hg-git is part of mercurial (or extension to that) whereas git-remote-hg is working on top of git directly | 14:02 |
mal | pvuorela: assuming the mirrow is updated via MRs you could also compare the mirror by downloading the source package and make the same diff, only difference would be that one method contains commit history and the other doesn't | 14:07 |
pvuorela | mal: unless package and version control have differences. e.g. documentation generated. didn't check, though. | 14:10 |
mal | pvuorela: yep, I'll do some testing, just to make sure what did you mean by the method of extrating the sources, so we exact the sources to a git project? btw, in the past some hg to git system was probably used as https://git.merproject.org/mer-core/libsdl has the history | 14:18 |
pvuorela | mal: yea, was probably done that way. with extracting i meant something like: git rm everything except rpm dir in libsdl, extract .tar.gz into a sub-directory, git add the directory. | 14:25 |
mal | pvuorela: yep, thought something like that | 14:28 |
*** Nokius_ is now known as Nokius | 21:03 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!