On 2015-05-13 21:17:20, Tiago Ilieve wrote: > Marcin, Hello > > All stuff required to build should be included within tarball. If something is > > missing I think you can fix it as you've got RW access to upstream, don't you? > > Yes I do, but I'm not sure if that's the case. The mentioned problem lies > here[1], where a `git` subprocess is called to return the last tag or commit, > which will be used later to build links compatible with both GitHub and Read The > Docs. Why not to check if workdir is a git repo and run those if it is and in other case just rely on some kind of marker which would tell that docs have to be build from the tarball without using git or downloading anything from the net? > >From what I've understood about the package repository structure, even if we > move the entire `.git/` folder to the chroot, so it can have access to its > history, the commits in there will be different. This is caused by the way > `git-buildpackage` imports the upstream tree on another branch. I think you're right so that's why build process shouldn't rely on git neither on access to the internet. snip > And yes, the intention is to work with the latest upstream. What happens is that > I had to build the package in the current repository state to understand how it > is being done. Np :-) > > From mine POV we can work together ;-) > > Nice! Can we discuss those details about packaging off-list? I guess not > everyone here is interested in knowing every tiny bit of it. :-) Sure, expect email from me. snip > I haven't used any *-buildpackage before, but the workflow described in the > README.source looks pretty straightforward to me. I'll reach you if I got stuck > with it. Np, I'll add gbp.conf later today or tomorrow and let you know it the email. -- |_|0|_| | |_|_|0| "Heghlu'Meh QaQ jajVam" | |0|0|0| -------- kuLa --------- | gpg --keyserver pgp.mit.edu --recv-keys 0x58C338B3 3DF1 A4DF C732 4688 38BC F121 6869 30DD 58C3 38B3
Attachment:
signature.asc
Description: Digital signature