Hello Joey,
here the onfiguration of my propellor repro.
[remote "origin"]
url = git+ssh://git.debian.org/git/users/picca/propellor
fetch = +refs/heads/*:refs/remotes/origin/*
I use this central repository to manage a bunch of computers. The problem I is that all these computer do have acces only from this url
[remote "origin"]
url = https://anonscm.debian.org/git/users/picca/propellor.git
fetch = +refs/heads/*:refs/remotes/origin/*
So I can use the deploy branch trick.
BUT
since I have clone of the first repository (in order to push via ssh) on all of these computer (let call them locals), I need to add the deploy branch on all of them.
So I tryed to add this directly in the propellor configuration.
& PropellorRepo.hasOriginUrl "https://anonscm.debian.org/git/users/picca/propellor.git"
But when I do the spin for the first time from the locals, the url put in the /usr/local/propellor/.git/config files is the one from the locals cloned repository which is the ssh adress and not the https one.
So the propellor spin stop with a timeout because it tryes to git fetch with the wrong url.
My question is is it normal ?
PS: I also tryed to add
& bootstrappedFrom (GitRepoUrl "https://anonscm.debian.org/git/users/picca/propellor.git")
but the result was the same.
Cheers
I don't know what deploy branch trick you're referring to using.
There is the deploy remote, which I think should do what you want:
When you have a deploy remote configured in your local repository, the first time you spin a new host, it will clone its /usr/local/propellor from the url of the deploy remote. On subsequent spins, it sends the url over, and the host's repository's url is updated too.
The PropellorRepo.hasOriginUrl property doesn't take effect until propellor runs with that configuration, which could be a bit of a chicken and egg situation. Except hmm, spin could look for that property and use repo url instead of the deploy remote's url. Yeah, I've done that now!