@hyeyoo confession:
deacde and a half years ago, i've been a Jenkins (Hudson back then) installation operator and still got the escalated stress response to this :)
@vbabka @liskin @hyeyoo @monsieuricon also, you'll get bullied if you do the initial clone with git clone instead of fetching the bundles
@hyeyoo @liskin @monsieuricon @vbabka the idea is that you get the mainline base from a bundle (even perhaps slightly outdated) and then fetch the rest the usual way -- hopefully a lot less data
@monsieuricon @vbabka @hyeyoo your endeavours and @liskin 's musings made me curious whether there's a way to optimize for shallow clones? Like tell a git repack to do a pack containing only the commits that would be needed for a shallow clone or something. But it doesn't seem to be the case.
(Confession: i'm probably hopelessly outdated when it comes to anything new in git for the last ten years or so)
@hyeyoo @liskin @lkundrak @monsieuricon @vbabka Yes, bundles are a lesser known git feature, but they are cool. They can also be used for efficient offline repo transfers; best thing you can pull from a bundle just like from a repo URL…