#general
- e
enough-analyst-54434
10/02/2015, 2:00 AMI think workflows are things you worry about for the 1st 2 weeks - e
enough-analyst-54434
10/02/2015, 2:00 AMand then they work - b
bored-art-40741
10/02/2015, 2:00 AMStrongly disagree - e
enough-analyst-54434
10/02/2015, 2:01 AMI've had 10+ in my life and they all have been different but fine - e
enough-analyst-54434
10/02/2015, 2:01 AMNono - not cd? - e
enough-analyst-54434
10/02/2015, 2:01 AMIts literally pull from jsmith - e
enough-analyst-54434
10/02/2015, 2:01 AMyou pull his branch from his machine - b
bored-art-40741
10/02/2015, 2:01 AMOh, I see - b
bored-art-40741
10/02/2015, 2:02 AMWell, I think there's a happier medium: origin is pristine and you have a "dirty" shared remote for branches - e
enough-analyst-54434
10/02/2015, 2:02 AMIts sortof ridiculous to use a server middleman for that sort of sharing - b
bored-art-40741
10/02/2015, 2:02 AMWhere you just nuke it with a week of heads up every once in a while - b
bored-art-40741
10/02/2015, 2:03 AMIf I could seamlessly yank a branch from a ref without having to add a remote, I'd be into it - b
bored-art-40741
10/02/2015, 2:03 AMWhich is probably possible in git, now that I think about it - e
enough-analyst-54434
10/02/2015, 2:03 AMyou can - a
ancient-stone-50795
10/02/2015, 2:03 AMWe have jobs that audit stale branches / old branches that haven’t been merged. - e
enough-analyst-54434
10/02/2015, 2:03 AMA remote can be an url - e
enough-analyst-54434
10/02/2015, 2:04 AMit does not need to be added with an alias - a
ancient-stone-50795
10/02/2015, 2:04 AMI was thinking we’d just add a little bit of scripting around git to make the second ‘dirty’ server a little easier to manage. - b
bored-art-40741
10/02/2015, 2:05 AMThere's a balance here, and it's definitely not something that requires git-fu on par with mine - f
freezing-motorcycle-14222
10/02/2015, 2:05 AMSo it's the branches that slow git down? - b
bored-art-40741
10/02/2015, 2:05 AMIn very large repos with lot of users and lots of upstream refs, it does - a
ancient-stone-50795
10/02/2015, 2:05 AMThere are different ways that git slows down. Lots of branches apparently slow down operations on the server side. - b
bored-art-40741
10/02/2015, 2:07 AMYeah, there are a bunch of variables you can push to make git become slow, but generally you have to push a couple at the same time to suffer - b
bored-art-40741
10/02/2015, 2:07 AMIn the case of the three large monorepos represented here, a large repo with lots of commits and also lots of divergent references on the primary upstream is bad for performance - b
bored-art-40741
10/02/2015, 2:08 AMThe last thing though is not necessary if you manage your workflows carefully - b
bored-art-40741
10/02/2015, 2:09 AMI guess I should say lots of named references upstream - b
bored-art-40741
10/02/2015, 2:10 AMSince otherwise they will be GCed by git - w
witty-crayon-22786
10/02/2015, 6:13 AM@enough-analyst-54434, @bored-art-40741 : middlemen are inevitable though. if your branch isn't backed up, it doesn't exist - e
enough-analyst-54434
10/02/2015, 6:26 AMBut you don't need git for backups - e
enough-analyst-54434
10/02/2015, 6:28 AMShard my branch commits in patch mailbox form off in a scalable store - say manhattan