That looks very promising. Thank you!
Hmm I'll need to figure out how to turn it on from Python rather than the command line, but I think I have enough to go on.
From: Jorge Vargas <address@hidden>
Date: Mon, 18 Jun 2012 14:54:25 +0100
To: Brent Tubbs <address@hidden>
Cc: "address@hidden" <address@hidden>
Subject: Re: [Fab-user] Is there an opt-in switch for the old line buffering behavior?
Date: Mon, 18 Jun 2012 14:54:25 +0100
To: Brent Tubbs <address@hidden>
Cc: "address@hidden" <address@hidden>
Subject: Re: [Fab-user] Is there an opt-in switch for the old line buffering behavior?
Hello Fabric users and devs,I'm working on a deployment orchestration tool that uses Celery workers to execute Fabric tasks. The actual functionality is going great, but the log output from remote commands is a bit wonky; it's printed with just one character per line. See http://pastebin.com/0SW99xL8 for an example.I've confirmed that with the old line buffering behavior from Fabric <=0.9 I can get normally-formatted responses in my Celery worker logs. (Example with a dummy project/task: http://pastebin.com/Y7jAhEqq.) So, a couple questions:
- Where in the codebase is the best place to look to re-enable line buffering? Is there just one place, or would I have to patch a bunch of stuff all over?
Did you saw ? I believe it does what you need.
- Would a env.buffer_lines option be of interest to upstream or anyone else, or should I anticipate keeping this as my own fork/monkeypatch?
As I said above the flag is already in the codebase. I'm not 100% sure how it's executed but i'm pretty sure it's stored in env just like any other variables in fabric. So try it as I think you don't need a patch at all. Do keep in mind it has some limitations http://docs.fabfile.org/en/1.4.0/usage/parallel.html#linewise-output which is why it was turned off by default
_______________________________________________
Fab-user mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/fab-user
