Internal repository of mozbase

RESOLVED WONTFIX

Status

Testing
Mozbase
RESOLVED WONTFIX
6 years ago
6 years ago

People

(Reporter: wlach, Unassigned)

Tracking

Trunk
Points:
---

Firefox Tracking Flags

(Not tracked)

Details

It was brought up on irc earlier today that we couldn't generate a talos.zip because it pulls files directly from mozbase on github, which was down:

http://hg.mozilla.org/build/talos/file/b666faec6abe/create_talos_zip.py

This behaviour might change eventually, but it's probably a good idea for us to have a copy of mozbase mirrored on an official mozilla server (with full history), as it's so important for everything we do and will only become more so in the future. I can definitely see other parts of automation wanting to pull it down in the future.

I don't know if there are legal issues or anything (i.e. contributor agreements) to worry about, and I don't know of anything important that's actually blocking on this. Nonetheless, I'll open this bug up for discussion purposes.
OS: Linux → All
Hardware: x86_64 → All

Comment 1

6 years ago
FWIW, hopefully talos.zip will be a thing of the past someday when we're using mozharness + talos + internal (e.g. mozbase) python package server.

Comment 2

6 years ago
So I'm (currently) of the opinion that an internal pypi is sufficient for this use-case.  Ignoring talos.zip, which should soon go away (hopefully), do you agree with this assertion?  We currently mirror mozbase to m-c for most test harnesses, and given that we continue to do this and talos continues to live outside of the tree (bug 787200), talos seems the only odd man out and will (again hopefully) not be for the long haul.  Neither the mirroring nor pypi is done with per-commit history.

The internal pypi documentation is linked to from https://wiki.mozilla.org/Auto-tools/Projects/MozBase#Mozbase_and_Mozharness

If we were to use an internal repository, is there any particular advantage of copying vs replacing, outside of the fact that some people prefer to work with git and others prefer hg?
(In reply to Jeff Hammel [:jhammel] from comment #2)
> So I'm (currently) of the opinion that an internal pypi is sufficient for
> this use-case.  Ignoring talos.zip, which should soon go away (hopefully),
> do you agree with this assertion?  We currently mirror mozbase to m-c for
> most test harnesses, and given that we continue to do this and talos
> continues to live outside of the tree (bug 787200), talos seems the only odd
> man out and will (again hopefully) not be for the long haul.  Neither the
> mirroring nor pypi is done with per-commit history.
> 
> The internal pypi documentation is linked to from
> https://wiki.mozilla.org/Auto-tools/Projects/MozBase#Mozbase_and_Mozharness

Yes, I think this bug is WONTFIX. The current reliability of github is quite good and as you say this will go away in the near future.
 
> If we were to use an internal repository, is there any particular advantage
> of copying vs replacing, outside of the fact that some people prefer to work
> with git and others prefer hg?

Well, even if the mozilla copy of mozbase came to be the "master", it could be useful to have a github mirror to ease forking, tracking, etc. Talos's github clone would be an example of that.
Status: NEW → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.