[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openembedded-core
Subject:    [OE-core] [PATCH 0/2] bitbake.conf: add GITHUB_GIT
From:       richard.purdie () linuxfoundation ! org (Richard Purdie)
Date:       2015-07-31 14:30:52
Message-ID: 1438353052.22462.80.camel () linuxfoundation ! org
[Download RAW message or body]

On Thu, 2015-07-30 at 08:34 -0300, Otavio Salvador wrote:
> On Thu, Jul 30, 2015 at 5:32 AM, Olof Johansson <olof.johansson at axis.com> wrote:
> > Excerpts from Robert Yang's message of 2015-07-30 09:09:07 +0200:
> >> PREMIRRORS is for all the recipes/protocols ?
> >
> > No, PREMIRRORS is quite expressive, for instance:
> >
> > PREMIRRORS = " \
> >   https?$://downloads.example.org$/.*   http://mirror.example.com/ \n \
> > "
> >
> > This should only premirror downloads.example.com files, fetched via http
> > or https. See the bitbake fetcher's uri_replace method in
> > bitbake/lib/bb/fetch2/__init__.py for the implementation.
> 
> Yes, you can.
> 
> But the mirrors scheme allows for, for example, map it for git:// and
> https:// protocols without changing the recipes.
> 
> I am in favor of consistency here and having a unique var for github helps.

I came to review this as Ross pulled it into mut and I reached the same
conclusion as Olof here. We have PREMIRRORS, we can also exclude
networks from the fetcher and therefore have a ton of ways of rewriting
this already.

Yes, we do want standardisation, but all the urls look standardised to
me already. 

IMO, we want to see less of these global variables, not more. There once
was a time when they might have been useful but the fetcher is pretty
good about dealing with this now so lets not complicate this more than
we need to.

Cheers,

Richard



[prev in list] [next in list] [prev in thread] [next in thread]