On Tue, May 19, 2015 at 01:15:04AM +0100, John Ballance wrote:
> Hi
>
> mea culpa.. as you found, I have no "" on the AB_HG line. a typo here.
It turns out that John was doing a syntax like:
AB_HG=https://blahblah/projectname nameondisc
and this was confusing things. This doesn't really fit with the idea of
AB_HG as a URL rather than a list of options to hg, though I can see the
benefits of providing extra options.
> (I've also got some mods for the fetcher to do a hg pull if the
> sources are there.. cuts about 15 mins off the load time. .. how can
> I usefully feed those back as I dont have update rights(as far as I
> can see)
Posting a patch here is the best way, it's best to make it an attachment
(just the raw output from 'svn diff', don't zip it or anything) so quoting
inline doesn't mangle it. (In another version control system, 'git
send-email' will do everything for you, but doesn't exist for SVN sadly).
One caveat though is that often we want to wipe out the previous state so
that we know a build will run from scratch, whereas any reliance on previous
state then makes a build a function on all the previous builds. So is it
possible to add a flag that would enable 'svn update'/'git pull'/'hg pull'
in fetch-program? That would mean we can either build afresh or, if you're
debugging or updating the build, just pulling recent changes. There's also
the issue that you want to remove unversioned files (like compiler output)
and you can't trust 'make clean' to do that
('git clean -xfd && git reset --hard' is a thing, but not all VCSes do that)
Theo
(my git bias may be showing)
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
No comments:
Post a Comment