On Mon Oct 19 20:33:19 UTC 2009 Robert Collins wrote:
> On Mon, 2009-10-19 at 11:09 +0000, James Westby wrote:
> > I'm working on a "cookbook" feature, that allows you to build
> > multiple recipes in one go, with some convention as to how
> > things are done. I could include the ability to specify a branch
> > from which the recipes are retrieved with that. I think that
> > might be a better layering, but I can see that for a single
> > recipe the overhead of the convention may put some people
> > off.
>
> I'd be happy putting the recipes in one branch and referring to them; I
> wouldn't want to build all the recipes at once as we're doing
> just-in-time builds triggered on commits, not on cron.
That's slightly different to what I had in mind then. I'll have
to think about the best way to integrate this.
Note that bzr-builder (of course) uses bzr's transports to retrieve
the file, so it would happily use a link to raw output from
loggerhead. We could also do
try:
transport.get(basename)
except 404:
b, rp = Branch.open_containing(dirname)
b.get_file_bytes(os.path.join(rp, basename))
On Mon Oct 19 20:33:19 UTC 2009 Robert Collins wrote:
> On Mon, 2009-10-19 at 11:09 +0000, James Westby wrote:
> > I'm working on a "cookbook" feature, that allows you to build
> > multiple recipes in one go, with some convention as to how
> > things are done. I could include the ability to specify a branch
> > from which the recipes are retrieved with that. I think that
> > might be a better layering, but I can see that for a single
> > recipe the overhead of the convention may put some people
> > off.
>
> I'd be happy putting the recipes in one branch and referring to them; I
> wouldn't want to build all the recipes at once as we're doing
> just-in-time builds triggered on commits, not on cron.
That's slightly different to what I had in mind then. I'll have
to think about the best way to integrate this.
Note that bzr-builder (of course) uses bzr's transports to retrieve
the file, so it would happily use a link to raw output from
loggerhead. We could also do
try: .get(basename) open_containing (dirname) get_file_ bytes(os. path.join( rp, basename))
transport
except 404:
b, rp = Branch.
b.
type thing, so you could specify
bzr dailydeb lp:foo/recipes/whatever
Thanks,
James