Danek Duvall wrote:
> On Wed, Sep 17, 2008 at 05:08:01PM -0500, Norm Jacobs wrote:
>
>
>> If hg is capable of it (and I haven't looked into it enough), we can add
>> hooks to force the upload of missing tarballs at commit (push) time
>>
>
> That's actually a good idea.
Yeah, I have one every so often, though they don't always pan out. ;-)
> Or at least to check to make sure that the
> tarball is in place. And it's completely doable from a hook.
>
I would prefer it to force the upload if it's missing and not just check
for it.
>
>> and download of tarballs when a workspace is cloned or pulled.
>>
>
> As long as it's intelligent enough to be able to copy or symlink or
> something from a local disk cache, that's fine. I don't want to download a
> gig or so of tarballs I already have each time I create a new clone of the
> gate, and I'm sure most people are sitting behind even thinner pipes
>
> The downside of having this hook in place is that each developer will need
> to make sure that the hook is installed in their hgrc -- there's no way to
> centralize hook distribution. It's one of the reasons hgsetup exists for
> ON. The push hook is run on the gate, so that's not an issue, but the pull
> hook will be. So I'd still advocate both belt and suspenders, and have
> some mechanism for the makefiles to download the sources if they're
> missing.
>
Sort of a shame that the pull hooks can't be in the gate as well,
though I understand the implications being able to do so. Yes, I would
want a target or tool to check/force tarball download as well, so we can
encourage people to download/link tarballs immediately after cloning or
updating a workspace if/when they don't have client side hooks installed.
>> Are there any mercurial experts here that want to step up to help solve
>> this? provide assistence? , ...
>>
>
> I can help bounce around ideas, and I'd be interested in some coding, but
> I've got loads on my plate already, so I'd be in the position of doing it
> as a break from my other work. Mark really ought to be the one to help
> here, I think.
>
I still see some potential holes in all of this. I would like it to at
least nominally function as though the tarballs are in the gate, even if
they really aren't. We need to explore this further, find the gaps, and
address them. I will send out a strawman for everyone to torch.
-Norm
