Re: Substitute for git-package.mk in pkgsrc proper?

[Paul Ripke <stix%stix.id.au@localhost>]

On Tue, Dec 05, 2023 at 08:18:01AM +0100, Martin Husemann wrote:
> On Tue, Dec 05, 2023 at 07:41:45AM +0100, Martin Husemann wrote:
> > On Tue, Dec 05, 2023 at 05:31:57PM +1100, Paul Ripke wrote:
> > > I'm looking at cleaning up wip/prusaslicer and dependencies for promotion
> > > into pkgsrc, and I'm stuck on wip/opencascade-git, where the fetch options
> > > are password protected download URL, or fetch via git.
> > 
> > File a bug report upstream, this is unpackable and plain stupid.

Yeah, that was also my first reaction.

> Let me rephrase this: it certainly is (technically) possible to
> pkg this with a fetch phase that clones the anon git and checks out
> the wanted state...
> 
> but: upstream clearly does not intend this to be packaged by itself.
> Instead applications using (parts of) this should redistribute the code
> of the used components (see the FAQ).

Yeah, I read though that: "It is allowed and expected that you distribute
only the libraries and resources used by your application". I guess I
didn't quite read it quite as strict, but I can see that interpretation.

> So maybe complain to the upstream of the application that you are trying
> to pkg that uses this? They should include the proper state of this
> code (that they tested against) with their sources.

So I browsed through the OCCT bug database to see if anyone else had had
similar complaints, and discovered there is a way to download a tarball
from their gitweb interface. That'll do me, seems to work fine. Committed.

Thanks!

-- 
Paul Ripke
"Great minds discuss ideas, average minds discuss events, small minds
 discuss people."
-- Disputed: Often attributed to Eleanor Roosevelt. 1948.