Packages based on CVS sources

[Michael Thomas]

Hans de Goede wrote:
> 
> 
> Rex Dieter wrote:
> 
>> Wart wrote:
>>
>>> What is the policy for building packages for software that does not yet
>>> have upstream source tarballs, but only maintains the software in CVS?
>>
>>
>> Makes it impossible to verify upstream sources for review.  For new
>> packages, definite bad idea.
>>
> 
> Erm,
> 
> Why does that make it hard to verify the sources? Download srpm, extract
> tar gz made by packager from CVS checkout. Do a cvs checkout yourself,
> next do a diff -ur --exclude=CVS, and there you have your comparision.

I hadn't thought of that...

>>> Am I allowed to pull the CVS sources and make my own source release?
>>> Obviously, I would rather that this is all done upstream, but in case
>>> they aren't responsive to such requests, what are my choices?
>>
>>
>> The only case I can think of that *may* be acceptable: to update an
>> existing package, with grave bug(s), but that should be a rare exception.
>>
> 
> Again, why? I have no personal interest in tclpro (I don't even know
> what it is).

It's a graphical debugger for Tcl scripts.  It was open-sourced when the
company that developed it (Scriptics / Ajuba) was acquired by Interwoven
in 2000.  A commercial version with updates was later released by
ActiveState.  Not much work has gone into the open-source version since
then, yet it's still functional and very useful.

 But I see no reason why a CVS checkout can't be build into
> a package, as long as you put the commands used todo the checkout in a
> comment in the spec file all is well.

I'll do that for now until I get some sort of response from the upstream
developers.  There's still a lot of work to be done before I'm ready to
submit it as a new package.  Hopefully there will be upstream tarballs
by then.

Thanks,

--Mike