[gentoo-dev] ARCH keywording for g-cpan?

[gentoo-dev] ARCH keywording for g-cpan?

[Michael Cummings]

[-- Attachment #1: Type: text/plain, Size: 1454 bytes --]

This is geared mostly towards arch devs, but I'm open to any comments :)

A new version of g-cpan is about a week away. Lot of feature additions, 
upgrades, code revisions (wouldn't recommend diff'ing the current to the new 
because I know there's only a few lines in common). It's still 100% perl (in 
fact, the old version was only 95%, so that's another improvement - no more 
system calls, ick), and the big reason for even going down the road of a 
rewrite was to add support for the intended split of dev-perl into new, 
smaller, manageable(!) categories.

In the past, g-cpan was lumped in with portage, so it received all of the 
keyword bliss of portage. When it was initially split from portage it also 
inherited that keyword batch, since the version in the split and the version 
in the current portage were identical. But herein lies my dilemma. While 100% 
perl code means there ~shouldn't~ be any issues on any platform that can 
install perl, I don't want to tick any arch's off by unmasking en masse. 
Would you all prefer a bug with recommended tests for the new version? Do you 
all even care? :) Let me know, on list, off list, just don't call me late at 
night, it makes the wife wonder.

Mike

-- 

-----o()o---------------------------------------------
Michael Cummings   |    #gentoo-dev, #gentoo-perl
Gentoo Perl Dev    |    on irc.freenode.net 
-----o()o---------------------------------------------

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] ARCH keywording for g-cpan?

[Aron Griffis]

[-- Attachment #1: Type: text/plain, Size: 1017 bytes --]

Michael Cummings wrote:	[Wed May 18 2005, 05:56:41AM EDT]
> Would you all prefer a bug with recommended tests for the new
> version? Do you all even care? :) Let me know, on list, off list,

For a project like this, where you wrote the code and you know there
shouldn't be cross-platform issues, I'd suggest marking stable on all
platforms.  That is what I do for keychain.

If you choose that route, you make things easier for everybody, but
you also risk the wrath of arch teams if your evaluation turns out to
be wrong.  For keychain it has never bitten me, but it's certainly
possible that someday I'll find a snippet of shell code that works
everywhere except SPARC.  When that happens, I'll owe an apology (and
a change of behavior) to the SPARC team for marking stable without
testing on their platform.

If you find other arch teams opposed to this approach, then at least
I can tell you to go ahead and mark stable when you see fit for alpha
and ia64.

Regards,
Aron

--
Aron Griffis
Gentoo Linux Developer


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]