[OPEN-ILS-DEV] Gentoo ebuild 20070215

[Eric Lesage]

On Thu, 22 Feb 2007, Don McMorris wrote:

> This is mostly for ELesage, but I'm sure it's relevant to others.
>
> I'm working on installing Evergreen inside a Gentoo VM.  I'm taking
> the opportunity to test the ebuild developed by Mr. Lesage
> (openils-portage-20070215.tar.gz ).
>
> I've run into trouble installing (via CPAN) DateTime::Format::ISO8601.
> The issue I've had seems to have been widely experienced by CPAN
> testers.

BTW, if you check the ebuild comments of :ISO8601, you'll see that the 
tests on this ebuild are disabled due to an error in t/02_examples.... :-)

I've just found there is a bug report for this on CPAN:
http://rt.cpan.org/Public/Bug/Display.html?id=24335

I'll see if my next release could include the suggested patch.


> As such, I tried emerge'ing DateTime-Format-ISO8601.  I got
> an ebuild error in DateTime-Format-Builder, regarding invalid ATOM
>> dev-lang/perl.

Indeed, there's a superfluous ">" present in the string. I'll correct this 
for my next release. It's strange that it used to work; perhaps the latest 
portage release checks more carefully.

> Just so I could continue, I've removed the dev-lang/perl portion from
> the DateTime-Format-Builder-0.78.07.ebuild.  In order for emerge to
> continue, I had to hack the Manifest to the following:
> EBUILD DateTime-Format-Builder-0.78.07.ebuild 737 RMD160
[etc]

For general information (maybe you knew), the command:
ebuild ebuildwhichchanged.ebuild digest
regenerates the Manifest for you.

My next tree should also include ebuilds for the libdbi 0.8.2(-1) released 
earlier this week. And of course I'll be updating for 1.0.2.4. I'll have 
to evaluate the requirements for 1.1 before including it in a release as 
well.


Also take note that Email:Send is now in the official portage tree, 
altough they use a slightly different versioning scheme. I'll keep my own 
version for now, until theirs get into stable for x86 and amd64; when 
that happens, my version will automatically stop being used, anyway.

> Now... onward!

Thanks for your bug report!

-- 
Eric Lesage