On Feb 16, 2012, at 8:05 AM, Andrew Benton wrote:
> On Thu, 16 Feb 2012 15:20:49 +0000
> Ken Moffat <[email protected]> wrote:
>
>> On Thu, Feb 16, 2012 at 07:46:07AM -0700, [email protected] wrote:
>>> +
>>> + <para>This package does not have a working testsuite.</para>
>>> +
>>
>> I was originally going to ask to have it minuted that this was
>> where I cracked and told the truth about a testsuite, after some
>> circumlocutions in earlier commits :-)
>>
>> And I was going to plead in mitigation that I've had it 'up to
>> here' with packages that need to be installed before their tests
>> will run (or, perhaps, since this is gnome, they expect you to be
>> running the previous stable version). For this one, I even went
>> back to a gnome session and reinstalled libpeas (the gnome devs
>> recommend seed as a dep for libpeas : that is circular, since seed
>> requires peas), but still there were two failures in the seed
>> tests.
>>
>> But instead, I'm going to suggest that we just stop mentioning
>> testsuites in BLFS.
>
> Or at least make it optional?
In principle, I think testsuites are awesome; I like the confidence they give
that the system works as advertised.
I mean, if you're compiling from source, you need some sort of validation that
the system works. I'm guessing you've all run into some practical downside.
What is the case against them in BLFS? I would think a
system-built-from-source would want some sort of validation. For my own
scripts, I solve the problem by adding comments in front of the tests. When
I'm debugging my scripts, the comments stay in. When I'm making a "master"
build (to deploy across VMs, for example), I do the full tested build.
Q
--
http://linuxfromscratch.org/mailman/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
