[PATCH] test case fixes for Solaris

[Matt Mackall]

On Wed, 2013-08-21 at 14:41 -0700, Danek Duvall wrote:
> Matt Mackall wrote:
> 
> > On Wed, 2013-08-21 at 13:50 -0700, Danek Duvall wrote:
> > > # HG changeset patch
> > > # User Danek Duvall <danek.duvall at oracle.com>
> > > # Date 1377115773 25200
> > > # Node ID 5e7ec26663958bc8fc51ee3ab517274268b4f498
> > > # Parent  6ac206fb6f27492a98f46bbff090407ee1b1de72
> > > test case fixes for Solaris
> > > 
> > > Solaris versions of grep, diff, tail, and sed are different from the GNU
> > > versions, causing some test cases to fail.  Work around these differences
> > > in OS-independent ways.
> > 
> > Horrors. How did these persist so long?
> 
> Because it's been a long time since I've done an upgrade of Mercurial in
> Solaris, and I haven't been running the bits in between myself (and I guess
> no one else is running the tests on Solaris, or cares).

I would have thought the Mercurial-on-Solaris community would be larger
than the set using Cadmium.

> > I think we probably shouldn't accept these without corresponding checks
> > in contrib/check-code.py otherwise these will just be recurring problems
> > that people on more mainstream Unices won't notice. Please separate out
> > each class of breakage into a separate patch and include a corresponding
> > check.
> 
> Okay.  Can I consider this confirmation that the workarounds I used were
> appropriate?

Well I can't say I was super-thrilled by them, as we've been trying to
remove filtery-hackery. I'd rather just replace the problem bits with
approaches that avoid the problem. Diff doesn't work portably? If the
files are small, how about cat?

But even where that's done, we'll want the check-code rule to avoid
creep-back.

>   - I know that GNU head and tail at some point in the past started
>     disallowing args like "-8" and +8", and required the use of -n, but
>     then I think backed off of that change.  So I'm not sure what the right
>     thing is here.

Probably complains on some machines still.

>   - I have no idea how to catch the problem in test-parse-date.t, since the
>     contextual link between the missing newline and the sed that might not
>     be able to recognize the last line of a file without it is tenuous.  Is
>     that something we can just leave to more frequent testing?
> 
> > Also, please start your patch summaries with a relevant keyword, like
> > "solaris:".
> 
> Sure.
> 
> Thanks,
> Danek


-- 
Mathematics is the supreme nostalgia of our time.