Mercurial 2.0 call for testing

[Danek Duvall]

Mads Kiilerich wrote:

> Danek Duvall wrote, On 10/17/2011 12:48 AM:
> >Matt Mackall wrote:
> >
> >>http://mercurial.selenic.com/release/mercurial-2.0-rc.tar.gz
> >I'm running on Solaris 11 (latest internal bits, which are pretty close to
> >final).
> 
> >test-hup is failing the usual way for me, where the "wait" ends up in the
> >output after "killed!", which I've never considered to be a real problem.
> 
> I have "never" seen that test fail on solaris, so it might be a real
> and unknown issue.
>
> >The cause seems to be using /bin/sh (ksh93) instead of bash as
> >HGTEST_SHELL.
> 
> There shouldn't be any bashisms in the tests - we just assume that
> /bin/sh is something like a posix bourne shell ...

Maybe, but honestly, I've seen enough weird problems with our ksh93 that I
don't completely trust it.  And zsh works just fine for this test, too, not
just bash.

> >test-run-tests has two additional lines in its output:
> >
> >     \x1b[?1034h (no-eol) (esc)
> >
> >which seem to be the result of some termcap weirdness.  If I set TERM=dumb
> >in the environment for that test, it passes.
> >
> >And test-url seems to have an extra blank line in its output at the top.
> >Also seems to be averted with TERM=dumb, and also doesn't seem serious.
> 
> I haven't seen these failures, but there might be a need for something
> similar to the first lines of test-doctest.py (because of strange things
> happening in pythons doctest).

I can submit a patch to do that, if there's interest.

Danek