[PATCH] tests: do not assume run-tests.py sets $BINDIR to a useful value

[Greg Ward]

On Tue, May 19, 2009 at 4:30 AM, Mads Kiilerich <mads at kiilerich.com> wrote:
> On 05/19/2009 04:33 AM, Greg Ward wrote:
>>
>> # HG changeset patch
>> # User Greg Ward<greg-hg at gerg.ca>
>> # Date 1242700406 14400
>> # Node ID f38ec158a7c087c5e8c6352fbdd093e4e658b9fb
>> # Parent  252232621165917755727729c7f0b9a1f1263668
>> tests: do not assume run-tests.py sets $BINDIR to a useful value.
>>
>> (It's not true when someone passes an incorrect --with-hg, and right
>> now nothing checks that --with-hg is correct.  And I'm about to send
>> patches that clarify use of --with-hg and drop $BINDIR from the
>> environment.)
[...]
> These PATH settings were introduced in.
> http://hg.intevation.org/mercurial/crew/rev/d596b1f2935a and
> http://hg.intevation.org/mercurial/crew/rev/660c8dd44060 in order to ensure
> that PATH didn't contain p4 or hgmerge and that hg handles this correctly.
>
> It was perhaps an unintended side-effect that BINDIR could be used for that,
> but the need must be adressed somehow.

Hmmmm.  I guess the easiest thing to do is fix my patch to
run-tests.py to restore BINDIR to the environment.  Alternately, what
about

hg=`which hg`
PATH=`dirname $hg` hg ...

in those two test scripts?

[...thinks for a moment...]

Nahh.  Dumb idea.  I'll restore $BINDIR.  Please disregard this patch.

Greg