On Tue, 28 Aug 2007, James Keenan via RT wrote:
> On Tue Aug 28 19:35:12 2007, [EMAIL PROTECTED] wrote:
> > James E Keenan wrote:
>
> > I'll put a TODO block around it until I can get tuit.
> >
>
> TODO block implemented in r20894.
Well, that TODO block is around the wrong test, (it's the test before
that) but it doesn't matter -- the test failure would also need to be
trapped by an eval. Without the eval, the test file still dies early.
The comment 'not working for all C compilers' is misleading too. The
problem is that the test assumes that passing '/usr/bin/gcc-3.3' to the
hints file to test for the compiler is valid and will pass. That's simply
not true. The hints file test that's getting triggered runs after the user
has selected a compiler. The hints file is allowed to assume the user
either selected the correct compiler or needs to know if it's the wrong
compiler. You can't feed it a random wrong value and expect it to still
pass.
You can test this for yourself by applying the linux hints file patch I
posted in RT #45025 and changinc CC to some compiler not present on your
system, e.g. /usr/bin/gcc-99.99, or some such.
--
Andy Dougherty [EMAIL PROTECTED]
