This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: NetWare update

>> >> gcc/testsuite/
>> >> 	* gcc.dg/20040813-1.c: xfail and don't execute dg-final for
>> >> 	*-*-netware.
>> >
>> >See the immediately preceeding skip-if.  Means you don't need
>> >the dg-error line at the end.
>> I'm not sure I understand what you're trying to say here.
>Seems clear to me.  Just add *-*-netware to make that line say:
>/* { dg-skip-if "No stabs" { mmix-*-* *-*-aix* *-*-netware } { "*" } {
"" } } */

This is what I considered obvious and thus tried first. It didn't work
as appearantly intended, as described earlier (i.e. it didn't prevent
dg-final from getting executed).
My non-understanding was mainly due to the reference to dg-error: The
test in question does not contain dg-error at all.

>> What I found
>> is that the dg-skip-if simply doesn't work
>It works for me.  Perhaps communication would be helped if you
>post the actual edit and "make check" output and the gcc.log for
>running just this test, perhaps with additional flags "-v -v -v"?
>(not necessary if the above line works for you, methinks)

>I get just a single line saying "UNSUPPORTED: gcc.dg/20040813-1.c"
>in gcc.log, with no "stdout" message.
>I don't know what could be wrong.  Are you using dejagnu-1.4.4
>as is the mandated version?

I've been using 1.4.3. It may be considered fine to hide somewhere in
the documentation that as of a certain date 1.4.4 is required, but it'd
be much more helpful if the testsuite run itself checked for this rather
than producing unpredictable results. (Beyond that I think that a
tolerance in revision number should generally be acceptable, but of
course there may be exclusions where something absolutely can't work
without the newest-latest-and-greatest).

Nevertheless it doesn't work any different with 1.4.4.

I'm attaching gcc.log as requested; I can't attach stdout resulting
from "-v -v -v" due to size limitations this list places on mails.

To me it boils down to (as said before) compile_flags not being set at
the time dg-skip-if (and from there check_conditional_xfail, where
compile_flags are used) gets evaluated, as can be seen in the attached
file, where "Compiler flags are: " is followed by nothing, which of
course doesn't match against "* * *". [Note that this is a cross
compiler, and there is no remote target (that is, no execution tests
ever get executed). But I think this should not affect the compile-only
test here.]


Attachment: gcc.log
Description: Binary data

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]