PING: [patch, fortran] PR 27997: Implement F2003-style array constructor with typespec

FX fxcoudert@gmail.com
Thu May 8 11:07:00 GMT 2008


>>>> Some of these fail on x86_64-pc-mingw32:
>
>  ping..

I was waiting for Daniel to look through it, as he's also working on
another patch. A quick look shows that all tests fail due to errors
similar to:

/tmp/build/gcc-svn/gcc/gcc/testsuite/gfortran.dg/array_constructor_type_5.f03:12:
Warning: Nonconforming tab character at (1)

There indeed appears to be tabs in some of these testcases. As Daniel
had a good testcase (he didn't see the failure), I think the diff or
his mailer screwed it (that has happened in the past). Another
possibility, which has (I think) happened to a lot of us in the past
is the deadly combination below:
  - make your patch & regtest it, it goes fine
  - write your testcases outside gfortran.dg (or after regtesting),
and run them manually: they go fine
These testcases might still fail when run in the testsuite framework,
because the testsuite is compiled with the -pedantic option...

Daniel, please make sure to check what's the issue, and if it's none
of the two above, we'll investigate ;-)

FX


-- 
FX Coudert
http://www.homepages.ucl.ac.uk/~uccafco/



More information about the Gcc-patches mailing list