[Patch, fortran] PR30878 - Rejects function f1; namelist /nml/ f1

[Paul Richard Thomas]

Duuuh!

On 5/9/07, Paul Richard Thomas <paul.richard.thomas@gmail.com> wrote:
> :ADDPATCH fortran:
>
> This PR comes about because gfortran throws an error upon an implicit
> result variable appearing in a namelist.  The patch consists of two
> steps: (i) Prevent the emission of the compilation error in symbol.c
> and resolve.c; and (ii) Add the requisite code to trans-io.c.  The
> comments and the ChangeLogs are sufficiently descriptive to follow
> what is going on.  The testcase is a mutilation of the reporter's.
>
> Bootstrapped and regtested on x86_ia64 - OK for trunk?
>
> Paul
>
> PS At the time of my first contribution to gfortran, the testsuite
> contain two testcases for NAMELIST.  There are now 33 of the
> blighters.  Please accept my apologies:)
>
> 2007-05-09  Paul Thomas  <pault@gcc.gnu.org>
>
>        PR fortran/30878
>        * resolve.c (resolve_fl_namelist): It is not an error if the
>        namelist element is the result variable of the enclosing
>        function.  Search for the symbol in current and all parent
>        namespaces for a potential conflict.
>        * symbol.c (check_conflict): Remove the conflict between
>        'in_namelist' and 'FL_PROCEDURE' because the symbol info
>        is not available to exclude function result variables.
>        * trans-io.c (nml_get_addr_expr): Use the fake result decl
>        if the symbol is an implicit result variable.
>
> 2007-05-09  Paul Thomas  <pault@gcc.gnu.org>
>
>        PR fortran/30878
>        * gfortran.dg/namelist_29.f90: New test.
>


-- 
"Success is the ability to go from one failure to another with no loss
of enthusiasm."  -  Winston Churchill
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: pr30878.diff
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20070509/f4221be1/attachment.ksh>