This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [gfortran,patch] Complete libgfortran namespace separation and version number
On Sat, Jan 20, 2007 at 04:01:33PM +0100, FX Coudert wrote:
> ping**2
>
> Since that time, the libgfortran version bump has been commited as
> part of the libgfortran reorganisation (rev. 120869). Please find
> attached the updated patch (wrt to today's retrunk). It is currently
> bootstrapping and regtesting on i686-linux, if you don't hear from me
> during the next two hours, then it will have bootstrapped and
> regtested fine :)
>
> [ http://gcc.gnu.org/ml/gcc-patches/2007-01/msg00567.html ]
>
> >:ADDPATCH fortran:
> >
> >Attached patch does the following three things:
> > - add an undocumented -fallow-leading-underscore option to gfortran,
> >for internal use only, until the ISO_C_BINDING patch is commited, that
> >allows symbol names to begin with an underscore (note: it does
> >explicitely ICE if one of these variables has implicit type)
> > - use this to prefix all libgfortran symbols names with _gfortran_
> >(this was not the case before, especially for specifics), achieving
> >complete namespace separation for the library
> > - bump the libgfortran version number to 4.3.0
> >
> >Bootstrapped & regtested on i686-linux. OK for mainline?
>
The patch is OK with two minor changes.
Index: gcc/fortran/symbol.c
===================================================================
--- gcc/fortran/symbol.c (revision 121000)
+++ gcc/fortran/symbol.c (working copy)
@@ -202,6 +202,12 @@
char letter;
letter = sym->name[0];
+
+ if (gfc_option.flag_allow_leading_underscore && letter == '_')
+ gfc_internal_error ("Option -fallow_leading_underscore is for use only by "
+ "gfortran developpers, and should not be used for "
+ "implicitely typed variables");
+
s/developpers/developers
s/implicitely/implicitly
--
Steve