[Patch, fortran] PR86248 - [7/8/9/10 Regression] LEN_TRIM in specification expression causes link failure

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

[Patch, fortran] PR86248 - [7/8/9/10 Regression] LEN_TRIM in specification expression causes link failure

Paul Richard Thomas
As far as I can tell, this is a 6-regression as well - ***sigh***

The patch is fundamentally very simple. Symbols that were marked with
the fn_result_spec flag that really were module parameters were having
the wrong name mangling applied to them. The rest of the patch is a
tidy up.

Regtested on FC30/x86_64 - OK for all the branches after a bedding in
period on trunk?

Cheers

Paul

2019-10-26  Paul Thomas  <[hidden email]>

    PR fortran/86248
    * resolve.c (flag_fn_result_spec): Correct a typo before the
    function declaration.
    * trans-decl.c (gfc_sym_identifier): Boost the length of 'name'
    to allow for all variants. Simplify the code by using a pointer
    to the symbol's proc_name and taking the return out of each of
    the conditional branches. Allow symbols with fn_result_spec set
    that do not come from a procedure namespace and have a module
    name to go through the non-fn_result_spec branch.

2019-10-26  Paul Thomas  <[hidden email]>

    PR fortran/86248
    * gfortran.dg/char_result_19.f90 : New test.
    * gfortran.dg/char_result_mod_19.f90 : Module for the new test.

submit.diff (6K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [Patch, fortran] PR86248 - [7/8/9/10 Regression] LEN_TRIM in specification expression causes link failure

Steve Kargl
On Sat, Oct 26, 2019 at 07:39:55PM +0100, Paul Richard Thomas wrote:

> As far as I can tell, this is a 6-regression as well - ***sigh***
>
> The patch is fundamentally very simple. Symbols that were marked with
> the fn_result_spec flag that really were module parameters were having
> the wrong name mangling applied to them. The rest of the patch is a
> tidy up.
>
> Regtested on FC30/x86_64 - OK for all the branches after a bedding in
> period on trunk?
>

OK.  Note you cannot wait too long for 7 or 9.  The 7 branch
is being closed, and the 9-branch is close to its next release.

--
Steve