Re: [patch, fortan] PR87103 - [OOP] ICE in gfc_new_symbol() due to overlong symbol name

[Andrew Benson <abenson at carnegiescience dot edu>]

On Wednesday, September 5, 2018 12:35:04 PM PDT Bernhard Reutner-Fischer 
wrote:
> On Wed, 5 Sep 2018 at 03:30, Jerry DeLisle <jvdelisle@charter.net> wrote:
> > On 09/04/2018 10:43 AM, Bernhard Reutner-Fischer wrote:
> > > On Tue, 4 Sep 2018 at 18:43, Andrew Benson <abenson@carnegiescience.edu> 
wrote:
> > >> As suggested by Janus, PR87103 is easily fixed by the attached patch
> > >> which
> > >> increases GFC_MAX_SYMBOL_LEN to 76 (sufficient to hold the maximum
> > >> allowed F08 symbol length of 63, plus a null terminator, plus the
> > >> "__tmp_class_" prefix).> > 
> > > This is so much wrong.
> > > Note that this will be fixed properly by the changes contained in the
> > > https://gcc.gnu.org/git/?p=gcc.git;a=shortlog;h=refs/heads/aldot/fortran
> > > -fe-stringpool branch.
> > > There we keep the GFC_MAX_SYMBOL_LEN at 63 proper but use an internal
> > > buffer double that size which in turn is sufficient to hold all
> > > compiler-generated identifiers.
> > > See gfc_get_string() even in current TOT.
> > > 
> > > Maybe we should bite the bullet and start to merge the stringpool
> > > changes now instead of this hack?
> > 
> > It all makes sense to me, please proceed. (my 2 cents worth)
> 
> Ok so i will reread the fortran-fe-stringpool series and submit it
> here for review.
> 
> Let's return to the issue at hand for a moment, though.
> I tested the attached alternate fix on top of the
> fortran-fe-stringpool branch where it fixes PR87103.
> Maybe somebody has spare cycles to test it on top of current trunk?
> 
> thanks,
> 
> [PATCH,FORTRAN] PR87103: Remove max symbol length check from gfc_new_symbol
> 
> gfc_match_name does check for too long names already. Since
> gfc_new_symbol is also called for symbols with internal names containing
> compiler-generated prefixes, these internal names can easily exceed the
> max_identifier_length mandated by the standard.
> 
> gcc/fortran/ChangeLog
> 
> 2018-09-04  Bernhard Reutner-Fischer  <aldot@gcc.gnu.org>
> 
> PR fortran/87103
> * expr.c (gfc_check_conformance): Check vsnprintf for truncation.
> * iresolve.c (gfc_get_string): Likewise.
> * symbol.c (gfc_new_symbol): Remove check for maximum symbol
> name length.  Remove redundant 0 setting of new calloc()ed
> gfc_symbol.

This patch tests successfully on the current trunk for me.

-- 

* Andrew Benson: http://users.obs.carnegiescience.edu/abenson/contact.html

* Galacticus: https://bitbucket.org/abensonca/galacticus