[Bug c/100773] New: ice in operator[], at vec.h:890

[Bug c/100773] New: ice in operator[], at vec.h:890

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

            Bug ID: 100773
           Summary: ice in operator[], at vec.h:890
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
  Target Milestone: ---

Someplace between gcc trunk dated 20210524 (experimental) (0f3cd532fa02d178)
and 20210526 (experimental) (0eac9c60ac1f28ee), this C code:

fts_build_mail() {
  int attempts = 2;
  while (fts_build_mail_real())
    if (--attempts == 0)
      break;
}

with compiler flag -O2 does this: 

$ /home/dcb/gcc/results/bin/gcc -c -O2 -w bug720.c
during GIMPLE pass: evrp
fts-build-mail.c: In function ‘fts_build_mail’:
fts-build-mail.c:633:1: internal compiler error: in operator[], at vec.h:890
  633 | }
      | ^
0x1789c8e vec<range_def_chain::rdc, va_heap, vl_embed>::operator[](unsigned
int)
        ../../trunk.git/gcc/vec.h:890
0x1789c8e vec<range_def_chain::rdc, va_heap, vl_ptr>::operator[](unsigned int)
        ../../trunk.git/gcc/vec.h:1461
0x1789c8e range_def_chain::register_dependency(tree_node*, tree_node*,
basic_blo
ck_def*)
        ../../trunk.git/gcc/gimple-range-gori.cc:179
0x177f4e6 fold_using_range::range_of_range_op(irange&, gimple*, fur_source&)
        ../../trunk.git/gcc/gimple-range.cc:447

[Bug tree-optimization/100773] [12 Regression] ice in operator[], at vec.h:890

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|ice in operator[], at       |[12 Regression] ice in
                   |vec.h:890                   |operator[], at vec.h:890
   Target Milestone|---                         |12.0

[Bug tree-optimization/100773] [12 Regression] ice in operator[], at vec.h:890

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

--- Comment #1 from David Binderman <dcb314 at hotmail dot com> ---
I will have a go at doing a git bisect. 51 commits to start with.

[Bug tree-optimization/100773] [12 Regression] ice in operator[], at vec.h:890

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

--- Comment #2 from David Binderman <dcb314 at hotmail dot com> ---
Seems fine at 4b1987f8ad382a6c. Trying 10b286ce335cca13.

[Bug tree-optimization/100773] [12 Regression] ice in operator[], at vec.h:890

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

--- Comment #3 from David Binderman <dcb314 at hotmail dot com> ---
Seems fine at 10b286ce335cca13. Trying b6bdd7a4cb41ee05.

[Bug tree-optimization/100773] [12 Regression] ice in operator[], at vec.h:890

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

--- Comment #4 from David Binderman <dcb314 at hotmail dot com> ---
Seems bad at b6bdd7a4cb41ee05, so the current range is
10b286ce335cca13..b6bdd7a4cb41ee05, or about 6 revisions.

[Bug tree-optimization/100773] [12 Regression] ice in operator[], at vec.h:890

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to David Binderman from comment #4)
> Seems bad at b6bdd7a4cb41ee05, so the current range is
> 10b286ce335cca13..b6bdd7a4cb41ee05, or about 6 revisions.

It is most likely b6bdd7a4cb41ee05 which introduced the interaction with the
other range patches.

And maybe even fixed with 
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=7f0cfeb1ac580edf629874146b349fe7ad7e7718

[Bug tree-optimization/100773] [12 Regression] ice in operator[], at vec.h:890

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |marxin at gcc dot gnu.org
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED

--- Comment #6 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #5)
> (In reply to David Binderman from comment #4)
> > Seems bad at b6bdd7a4cb41ee05, so the current range is
> > 10b286ce335cca13..b6bdd7a4cb41ee05, or about 6 revisions.
> 
> It is most likely b6bdd7a4cb41ee05 which introduced the interaction with the
> other range patches.
> 
> And maybe even fixed with 
> https://gcc.gnu.org/git/?p=gcc.git;a=commit;
> h=7f0cfeb1ac580edf629874146b349fe7ad7e7718

Yes, this fixed that.

[Bug tree-optimization/100773] [12 Regression] ice in operator[], at vec.h:890

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100773

--- Comment #7 from David Binderman <dcb314 at hotmail dot com> ---
(In reply to Martin Liška from comment #6)
> Yes, this fixed that.

I've just checked and yes, I agree.