[Bug middle-end/110200] New: genmatch generating questionable code with convert and !

[Bug middle-end/110200] New: genmatch generating questionable code with convert and !

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110200

            Bug ID: 110200
           Summary: genmatch generating questionable code with convert and
                    !
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Keywords: build
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: pinskia at gcc dot gnu.org
  Target Milestone: ---

Add these two patterns to match.pd:
```
/* (convert)(a ? b : CST) -> (a ? (convert)b : CST1) */
(simplify
 (convert (cond @0 @1 CONSTANT_CLASS_P@2))
 (cond @0 (convert @1) (convert! @2)))

 /* (convert)(a ? CST : b) -> (a ? CST1 : (convert)b) */
(simplify
 (convert (cond @0 CONSTANT_CLASS_P@1 @2))
 (cond @0 (convert! @1) (convert @2)))
```

And the code produced by genmatch for generic is:
```
                {
                  tree _o1[1], _r1;
                  _o1[0] = captures[2];
                  if (TREE_TYPE (_o1[0]) != type)
                    _r1 = fold_build1_loc (loc, NOP_EXPR, type, _o1[0]);
                    if (EXPR_P (_r1))
                      goto next_after_fail867;
                  else
                    _r1 = _o1[0];
                  res_op2 = _r1;
                }

```

I think there is a missing `{`/`}` pair around the fold_build1_loc/if/goto
statements.

[Bug middle-end/110200] genmatch generating questionable code with convert and !

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110200

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2023-06-12
             Status|UNCONFIRMED                 |ASSIGNED
     Ever confirmed|0                           |1
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed, mine.

[Bug middle-end/110200] genmatch generating questionable code with convert and !

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110200

--- Comment #2 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Richard Biener <rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:820d1aec89c43dbbc70d3d0b888201878388454c

commit r14-1706-g820d1aec89c43dbbc70d3d0b888201878388454c
Author: Richard Biener <rguenther@suse.de>
Date:   Mon Jun 12 10:17:26 2023 +0200

    middle-end/110200 - genmatch force-leaf and convert interaction

    The following fixes code GENERIC generation for (convert! ...)
    which currently generates

      if (TREE_TYPE (_o1[0]) != type)
        _r1 = fold_build1_loc (loc, NOP_EXPR, type, _o1[0]);
        if (EXPR_P (_r1))
          goto next_after_fail867;
      else
        _r1 = _o1[0];

    where obviously braces are missing.

            PR middle-end/110200
            * genmatch.cc (expr::gen_transform): Put braces around
            the if arm for the (convert ...) short-cut.

[Bug middle-end/110200] genmatch generating questionable code with convert and !

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110200

--- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-13 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:73ae34bb693038829c05bed30d7ac623e67bde2e

commit r13-7439-g73ae34bb693038829c05bed30d7ac623e67bde2e
Author: Richard Biener <rguenther@suse.de>
Date:   Mon Jun 12 10:17:26 2023 +0200

    middle-end/110200 - genmatch force-leaf and convert interaction

    The following fixes code GENERIC generation for (convert! ...)
    which currently generates

      if (TREE_TYPE (_o1[0]) != type)
        _r1 = fold_build1_loc (loc, NOP_EXPR, type, _o1[0]);
        if (EXPR_P (_r1))
          goto next_after_fail867;
      else
        _r1 = _o1[0];

    where obviously braces are missing.

            PR middle-end/110200
            * genmatch.cc (expr::gen_transform): Put braces around
            the if arm for the (convert ...) short-cut.

    (cherry picked from commit 820d1aec89c43dbbc70d3d0b888201878388454c)

[Bug middle-end/110200] genmatch generating questionable code with convert and !

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110200

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:d5f72834a4817b180625a540b99f5c1934c2e0b8

commit r12-9694-gd5f72834a4817b180625a540b99f5c1934c2e0b8
Author: Richard Biener <rguenther@suse.de>
Date:   Mon Jun 12 10:17:26 2023 +0200

    middle-end/110200 - genmatch force-leaf and convert interaction

    The following fixes code GENERIC generation for (convert! ...)
    which currently generates

      if (TREE_TYPE (_o1[0]) != type)
        _r1 = fold_build1_loc (loc, NOP_EXPR, type, _o1[0]);
        if (EXPR_P (_r1))
          goto next_after_fail867;
      else
        _r1 = _o1[0];

    where obviously braces are missing.

            PR middle-end/110200
            * genmatch.cc (expr::gen_transform): Put braces around
            the if arm for the (convert ...) short-cut.

    (cherry picked from commit 820d1aec89c43dbbc70d3d0b888201878388454c)

[Bug middle-end/110200] genmatch generating questionable code with convert and !

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110200

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:5bf1c33a5e619564e8602601cc1beebd99e1fb9e

commit r11-10856-g5bf1c33a5e619564e8602601cc1beebd99e1fb9e
Author: Richard Biener <rguenther@suse.de>
Date:   Mon Jun 12 10:17:26 2023 +0200

    middle-end/110200 - genmatch force-leaf and convert interaction

    The following fixes code GENERIC generation for (convert! ...)
    which currently generates

      if (TREE_TYPE (_o1[0]) != type)
        _r1 = fold_build1_loc (loc, NOP_EXPR, type, _o1[0]);
        if (EXPR_P (_r1))
          goto next_after_fail867;
      else
        _r1 = _o1[0];

    where obviously braces are missing.

            PR middle-end/110200
            * genmatch.c (expr::gen_transform): Put braces around
            the if arm for the (convert ...) short-cut.

    (cherry picked from commit 820d1aec89c43dbbc70d3d0b888201878388454c)

[Bug middle-end/110200] genmatch generating questionable code with convert and !

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110200

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed.

[Bug middle-end/110200] genmatch generating questionable code with convert and !

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110200

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |11.4.1, 12.3.1, 13.1.1
      Known to fail|                            |11.4.0, 12.3.0, 13.1.0
   Target Milestone|---                         |11.5