[Bug fortran/103471] New: [9/10/11/12 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[Bug fortran/103471] New: [9/10/11/12 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

            Bug ID: 103471
           Summary: [9/10/11/12 Regression] ICE in gfc_typenode_for_spec,
                    at fortran/trans-types.c:1114
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Some time ago, r8 was corrected, between 20191115 (ICE)
and 20200507 (error), but the ICE appeared again in r9 :
(follow-up of pr93484)


$ cat z1.f90
program p
   implicit none
   print *, [real(x(k))]
end


$ gfortran-8 -c z1.f90
z1.f90:3:21:

    print *, [real(x(k))]
                     1
Error: Symbol 'k' at (1) has no IMPLICIT type


$ gfortran-12-20211128 -c z1.f90
z1.f90:1:9:

    1 | program p
      |         1
internal compiler error: in gfc_typenode_for_spec, at
fortran/trans-types.c:1114
0x8d8bb0 gfc_typenode_for_spec(gfc_typespec*, int)
        ../../gcc/fortran/trans-types.c:1114
0x8d8db6 gfc_sym_type(gfc_symbol*, bool)
        ../../gcc/fortran/trans-types.c:2272
0x875938 gfc_get_symbol_decl(gfc_symbol*)
        ../../gcc/fortran/trans-decl.c:1813
0x87830f generate_local_decl
        ../../gcc/fortran/trans-decl.c:5848
0x834652 do_traverse_symtree
        ../../gcc/fortran/symbol.c:4174
0x87944c generate_local_vars
        ../../gcc/fortran/trans-decl.c:6054
0x87944c gfc_generate_function_code(gfc_namespace*)
        ../../gcc/fortran/trans-decl.c:7579
0x7fc90e translate_all_program_units
        ../../gcc/fortran/parse.c:6638
0x7fc90e gfc_parse_file()
        ../../gcc/fortran/parse.c:6925
0x84967f gfc_be_parse_file
        ../../gcc/fortran/f95-lang.c:216

[Bug fortran/103471] [9/10/11/12 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |9.5
           Priority|P3                          |P4

[Bug fortran/103471] [9/10/11/12 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |markeggleston at gcc dot gnu.org,
                   |                            |marxin at gcc dot gnu.org
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2021-11-30
     Ever confirmed|0                           |1

--- Comment #1 from Martin Liška <marxin at gcc dot gnu.org> ---
Started with r7-4028-g87c9fca50cbe7ca9, was fixed in r10-7369-gc38daa7976886a59
and started again in r11-693-g3ea6977d0f1813d9.

[Bug fortran/103471] [9/10/11/12 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kargl at gcc dot gnu.org

--- Comment #2 from kargl at gcc dot gnu.org ---
An error is being queued in symbol.c(gfc_set_default_type), but
error reporting has been disabled for some reason.  Changing 
gfc_error() to gfc_error_now() restores emitting the error
message, but is this correct!

diff --git a/gcc/fortran/symbol.c b/gcc/fortran/symbol.c
index 179f6029ca3..43691710120 100644
--- a/gcc/fortran/symbol.c
+++ b/gcc/fortran/symbol.c
@@ -303,11 +303,11 @@ gfc_set_default_type (gfc_symbol *sym, int error_flag,
gfc_namespace *ns)
        {
          const char *guessed = lookup_symbol_fuzzy (sym->name, sym);
          if (guessed)
-           gfc_error ("Symbol %qs at %L has no IMPLICIT type"
+           gfc_error_now ("Symbol %qs at %L has no IMPLICIT type"
                       "; did you mean %qs?",
                       sym->name, &sym->declared_at, guessed);
          else
-           gfc_error ("Symbol %qs at %L has no IMPLICIT type",
+           gfc_error_now ("Symbol %qs at %L has no IMPLICIT type",
                       sym->name, &sym->declared_at);
          sym->attr.untyped = 1; /* Ensure we only give an error once.  */
        }

[Bug fortran/103471] [10/11/12/13 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|9.5                         |10.4

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 9 branch is being closed

[Bug fortran/103471] [10/11/12/13 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.4                        |10.5

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 10.4 is being released, retargeting bugs to GCC 10.5.

[Bug fortran/103471] [11/12/13/14 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.5                        |11.5

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 10 branch is being closed.

[Bug fortran/103471] [11/12/13/14 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

Paul Thomas <pault at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pault at gcc dot gnu.org

--- Comment #6 from Paul Thomas <pault at gcc dot gnu.org> ---
The current mainline and 13-branches now power through to the fatal error at
trans-decl.cc:1800.

Paul

[Bug fortran/103471] [11/12/13/14 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |anlauf at gcc dot gnu.org

--- Comment #7 from anlauf at gcc dot gnu.org ---
(In reply to Paul Thomas from comment #6)
> The current mainline and 13-branches now power through to the fatal error at
> trans-decl.cc:1800.

True, but the location does not look nice:

pr103471.f90:1:9:

    1 | program p
      |         1
Fatal Error: k at (1) has no default type
compilation terminated.


With the following tentative - but otherwise untested - change:

diff --git a/gcc/fortran/trans-decl.cc b/gcc/fortran/trans-decl.cc
index e160c5c98c1..602f35f0cab 100644
--- a/gcc/fortran/trans-decl.cc
+++ b/gcc/fortran/trans-decl.cc
@@ -1797,7 +1797,9 @@ gfc_get_symbol_decl (gfc_symbol * sym)
     }

   if (sym->ts.type == BT_UNKNOWN)
-    gfc_fatal_error ("%s at %C has no default type", sym->name);
+//    gfc_fatal_error ("%s at %C has no default type", sym->name);
+    gfc_fatal_error ("%s at %L has no default type", sym->name,
+                    &sym->declared_at);

   if (sym->attr.intrinsic)
     gfc_internal_error ("intrinsic variable which isn't a procedure");


we are back at:

pr103471.f90:3:21:

    3 |    print *, [real(x(k))]
      |                     1
Fatal Error: k at (1) has no default type
compilation terminated.


This looks more user friendly.

[Bug fortran/103471] [11/12/13/14 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

--- Comment #8 from Paul Thomas <pault at gcc dot gnu.org> ---

> This looks more user friendly.

Also true. I have put it on to regtest but I think that it might be a good idea
to understand how the symbol evades resolution :-)

Paul

[Bug fortran/103471] [11/12/13/14 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

--- Comment #9 from Paul Thomas <pault at gcc dot gnu.org> ---

> This looks more user friendly.

Also true. I have put it on to regtest but I think that it might be a good idea
to understand how the symbol evades resolution :-)

Paul

[Bug fortran/103471] [11/12/13/14 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

--- Comment #10 from Paul Thomas <pault at gcc dot gnu.org> ---
(In reply to Paul Thomas from comment #9)
> > This looks more user friendly.
> 
> Also true. I have put it on to regtest but I think that it might be a good
> idea to understand how the symbol evades resolution :-)
> 
> Paul

It regtests OK.

Paul

[Bug fortran/103471] [11/12/13/14 Regression] ICE in gfc_typenode_for_spec, at fortran/trans-types.c:1114

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

--- Comment #11 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Paul Thomas <pault@gcc.gnu.org>:

https://gcc.gnu.org/g:f17d31e709af9b2d488adecd6cd040dfc1f23b04

commit r14-10059-gf17d31e709af9b2d488adecd6cd040dfc1f23b04
Author: Paul Thomas <pault@gcc.gnu.org>
Date:   Sun Apr 21 17:24:24 2024 +0100

    Fortran: Detect 'no implicit type' error in right place [PR103471]

    2024-04-21  Paul Thomas  <pault@gcc.gnu.org>

    gcc/fortran
            PR fortran/103471
            * resolve.cc (resolve_actual_arglist): Catch variables silently
            set as untyped, resetting the flag so that gfc_resolve_expr can
            generate the no implicit type error.
            (gfc_resolve_index_1): Block index expressions of unknown type
            from being converted to default integer, avoiding the fatal
            error in trans-decl.cc.
            * symbol.cc (gfc_set_default_type): Remove '(symbol)' from the
            'no IMPLICIT type' error message.
            * trans-decl.cc (gfc_get_symbol_decl): Change fatal error locus
            to that of the symbol declaration.
            (gfc_trans_deferred_vars): Remove two trailing tabs.

    gcc/testsuite/
            PR fortran/103471
            * gfortran.dg/pr103471.f90: New test.

[Bug fortran/103471] Missed no IMPLICIT type errors

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103471

Paul Thomas <pault at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[11/12/13/14 Regression]    |Missed no IMPLICIT type
                   |ICE in                      |errors
                   |gfc_typenode_for_spec, at   |
                   |fortran/trans-types.c:1114  |

--- Comment #12 from Paul Thomas <pault at gcc dot gnu.org> ---
The regression has been cured in the above commit and some of the missed error
messages have been fixed.

Compared with the submission to the list and discussed with Harald Anlauf, an
extra chunk has been added in resolve.cc(resolve_actual_arglist) to catch
untyped actual arguments.

See the testcase commented out line for an example interference between
different statements with untyped symbols.

I am keeping this one open having changed the summary.

Paul