[Bug c++/98528] New: g++.dg/modules/hello-1 FAILs

[Bug c++/98528] New: g++.dg/modules/hello-1 FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

            Bug ID: 98528
           Summary: g++.dg/modules/hello-1 FAILs
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ro at gcc dot gnu.org
                CC: nathan at gcc dot gnu.org
  Target Milestone: ---
            Target: i386-pc-solaris2.11, sparc-sun-solaris2.11,
                    i586-unknown-freebsd11.4, x86_64-unknown-freebsd11.4

The new g++.dg/modules/hello-1 test FAILs on Solaris and FreeBSD:

+UNRESOLVED: g++.dg/modules/hello-1 execute
+UNRESOLVED: g++.dg/modules/hello-1 execute
+FAIL: g++.dg/modules/hello-1 link
+FAIL: g++.dg/modules/hello-1 link

Undefined                       first referenced
 symbol                             in file
_ZL16__gthread_active               /var/tmp//ccAd8Hnb.o
ld: fatal: symbol referencing errors

This should be extern "C" __gthread_active instead, there's no mangled version
in libgcc.

Besides, the UNRESOLVED and FAIL lines are identical for both the -std=c++17
and
c++2a.  However, different tests are supposed to have different test names.

[Bug c++/98528] g++.dg/modules/hello-1 FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

Rainer Orth <ro at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |11.0

[Bug c++/98528] g++.dg/modules/hello-1 FAILs

[nathan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

--- Comment #1 from Nathan Sidwell <nathan at gcc dot gnu.org> ---
Could you provide the -save-temps .ii files please?  I suspect something
different in the solaris system headers.

it seems to be losing the c++17/2a distinguisher in the test name ...

[Bug c++/98528] g++.dg/modules/hello-1 FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

--- Comment #2 from Rainer Orth <ro at gcc dot gnu.org> ---
Created attachment 49888
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=49888&action=edit
i386-pc-solaris2.11 32-bit hello-1_[ab].ii

[Bug c++/98528] g++.dg/modules/hello-1 FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

--- Comment #3 from Rainer Orth <ro at gcc dot gnu.org> ---
(In reply to Nathan Sidwell from comment #1)
> Could you provide the -save-temps .ii files please?  I suspect something
> different in the solaris system headers.

Sure, attached.

> it seems to be losing the c++17/2a distinguisher in the test name ...

I believe I've seen the same on other failing modules tests on gcc-testresults
as well.

[Bug c++/98528] [11 Regression] g++.dg/modules/hello-1 FAILs

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
Any update?

[Bug c++/98528] [modules] g++.dg/modules/hello-1 FAILs

[nathan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

Nathan Sidwell <nathan at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
            Summary|[11 Regression]             |[modules]
                   |g++.dg/modules/hello-1      |g++.dg/modules/hello-1
                   |FAILs                       |FAILs
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-02-26
           Assignee|unassigned at gcc dot gnu.org      |nathan at gcc dot gnu.org

[Bug c++/98528] [modules] g++.dg/modules/hello-1 FAILs

[ro at CeBiTec dot Uni-Bielefeld.DE]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

--- Comment #5 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> ---
> --- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
> Any update?

At least on Solaris, the failures are gone since 20210113.

However, gcc-testresults still shows similar failures for
i586-unknown-freebsd12.2 and pru-unknown-elf, though it's impossible to
tell what exactly is wrong there.

[Bug c++/98528] [modules] g++.dg/modules/hello-1 FAILs

[nathan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

--- Comment #6 from Nathan Sidwell <nathan at gcc dot gnu.org> ---
do you have preprocessed source for those other failing platforms?  my guess is
they have different system headers, particularly around threading?

[Bug c++/98528] [modules] g++.dg/modules/hello-1 FAILs

[ro at CeBiTec dot Uni-Bielefeld.DE]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

--- Comment #7 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> ---
> --- Comment #6 from Nathan Sidwell <nathan at gcc dot gnu.org> ---
> do you have preprocessed source for those other failing platforms?  my guess is
> they have different system headers, particularly around threading?

Unfortunately not: I just saw the failures on gcc-testresults.  I don't
have any idea about PRU at all; maybe Gerald is able to help with FreeBSD?

[Bug c++/98528] [modules] g++.dg/modules/hello-1 FAILs

[nathan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98528

Nathan Sidwell <nathan at gcc dot gnu.org> changed:

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

--- Comment #8 from Nathan Sidwell <nathan at gcc dot gnu.org> ---
ok, thanks for the explanation.  I'll close this one as what was reported has
been fixed :)