[Bug modula2/113749] New: [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[Bug modula2/113749] New: [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[doko at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

            Bug ID: 113749
           Summary: [14 Regression] m2 enabled build times out on i686-gnu
                    (GNU Hurd)
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: modula2
          Assignee: gaius at gcc dot gnu.org
          Reporter: doko at gcc dot gnu.org
  Target Milestone: ---

seen with trunk, works with the gcc-13 branch on i686-gnu:

[...]
test -d m2/gm2-libs-boot || /bin/bash ../../src/gcc/../mkinstalldirs
m2/gm2-libs-boot
m2/boot-bin/mc --olang=c++ --h-file-prefix=G -I../../src/gcc/m2/gm2-libs
-I../../src/gcc/m2/gm2-comp
iler -I../../src/gcc/m2/gm2-libiberty -I../../src/gcc/m2/gm2-gcc --quiet 
--gcc-config-system -o=m2/
gm2-libs-boot/GASCII.h ../../src/gcc/m2/gm2-libs/ASCII.def
E: Build killed with signal TERM after 180 minutes of inactivity

[Bug modula2/113749] [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |14.0

[Bug modula2/113749] [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[gaius at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

--- Comment #1 from Gaius Mulley <gaius at gcc dot gnu.org> ---
Thanks for the bug report - I suspect this requires the same bug fix as
PR 112920 - gm2 hangs in the type resolver.  Although the hang in this PR is in
the bootstrap tool - they both use the same type resolving algorithm.

[Bug modula2/113749] [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[gaius at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

Gaius Mulley <gaius at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2024-02-22
             Status|UNCONFIRMED                 |ASSIGNED
     Ever confirmed|0                           |1

--- Comment #2 from Gaius Mulley <gaius at gcc dot gnu.org> ---
Confirmed.

[Bug modula2/113749] [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[samuel.thibault@ens-lyon.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

--- Comment #3 from Samuel Thibault <samuel.thibault@ens-lyon.org> ---
I don't think it is a regression. We noticed it only recently in Debian only
because the configuration files got bogus. See for instance

https://buildd.debian.org/status/fetch.php?pkg=gcc-13&arch=hurd-i386&ver=13.2.0-10&stamp=1705711375&raw=

which fails the same:

m2/boot-bin/mc --olang=c++ --h-file-prefix=G -I../../src/gcc/m2/gm2-libs
-I../../src/gcc/m2/gm2-compiler -I../../src/gcc/m2/gm2-libiberty -I
E: Build killed with signal TERM after 180 minutes of inactivity

And we can trace it back to this commit in debian:

commit ac0e3d134553e01c04b6a1b7f9c855fcb747dd6f
Author: Matthias Klose <doko@ubuntu.com>
Date:   Tue Oct 8 10:34:25 2019 +0200

      * Disable gm2 on hurd-i386, mc hangs there (Samuel Thibault). Closes:
#940600.

That being said, it'd still be good to fix it, sure :)

[Bug modula2/113749] [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[gaius at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

--- Comment #4 from Gaius Mulley <gaius at gcc dot gnu.org> ---
Created attachment 57516
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57516&action=edit
Proposed fix

Here is a proposed patch.  The bug fix changes the FIO module to use the target
O_RDONLY, O_WRONLY, SEEK_SET and SEEK_END (now available from the module
wrapc).

I've rebuilt the bootstrap tools mc and pge as they have their own wrapc and C
translations of FIO.

[Bug modula2/113749] [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[gaius at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

Gaius Mulley <gaius at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #57516|0                           |1
        is obsolete|                            |

--- Comment #5 from Gaius Mulley <gaius at gcc dot gnu.org> ---
Created attachment 57517
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57517&action=edit
Proposed fix v2

Correction - to include wrapc.o when linking the pge tool.

[Bug modula2/113749] [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

--- Comment #6 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Gaius Mulley <gaius@gcc.gnu.org>:

https://gcc.gnu.org/g:d1b241b9506cdc0ebd3f43d12cf77d7c33271342

commit r14-9167-gd1b241b9506cdc0ebd3f43d12cf77d7c33271342
Author: Gaius Mulley <gaiusmod2@gmail.com>
Date:   Sun Feb 25 11:08:37 2024 +0000

    PR modula2/113749 m2 enabled build times out on i686-gnu-hurd

    The bug fix changes the FIO module to use the target O_RDONLY,
    O_WRONLY, SEEK_SET and SEEK_END (now available from the module wrapc).
    Also rebuilt are the bootstrap tools mc and pge as they have their
    own wrapc and C translations of FIO.

    gcc/m2/ChangeLog:

            PR modula2/113749
            * Make-lang.in (BUILD-PGE-O): Add m2/pge-boot/Gwrapc.o.
            * gm2-libs-ch/wrapc.c (wrapc_SeekSet): New function.
            (wrapc_SeekEnd): Ditto.
            (wrapc_ReadOnly): Ditto.
            (wrapc_WriteOnly): Ditto.
            * gm2-libs/FIO.mod (SEEK_SET): Remove.
            (SEEK_END): Remove.
            (UNIXREADONLY): Remove.
            (UNIXWRITEONLY): Remove.
            (ConnectToUnix): Replace UNIXWRITEONLY with a call to WriteOnly.
            Replace UNIXREADONLY with a call to ReadOnly.
            (SetPositionFromBeginning): Replace SEEK_SET with a call to
            SeekSet.
            (SetPositionFromEnd): Replace SEEK_END with a call to
            SeekEnd.
            * gm2-libs/wrapc.def (SeekSet): New procedure function.
            (SeekEnd): New procedure function.
            (ReadOnly): New procedure function.
            (WriteOnly): New procedure function.
            * mc-boot-ch/Glibc.c (BUILD_MC_LIBC_TRACE): Undef.
            (check_init): New function.
            (tracedb): Ditto.
            (tracedb_open): Ditto.
            (tracedb_result): Ditto.
            (libc_read): Ditto.
            (libc_write): Ditto.
            (libc_close): Ditto.
            (libc_creat): Ditto.
            (libc_open): Ditto.
            (libc_lseek): Ditto.
            * mc-boot-ch/Gwrapc.c (wrapc_SeekSet): New function.
            (wrapc_SeekEnd): Ditto.
            (wrapc_ReadOnly): Ditto.
            (wrapc_WriteOnly): Ditto.
            * mc-boot/GDynamicStrings.cc: Rebuilt.
            * mc-boot/GFIO.cc: Ditto.
            * mc-boot/GIndexing.cc: Ditto.
            * mc-boot/GM2Dependent.cc: Ditto.
            * mc-boot/GM2EXCEPTION.cc: Ditto.
            * mc-boot/GPushBackInput.cc: Ditto.
            * mc-boot/GRTExceptions.cc: Ditto.
            * mc-boot/GRTint.cc: Ditto.
            * mc-boot/GSArgs.cc: Ditto.
            * mc-boot/GStdIO.cc: Ditto.
            * mc-boot/GStringConvert.cc: Ditto.
            * mc-boot/GSysStorage.cc: Ditto.
            * mc-boot/Gdecl.cc: Ditto.
            * mc-boot/Gkeyc.cc: Ditto.
            * mc-boot/Glibc.h: Ditto.
            * mc-boot/GmcComment.cc: Ditto.
            * mc-boot/GmcComp.cc: Ditto.
            * mc-boot/GmcDebug.cc: Ditto.
            * mc-boot/GmcMetaError.cc: Ditto.
            * mc-boot/GmcStack.cc: Ditto.
            * mc-boot/GmcStream.cc: Ditto.
            * mc-boot/GnameKey.cc: Ditto.
            * mc-boot/GsymbolKey.cc: Ditto.
            * mc-boot/Gvarargs.cc: Ditto.
            * mc-boot/Gwrapc.h: Ditto.
            * mc/decl.mod (getSymName): Add pointerref clause.
            * mc/mcStream.mod (copy): Check for an error after every read.
            * mc/varargs.mod (copy): Rewrite pointer arithmetic using INC to
            avoid type incompatibility.
            * pge-boot/GDynamicStrings.cc: Rebuilt.
            * pge-boot/GDynamicStrings.h: Ditto.
            * pge-boot/GFIO.cc: Ditto.
            * pge-boot/GFIO.h: Ditto.
            * pge-boot/GIO.cc: Ditto.
            * pge-boot/GIndexing.cc: Ditto.
            * pge-boot/GM2Dependent.cc: Ditto.
            * pge-boot/GM2EXCEPTION.cc: Ditto.
            * pge-boot/GNameKey.cc: Ditto.
            * pge-boot/GPushBackInput.cc: Ditto.
            * pge-boot/GRTExceptions.cc: Ditto.
            * pge-boot/GStdIO.cc: Ditto.
            * pge-boot/GSymbolKey.cc: Ditto.
            * pge-boot/GSysStorage.cc: Ditto.
            * pge-boot/Glibc.h: Ditto.
            * pge-boot/Gwrapc.cc: Ditto.
            * pge-boot/Gwrapc.h: Ditto.

    libgm2/ChangeLog:

            PR modula2/113749
            * libm2pim/wrapc.cc: Include fcntl.h.
            (SeekSet): New function.
            (SeekEnd): Ditto.
            (ReadOnly): Ditto.
            (WriteOnly): Ditto.

    Signed-off-by: Gaius Mulley <gaiusmod2@gmail.com>

[Bug modula2/113749] [14 Regression] m2 enabled build times out on i686-gnu (GNU Hurd)

[gaius at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113749

Gaius Mulley <gaius at gcc dot gnu.org> changed:

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

--- Comment #7 from Gaius Mulley <gaius at gcc dot gnu.org> ---
I'm optimistically closing the PR - thanks for the report - please re-open if
the problem persists.  I've seen cc1gm2 build on a debian i686 hurd system.

Interestingly the bug fix might well be applicable for other non Unix hosts.