[Bug bootstrap/102675] New: [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[Bug bootstrap/102675] New: [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[gerald at pfeifer dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

            Bug ID: 102675
           Summary: [12 regression] Bootstrap fails in libsanitizer:
                    'MD5_DIGEST_STRING_LENGTH' was not declared in this
                    scope
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gerald at pfeifer dot com
                CC: hjl.tools at gmail dot com
  Target Milestone: ---
              Host: *-*-freebsd*
            Target: *-*-freebsd*
             Build: *-*-freebsd*

This happens on all FreeBSD platforms and versions:

CC-HEAD/libsanitizer/sanitizer_common/sanitizer_platform_limits_freebsd.cpp:370
:36: error: 'MD5_CTX' was not declared in this scope
  370 | const unsigned MD5_CTX_sz = sizeof(MD5_CTX);
      |                                    ^~~~~~~


GCC-HEAD/libsanitizer/sanitizer_common/sanitizer_platform_limits_freebsd.cpp:371
:36: error: 
'MD5_DIGEST_STRING_LENGTH' was not declared in this scope; did you mean 
'SHA256_DIGEST_STRING_LENGTH'?
  371 | const unsigned MD5_return_length = MD5_DIGEST_STRING_LENGTH;
      |                                    ^~~~~~~~~~~~~~~~~~~~~~~~
      |                                    SHA256_DIGEST_STRING_LENGTH


It was introduced by commit 2e3d50c09519d1b4899845b21843bae66ecffc2f
Author: H.J. Lu <hjl.tools@gmail.com>
Date:   Wed Oct 6 10:24:24 2021 -0700

    libsanitizer: Merge with upstream

    Merged revision: fdf4c035225de52f596899931b1f6100e5e3e928


I believe the problem is that this adds #include <md5.h> and some
dependencies on constants defined in FreeBSD's /usr/include/md5.h,
where GCC features it's on $GCC_SOURCE/include/md5.h which does not
provide the required constants and types.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-10-09
             Status|UNCONFIRMED                 |NEW

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to Gerald Pfeifer from comment #0)
> This happens on all FreeBSD platforms and versions:
> 
> I believe the problem is that this adds #include <md5.h> and some
> dependencies on constants defined in FreeBSD's /usr/include/md5.h,
> where GCC features it's on $GCC_SOURCE/include/md5.h which does not
> provide the required constants and types.

That file is FreeBSD specific.  Can you use a local patch to force
/usr/include/md5.h, like

#include_next <md5.h>

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[gerald at pfeifer dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #2 from Gerald Pfeifer <gerald at pfeifer dot com> ---
(In reply to H.J. Lu from comment #1)
> That file is FreeBSD specific.  Can you use a local patch to force
> /usr/include/md5.h, like
> 
> #include_next <md5.h>

I tried replacing #include <md5.h> by #include_next <md5.h>; that did
not lead to any change.

Using #include "/usr/include/md5.h" restored bootstrap.

This confirms my hypothesis that it's GCC's md5.h being picked up that
is the issue for this regression.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to Gerald Pfeifer from comment #2)
> (In reply to H.J. Lu from comment #1)
> > That file is FreeBSD specific.  Can you use a local patch to force
> > /usr/include/md5.h, like
> > 
> > #include_next <md5.h>
> 
> I tried replacing #include <md5.h> by #include_next <md5.h>; that did
> not lead to any change.
> 
> Using #include "/usr/include/md5.h" restored bootstrap.
> 
> This confirms my hypothesis that it's GCC's md5.h being picked up that
> is the issue for this regression.

Does this file use any GCC specific header files?  If not, you can
filter out GCC specific -Ixxx from CXX and CXXFLAGS for this file.
Or can you prepend -Isystem-headr-directory to CXXFLAGS for this file?

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #4 from H.J. Lu <hjl.tools at gmail dot com> ---
Another possibility is to add a configure test to locate the system
<md5.h> and include it instead of <md5.h>.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
            Version|unknown                     |12.0
   Target Milestone|---                         |12.0

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
or adjust the code to the libiberty provided md5.h which is available
everywhere

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

kargl at gcc dot gnu.org changed:

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

--- Comment #6 from kargl at gcc dot gnu.org ---
Just ran into this issue. :-((In reply to H.J. Lu from comment #4)
> Another possibility is to add a configure test to locate the system
> <md5.h> and include it instead of <md5.h>.

Another possibility is to revert your patch so that others can bootstrap.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #7 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to kargl from comment #6)
> 
> Another possibility is to revert your patch so that others can bootstrap.

The problem won't go away as long as we keep syncing with upstream.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #8 from H.J. Lu <hjl.tools at gmail dot com> ---
Does sanitizer_common/sanitizer_platform_limits_freebsd.cpp need any header
files from GCC?

If yes, why aren't they needed in compiler-rt?
If no, can you filter out these -I options in Makefile?

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[gerald at pfeifer dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #9 from Gerald Pfeifer <gerald at pfeifer dot com> ---
(In reply to H.J. Lu from comment #8)
> Does sanitizer_common/sanitizer_platform_limits_freebsd.cpp need any header
> files from GCC?

>From what I found, that does not appear to be the case.

> If yes, why aren't they needed in compiler-rt?
> If no, can you filter out these -I options in Makefile?

How do I go about that? (I did see your earlier suggestions, just ran 
out of time and, frankly, the experience on how to practically do this.)

If you have a prototype patch to try (and maybe tweak) I'll happily do.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |WAITING

--- Comment #10 from H.J. Lu <hjl.tools at gmail dot com> ---
(In reply to Gerald Pfeifer from comment #9)
> 
> If you have a prototype patch to try (and maybe tweak) I'll happily do.

Please send me all Makefiles under libsanitizer and the failed command line.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[gerald at pfeifer dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #11 from Gerald Pfeifer <gerald at pfeifer dot com> ---
Created attachment 51657
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51657&action=edit
$WRKDIR/i586-unknown-freebsd11.4/libsanitizer/Makefile

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[gerald at pfeifer dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #12 from Gerald Pfeifer <gerald at pfeifer dot com> ---
Created attachment 51658
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51658&action=edit
$WRKDIR/i586-unknown-freebsd11.4/libsanitizer/sanitizer_common/Makefile

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #13 from H.J. Lu <hjl.tools at gmail dot com> ---
Created attachment 51659
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51659&action=edit
A patch

Please try this.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[gerald at pfeifer dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

Gerald Pfeifer <gerald at pfeifer dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW

--- Comment #14 from Gerald Pfeifer <gerald at pfeifer dot com> ---
(In reply to H.J. Lu from comment #13)
> Please try this.

Thank you, H.J., and sorry it took me to run tests since I ran into
unrelated problems.

Good news: with your patch GCC is back to bootstrap on FreeBSD!


(Will you apply this to GCC and push it upstream to libsanitizer?)

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch

--- Comment #15 from H.J. Lu <hjl.tools at gmail dot com> ---
A patch is posted at

https://gcc.gnu.org/pipermail/gcc-patches/2021-October/582945.html

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |unlvsur at live dot com

--- Comment #16 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 103203 has been marked as a duplicate of this bug. ***

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

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

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

--- Comment #17 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
The include/md5.h part of this has been committed as
r12-4936-g44d9d55c6d0e3a1e26427662d30f350a80282634
The other change hasn't been applied from what I can see.

I wonder if we couldn't do instead:
2021-11-16  Jakub Jelinek  <jakub@redhat.com>

        PR bootstrap/102675
        * sanitizer_common/Makefile.am: Use -DUSE_SYSTEM_MD5 in AM_CXXFLAGS
        of sanitizer_platform_limits_freebsd.cpp.
        * sanitizer_common/Makefile.in: Regenerated.

--- libsanitizer/sanitizer_common/Makefile.am.jj        2021-11-05
00:43:22.647623646 +0100
+++ libsanitizer/sanitizer_common/Makefile.am   2021-11-16 12:29:58.574930436
+0100
@@ -17,6 +17,7 @@ AM_CXXFLAGS += -DSANITIZER_LIBBACKTRACE
 endif
 AM_CCASFLAGS = $(EXTRA_ASFLAGS)
 ACLOCAL_AMFLAGS = -I m4
+sanitizer_platform_limits_freebsd.lo: AM_CXXFLAGS += -DUSE_SYSTEM_MD5

 noinst_LTLIBRARIES = libsanitizer_common.la

--- libsanitizer/sanitizer_common/Makefile.in.jj        2021-11-05
00:43:22.647623646 +0100
+++ libsanitizer/sanitizer_common/Makefile.in   2021-11-16 12:30:58.611088913
+0100
@@ -796,6 +796,7 @@ uninstall-am:

 .PRECIOUS: Makefile

+sanitizer_platform_limits_freebsd.lo: AM_CXXFLAGS += -DUSE_SYSTEM_MD5

 # Tell versions [3.59,3.63) of GNU make to not export all variables.
 # Otherwise a system limit (for SysV at least) may be exceeded.

and avoid the need of a local patches for this - the Makefiles are owned by
gcc, not imported from compiler-rt, and we require GNU make for building gcc
AFAIK.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[gerald at pfeifer dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #18 from Gerald Pfeifer <gerald at pfeifer dot com> ---
Thank you, Jakub!

I finished testing/preparing the second part of the patch, just did not get
to push.  I'll give your suggested patch a try tonight!

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[unlvsur at live dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #19 from cqwrteur <unlvsur at live dot com> ---
(In reply to Jakub Jelinek from comment #17)
> The include/md5.h part of this has been committed as
> r12-4936-g44d9d55c6d0e3a1e26427662d30f350a80282634
> The other change hasn't been applied from what I can see.
> 
> I wonder if we couldn't do instead:
> 2021-11-16  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR bootstrap/102675
> 	* sanitizer_common/Makefile.am: Use -DUSE_SYSTEM_MD5 in AM_CXXFLAGS
> 	of sanitizer_platform_limits_freebsd.cpp.
> 	* sanitizer_common/Makefile.in: Regenerated.
> 
> --- libsanitizer/sanitizer_common/Makefile.am.jj	2021-11-05
> 00:43:22.647623646 +0100
> +++ libsanitizer/sanitizer_common/Makefile.am	2021-11-16 12:29:58.574930436
> +0100
> @@ -17,6 +17,7 @@ AM_CXXFLAGS += -DSANITIZER_LIBBACKTRACE
>  endif
>  AM_CCASFLAGS = $(EXTRA_ASFLAGS)
>  ACLOCAL_AMFLAGS = -I m4
> +sanitizer_platform_limits_freebsd.lo: AM_CXXFLAGS += -DUSE_SYSTEM_MD5
>  
>  noinst_LTLIBRARIES = libsanitizer_common.la
>  
> --- libsanitizer/sanitizer_common/Makefile.in.jj	2021-11-05
> 00:43:22.647623646 +0100
> +++ libsanitizer/sanitizer_common/Makefile.in	2021-11-16 12:30:58.611088913
> +0100
> @@ -796,6 +796,7 @@ uninstall-am:
>  
>  .PRECIOUS: Makefile
>  
> +sanitizer_platform_limits_freebsd.lo: AM_CXXFLAGS += -DUSE_SYSTEM_MD5
>  
>  # Tell versions [3.59,3.63) of GNU make to not export all variables.
>  # Otherwise a system limit (for SysV at least) may be exceeded.
> 
> and avoid the need of a local patches for this - the Makefiles are owned by
> gcc, not imported from compiler-rt, and we require GNU make for building gcc
> AFAIK.

Jakub. can you also fix the PR100017 today?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100017

Just add -nostdinc++ in the configure and configure.ac based on yujie yang's
patch?

It has been reported for 7 months for such a simple bug.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[gerald at pfeifer dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #20 from Gerald Pfeifer <gerald at pfeifer dot com> ---
Thank you, Jakub!

I finished testing/preparing the second part of the patch, just did not get
to push.

So I went ahead and gave your suggested patch a try - and it passes bootstrap,
so good to go I'd say! Thanks.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

--- Comment #21 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:edd2249b17fb7150da66b03adb1a4e381926047d

commit r12-5367-gedd2249b17fb7150da66b03adb1a4e381926047d
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu Nov 18 10:06:01 2021 +0100

    libsanitizer: Fix bootstrap on FreeBSD [PR102675]

    Here is an alternative to the patch changing a file imported from
    compiler-rt upstream, so that we don't need to cary a local patch for that
    particular problem.

    2021-11-18  Jakub Jelinek  <jakub@redhat.com>

            PR bootstrap/102675
            * sanitizer_common/Makefile.am: Use -DUSE_SYSTEM_MD5 in AM_CXXFLAGS
            of sanitizer_platform_limits_freebsd.cpp.
            * sanitizer_common/Makefile.in: Regenerated.

[Bug bootstrap/102675] [12 regression] Bootstrap fails in libsanitizer: 'MD5_DIGEST_STRING_LENGTH' was not declared in this scope

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102675

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

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

--- Comment #22 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Hopefully fixed.