Re: official lang/gcc14 package builds are broken, at least for armv7: "error: attempt to use a poisoned identifier" and more

On Sep 17, 2024, at 11:40, Dimitry Andric <dim@FreeBSD.org> wrote:

> This is because of gcc's weird identifier poisoning system, and them not correctly marking all used standard C++ headers for inclusion before the poisoning takes place.
> 
> Try the attached patch, which I have been using for a long time now (months, at least since the bootstrap options were deleted, but which I reverted locally).
> 
> -Dimitry
> 
> <lang__gcc14-devel-fix-plugins-1.diff>
> 
>> On 17 Sep 2024, at 16:24, Mark Millard <marklmi@yahoo.com> wrote:
>> 
>> Also: no member named 'fancy_abort' in namespace 'std'
>> 
>> Such is associated with not using STANDARD_BOOTSTRAP for armv7, given GCC's historic lack of keeping builds of everything working for just direct use of clang++/libc++ or the like.
>> 
>> The specific change that got back into this issue was:
>> 
>> -LANGUAGES:= c,c++,objc,fortran
>> +LANGUAGES:= c,c++,objc,fortran,jit
>> 
>> It is the jit part of the build that fails to build as stands.
>> 
>> See:
>> 
>> https://pkg-status.freebsd.org/ampere2/data/main-armv7-default/p13f486017d78_s3df987c99/logs/errors/gcc14-14.2.0_1.log
>> 
>> Example:
>> 
>> In file included from /wrkdirs/usr/ports/lang/gcc14/work/gcc-14.2.0/gcc/jit/dummy-frontend.cc:23:
>> In file included from /wrkdirs/usr/ports/lang/gcc14/work/gcc-14.2.0/gcc/jit/jit-playback.h:24:
>> In file included from /usr/include/c++/v1/string:594:
>> In file included from /usr/include/c++/v1/__memory_resource/polymorphic_allocator.h:20:
>> In file included from /usr/include/c++/v1/tuple:1455:
>> In file included from /usr/include/c++/v1/exception:82:
>> /usr/include/c++/v1/__exception/exception_ptr.h:111:3: error: no member named 'fancy_abort' in namespace 'std'; did you mean simply 'fancy_abort'?
>> 111 | std::abort();
>> | ^~~~~
>> /wrkdirs/usr/ports/lang/gcc14/work/gcc-14.2.0/gcc/system.h:808:13: note: 'fancy_abort' declared here
>> 808 | extern void fancy_abort (const char *, int, const char *)
>> | ^
>> In file included from /wrkdirs/usr/ports/lang/gcc14/work/gcc-14.2.0/gcc/jit/dummy-frontend.cc:23:
>> In file included from /wrkdirs/usr/ports/lang/gcc14/work/gcc-14.2.0/gcc/jit/jit-playback.h:26:
>> In file included from /usr/include/c++/v1/vector:325:
>> In file included from /usr/include/c++/v1/__format/formatter_bool.h:20:
>> In file included from /usr/include/c++/v1/__format/formatter_integral.h:35:
>> /usr/include/c++/v1/locale:288:32: error: attempt to use a poisoned identifier
>> 288 | __status = (unsigned char*)malloc(__nkw);
>> | ^
>> /usr/include/c++/v1/locale:1385:24: error: attempt to use a poisoned identifier
>> 1385 | __ob = (char_type*)malloc(2 * static_cast<size_t>(__nc) * sizeof(char_type));
>> | ^
>> /usr/include/c++/v1/locale:2429:31: error: attempt to use a poisoned identifier
>> 2429 | _Tp* __t = (_Tp*)std::realloc(__owns ? __b.get() : 0, __new_cap);
>> | ^
>> /usr/include/c++/v1/locale:2632:24: error: attempt to use a poisoned identifier
>> 2632 | __h.reset((char*)malloc(static_cast<size_t>(__wn - __wb.get() + 2)));
>> | ^
>> /usr/include/c++/v1/locale:2910:28: error: attempt to use a poisoned identifier
>> 2910 | __hd.reset((char_type*)malloc(static_cast<size_t>(__n) * sizeof(char_type)));
>> | ^
>> /usr/include/c++/v1/locale:2936:28: error: attempt to use a poisoned identifier
>> 2936 | __hw.reset((char_type*)malloc(__exn * sizeof(char_type)));
>> | ^
>> /usr/include/c++/v1/locale:2974:27: error: attempt to use a poisoned identifier
>> 2974 | __h.reset((char_type*)malloc(__exn * sizeof(char_type)));
>> | ^
>> 

In my environment (that somewhat predates the change that made the
ampere2 based build fail), I set up a poudriere-devel test for
lang/gcc14 ( not lang/gcc14-devel ):

# poudriere ports -l
PORTSTREE METHOD TIMESTAMP           PATH
default   null   2021-04-18 02:05:47 /usr/ports

# ~/fbsd-based-on-what-commit.sh -C /usr/ports/
d13cce06daa6 (HEAD -> main, freebsd/main, freebsd/HEAD) graphics/gmic-qt: Update to 3.4.1
Author:     Jason E. Hale <jhale@FreeBSD.org>
Commit:     Jason E. Hale <jhale@FreeBSD.org>
CommitDate: 2024-09-02 04:05:53 +0000
branch: main
merge-base: d13cce06daa6361cdbd6b97b768ab80aa60b8677
merge-base: CommitDate: 2024-09-02 04:05:53 +0000
n675562 (--first-parent --count for merge-base)

(Note: I've made no commits.)

# git -C /usr/ports/ status lang/gcc14
On branch main
Your branch is up to date with 'freebsd/main'.

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)
modified:   lang/gcc14/Makefile

Untracked files:
  (use "git add <file>..." to include in what will be committed)
lang/gcc14/files/patch-gcc_jit_dummy-frontend.cc
lang/gcc14/files/patch-gcc_jit_jit-builtins.cc
lang/gcc14/files/patch-gcc_jit_jit-playback.cc
lang/gcc14/files/patch-gcc_jit_jit-recording.cc
lang/gcc14/files/patch-gcc_jit_libgccjit.cc

no changes added to commit (use "git add" and/or "git commit -a")

(Note, below the whitespace details might not be preserved,
so the text is only suggestive.)

# git -C /usr/ports/ diff lang/gcc14
diff --git a/lang/gcc14/Makefile b/lang/gcc14/Makefile
index 53bac89a7f14..17029847fa7c 100644
--- a/lang/gcc14/Makefile
+++ b/lang/gcc14/Makefile
@@ -73,7 +73,7 @@ CONFIGURE_TARGET=     x86_64-portbld-${OPSYS:tl}${OSREL}
 CONFIGURE_ARGS+=       --with-abi=elfv2
 .endif
 
-LANGUAGES:=    c,c++,objc,fortran
+LANGUAGES:=    c,c++,objc,fortran,jit
 TARGLIB=       ${PREFIX}/lib/gcc${SUFFIX}
 TARGLIB32=     ${PREFIX}/lib32 # The version information is added later
 LIBEXEC=       ${PREFIX}/libexec/gcc${SUFFIX}

And, from the content of your patch:

# head -30 /usr/ports/lang/gcc14/files/patch-gcc_jit_*.cc
==> /usr/ports/lang/gcc14/files/patch-gcc_jit_dummy-frontend.cc <==
--- gcc/jit/dummy-frontend.cc.orig	2024-01-14 22:32:46 UTC
+++ gcc/jit/dummy-frontend.cc
@@ -18,6 +18,8 @@ along with GCC; see the file COPYING3.  If not see
 <http://www.gnu.org/licenses/>.  */
 
 #include "config.h"
+#define INCLUDE_STRING
+#define INCLUDE_VECTOR
 #include "system.h"

==> /usr/ports/lang/gcc14/files/patch-gcc_jit_jit-builtins.cc <==
--- gcc/jit/jit-builtins.cc.orig	2024-01-14 22:32:46 UTC
+++ gcc/jit/jit-builtins.cc
@@ -18,6 +18,8 @@ along with GCC; see the file COPYING3.  If not see
 <http://www.gnu.org/licenses/>.  */
 
 #include "config.h"
+#define INCLUDE_STRING
+#define INCLUDE_VECTOR
 #include "system.h"
 #include "coretypes.h"
 #include "target.h"

==> /usr/ports/lang/gcc14/files/patch-gcc_jit_jit-playback.cc <==
--- gcc/jit/jit-playback.cc.orig	2024-01-14 22:32:46 UTC
+++ gcc/jit/jit-playback.cc
@@ -20,6 +20,8 @@ along with GCC; see the file COPYING3.  If not see
 
 #include "config.h"
 #define INCLUDE_MUTEX
+#define INCLUDE_STRING
+#define INCLUDE_VECTOR
 #include "libgccjit.h"
 #include "system.h"
 #include "coretypes.h"

==> /usr/ports/lang/gcc14/files/patch-gcc_jit_jit-recording.cc <==
--- gcc/jit/jit-recording.cc.orig	2024-01-14 22:32:46 UTC
+++ gcc/jit/jit-recording.cc
@@ -20,6 +20,8 @@ along with GCC; see the file COPYING3.  If not see
 
 #include "config.h"
 #define INCLUDE_SSTREAM
+#define INCLUDE_STRING
+#define INCLUDE_VECTOR
 #include "system.h"
 #include "coretypes.h"
 #include "tm.h"

==> /usr/ports/lang/gcc14/files/patch-gcc_jit_libgccjit.cc <==
/jit/libgccjit.cc.orig	2024-01-14 22:32:46 UTC
+++ gcc/jit/libgccjit.cc
@@ -20,6 +20,8 @@ along with GCC; see the file COPYING3.  If not see
 
 #include "config.h"
 #define INCLUDE_MUTEX
+#define INCLUDE_STRING
+#define INCLUDE_VECTOR
 #include "system.h"
 #include "coretypes.h"
 #include "timevar.h"

The result for poudriere bulk lang/gcc14 was:

[00:06:55] [01] [00:00:00] Building   lang/gcc14 | gcc14-14.2.0
[00:28:00] [01] [00:21:05] Finished   lang/gcc14 | gcc14-14.2.0: Success ending TMPFS: 1.61 GiB

So it looks like adding your files/patch-gcc_jit_*.cc
files to lang/gcc14 would allow the ampere2 builds
of lang/gcc14 for armv7 to work without use of
STANDARD_BOOTSTRAP (at least for now).


( The "ending TMPFS: . . . GiB" is something I add to
poudirere for helping to find packages that I might
want to add to TMPFS_BLACKLIST .)


The context has:

# poudriere jail -jmain-armv7 -i
Jail name:         main-armv7
Jail version:      15.0-CURRENT
Jail arch:         armv7
Jail method:       pkgbase
Jail mount:        /usr/local/poudriere/jails/main-armv7-poud
Jail fs:            Jail updated:      2024-08-27 21:47:56
Jail pkgbase:      disabled

The pkgbase system for the jail is from the official
materials at the time it was last updated. It is
not my personal world build at all.


Thanks,
Mark

===
Mark Millard
marklmi at yahoo.com