[Bug]: sed/gawk does not compile under clang-21

[j2kun]

What happened?

clang version 21.1.8

In file included from external/sed+/sed/compile.c:19:
In file included from external/sed+/sed/sed.h:25:
external/sed+/sed/utils.h:30:3: error: expected function body after function declarator
   30 |   _GL_ATTRIBUTE_FORMAT_PRINTF_STANDARD (1, 2);
      |   ^
external/sed+/sed/utils.h:42:38: error: expected function body after function declarator
   42 |                    const char *mode) _GL_ARG_NONNULL ((1, 2, 3, 4));
      |                                      ^
external/sed+/sed/compile.c:263:9: error: call to undeclared function 'panic'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
  263 |         panic ("Called savchar with unexpected pushback (%x)",
      |         ^
external/sed+/sed/compile.c:1623:13: error: call to undeclared function 'panic'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
 1623 |             panic (_("can't find label for jump to `%s'"), go->name);

Not sure what to do here. @UebelAndre would it make sense to disable -Wimplicit-function-declaration in the BUILD file for the sed module?

Version

Development (host) and target OS/architectures:

Output of bazel --version: 8.4.0

Version of relevant rules from the WORKSPACE or MODULE.bazel file:

Language(s) and/or frameworks involved:

How to reproduce

Minimal MODULE.bazel

module(name = "sed_test")

bazel_dep(name = "rules_cc", version = "0.2.13")
bazel_dep(name = "toolchains_llvm_bootstrapped", version = "0.5.7")

register_toolchains("@toolchains_llvm_bootstrapped//toolchain:all")

bazel_dep(name = "sed", version = "4.9.bcr.3")

Then

bazel build @sed

Any other information?

No response

[j2kun]

I actually get a similar build failure for gawk@5.3.2.bcr.3

ERROR: /home/jkun/.cache/bazel/_bazel_jkun/886f3804b421b5442165b2f8eb57dad5/external/gawk+/BUILD.bazel:52:11: Compiling support/regex.c failed: (Exit 1): clang failed: error executing CppCompile command (from target @@gawk+//:support) external/toolchains_llvm_bootstrapped++_repo_rules+llvm-toolchain-minimal-21.1.8-linux-amd64/bin/clang -target x86_64-linux-gnu -no-canonical-prefixes -Wno-builtin-macro-redefined '-D__DATE__="redacted"' ... (remaining 62 arguments skipped)

In file included from external/gawk+/support/regex.c:71:
external/gawk+/support/regexec.c:188:36: error: call to undeclared function '_REGEX_NELTS'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
  188 |          size_t nmatch, regmatch_t pmatch[_REGEX_NELTS (nmatch)], int eflags)
      |                                          

[j2kun]

A little bit of digging on sed: the _GL_ARG_NONNULL macro is part of gnulib and imported into sed in lib/arg-nonnull.h during ./bootstrap. It contains

/* _GL_ARG_NONNULL((n,...,m)) tells the compiler and static analyzer tools
   that the values passed as arguments n, ..., m must be non-NULL pointers.
   n = 1 stands for the first argument, n = 2 for the second argument etc.  */
#ifndef _GL_ARG_NONNULL
# if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 3) || defined __clang__
#  define _GL_ARG_NONNULL(params) __attribute__ ((__nonnull__ params))
# else
#  define _GL_ARG_NONNULL(params)
# endif
#endif

How are gnulib dependencies supposed to be resolved in Bazel?

[j2kun]

Ah, ok so the particular source fetched for sed (https://ftp.gnu.org/gnu/sed/sed-4.9.tar.xz) has the gnulib sources already vendored in.

So the question is why, under toolchains_llvm_bootstrapped, does it not properly find the macro

[mlyoung101]

I'm not sure if this is 100% related, but I get a similar-in-vain issue with sed, under Clang 20.

ERROR: /home/matt/.cache/bazel/_bazel_matt/b1099a6bc90ff7286df31242b0202cf4/external/sed+/BUILD.bazel:279:11: Compiling lib/memchr.c [for tool] failed: (Exit 1): cc_wrapper.sh failed: error executing CppCompile command (from target @@sed+//:lib/sed) external/toolchains_llvm++llvm+llvm_toolchain/bin/cc_wrapper.sh -U_FORTIFY_SOURCE '--target=x86_64-unknown-linux-gnu' -U_FORTIFY_SOURCE -fstack-protector -fno-omit-frame-pointer -fcolor-diagnostics ... (remaining 52 arguments skipped)

Use --sandbox_debug to see verbose messages from the sandbox and retain the sandbox build root for debugging
external/sed+/lib/memchr.c:59:1: error: expected identifier or '('
   59 | __memchr (void const *s, int c_in, size_t n)
      | ^
external/sed+/lib/memchr.c:54:19: note: expanded from macro '__memchr'
   54 | # define __memchr memchr
      |                   ^
/usr/include/string.h:122:3: note: expanded from macro 'memchr'
  122 |   __glibc_const_generic (S, const void *, memchr (S, C, N))
      |   ^
/usr/include/sys/cdefs.h:838:3: note: expanded from macro '__glibc_const_generic'
  838 |   _Generic (0 ? (PTR) : (void *) 1,                     \
      |   ^
1 error generated.

This appears to be with sed 4.9.bcr.3