[Bug]: _GNU_SOURCE macro defined twice in the codebase

[rdmark]

Contact Details

No response

Version

5.7.2

Description

The _GNU_SOURCE macro is defined twice under the exact same #ifdef logic in the WolfSSL codebase:

• wolfssl/src/ssl.c

  Line 31 in 398f8c9

  #define _GNU_SOURCE

• wolfssl/src/bio.c

  Line 30 in 398f8c9

  #define _GNU_SOURCE

This means that you will always get compiler warnings. I believe this is harmless, but adds noise to the log that may distract from actual bugs...

Reproduction steps

1. Build the library (e.g. embedded in the Netatalk build system)
2. Observe the compiler log

Relevant log output

[206/318] Compiling C object libatalk/ssl/libssl.a.p/src_ssl.c.o
In file included from ../libatalk/ssl/wolfcrypt/src/misc.c:29,
                 from ../libatalk/ssl/src/ssl.c:45:
./config.h:800: warning: "_GNU_SOURCE" redefined
  800 | #define _GNU_SOURCE 1
      | 
../libatalk/ssl/src/ssl.c:31: note: this is the location of the previous definition
   31 |     #define _GNU_SOURCE
      | 
In file included from ../libatalk/ssl/src/x509.c:24,
                 from ../libatalk/ssl/src/ssl.c:23296:
./config.h:800: warning: "_GNU_SOURCE" redefined
  800 | #define _GNU_SOURCE 1
      | 
In file included from ../libatalk/ssl/src/ssl.c:21882:
../libatalk/ssl/src/bio.c:30: note: this is the location of the previous definition
   30 |     #define _GNU_SOURCE
      |

[rdmark]

Proposing #7958 as a naive solution, but I would be curious if there is a better way to avoid the collision (define it in a header?)

[rdmark]

#7971 resolves all my issues. Thank you very much for your support!