Is --float broken?

[pmatos]

If I generate a program with --float using cmith 2.4.0, and then compile with gcc 14.1.1, I see that we are passing float * to transparent_crc_bytes as first argumnet which should have type char *. GCC definitely does not like that.

Attaching csmith generated file.
test1.c.gz

test1.c:1617:28: error: passing argument 1 of ‘transparent_crc_bytes’ from incompatible pointer type [-Wincompatible-pointer-types]
 1617 |     transparent_crc_bytes (&g_2320, sizeof(g_2320), "g_2320", print_hash_value);
      |                            ^~~~~~~
      |                            |
      |                            const float *

[pmatos]

ping. @eeide are you still working on this?

[pmatos]

Same as #129
Interestingly clang happily compiles it as seen in #135

[eeide]

Hi, @pmatos! I personally haven't made any substantive code changes to Csmith in a long time.

The most recent changes to Csmith have been from people offering patches through pull requests. A patch for the issue you pointed out would be most welcome!

[pmatos]

Hi, @pmatos! I personally haven't made any substantive code changes to Csmith in a long time.

The most recent changes to Csmith have been from people offering patches through pull requests. A patch for the issue you pointed out would be most welcome!

I see. Thanks for the quick reply.