latest Master compiling error: "unknown type name ‘bool’" about darktable HOT 13 CLOSED

Should be fixed now, thanks for reporting.

from darktable.

I'm getting the next one now:

/home/s7habo/darktable/src/common/variables.c: In function ‘_is_number’:
/home/s7habo/darktable/src/common/variables.c:314:37: error: ‘false’ undeclared (first use in this function)
  314 |   if(!g_ascii_isdigit(*str)) return false;  // don't take empty strings
      |                                     ^~~~~
compilation terminated due to -Wfatal-errors.
gmake[2]: *** [bin/CMakeFiles/lib_darktable.dir/build.make:1146: bin/CMakeFiles/lib_darktable.dir/common/variables.c.o] Fehler 1
gmake[2]: *** Auf noch nicht beendete Prozesse wird gewartet …
gmake[1]: *** [CMakeFiles/Makefile2:2431: bin/CMakeFiles/lib_darktable.dir/all] Fehler 2
gmake: *** [Makefile:156: all] Fehler 2

from darktable.

I'm getting the next one now:

Hopefully fixed now...

from darktable.

Yes, everything went well. Thanks for the quick fix!

from darktable.

Curious how this got past the CI?

from darktable.

And it was also compiling fine for me.

from darktable.

I've seen it as well on my Windows machine, while the Windows CI was passing, very strange...

from darktable.

There's nothing strange. C90 does not support the boolean data type. C99 does. If you look at the build logs on all CI platforms, you will see -std=gnu99 on the command line of the compiler call. Apparently, for some reason, @s7habo's gcc (which was gcc12, whereas we at CI run newer compilers) started in C90 mode. This seems to be the default gcc12 setting in Ubuntu...

from darktable.

Mine is GCC 13.2 on Windows. Same as CI. I don't think I do anything differently, no special compiler flags...

from darktable.

But C99 standard is enabled in dt's cmake:

So it's not about C standard version IMHO, but maybe about support of said feature (bool, true and false) by various compiler versions.

from darktable.

Or, more likely, it's just about including <stdbool.h>. On some systems it may be included by some other includes, on others - not.

from darktable.

Still doesn't explain why my local Windows MSYS2 build failed while the (presumably identically set up) CI didn't.

from darktable.

Well, there I can’t help, obviously they are different. E.g. enabled features and dependencies (and their versions) can affect includes, not just base build environment.

from darktable.