Hello ntfs-3g maintainers, please excuse my probably misdesigned action. Matthias Klose filed this bug report primarily because package partclone, which is built correctly with gcc-13, FTBFS with gcc-14.
I could build it with gcc-14, by applying this simple patch to file
/usr/include/ntfs-3g/ntfstime.h:
---------------8<-----------------------------------
--- /usr/include/ntfs-3g/ntfstime.h.orig 2024-07-06 16:58:02.341675225
+0200
+++ /usr/include/ntfs-3g/ntfstime.h 2024-07-06 16:58:46.110108507 +0200
@@ -35,6 +35,7 @@
#endif
#include "types.h"
+#include <time.h>
/*
* assume "struct timespec" is not defined if st_mtime is not defined
---------------8<-----------------------------------
That means that the bug might be fixed in two ways:
1 - by defining HAVE_TIME_H somewhere outside the file ntfstime.h which
would cause the inclusion of time.h, because of lines 27--29 in
file ntfstime.h
2 - by enforcing the inclusion of the file time.h as I did
Enforcing uncontionnally the inclusion of the file time.h as I did is
probably bad. However why should a program like partclone define the
macro HAVE_TIME_H just to be able to make use of ntfs-3g in addition to
many other filesystems, which do not require such a definition?
I suspect that this issue might be addressed, by enhancing the way
dh_auto_configure works by default.
Thank you in advance for your enlightenments, your suggestions.
Best regards, Georges.
signature.asc
Description: PGP signature
