- Jan 13, 2020
-
-
Thomas Jahns authored
-
Thomas Jahns authored
* For some reason gcc 9.1 does not detect that status is never used uninitialized.
-
Thomas Jahns authored
* Given how many lines documenting the confusing pattern and how many compiler warnings are eliminated by this change...
-
Thomas Jahns authored
* Port this upstream and turn it into a cgribexlib update.
-
Thomas Jahns authored
* AM_CONDITIONAL must be defined in all control flows, or configure cannot finish successfully.
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
* There is no caller for the newly removed function stubs anyway.
-
Thomas Jahns authored
-
Thomas Jahns authored
* GCC 8.x and newer warn about the previously used "void (*)()", use "void (*)(void)" instead.
-
Thomas Jahns authored
-
-
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
* collSpec is meant to be a dumb data collection, none of its member should be subject to change.
-
Thomas Jahns authored
* gcc 6.4 does not realize posix_memalign initializes ptr on some Linux systems.
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
Thomas Jahns authored
-
-
Thomas Jahns authored
-
Thomas Jahns authored
* The previous implementation had the destructor also release the mapping. * That was problematic, because the internals of objects depended on the external mapping unnecessarily. * This also fixes a potential double-free issue discovered by Jan Frederik Engels.
-
Thomas Jahns authored
-
Thomas Jahns authored
* Not yet created resource handles are by definition unused.
-
Thomas Jahns authored
* This applies to NetCDF where the first write contains invariants that are not present in later writes.
-