[committed 0/2] CRIS: Fix compilation warnings that recent gcc treats as errors
Hans-Peter Nilsson <[email protected]>
| Newsgroups | gmane.comp.lib.newlib |
|---|---|
| Message-ID | <[email protected]> |
As you've seen in other patch submissions, there was recently a change in gcc such that implicit function declarations and type mismatches are now errors, no longer just warnings. Problems are in port-specific code, as I don't see any breakage for arm-eabi. Patches for some other ports are in, and I guess lots more will trickle in as port maintainers and other people find out and fix. This two-set of patches handles newlib and libgloss for cris-elf. A gcc regtest round shows no regressions compared to before the "breaking" gcc commit. Caveat: a patch-set adding a _getentropy stub from August (IIRC) causes lots of FAILs; some of those are worked around in gcc to get better coverage. Film at 11. Anyway, a regtest-round with these patches applied to a newlib checkout before the _getentropy patches shows no regressions for cris-elf+cris-sim, and working around *some* of the breakage with a slightly more recent newlib checkout (dab15f6740ff) shows that to be the only issue. brgds, H-P