On Sep 16, 2012, at 10:31 , Richard W.M. Jones wrote:
... and why they are (probably) not errors.
[snip]
> prohibit_magic_number_exit
> po-docs/ja.po:59861:"might be called indirectly from L<exit(3)>, which can
cause unexpected "
> po-docs/libguestfs-docs.pot:50902:"might be called indirectly from
L<exit(3)>, which can cause unexpected "
> po-docs/uk.po:57345:"might be called indirectly from L<exit(3)>, which can
cause unexpected "
> src/guestfs.pod:2028:callback might be called indirectly from L<exit(3)>, which
can cause
> tests/mount-local/test-parallel-mount-local.c:104: exit (77);
> tests/mount-local/test-parallel-mount-local.c:110: exit (77);
> tests/regressions/rhbz790721.c:79: exit (77);
> maint.mk: use EXIT_* values rather than magic number
> make: *** [sc_prohibit_magic_number_exit] Error 1
I think exit (77) is acceptable. To automake, this means that the
test has been skipped, and there is no constant for it. However
writing a regexp that matches all numbers except "77" is quite hard.
[snip]
Out of curiosity, why exit code 77 (permission denied) instead of exit code 75 (temporary
failure, user is invited to retry) [1]? Is this just an automake-ism?
--Alex
[1] For reference, here's a sysexits listing, consistent with what I saw in
/usr/include/sysexits.h on a Fedora 17 VM:
<
http://www.opensource.apple.com/source/Libc/Libc-320/include/sysexits.h>