Preparing for the next release: call for testing
Dmitry V. Levin
ldv at altlinux.org
Thu Feb 26 20:35:22 UTC 2015
On Thu, Feb 26, 2015 at 01:10:39AM -0500, Mike Frysinger wrote:
> a few signed warnings have crept in:
> socketutils.c: In function 'receive_responses':
> socketutils.c:145:110: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
> netlink_inet_diag.c: In function 'check_responses':
> netlink_inet_diag.c:67:112: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
> netlink_unix_diag.c: In function 'check_responses':
> netlink_unix_diag.c:75:112: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
The whole -Wsign-compare thing is a regression introduced by commit
v4.9-8-g3ed5d02 :)
On some platforms, any use of NLMSG_OK macro results to these
-Wsign-compare warnings. The way how NLMSG_OK is defined in
<linux/netlink.h> leaves no room for a fix.
> guess this should be declared differently ? maybe as a noinst_PROGRAM ?
> Makefile.am:411: warning: variable 'ioctlsort_LDFLAGS' is defined but no program or
> Makefile.am:411: library has 'ioctlsort' as canonical name (possible typo)
> although the build settings look a bit dicey ... it's mixing target compiler &
> settings and the build compiler. this leads to all cross-compiles failing:
> ./ioctlsort0 > ioctlent0.h
> /bin/sh: ./ioctlsort0: cannot execute binary file: Exec format error
After the big change in ioctl parser (v4.9-222-gdf7aa2b) ioctlsort is used
at build time. The trick is that ioctlsort is now being compiled with
target headers, so all cross-compiling configurations probably need some
local ioctlsort_* adjustments.
See also https://sourceforge.net/p/strace/mailman/message/33285976/.
--
ldv
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://lists.strace.io/pipermail/strace-devel/attachments/20150226/0b183d44/attachment.bin>
More information about the Strace-devel
mailing list