<div dir="ltr">Yeah, assert() should be marked noreturn (directly or indirectly) in general or bad things happen. Those assert()s really want to be assert()s or calls to abort() or something: there is nothing good that can happen after returning from those places.</div> <div class="gmail_quote"><div dir="ltr">On Sun, Oct 16, 2016 at 11:19 PM Alan Coopersmith <<a href="mailto:alan.coopersmith@oracle.com">alan.coopersmith@oracle.com</a>> wrote: </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On 10/ 8/16 12:58 PM, Pablo Cholaky wrote: > I think this patch is currently harmless and makes support to clang to properly > compile it. clang doesn't complain on the current code for me (probably because my libc headers define the function called on a failed assert with the noreturn attribute), but adding this: > assert(0); > + return -1; causes clang to report two new warnings for me: ./xcb_bitops.h:210:10: warning: integer constant not in range of enumerated type 'xcb_image_order_t' (aka 'enum xcb_image_order_t') [-Wassign-enum] return -1; ^ ./xcb_bitops.h:210:10: warning: implicit conversion changes signedness: 'int' to 'xcb_image_order_t' (aka 'enum xcb_image_order_t') [-Wsign-conversion] return -1; ~~~~~~ ^~ Admittedly I run clang with more warnings enabled than the default build, to help find issues during development. -- -Alan Coopersmith- <a href="mailto:alan.coopersmith@oracle.com" class="gmail_msg" target="_blank">alan.coopersmith@oracle.com</a> Oracle Solaris Engineering - <a href="http://blogs.oracle.com/alanc" rel="noreferrer" class="gmail_msg" target="_blank">http://blogs.oracle.com/alanc</a> _______________________________________________ Xcb mailing list <a href="mailto:Xcb@lists.freedesktop.org" class="gmail_msg" target="_blank">Xcb@lists.freedesktop.org</a> <a href="https://lists.freedesktop.org/mailman/listinfo/xcb" rel="noreferrer" class="gmail_msg" target="_blank">https://lists.freedesktop.org/mailman/listinfo/xcb</a></blockquote></div>