Hi, > Good point. I'm not too sure either; it depends what the current > needs are, or what they would be if the code was nicer. Options > include: just using atexit(3); ...
With good timing, this just cropped up. With gcc 4.9.2, `-O -D_FORTIFY_SOURCE=2', on 1.6-branchpoint-1494-g23645d5, I get uip/sendsbr.c: In function ‘sendsbr’: uip/sendsbr.c:749:24: error: variable ‘host’ might be clobbered by ‘longjmp’ or ‘vfork’ [-Werror=clobbered] const char *addr, *host; I guess because it's inlining handle_sendfrom into the function that calls setjmp(3), and adios() calls that `done' function pointer which has been set to do a longjmp() in armed_done(). Though I think the coder is allowed to disregard inlining effects; perhaps some gcc's get this wrong. -- Cheers, Ralph. https://plus.google.com/+RalphCorderoy _______________________________________________ Nmh-workers mailing list Nmh-workers@nongnu.org https://lists.nongnu.org/mailman/listinfo/nmh-workers