This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: [PATCH, RFA] Fix handle_stop_sig lossage
- To: Mark Kettenis <kettenis at wins dot uva dot nl>
- Subject: Re: [PATCH, RFA] Fix handle_stop_sig lossage
- From: jtc at redback dot com (J.T. Conklin)
- Date: 07 Feb 2001 14:36:07 -0800
- Cc: gdb-patches at sources dot redhat dot com, ezannoni at cygnus dot com
- References: <200102070800.f1780px04634@delius.kettenis.local>
- Reply-To: jtc at redback dot com
>>>>> "Mark" == Mark Kettenis <kettenis@wins.uva.nl> writes:
Mark> The attached patch makes GDB build again on Solaris 2.6 and
Mark> FreeBSD/Alpha (and probably a couple of more systems). The problem is
Mark> that the fragment that defines STOP_SIGNAL if SIGSTP is defined in
Mark> event-top.h can appear before <signal.h> is included, whereas a
Mark> similar fragment exists in top.c, where <signal.h> is explicitly
Mark> included. This leads to undefined references to handle_stop_sig.
Mark> I suspect the roblem was somehow introduced by the recent
Mark> signals.h removal patch.
I suspect the same thing. I checked the diff, and event-top.c was the
one file where I not only changed "signals.h" to <signal.h>, but also
moved it. I'm not sure why, perhaps a misguided sense of aesthetics.
After all, there shouldn't be any hidden dependencies in header file
ordering :-).
A quick fix would be to change the order back the way it was. But I
personally like your patch better, since it eliminates header file
ordering issues (at least in this instance).
--jtc
--
J.T. Conklin
RedBack Networks