From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Kettenis To: eliz@is.elta.co.il Cc: gdb-patches@sources.redhat.com Subject: Re: [PATCH] Cleanup i386-tdep.c Date: Mon, 19 Feb 2001 08:03:00 -0000 Message-id: <200102191602.f1JG2f622103@debye.wins.uva.nl> References: <200102181017.f1IAHmd11648@delius.kettenis.local> <4331-Sun18Feb2001223249+0200-eliz@is.elta.co.il> X-SW-Source: 2001-02/msg00360.html Date: Sun, 18 Feb 2001 22:32:49 +0200 From: "Eli Zaretskii" > Date: Sun, 18 Feb 2001 11:17:48 +0100 > From: Mark Kettenis > > Please create i386-nat.c (and i386-nat.h, for the exported > interfaces). Is it okay to use preprocessor symbols, like i386-tdep.c does, to get the watchpoint code included only in those x86 ports which support this kind of watchpoints? I mean, i386-nat.c, once created, will probably be added to NATDEPFILES of every native x86 platform, right? However, some of them might not support watchpoints through debug registers (I think Solaris doesn't). In principle, I think we don't want to introduce any additional preprocessor symbols if we can avoid it. Since we only include i386-nat.c in NATDEPFILES for configurations that actually use code from that file, we don't need them yet. If we ever add more code to i386-nat.c and we want to elide the unused debug register code, we can always introduce a preprocessor symbol at that moment. Mark