Date: Tue, 11 May 2004 19:54:58 +0100 From: Colin Percival <colin.percival@wadham.ox.ac.uk> To: obrien@FreeBSD.org Cc: Nate Lawson <nate@root.org> Subject: Re: cvs commit: src/sys/dev/led led.h Message-ID: <6.1.0.6.1.20040511194832.03e5e988@popserver.sfu.ca> In-Reply-To: <20040511181554.GA13486@dragon.nuxi.com> References: <20040510115040.0C9B516A53A@hub.freebsd.org> <20040511100208.C75906@root.org> <20040511181554.GA13486@dragon.nuxi.com>
next in thread | previous in thread | raw e-mail | index | archive | help
At 19:15 11/05/2004, David O'Brien wrote: >On Tue, May 11, 2004 at 10:02:59AM -0700, Nate Lawson wrote: >> Do we even want __FBSDID in .h files? It doesn't seem to make sense since >> it's not a compilation unit (i.e. no linkage). > >It actually does make sense, as a header's contents does wind up in a .o >eventually :-) And header contents can be the cause of problems as much >as a .c file. There is no problem having multiple __FBSDID in either >a.out or ELF objects. However, there are problems with having a header file's __FBSDID tags end up in a several binaries. Quite apart from the resulting bloat, most changes to header files don't actually result in many binaries being modified; including compilable $Id$ tags in the headers would result in lots of spurious binary changes. This would make me (and anyone who uses FreeBSD Update) very unhappy. Colin Percival
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6.1.0.6.1.20040511194832.03e5e988>