Date: Fri, 22 May 1998 15:02:51 +1000 From: Bruce Evans <bde@zeta.org.au> To: eivind@yes.no, gibbs@plutotech.com, jkh@time.cdrom.com Cc: cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-sys@FreeBSD.ORG, jkh@FreeBSD.ORG Subject: Re: cvs commit: src/sys/pci if_tl.c if_tlreg.h Message-ID: <199805220502.PAA03138@godzilla.zeta.org.au>
index | next in thread | raw e-mail
>> > No, no, no. Create an osreldate.h file instead. Please? Using __FreeBSD__ >> > doesn't work because it's tied to the compiler, not the source. No. Neither should be used. >> 1. The osreldate.h file just isn't created in time and if we want >> our SNAP builds to work for now, this change has to happen. >> If we want to work out how to get the osreldate.h file bootstrapped >> into place at a later time then that's fine with me, but there's >> no use holding up the show until then. > >That was the change I installed and backed out so fast earlier today; >it unfortunately didn't work. Fortunately :-). >> 2. Absolutely *nothing* else in the system that I've been able to >> find uses __FreeBSD_version directly in the way that this driver The isp driver uses it. >> did. If this is really a step forward then I'd suggest that >> a whole lot more needs changing if self-consistency is also your >> goal (and I think it should be, especially where this kind of >> stuff is concerned, since other driver writers will imitate it). It's a step backwards. We use cvs for version control here. >The reason nothing else use __FreeBSD_version directly in this way is >that we don't provide it to kernel files, and that _suck_. It is one >of the biggest hurdles for people attempting to do development outside >of the FreeBSD tree proper. That may be so, but when the sources are committed to the tree proper, they shouldn't have ifdef version spaghetti. Brucehelp
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199805220502.PAA03138>