From owner-freebsd-current Sun May 12 22:46:02 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id WAA19072 for current-outgoing; Sun, 12 May 1996 22:46:02 -0700 (PDT) Received: from grumble.grondar.za (root@grumble.grondar.za [196.7.18.130]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id WAA19067 Sun, 12 May 1996 22:45:56 -0700 (PDT) Received: from grumble.grondar.za (mark@localhost [127.0.0.1]) by grumble.grondar.za (8.7.5/8.7.3) with ESMTP id HAA05505; Mon, 13 May 1996 07:45:45 +0200 (SAT) Message-Id: <199605130545.HAA05505@grumble.grondar.za> To: Bill Paul , current@freebsd.org Subject: Re: Make world breakage... Date: Mon, 13 May 1996 07:45:41 +0200 From: Mark Murray Sender: owner-current@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Sorry for this spam, guys. I had an environment variable set from from doing a ".script" about three days ago. I feel like such a wally.... M Mark Murray wrote: > [copied to current] > > Bill Paul wrote: > > > Bill Paul wrote: > > > > > Looks like some problem with "inline"s? > > > > > > > > > > Script started on Sat May 11 13:41:51 1996 > > > > > bash# make > > > > > gcc -W -Wall -pedantic -ansi -O -pipe -O -DLIBC_RCS -DSYSLIBC_RCS -D_ _D > BINT > > > ERFACE_PRIVATE -DPOSIX_MISTAKE -I/a/src/lib/libc/locale -DYP -c /a/src/li b/ > libc > > > /gen/getpwent.c -o getpwent.o > > > > > > > > [chop] > > > > > > > > Well, I'll fix this of course, but, I think I'm confused: have we > > > > always been building the world with '-pedantic -ansi?' > > Hmm... I fixed this by hand, and like you said, it fell over left, right > and centre with lots more inlines. I did a hunt for -pedantic -ansi, > and cant find where it is creeping in. I am as confused as hell, now. > > > > Not sure. Must be? I just checked my make.conf for nonstandard CFLAGS, > > > and that is commented out. I check my sources out of CVS, so I can see > > > if anything else is different, and so far everything (except this) is > > > peachy... > > > > But I don't understand... there are other places in libc that use > > the 'inline' keyword which should also fail (look in libc/stdlib/getenv.c, > > among other places). At least, they fail when I try to compile with > > with -pedantic -ansi. If we've been using -pedantic -ansi all along, > > then these should also have failed, unless they're doing something > > magic that I just can't see. Have you tried building these to see > > what happens? If they barf too, then you'd assume someone would have > > noticed it a long time ago. Also, I made these changes before > > the last SNAP went out: if they don't compile, then how'd they manage > > to get a release built? > > Oink? I haven't a clue... > > > If you change the 'inline' keywords in getpwent.c to '__inline__' then > > it will compile (though there will be other non-fatal warnings). Try > > this and see if it completes normally. My money says it'll die somewhere > > else. > > > > Note that ypserv also uses the 'inline' keyword, so it too may barf > > (if you are indeed using the -pedantic -ansi flags for everything). > > > > I'm still confused. > > Me too. Big time. > > M > > -- > Mark Murray > 46 Harvey Rd, Claremont, Cape Town 7700, South Africa > +27 21 61-3768 GMT+0200 > Finger mark@grondar.za for PGP key -- Mark Murray 46 Harvey Rd, Claremont, Cape Town 7700, South Africa +27 21 61-3768 GMT+0200 Finger mark@grondar.za for PGP key