Date: Thu, 27 Jun 1996 17:24:47 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@GndRsh.aac.dev.com> To: jkh@time.cdrom.com (Jordan K. Hubbard) Cc: nate@mt.sri.com, scott@statsci.com, current@FreeBSD.ORG Subject: Re: Building inside of /usr/src? Message-ID: <199606280024.RAA13727@GndRsh.aac.dev.com> In-Reply-To: <7235.835914709@time.cdrom.com> from "Jordan K. Hubbard" at "Jun 27, 96 03:31:49 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
> > Well, you already took it out, so I guess comments don't matter, but > > there's *probably* a reason for doing it. You're best bet would be to > > remove it *locally* and then test the change to see if bad things happen > > before spamming the repository making us find all the bugs. :( > > I did test it and I carefully followed the logic of the code before > making the commit (that's why the commit followed so quickly - it > became very quickly clear that this behavior was now a bug in the new > system). > > There's absolutely *no* reason to ignore the output of getcwd() and it > does check the directory returned for errors / permission problems. > > The only thing I can figure out was that someone found AMD's path > mangling to be non-intuitive (though functional) or something and > figured they'd substitite in the value of $PWD. I can think of no > other reason for it and, in fact, it could easily be argued that they > weren't thinking straight when they made this change as it makes it > far too easy to confuse the bejeezus out of make by using a bogus > shell or setting the value of PWD yourself. If you read the comments above this section of code, and have worked on asymetric AMD managed user home directory systems, and used pmake in parallel mode on these systems it would be clear as to why it if preferintial to use $PWD if $PWD infact resolves to the sameplace as getcwd(). I may have one AMD mounting /home/* onto /net/blah/* and another that mounts them to /auto/* and yet another that mounts to /tmp_mnt. In this case $PWD (ie /home/rgrimes) always correctly resolves to my home directory on _ALL_ machines, but /tmp_mnt/gndrsh/A/rgrimes will not resolve correctly on my Apollo systems (it's /net/A/rgrimes overthere). This ``feature'' of make is used when launching builds for architecture X when running it from a machine of arch Y via a remote make. Please back out your commit... there was, and is, a good reason for doing what it does. The brokeness is in you new .mk stuff if any place. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199606280024.RAA13727>