Date: Sat, 30 May 1998 17:06:50 -0500 (CDT) From: Richard Wackerbarth <rkw@dataplex.net> To: Bruce Evans <bde@zeta.org.au> Cc: current@FreeBSD.ORG Subject: Re: elf vs. bsd.*.mk Message-ID: <l0313030cb195eb594b01@[208.2.87.10]> In-Reply-To: <199805302148.HAA15922@godzilla.zeta.org.au>
next in thread | previous in thread | raw e-mail | index | archive | help
At 9:48 PM -0000 5/30/98, Bruce Evans wrote:
>>>We want them defined for all bsd.*.mk files if they are "constant".
>>>BINDIR can not be defaulted because it varies within a single world,
>>>and LIBDIR is variable if you want it to put {aout,elf} in it.
>>
>>Virtually NOTHING is constant. Quite often if may be desirable for
>>it to take on a default value because we have not specified any
>>overriding value.
>
>Virtually everything is "constant", i.e. has the same values for all
>sub-makes.
>
>>I SHOULD be able to simultaneously compile, from the same source tree,
>>systems for two different machine architectures and/or variations
>>of compile parameters.
>
>Not in a single invocation of make.
I agree. However, my concern is that the prior tendancy has been to
expand the "invariant within a tree of sub-makes" into "globally constant
across all uses". For example, there have been numerous assumptions
that the source tree is ALWAYS in /usr/src. Here are still places that
effectively assume that /x/y/z/../../y/z is the same as /x/y/z. In the
presence of symbolic links, this may well be false.
Richard Wackerbarth
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?l0313030cb195eb594b01>