httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rodent of Unusual Size <Ken.C...@Golux.Com>
Subject Re: [PATCH] lets do 1.2.5
Date Thu, 08 Jan 1998 15:56:59 GMT
Martin Kraemer wrote:
> On Wed, Jan 07, 1998 at 01:51:19PM -0500, Jim Jagielski wrote:
> >
> > Nope. Not in the CC variable. Someplace else, yes, but not CC
> I disagree with your view, and with Ken's statement that $(CC) must be a
> possible argument to which(1). There is no prior art where $(CC) is
> `which'able or `which'ed, but (as Mark explained) there are some switches
> which can be seen (together with the executable name)
> as "the atomic compiler call" for a certain situation.

I'm not going to argue the 'no prior art' claim because I consider
it indefensible - and not worth the effort.  How many compilers can
dance on the head of a pinout, anyway?

> If we were to impose the restriction that this be a single word, then we
> force every user who happens to have such a compiler to write his own
> wrapper script around cc, only to be able to use it with apache.
> OTOH, the patch (as posted on the list) simply replaces "make CC=$(CC)"
> by "make CC='$(CC)'" which is neither error prone nor incorrect, and is
> the same mechanism as is used for AUX_CFLAGS.

I'm rescinding my objection too; I'm not totally convinced, but I'm
persuaded enough to say 'okey.'  I think we should somehow encourage
users to only stick the stuff in CC that can't be put into *FLAGS.
Hopefully that will be default behaviour for most people.

So +1 on changing $(CC) to '$(CC)' for 1.2.6.

Marc, concerning the discussion about this for 1.3 that you mentioned
to Jim..  I remember the discussion about changing $(CC) to '$(CC)'
and the arguments for and against were posed differently that time than
they have been this.  So it's reasonable that Jim and I regarded
this as a new discussion.

#ken	P-)}

View raw message