# [EMAIL PROTECTED] / 2002-10-14 16:09:47 +0300: > On Mon, Oct 14, 2002 at 02:14:23PM +0200, Roman Neuhauser wrote: > > # [EMAIL PROTECTED] / 2002-10-14 14:36:22 +0300: > > > On Mon, Oct 14, 2002 at 11:57:18AM +0200, Roman Neuhauser wrote: > > > > In any event, could someone point me to a place in the make(1) > > > > man page where it says that the S and C modifiers dereference > > > > variables given in both the pattern and replacement, while the N > > > > and M modifiers insist on being fed literal strings? I couldn't > > > > find it. Thanks! :) > > > > > > The last paragraph of the S modifier talks about variable > > > expansion inside the "old_string" and "new_string". There is no > > > similar talk in the M and N descriptions. Even the OpenBSD's > > > make(1) is subject to this same restriction. > > > > Ok, so that's the missing part. Anyway, is there a technical > > reason why this is so? > > > Yes, there is. You forgot to send us the patch. :-)
:) carefully, there. i'm only looking for a [tng]roff tutorial so that I can start sending patches for the man page. :) aside: not to insult anyone or anything like that, plus I'm aware make is not a programming language interpreter, but it's really braindamaged. consider this: .for FOO in foo bar baz FOO=${FOO} .if !empty(FOO:S/baz//) BAR+= ${FOO} .endif .endfor all: @echo BAR: ${BAR} comment out the second line to break it. -- If you cc me or take the list(s) out completely I'll most likely ignore your message. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message