Joerg Schilling <[EMAIL PROTECTED]> wrote on 23 May 2006 22:22:09 GMT:
> In article <[EMAIL PROTECTED]>, Alan Mackenzie  <[EMAIL PROTECTED]> wrote:
>>Joerg Schilling <[EMAIL PROTECTED]> wrote on 23 May 2006 19:14:34 GMT:

>>> GNU make has many bugs that prevent GNU make from being POSIX compliant.

>>OK.

>>> Some of the bugs are related to the makefile parser and this causes real 
>>> problems.

>>That is, problems which stop you getting your work done, not just
>>special boundary cases you can dream up to break things.

> As many people who are on e.g. Linux and don't know a more POSIX
> compliant make program although there is "smake", these people write
> makefiles that only work in case that the GNU make bugs are present.
> This is one of the most nasty reasons for non-portable programs today.

Thanks, I didn't know that.  I get the feeling you've expended a fair bit
of effort trying to get the GNU make to come into line.

I saw your reply to Ray Ingles.  I can't really see that the thing about
lots of escaped newlines would cause a problem in practice.  Has even a
single person in the entire history of programming ever, in all
seriousness, written consecutive escaped newlines like that?  I mean,
even the shell has problems with WS.  For example, if you call the shell
script foo thusly: foo "a   b"

and foo calls bar so:

   bar "$1"

, bar will see "a b".

But the stuff about $< and $* looks more irksome altogether.  Are you
sure that these differences aren't perhaps more a matter of the POSIX
spec not being 100% unambiguous?

Presumably it is possible, even if not nice, to bridge the differences
between these makes with autoconf and things like that.

-- 
Alan Mackenzie (Munich, Germany)
Email: [EMAIL PROTECTED]; to decode, wherever there is a repeated letter
(like "aa"), remove half of them (leaving, say, "a").

_______________________________________________
gnu-misc-discuss mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/gnu-misc-discuss

Reply via email to