bug-make
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Built make-3.80rc1 on Cygwin - brief


From: Soren Andersen
Subject: Built make-3.80rc1 on Cygwin - brief
Date: Sat, 31 Aug 2002 07:20:12 UT

Hello,

Just a brief report. BTW, I am trying unsuccessfully to use
gmane -- www.gmane.org -- to POST to this List; I can READ msgs just
fine using gmane (which is the only way I'll ever read this List;
somewhere around the 18th developer's List I subscribed to, some time
back, I became completely overloaded with List email traffic on
conventional MUAs...)

I have built make-3.80rc1 on Cygwin (the posix-emulation layer / Gnu
tools porting platform for Win32). Doing so required selective care in
invoking targets from the autof__ked, uh, automake-based, that is,
Makefile ;-). And I used the configure options to build the slimmest
`make' I could, avoiding nls.

The make I built fails some of the test suite (in fact one test scripts
*dies*) because the test "rig" for make is buggy and needed / needs a
good going-over to update certain things ("Windows" hasn't been the
id string for Perl's notion of $^O on Win32 for years now!), and also
to bring it into conformance with modern standards of Perl scripting.
It's a decade old. There's no "-w" and no "use strict" anywhere in
sight. It's dangerous in that it will easily give misleading results
(either positive or negative) because so many failues are being allowed
to happen silently -- undefined variables and so on that would be
caught by upgrading to stricter Perl coding standards.

In fact, when the test 'rig' is fixed as I have done, make passes all
test on Cygwin. Congratulations, Paul Smith, on a very promising
release of `make' that I am eager to begin using.

  Cygwin -and- MinGW user,
     Soren Andersen
   -- 'perlspinr' --


-- 
http://fastmail.fm - You've just been FastMailed!




reply via email to

[Prev in Thread] Current Thread [Next in Thread]