Re: MSVC build test results

[Paul D. Smith]

%% Eli Zaretskii <address@hidden> writes:

  ez> I think POSIX::getcwd should exist on Windows, but I cannot test where
  ez> I'm typing this.  Can you post a short script to test that?

 perl -e 'use POSIX qw(getcwd); my $d = getcwd(); print "$d\n";'

should do it...

  >> No, it's trying to test that backslash handling works properly in
  >> strings that are eval'd... maybe we could quote it so the backslash
  >> actually was preserved instead of removed by the shell.  That might
  >> be a better test anyway since the current test can't determine
  >> between make incorrectly removing the backslash and the shell
  >> correctly removing it.

  ez> If you propose a simple Makefile to test this, I can run it on
  ez> Windows and tell you whether it works as you'd expect.

I checked in a new version of the eval regression test yesterday that
changes this... can you see if that works?

  >> But, my question still remains: why doesn't it print the error.  It
  >> would be nice if someone invoked "make -l 1.0" or something on a Windows
  >> system and told me what it printed.

  ez> It says:
  ez>   gnumake: *** No targets specified and no makefile found.  Stop.

  ez> In other words, the message isn't printed.

  ez> Can you tell me why you think the message about load limits should
  ez> be printed on Windows?  All I see in the source is that it sets
  ez> things up as if there were zero load.

Hrm.  I thought there was a message that was printed by make on systems
that didn't have this feature.  Certainly the regression tests expect
that to be so.  I'll look at the code.

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <address@hidden>          Find some GNU make tips at:
 http://www.gnu.org                      http://make.paulandlesley.org
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist