Re: GNU make 3.81rc1 released -- please test.

[Steve James]

On Monday 20 February 2006 04:25, address@hidden wrote:
> Hi all;
>
> Please find the first release candidate for GNU make 3.81, 3.81rc1,
> available now for download from ftp://alpha.gnu.org/gnu/make:
>
>   c907a044ebe7dff19f56f8dbb829cd3f 
> ftp://alpha.gnu.org/gnu/make/make-3.81rc1.tar.bz2
> d9e1fd176b32da0edf5c43b6cca9f493 
> ftp://alpha.gnu.org/gnu/make/make-3.81rc1.tar.gz
>
>
> I know that there are still outstanding bugs and enhancement requests,
> but this is the functionality I'm going to release with 3.81.  Please
> test this and report any _critical_ bugs or issues: core dumps, new
> regressions, etc.  I'm anticipating a one week cycle, so, unless enough
> issues are found that another release candidate is warranted, 3.81 will
> be released next weekend.
>
>
> Thanks, and have fun!

I've been using 3.81beta3 up to now. Here's what happened with 3.81rc1...

Linux
================================
  Linux ivybridge.beachsolutions.plus.com 2.6.11-12mdk #1 Mon Jun 27 23:18:37 
MDT2005 i686 Intel(R) Pentium(R) 4 CPU 2.50GHz unknown GNU/Linux

348 Tests in 96 Categories Complete ... No Failures :-)

My make system still works. Hooray.


Solaris
================================
  SunOS ventnor 5.8 Generic_117350-29 sun4u sparc SUNW,Sun-Blade-100 Solaris
  (compile warning: glob.c:197: warning: passing arg 0 of `my_realloc' from 
incompatible pointer type)

347 Tests in 96 Categories Complete ... No Failures :-)

My make system still works. Hooray.

================================
  CYGWIN_NT-5.1 derriford 1.5.18(0.132/4/2) 2005-07-02 20:30 i686 unknown 
unknownCygwin
  gcc.exe (GCC) 3.4.2 (mingw-special)
  #define HAVE_CYGWIN_SHELL
  (compile warning: sub_proc.c:708: warning: dereferencing type-punned pointer 
will break strict-aliasing rules)
  (compile warning: ./glob/glob.c:197: warning: passing arg 1 of `my_realloc' 
from incompatible pointer type)

31 Tests in 18 Categories Failed (See .diff files in work dir for details) :-(

Hmm, I've never seen the Windows tests pass, so I carried on regardless. Maybe 
the following is academic since the tests are already telling me there are 
problems. I persevered because the tests for 3.81beta3 didn't pass either, 
yet the executable is still usable...


This has stopped working on Windows (but still works for Solaris & Linux):
   escaped_newline:=\\\\\\\n
   echo -e blah blah $(escaped_newline) more stuff
Now "\n" appears in the output. SHELL=/bin/bash. This is all Cygwin.


This line doesn't work on Windows any more

   error_text += $(shell $(PERL) -e 'foreach (@ARGV) {print qq/*** $$_ is a 
prerequisite file but it does not exist/ unless -f }' $(error_checkReqFiles))

and get the following message from make. Not seen this before:

  process_begin: CreateProcess(NULL, /usr/bin/perl -w -e "hello foreach 
(@ARGV) {print qq/*** $_ is a prerequisite file but it does not exist/ unless 
-f }" /home/build/tools/jellyfish1/x86-pc-win32/flexsdk/bin/lmnewgen.exe, ...) 
failed.

The shell command has expanded to the correct value and works if entered 'by 
hand'. The single quotes in the script have mysteriously changed to double 
ones in the error message?

So maybe that's of help?
Cheers, Steve.