Re: make3.81-beta4 mingw not failing on error?

[Matt England]

At 3/7/2006 09:30 PM, Earnie Boyd wrote:
> I suspect that you are running into issues of using more than one runtime 
> MSYS and MSVCRT.  MSYS provides its own versions of make so that it can 
> control the processes.  You've used the MSYS toolset to build a new 
> version of make for a target that uses the MSVCRT runtime and MSYS does 
> not have control of it.  Use your new version of make without MSYS in the 
> PATH and use the CMD.EXE shell.  If you still have problems then there is 
> an issue with make; else it will be impossible for you to prove there is one.

Interesting.  Ok, I'll trying running the make from cmd.exe without msys in 
the path after I get done with my Debian-based mingw cross compile work.

> You've used the MSYS toolset to build a new version of make for a target 
> that uses the MSVCRT runtime and MSYS does not have control of it.

Is there a means by which I could build a 3.81rc1-based make that has 
proper "control" of stuff (MSVCRT, etc) so I don't have these 
problems?  ie, what procedures do the official MSYS people use to build 
their make?

And boy, would I love them (the MSYS folks) to build a new, 3.8x-base make, 
so it could support my heavily-$(eval)-ed Makefiles.  :)

I also begin copying the mingw-users list; for those reading from that 
list, you can see the context of what led up to this discussion by reading 
from the main email that started this thread:

http://lists.gnu.org/archive/html/make-w32/2006-03/msg00014.html

-Matt