[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Tue, 25 Jan 2005 19:01:49 -0600
I've just started
using gnu make. My goal is to get it to work on Windows XP without a unix shell.
I noticed a conversation a few months ago where David Baird offered a patch so
that "SHELL = cmd.exe" would work in a makefile. I tried his fix, and I see that
if I set SHELL = cmd.exe in a makefile, that make has already done a search for
sh.exe, and if found, sets default_shell. Setting SHELL in the makefile doesn't
override this, so sh.exe is still used. I've tried a few hacks to get around
this, but have not found anything satisfactory. Now I see a patch that attempts
to use -j for cmd.exe, but before I work on that, I know I have to fix
the SHELL problem, because I can't guarantee that all systems that run my
makefiles will be without a sh.exe, and the behavior of the makefile will vary
based on that.
I get the feeling
that CMD.EXE is thought of as not much better than COMMAND.COM, and that's just
not the case. You don't have to force CMD.EXE to use batch mode (although it
still requires shell switches to have a "/" rahter than a "-"), but fixing all
of the CMD.EXE stuff looks like a big job, but before I start I want to
see if it is already being done.
Is anyone else in
this area of the code? I'd like to work together on
- CMD.EXE support,