Re: Make CVS HEAD version build out of the box

[Yongwei Wu]

On 02/10/2007, Eli Zaretskii <address@hidden> wrote:
> > Date: Mon, 1 Oct 2007 22:36:41 +0800
> > From: "Yongwei Wu" <address@hidden>
> > Cc: address@hidden
> >
> > > Beware: the make.exe binary produced in this way is not guaranteed to
> > > behave like intended.  This is because this way of building a native
> > > Windows port of Make is not officially supported.  As one caveat, you
> > > will see that config.h.W32.template has near its end a few symbols
> > > whose definition changes Make's behavior wrt program invocation with
> > > various shells you might have installed: I'm not sure what the
> > > configure script does with those definitions, so please at least look
> > > at those symbols and verify that the binary you produced will behave
> > > as you'd like it to.
> >
> > I checked config.h, and the only things I have some doubts are:
>
> I didn't mean config.h, I meant specifically config.h.W32.template.
> (Because you ran the configure script, it used the general config.h.in
> rather than the W32 template, to create config.h, and config.h.in does
> not have those symbols.)  There are several symbols near the end of
> the template that you need to consider:
>
>  BATCH_MODE_ONLY_SHELL
>  HAVE_CYGWIN_SHELL
>  HAVE_MKS_SHELL
>
> Please read the comments that describe each one of them (and the code
> that uses them, if the comments are not clear enough).

OK, I did compare config.h.W32.template and my config.h, and found
nothing significant other than those I mentioned last time.  I do not
think I need the settings you mentioned.  I want a clean native Win32
make.exe, like the one provided in MinGW.

I can confirm that the PATH_SEPARATOR_CHAR setting in the config.h by
my ./configure procedure is wrong.  Earnie, maybe you want to have a
check whether this issue exists in the MinGW build.  I hope not.

Best regards,

Yongwei

-- 
Wu Yongwei
URL: http://wyw.dcweb.cn/