Re: it looks like make --win32 doesn't solve the problem.

[Eli Zaretskii]

> From: Eric Hanchrow <address@hidden>
> Date: Thu, 01 Jun 2006 10:32:09 -0700
> 
> I have a set of pages that describe various failures I've encountered
> with various methods of building Emacs -- it's at 
> http://www.emacswiki.org/cgi-bin/wiki/BuildingOnWindowsTableofContents

I try to analyze your report below, but I'd really like to hear the OP
of this thread to confirm that his attempt failed with the same error
message.  Otherwise, perhaps there's another problem somewhere.

> The one that seems most relevant to this thread is 
> http://www.emacswiki.org/cgi-bin/wiki/BuildingCvsWThirtyTwoCygwin

I believe you are referring to this error message:

     "../nt/oo-spd/i386/addsection" "oo-spd/i386/temacs.bin" 
"oo-spd/i386/temacs.exe" EMHEAP 16
     '..' is not recognized as an internal or external command, operable 
program or batch file.

> I believe those pages have sufficient detail to allow someone to
> reproduce what I did, and see what I saw.

I don't have a Cygwin port of Make (and don't plan on installing one),
but this error message looks very strange to me, because there's
nothing wrong with the command which causes it.  The error message
clearly comes from cmd.exe, but on my machine cmd.exe runs such
commands without any problems, both from the command prompt and from
within (a MinGW-built) Make.

It is possible that Cygwin's Make invokes that command like this:

  cmd.exe /c "../nt/oo-spd/i386/addsection" "oo-spd/i386/temacs.bin" 
"oo-spd/i386/temacs.exe" EMHEAP 16

When I type such a command from the cmd prompt, I indeed see the above
error message.  But if this is how the Cygwin port of Make invokes
commands under the --win32 switch, then that's a bug in the Cygwin
port: it should mirror the forward slashes in the first argument of
the cmd.exe /c command line, because cmd treats forward slashes as
switch characters.  Please report this bug to the Cygwin maintainers.

(To avoid the immense complications of command arg quoting on Windows,
the native MinGW/MSVC port of Make puts the command on a batch file
and invokes cmd to run that batch file, so such problems never happen
with the native port.)

lisp/makefile.w32-in had a small problem with cmd as the shell in some
specific cases (I think they are unrelated to bootstrap), but I fixed
that now.