Re: [PATCH] Use UTF-8 active code page for Windows host.

[Costas Argyris]

Here is the patch with the Basic.mk.template and mk/Windows32.mk

changes.    I tried to keep most of the changes in the Windows-specific

file, but something had to happen in the general one as well, as far as

I could tell.

I deliberately sent only the changes relevant to the Basic.mk approach

here to facilitate review.

The changes are working as far as I can tell, with both

.\gnumake.exe -f Basic.mk TOOLCHAIN=msvc

and

.\gnumake.exe -f Basic.mk TOOLCHAIN=gcc

(and without TOOLCHAIN at all, of course, defaulting to msvc)

First I had to run

.\bootstrap.bat

to re-generate Basic.mk after changing Basic.mk.template and mk/Windows32.mk

These produce UTF-8-supporting gnumake.exe binaries in their corresponding

WinRel and GccRel folders.

As you will see, I haven't implemented checking if windres is available on

the path for the TOOLCHAIN=gcc case, as I have never come across a

case where a gcc distribution for Windows doesn't include windres (in all

cases I have seen, it gets distributed with binutils because it needs the

assembler and linker that are not part of gcc, and windres is part of binutils

so it's also there).

Tcc is not supported with the Basic.mk approach anyway, so no need to

worry about the fact that it doesn't ship with windres.

MSVC always has its own resource compiler available.

If you still need the check for windres for the gcc case, please let me know.

On Wed, 17 May 2023 at 14:10, Paul Smith <psmith@gnu.org> wrote:

On Wed, 2023-05-17 at 12:47 +0100, Costas Argyris wrote:
> However, when trying to prepare the new patch I realized that
> Basic.mk is an untracked file which is listed in .gitignore, so how
> would you like me to show you these latest changes?

The file to be changed is Basic.mk.template

Sorry I have been out of touch this week; I likely will not have much
time until the weekend.  I'm looking forward to the patch though!

For your initial question I understand the issue and I'm not sure how
it works for me I'll have to try it again.  It does seem wrong that we
replace the running instance of make; that can never work on Windows
(it works on most POSIX systems).

Basic-mk.patch
Description: Binary data