Re: [Gcl-devel] ANSI on Windows - 156 out of 4569 tests failed

[Mike Thomas]

Hi Camm.

> Greetings!  This is great to hear, Mike!  When did the ansi image
> start building on Windows?  Last I heard the relocations died at
> braid.o.

It is possible to build an ANSI image provided (quoted from "readme.mingw):

"A PCL/ANSI build ("configure --enable-ansi --enable-custreloc") can be
obtained by adding "(si::use-fast-links nil)" to the end of the list of
commands in the SETUP makefile macro in "pcl/makefile"."

To recap on the status of the Mingw port:

1. The ANSI "braid.o" link problem still exists (as does a workaround).

2. The C_INCLUDE_PATH and related Windows/POSIX shell script path problems
remain - Earnie Boyd, the maintainer of MSYS, is aware of this and is
working on it.  He even downloaded GCL!

3. BFD is still a problem.   The latest development is that the
"bfd_openr()" call on the ".o" file gives an error number 5 which is
"bfd_error_invalid_operation", although the program doesn't fail until the
usual place and indeed seems to .

4. Stratified Garbage collection does not work and sits at the bottom of my
list.

5. I have got gcl-tcl to the point where it launches the server but a
failure occurs in the stream writing code.

Cheers

Mike Thomas.