RE: [Gcl-devel] 2.6.2

[Mike Thomas]

Hi Vadim/Camm.

Some slightly better news on MinGW32 gcc 3.3.3/3.4.0; the MinGW32 "compiler"
problems are at least partly alleviated by using binutils 2.14.90 rather
than 2.15.90.

A few days ago I wrote in response to Vadim's test report:

| > I've reproduced that error with gcc 3.3.3 and binutils 2.15.90.

I did another test using binutils 2.14.90 instead of 2.15.90 with both gcc
3.3.3 and 3.4.0.

The gcc 3.3.3 Maxima "binary-gcl/specfn.o" problem goes away and with gcc
3.4.0 GCL now builds itself and Maxima - previously the GCL build failed.

The only fly in the ointment is that both these builds of Maxima now crash
the very first regression test in "make check".

Having said that, at least now both of the problematic C compilers are
producing identical results with GCL and Maxima which does simplify matters:

===================================================================
....
(C1)                            RUN_TESTSUITE()

Running tests in rtest1.mac:
Unrecoverable error: Segmentation violation..

This application has requested the Runtime to terminate it in an unusual
way.
Please contact the application's support team for more information.
===================================================================

In all these tests I used the recently released stable MinGW32 runtime 3.3.
Separate tests with that new runtime and retaining binutils 2.15.90 failed
to make any change for the better.

On that basis I am ruling out the new runtime as a contributor to those
specific problems cured by the change in binutils version.

Best wishes for a pleasant weekend to everyone.

Mike Thomas