Re: [boost] g++2.95 patch for Test.lib

From: John Maddock (john_at_[hidden])
Date: 2003-12-23 06:34:36

> Gennadiy emailed me a couple of days ago saying he is on vacation and will
> be out of email contact, and that he would like help with issues affecting
> Linux, which he does not have access to.
>
> So I'd say the answer is Yes.

Unfortunately further testing revealed this to be an "anti patch", the
following I managed to send to the wrong address yesterday (should have come
here):

John

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

> The following patch gets the last regressing Boost.test test program
> compiling with gcc2.95.3,

Strangely on Red Hat 9 that test (test_tools_test) is failing with all gcc
versions for me at runtime, even though the regression tests at
http://boost.sourceforge.net/regression-logs/cs-Linux/developer_result_page.html#test
show it passing,

... looking at it further it seems that the preprocessor directives are
messing up the test somehow (removing them results in the test passing),
so... patch withdrawn.

... and further testing, if I so much as add a single newline to that file,
the test starts failing at runtime !

What is it with this file?

John.