[boost] Re: Regressions against 1.30.2

From: David Abrahams (dave_at_[hidden])
Date: 2003-11-28 17:11:14

Beman Dawes <bdawes_at_[hidden]> writes:

> Looking at the regressions against 1.30.2 for Win32 and Linux from
> http://tinyurl.com/jvyc and http://tinyurl.com/k36t, there are a bunch
> of regressions. See below.
>
> Before looking at the detailed list, bear in mind that almost all of
> the Linux failures are new gcc 2.95.3 fails. My guess is that if type
> traits could be fixed on gcc 2.95.3, many of the other regressions on
> that platform would disappear.
>
> On the Win32 tests, the regressions are more varied. The Date-time and
> Regex libraries need a bit of attention, but I'm hoping all Boost
> developers will scan the list for regressions in their own libraries.
>
> Config:      Linux gcc 2.95.3 two tests
> Conversion:  Linux Intel 7.1 lexical_cast_test
> CRC:         Linux gcc 2.95.3 crc_test
> Date-time:   Win32 Borland 5.6.4 and Codewarrior
> Disjnt sets: Linux gcc 2.95.3
> Dyn bit set: Linux gcc 2.95.3
> Filesystem:  Win32 Borland/Intel (But not reproduced on Beman's Win32
> tests)
> Format:      Win32 Borland 5.5.1, Codewarrior, VC++ 6.0
> Function:    Linux gcc 2.95.3 three tests
>               Win32 Codewarrior function_n_test
> Graph:       Linux gcc 2.95.3
> Iterator:    All platforms failing iterator_categories test
                                     ^^^^^^^^^^^^^^^^^^^^^^^^
Sorry, what test is that?  I don't see anything by that name.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com