[Boost-testing] [Bronek_office][como-4_3_3-vc7_1] The most broken platform in town?

From: Joaquín Mª López Muñoz (joaquin_at_[hidden])
Date: 2006-07-17 08:49:45

I don't think so, despite its being punctered by countless red cells.
My opinion is that the aforementioned environment is having some
sort of misconfiguration problem by which bjam is not able to
find the libs to link against. So, the many red cells we're
seeing are just those testcases requiring some sort of linking (to
Boost.Test, Boost.Serialization, Boost.DateTime etc.) Please have
a look at the last post here about this problem:

http://lists.boost.org/boost-testing/2006/06/3084.php

where Bronek refers to a bjam run he did by hand: a casual inspection
seems to reveal that this manual run is actually able to link
successfully in places where the regression tool fails!

I'd be extremely grateful if this could be given a little investigation,

since as it stands now the toolset is contributing perhaps more than
50% of total regressions. The mentioned post also contains some
open questions regarding what could be a problem with toolset tags in
lib names.

Thank you very much,

Joaquín M López Muñoz
Telefónica, Investigación y Desarrollo