From: John Maddock (john_maddock_at_[hidden])
Date: 2002-07-17 05:19:29
> I would find the following useful: a table which gave for each
> compiler deficiency of importance to us, and for each boost-supported
> compiler the information of whether the deficiency is present or not.
> In other words, something like (one table per platform):
> MCWPro8 Gcc2.95 GCC3.02
> BOOST_NO_TEMPLATE_TEMPLATES X
> BOOST_NO_ARGUMENT_DEPENDENT_LOOKUP X ?
See the output of the config_info program
(libs/config/test/config_info.cpp), if you don't want to run it yourself
then the regression logs have the output from this listed before anything
> We can, or rather should, be able to get this information from reading
> the headers, but this whould be clearer and easier to use (if
> politically incorrect :-) ).
> A similar layer for the library aspect (BOOST_NO_STDC_NAMESPACE...)
> would be nice also but more complex (several libraries per compiler), so
> perhaps less usefull.
> This also raises a question: do we, or should we, have a battery of
> tests for detecting those deficiencies we care about? It seems to me
> that these macros are defined or not more or less as we request them, as
> we get bitten by the deficiencies. I do not advocate that the resulting
> config files be automatically built from these tests (there are always
> some unique compiler quirks, after all...), but at least then we would
> stand a fighting change of not forgeting some.
We've got them, along with a configure script if you want it - please refer
to the config docs, particularly the section on testing your current config.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk