From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-04-13 14:53:44
At 03:01 PM 4/13/2001, Thompson, Todd L wrote:
>I think it would be helpful in the compiler table if there was an
>indication if a workaround was required or not.
That is an interesting idea, but hard to do without human analysis.
It might be possible to run the table normally, then run it with a version
of boost/config.hpp that claimed compliance for all compilers, then look at
the differences. But that wouldn't detect individual tests which contained
workarounds for specific compilers.
Just looking at boost/config.hpp is interesting, but also misleading in
that a lot of effort goes into compilers popular with Boost developers, but
other compilers hardly get mentioned at all. So boost/config.hpp only
hints at compliance.
Is it just that Boost developers are a biased sample? Surely someone is
interested in C++ for traditional mainframes? I had a client a few years
ago that used C++ for embedded systems, so I know they exist, but Boost
testing doesn't seem to be attracting them either. Drifting off topic, but
it would be a nice contribution to Boost if someone wanted to run the tests
on something other that Windows/Linux/Unix-like systems.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk