[Boost-testing] Re: 'required toolsets' was (Re: [iostreams] need help understanding regressionre port)

From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2005-02-15 16:13:18

Jeff Garland writes:
> On Fri, 11 Feb 2005 17:47:07 -0600, Aleksey Gurtovoy wrote
>
>> That's because the page only lists failures for "required" toolsets -
>> - the ones that we (Boost developers) commit to keep in accordance with
>> "no regression" rule. The toolsets are enumerated in
>> http://cvs.sourceforge.net/viewcvs.py/boost/boost/status/explicit-
>> failures-markup.xml and are highlighted as bold on the library and 
>> summary report pages.
>
> That's interesting -- was there a discussion I missed about the
> 'required set' of compilers / standard libraries?

I'm not sure if you've missed them, but there was a few which were
summarized here:

    http://article.gmane.org/gmane.comp.lib.boost.devel/95208
    http://article.gmane.org/gmane.comp.lib.boost.devel/111277

> Somehow I was under the impression that each library author was
> still pretty much deciding what to do with these.

Yep, but it doesn't contradict what I said. Remeber that "no
regressions" rule still allows you to mark up the toolset as
"unusable".

-- 
Aleksey Gurtovoy
MetaCommunications Engineering