Vladimir Prus writes:
Jeff Garland wrote:
Looking at the boost-wide reports today, testmicrosec_time_clock is marked as failing red on a whole series of compilers. However, they should be 'green' (and have been in past reports) as they have been explicitly marked as failing:
I noted the same thing.
http://boost.sourceforge.net/regression-logs/cs-win32_metacomm/developer/pro...
has only a couple of yellow cells and boost-wide reports have a huge number of yellow ones.
There *was* a configuration problem with Boost-wide reports that did lead to a stale markup, but in general the comparison above (or at least the wording) is not valid -- there are more platforms in the Boost-wide reports than in any platform-specific ones, and therefore counting overall number of cells of a particular color doesn't make much sense.
While we're at it, I have two questions about boost-wide regression reports
1. How they are related with individual reports? Are they generated from individual reports.
They are generated from the same logs, yes.
2. Old style reports showed warnings, for example see:
http://boost.sourceforge.net/regression-logs/cs-Linux.html
New style reports do not. I find it not very good. I'd really want boost to be warning-free on gcc 3.4 and fixed some errors already. But I'm not sure the above page is up-to-date, and it's hard to search for warning in it. Maybe, new style reports could show warnings too?
Put on the TODO list. -- Aleksey Gurtovoy MetaCommunications Engineering