|
Boost : |
From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-12-08 16:06:46
At 08:41 PM 12/2/2001, David Abrahams wrote:
>Me, too. We could also formalize my template instantiation speed test and
>try it with more compilers:
>
>http://users.rcn.com/abrahams/instantiation_speed/index.html
I'd like to see both Jens' compile time results and your instantiation
speed test more widely available.
Issues to be resolved:
* Are we going to change over to using Boost.Jam to run the regression
tests? I'd like to see that happen; it is taking too long already to run
the tests, and that can only get much worse. No point in getting new
tables set up under regression.cpp only to change to Boost.Jam right away.
* The compile timings should somehow report failed compiles differently
from successful compiles. We don't want fast failure to look better than
slow success:-)
* Where do these timings live? CVS (like the current tables) or
separately (like the current regression logs).
* Where do we link to these new tables? Added columns in the
http://www.boost.org/status/compiler_status.html table seem like the
obvious place.
--Beman
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk