Re: [boost] PREPROCESSOR library is now in CVS

From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-11-27 16:24:02

At 03:45 PM 11/27/2001, Jens Maurer wrote:

 >Beman Dawes wrote:
 >> Boost doesn't want to get into the compiler conformance and performance
 >> testing business.  But it might be nice to figure out a way to shine 
some
 >> light on compilers with less than robust preprocessors, very slow
 >> preprocessors, less than robust template engines, very slow template
 >> instantiation, and some of the other QOI issues impacting Boost users.
 >
 >I'll add a "execution time" output to the regression test engine,
 >so that, in addition to compiler error messages, the total time
 >spent compiling is shown in the full regression logs.
 >
 >Anyone in favor of having a separate HTML page with execution
 >times?

Are you talking about compile time, test program execution time, or both?

Regardless, you might want to do some experimentation first to see if there 
are meaningful differences between compilers.  Since the regression tests 
aren't designed for timing, they may execute so quickly that the time is 
below the available timer resolution.  We need to feel our way a bit, and 
make sure we are reporting useful information.

 >All times are only meaningful when compared to executions on the
 >same platform (due to different underlying hardware), so Windows
 >vs. Linux comparisons don't make much sense.

Agreed.

--Beman