From: Jens Maurer (Jens.Maurer_at_[hidden])
Date: 2001-09-23 13:39:07
Daryle Walker wrote:
> I think the regression tests currently result in a custom web page being
> made. Maybe the regression tests should return precise XML instead so it
> can be analyzed by other applications. We would need something to create
> the current HTML version too, though.
The problem is that parsing XML the easy way, i.e. employing some
library such as Xerces (xml.apache.org), requires third-party software.
I feel that third-party software should not be mandatory in order to
be able to use boost. That was also one of the reasons for replacing
the Python-based regression test driver with a C++ one.
It's certainly an option to output XML in addition to HTML, as long
as boost tools never need to read it back in again. It would also
be possible to generate HTML from XML by means of XSLT (a transformation
tool/specification), but that of course also requires third-party
Please submit a suitable DTD for review.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk