Re: rcs tags and tests

Dumas Patrice wrote:
> [@ifhtml tags are] indeed a good thing, however it doesn't solve the problem here, as
> we use third part manuals as tests (but maybe this is not a good idea ?).
> 
> A sed script which removes the $Id ...$ string seems to be the best. Do
> you agree on that ?

sed scripts are fine for now and probably for good with some of the 
manuals.  There may be exceptions, as with the CVS manual.

With the case of the CVS manual in particular, the RCS keywords are 
integral parts of the generated manual, however, and I would like to be 
able to confirm in the test scripts that they survive texi2html intact. 
  This will, unfortunately, have to wait for me to fix the @ifhtml tags 
in texi2html and then the CVS manual itself to use them.

As for third party manuals in general, I like the idea a lot - they are 
already large, freestanding sets of data we don't need to do extra work 
to produce and probably run texi2html through valid real-world scenarios 
we may not have come up with even with careful planning.  If we notice 
features that aren't being tested in the 3rd party manuals, however, 
then we should maybe introduce them into our own test documents.

And, of course, there is always the solution of asking the 3rd parties 
to patch their manuals with @ifhtml tags to protext their keywords when 
necessary.  :)

Derek