On Fri, 8 Oct 2004, Tim Jenness wrote:
> On Fri, 8 Oct 2004, Peter W. Draper wrote:
>
> > may have seen the odd less Solaris-specific delta's being committed (for
> > instance Tim seems to have the habit of adding CNF_PAR after statement
> > definitions, not quite legal anyway you look at it). So I now plan to go
>
> It wasn't really my fault. do-inscnf from EXTREME did it and I forgot to
> check for these include file variants (and it build fine on linux).
So it was really my fault - guilty. INCLUDE 'CNF_PAR' is added after
the last existing INCLUDE line - inscnf can't work out which includes
contain parameter declarations only and which ones contain other
things. Putting the CNF_PAR line after the *first* existing INCLUDE
line (probably SAE_PAR) would be a better strategy though. I've just
taken a quick look at inscnf.c and don't sufficiently have my lex head on
to do it with confidence of not breaking anything. Since all the
CNF_PARs are done now (I think?) - and I'm busy with pre-ADASS
business - it's probably not worth fixing this urgently. However if
anyone's planning to do a load of do-inscnf runs in the future let
me know and I'll try to revisit it.
Mark
--
Mark Taylor Starlink Programmer Physics, Bristol University, UK
[log in to unmask] +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/
|