Need help with bizarre problem - maybe a new bug?



--- Wolfgang Jenkner  wrote:

[snip]

> I've just committed a fix for this.  The problem was due to a
> somewhat sloppy use of varlist and genvar in $letrat.  For more 
> details, please read the commit message, which should appear at
> 
> http://sourceforge.net/mailarchive/forum.php?forum_id=33117
> 
> I also fixed the problem that CRE expressions kept popping up where
> they shouldn't.

Sweet!!  Thanks Wolfgang - you get hacker points for that one!  Gotta
love those bugs that vanish in the debugging environment.

> So, you can eliminate the calls to `ratsimp' and `ratdisrep' which
> you inserted as work-arounds in `processunits' (this would be the
> best test for all this).

OK, I'll give it a shot.

> (By the way, shouldn't `currentdoeval' be initialized to `true' when
> loading `unit.mac'?)

Yep, it sure should.  Thanks for catching that.

I'm at work now, so I can't commit any fixes, but I'll get them in
there ASAP.

processunits is going to get redone again, at least to some extent,
when I incorporate dimension.mac's ability to handle things like
'diff(x,t)->%length/%time, or 'diff(x,t)->m/s.  Since these are
inherantly special cases, part has to come into play.  At this point,
ordergreat will also enter the picture to make sure units end up at the
end of an expression.  I need to probe mactex as well - aside from
setting up greek letters for things like Ohm, I'm guessing
(expression)*(units) will be possible in TeX output, which won't be a
universal solution but hopefully will be "good enough".  Possibly there
is some way to allow separation of things in an a*(b/c) vs. (a*b)/c way
so that they are displayed separately on the text output level (I'm
sure with enough work one could be added) but I'm inclined to come back
to that later after the more functional parts are done.

CY

P.S.  Is the implimentation of ordergreat and orderless really as dumb
as I think it is, or are there reasons it's done that way?  I can
always write my own function to manipulate those lists but I wonder if
the better course of action would be to improve/replace the current ones.

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com