mailing-list for TeXmacs Users

["Andrey G. Grozin" <address@hidden>]

On Sat, 12 Jan 2008, Vadim V. Zhytnikov wrote:
> >  Another solution might be to simply anticipate some of
> >  the upcoming version numbers ;^)
> Jokes aside. IMHO this is really the best approach.
> I maintain Maxima nad TeXmacs for ALT Linux distro for
> several years. And I always build TeXmacs with tiny patch
> which modifies behaviour of tm_maxima in simple way.
> If Maxima version doesn't match known version tm_maxima
> doesn't barf "Unknown Maxima version" but tries to start
> Maxima with the latest version of interface package -
> texmacs-maxima-5.11.0.lisp at present.
It is easy to do this in tm_maxima (and I think I'll do this change). 
However, the logic of
maxima_detect help
will be broken. Currently, if a maxima which understands --list-avail is 
detected (this means 5.9.0 or later),
maxima --list-avail
is called, and the result is grepped for good versions (meaning 5.9.3 or 
later) which understand -d. The list of good versions should be known at 
this point, and should be updated when a new version appears. It would be 
a mistake to grep for bad versions (5.9.0, 5.9.1, 5.9.2 - their list is 
finite and never changes). Suppose a user has 5.9.0 and 5.14.0 installed. 
Then
maxima -d
will work, because the default is the latest version. But grepping for 
5.9.0 will give a false positive.

Perhaps, I'll have to parse the output of
maxima --list-avail
separate individual components of versions, and compare them as numbers - 
if a version >= 5.9.3 is found, everything's OK (with a natural numerical 
interpretation of >=, which can be implemented in sh, with some effort. 
And don't forget rc1, cvs and other suffixes.)

Andrey