[Axiom-developer] Re: mathml hex(10) patch

[Waldek Hebisch]

Arthur Ralfs wrote:
> From: Waldek Hebisch <address@hidden>
> > You wrote:
> > > Tim, Waldek,<br><br>This patch is a band-aid for the 
> > hex(10) bug.
> > <snip>
> > > @@ -1285,6 +1307,10 @@
> > >            -- where it arises.  Removed 2007-02-14
> > >            concat(concat("<mtext>",str),"</mtext>")
> > >   -- if we get to here does that mean it's a variable?
> > > + -- test for something like #\A and strip off #\
> > > + str.1 = char "#" and str.2 = char "\" =>
> > > 
> > +            u : US := segment(3,len)$US
> > > 
> > +            concat ["<mi>",str.u,"</mi>"]
> > >          concat 
> > ["<mi>",str,"</mi>"]>        l : L E := (expr pretend L E)
> > >        null l => blank
> > > 
> > 
> > I wonder if we can get here with a string -- for string this is
> > clearly wrong.  AFAICS the start of 'hex(10)' problem is in
> > the line:
> > 
> >     stringify expr == (object2String$Lisp expr)@S
> > 
> > object2String converts character using _Lisp_ rules:
> > 
> > object2String x ==
> >   STRINGP x => x
> >   IDENTP x  => PNAME x
> >   NULL x    => '""
> >   PAIRP  x  => STRCONC(object2String first x, 
> > object2String rest x)
> >   WRITE_-TO_-STRING x
> > 
> > so you get '#\A' here.  I am not sure if we ever want '#\A' in
> > "math" output -- if no we could just use a function which would
> > convert characters to length 1 strings.
> > 
>  Do you think we can rewrite the stringify function to do the
> right thing?  I'm not sure how to proceed.  I copied stringify
> from the TeX domain.
> 

The following add a 'mathObject2String' function to convert characters
to length one strings and all other object using old rules.

--- trunk.bb/src/interp/format.boot     2007-12-16 17:25:14.000000000 +0100
+++ trunk/src/interp/format.boot        2007-12-16 22:22:40.000000000 +0100
@@ -706,6 +706,10 @@
     concat("attribute: ",form2String form)
   form2String x
 
+mathObject2String x ==
+  CHARACTERP x => COERCE([x],'STRING)
+  object2String x
+
 object2String x ==
   STRINGP x => x
   IDENTP x  => PNAME x

given this function it should be enough to change stringify to call
mathObject2String (instead of object2String).

Concerning TeX: TeX output has the same problem...

-- 
                              Waldek Hebisch
address@hidden