[Therion] scaling a line thickness in Metapost

[Tarquin Wilton-Jones]

> So maybe
> 
> def l_wall_sand_SKBB (expr P, sc) =
> 
>                 if unknown sc: sc:=1; fi;  %assuming parsing sc can be
> optional


Right now it doesn't work - it claims that I am asking for non-existent
parameters. I guess it would need a change in Therion to make it work.

For now, you can make an equivalent like this:
-attr scale xs

Then in your metapost:
string scale;
if known ATTR_scale:
  scale:=ATTR_scale;
else:
  scale:="m";
fi;

This is reinventing the wheel since "-scale" is a real parameter that
seems to exist but doesn't seem to get put anywhere useful. I would
prefer not to reinvent the wheel, when something already exists for that
purpose. But at least -attr exists to allow you to pass custom data into
functions.

The issue I have right now is that every bit of "def" code seems to
share its variables with every other bit of "def" code. So each time you
say "string scale", you are forcing *every* "scale" variable in all of
Therion's Metapost code to be a string, even if some other piece of code
is expecting it to be a number or whatever. (I can see that this same
issue will creep into other people's Metapost examples on the wiki, and
they will end up randomly trampling variables in other pieces of code.
For example, Andrew's wonderful boulder code uses "boolean overlap;"
without locally scoping that variable.)

You can call the variable "string l_u_foo_scale;" which will then be
very unlikely to clash with existing variables, but there has to be a
way to make variables local within a def. (Apparently you can use "save"
within a begingroup section to make variables local, so I am playing
with that right now to see if I can work out how to use it. You cannot
use vardef instead of def, because Therion relies on you having global
variables like "T" and "u" within your def. It would be so much nicer if
Metapost had a simple way to say "this variable is the local one, not a
higher scoped one" like most other languages I know. It over-complicates
this part.)

> I think scale m = 1, with the other values each scaled up or down by
> root 2. Therefore xs = 0.5, xl = 2.0


Thanks for that info :)