[bug #66054] [troff] permit special characters as source hyphenation codes

[G. Branden Robinson]

Follow-up Comment #4, bug #66054 (group groff):

[comment #3 comment #3:]
> [comment #2 comment #2:]
> > 2.  `.hcode \['a] \['a]` should have altered `\['a]`'s hyphenation code.
> 
> I might be misunderstanding the semantics of assigning a character's hcode
to itself.  I understood it to mean "make this character eligible for
hyphenation (if it wasn't already)."

Basically, yes.  There's a subtlety: you might do this to move a letter from
an equivalence class with another letter to its own equivalence class.

For example, by default, the formatter internally does the following.

.hcode B b

Observe.


$ printf '.ll 1n\nABBOT\n.pl \\n(nlu\n' | nroff -Wbreak
AB‐
BOT
$ printf '.hcode B B\n.ll 1n\nABBOT\n.pl \\n(nlu\n' | nroff -Wbreak
ABBOT


>  The actual code assigned, as you've pointed out, is (or can be) arbitrary.

It can be any value from 1..255.
 
> So in the case above, \['a] already had an hcode.  (Granted, it _shouldn't_
have, due to problem 1, but that's already been identified as a separate
problem.)  Thus, ".hcode \['a] \['a]" is effectively a no-op.

It wouldn't necessarily be if we had a larger pool of hyphenation code values
to draw from.  But we're limited to fewer than 255 in practice, because it's
hard/impossible to use any of the C0 or C1 controls, or the space, as
hyphenation code values.

> Maybe the root problem is that ".hcode a b" and ".hcode a a" use the same
syntax but do distinctly different things.  It obscures to humans writing roff
code what they're really asking for.

I won't say this critique is invalid, but it's not something I have a
hankering to overturn for _groff_ 1.24.

I'm also not sure I agree.  I think the semantics can be defended if one
understands what an "equivalence class" is.  Doesn't *everyone* take a
university course in discrete mathematics?  ;-)




    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?66054>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/

signature.asc
Description: PGP signature