Re: Need background on design of groff character classes

On 17.11.25 23:43, Dave Kemper wrote:

> And this would work to set its end-of-sentence property--but it would

> have the side effect of overwriting the flag (4) that \[em] already

> had set.

...

> The .cflags request only overwrites all existing flags; it does not

> have a syntax to preserve flags while adding a new one.

Instead, an AND and an OR operation seem a minimal implementational requirement, to meet your needs. Would comfortable syntactic sugar be to simply replace 's' with [+-], to explicate that the crude bulk overwrite is now a bitwise set/clear?  I.e.:

.cflag+ [flag] target

.cflag- [flag] target

Then replace one implementing line of 'C' with two cases, something like

flags &= ~(flag) ;

flags |= (flag) ;

ISTR that this flag structuring already exists:

#define thisflag  (1<<4)

#define thatflag  (1<<9)

for a miscible flag set.

If .class needs to set/clear multiple flags on entry/exit, then I'd almost be tempted to e.g.:

#define Class_alpha (flagA | flagD | flagG)

and stuff that through .cflag[+-] for two for the price of one.

(OK, that wouldn't implicitly handle exit where two classes shared a flag or two. Non-overlap would be a prerequisite for a practicable implementation of such (over)simplifiction.
Currently we have a big hammer, but not every problem is a nail.

OK, backward compatibility, and all that good stuff, so just add the new one alongside, for zero hysterical raisins?

Well, that's hopefully relevant, if I haven't misunderstood the issue, or missed a hidden nuance - neither entirely improbable. (The less one knows, the simpler things appear.)

Erik