Follow-up Comment #13, bug #68420 (group groff): At 2026-06-08T00:08:00-0400, Dave wrote: > Follow-up Comment #12, bug #68420 (group groff): > > [comment #7 comment #7:] >> CSTR #54 did not specify what happens when the control, >> no-break control, and escape characters are not all distinct. > ... >> In _my opinion_, because the behavior is unspecified, no >> reliable prediction can be made. > > This strikes me as somewhat a mischaracterization of CSTR #54. To my > knowledge, it doesn't claim to be a specification.
I agree. But I did not say CSTR #54 was a _specification_. I said that
it "did not specify" an aspect of behavior.
At the same time, the groff has, over the years, enjoyed some
participation by folks who treat CSTR #54 as the acme of documentation.
> One could argue that AT&T troff itself is the specification; indeed,
> (1) many details about troff cannot be answered by CSTR #54, and AT&T
> troff's behavior has been used to fill in these gaps; and (2) when the
> behavior of AT&T troff differs from what CSTR #54 documents, it's the
> latter generally considered to be in error (unless the "behavior is so
> anomalous that it must be considered a bug," as the groff manual
> states at one point).
This road leads to Hyrum's Law.
Not that that has stopped the aforementioned mailing list participants
from aggressively casting golden calves.
And your point leaves unaddresses the question of what to do when
distinct AT&T troff implementations disagree _with each other_, as
Carsten Kunze discovered at least 12 years ago, some unknown (presumably
AT&T-employed) engineer discovered perhaps in the time frame between DWB
2.0 and 3.3, and as we just revisited yesterday in bug #42675.
> So the behavior in question is _undocumented_, but potentially still
> _specified_, because AT&T troff did _something_ in this situation.
I'm not willing to elevate an implementation to specification status.
> I don't know what that something is, so it too may fall in the
> too-anomalous-to-clone bucket. That is, the above are general
> observations, not ironclad conclusions about this ticket. I leave the
> exercise of applying the general to the specific to Martin and/or the
> FreeBSD folk who authored the unusual usage.
ProofPoint, Inc. is all (or mostly) FreeBSD people? I didn't know that.
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?68420>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
