[bug #61909] [tbl] does not suppress inter-sentence space outside of text blocks

[G. Branden Robinson]

Follow-up Comment #2, bug #61909 (project groff):

Hi, Dave!

[comment #1 comment #1:]
> This is certainly worth documenting as a groff difference.  But it's
> not clear that it qualifies as a bug in groff rather than a bug in
> Heirloom troff.

I have no problem characterizing Heirloom Doctools troff as buggy, or
even Unix V7 troff as buggy (see bug #61597), but it's hard for me to
rationalize groff tbl's present behavior here.

> In particular, groff's system seems to give the user options that
> Heirloom's does not.

True.

> A groff user who doesn't want sentence spacing
> within a table need only invoke a .ss request before the .TS/.TE block
> to disable it.

Acknowledged.

> But how is an Heirloom user who _does_ want such spacing to achieve
> it?

By making the table entry a text block.

> So making the groff behavior match Heirloom's would seem to give the
> user less control over the table rendering.

Things that go into tbl(1) table entries (that are not text blocks) tend
strongly to _not_ be sentences.  It is surprising to me that each of the
default text formatting operations (enumerated in the man page) as
"filling, hyphenation, breaking, and adjustment" are disabled, but
supplementation with inter-sentence space is not.

Until the last few years, the phenomenon of inter-sentence space has
seen little discussion in *roff circles, it seems.  I am inclined to
think that people haven't complained about this because either (1) they
hadn't noticed or (2) they decided they just didn't understand tbl(1).

For the former group, this change will cause no problems.  For the
latter, I have a great deal of sympathy because I have struggled to
acquire competence with tbl until I undertook the task of rewriting our
man page for it.

A third group, one that takes Bjarni's belt-and-suspenders approach
to *roff input, will also notice no change, because they were
suppressing inter-sentence space all along.

> (I focus on Heirloom rather than Unix V7 because tbl-iss.roff as
> posted here is not legal V7 troff input: V7 troff's .ss request took
> only one argument.

And nroff ignores it anyway...(screenshot from V7 version of CSTR #54
attached).

> That it accepts the second without complaint only goes to show how
> long bug #61450 has been around.)

It has been around a long time, but in this case, it's ignored for the
same reason `ps` is.  The Teletype Model 37 just couldn't do that.

So should we look at V7 troff?  We can ask John Gardner to bust out his
cat-stream interpreter in JavaScript again.  I wish I had an easier time
using that thing.  Its generational clash of technologies makes it
perversely satisfying.

Regards,
Branden


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?61909>

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