[bug #62265] [tbl] tables in nroff mode seem to want one too many vees

[G. Branden Robinson]

Follow-up Comment #4, bug #62265 (project groff):

[comment #3 comment #3:]
> The detailed analysis in http://lists.gnu.org/r/groff/2022-07/msg00223.html
suggests that this and bug #62471 are respectively the vertical and horizontal
manifestations of the same bug.

Unfortunately there were some analytical problems with that rambling epic. 
Not only did I correct one point
<https://lists.gnu.org/archive/html/groff/2022-07/msg00267.html>, but it
appears another must fall--this one.

The problem with my original submission (comment #0) was that I forgot about
the paragraph distance inserted before the section heading.  (At present, this
is implemented by the `an-break-paragraph` macro.)  Worse, I helped mislead
myself by putting the `.br` _after_ the two empty requests I use to annotate
the expected presence of vertical space.

However in this case there is the additional complication that the horizontal
rule _does_ eat an additional vee--but only on nroff devices.

Here is a new exhibit that works as I expect.  I have instrumented it so that
the interested reader can play with the `PL` register and see if they can
contrive to break the poor man's keep implemented with the `ne` request.


$ cat EXPERIMENTS/sh-then-table.man 
.if r PL .pl \n(PLu
.TH sh\-then\-table 1 2022-07-30 "groff test suite"
.SH Name
sh\-then\-table \- experiment
.SH Description
Sed ut perspiciatis,
unde omnis iste natus error sit voluptatem accusantium doloremque
laudantium,
totam rem aperiam eaque ipsa,
quae ab illo inventore veritatis et quasi architecto beatae vitae dicta
sunt,
explicabo.
.
.
.br
.if t .ne 4v
.if n .ne 6v \" account for \n[PD] and horizontal rule
.SH Options
.TS
Lb Lb.
Flag    Description
_
\-a     Do it all.
.TE
$ for n in $(seq 19 26); do read -p "PL=${n}v? "; ./build/test-groff -t -man
-rPL=${n}v -rcR=0 -Tascii EXPERIMENTS/sh-then-table.man; done


I think there are 3 issues here.

1. I boneheadedly misled myself as noted above.  Not much can be done about
that.  Well, others might do the same.  See item 3, below.

2. The tbl(1) man page could add some cautions to nroff-oriented tbl users
about the vertical size of rules and box borders.

3. The man(7) package could really use keep macros.  I'm really itching to
scandalize Ingo by adding yet another feature.  But this one would be really
useful.  With ms(7)-like KS/KE macros, the user would never need to count
lines--as I have done, erroneously, multiple times--and nothing bad happens if
an implementation doesn't support them.  Well, nothing worse than would have
happened anyway.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?62265>

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