Re: Differences in `ne` and `bp` line-breaking behavior

[G. Branden Robinson]

At 2024-12-02T13:05:50-0500, Douglas McIlroy wrote:
> > I have discovered recently that `ne` and `bp` behave differently in
> > regards to pending input lines. `bp` breaks such lines, while `ne`
> > does not. In practice this means that `ne` does not behave like a
> > conditional `bp` as one would reasonably expect.
> 
> I invented .ne 55 years ago and have never heard a complaint about its
> design before. It is not a conditional .bp, because that would case a
> line break, which .ne never does, nor should.

Thanks for speaking to the issue, Doug.

I came up with the following demonstrator.  It produces the same (nroff)
output on groff 1.22.3, groff Git, DWB 3.3, and Heirloom Doctools.[1]

$ cat ATTIC/need.roff
.na
.de TT
Pilaf\(emI'm trapped under rice!  (nl=\\n(nl)
..
.wh \n(.pu/2u TT \" Plant trap at half the page length.
.ne 1v
(nl=\n(nl)
Hold my beer
.ne 6i
and watch this.
.br

I encourage the interested reader to predict the output of the
foregoing.  I'll use the rest of this message as "spoiler space" before
disclosing my output below.[2]

I therefore think our description of `ne` in groff(7) pretty badly needs
revision.

Here's what it says now.

       .ne       Break page if distance to next page location trap is
                 less than one vee.
       .ne d     Break page if distance to next page location trap is
                 less than distance d (default scaling unit v).

Here's what I propose.

       .ne       Advance drawing position to the next vertical position
                 trap if it is nearer than one vee.
       .ne d     Advance drawing position to the next vertical position
                 trap if it is nearer than distance d (default scaling
                 unit v).

How do people feel about that wording?

Returning to the exercise for the reader, I'll offer a hint by quoting
our Texinfo manual.

---snip---
Output line properties like page offset, indentation, adjustment, and
even the location of its text baseline, are not determined until the
line has been broken.  An output line is said to be "pending" if some
input has been collected but an output line corresponding to it has not
yet been written; such an output line is also termed "partially
collected".  If no output line is pending, it is as if a break has
already happened; additional breaks, whether explicit or implicit, have
no effect.  If the vertical drawing position is negative--as it is when
the formatter starts up--a break starts a new page (even if no output
line is pending) unless an end-of-input macro is being interpreted.
*Note End-of-input Traps::.
---end snip---

These facts are important to know; it'd be good to get that paragraph
into groff(7) somewhere--or roff(7), since as far as I know they apply
to pretty much every *roff ever.

Regards,
Branden

[1] ...with a difference in how the em dash is rendered on a terminal.
    DWB nroff writes a single hyphen; Heirloom and groff write two.

[2] Answer to exercise:

$ nroff -T ascii ATTIC/need.roff | nl -ba
     1
     2
     3
     4
     5
     6
     7
     8
     9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    29
    30
    31
    32
    33
    34  (nl=0) Hold my beer Pilaf--I'm trapped under rice!  (nl=1320) and
    35  watch this.
    36
    37
    38
    39
    40
    41
    42
    43
    44
    45
    46
    47
    48
    49
    50
    51
    52
    53
    54
    55
    56
    57
    58
    59
    60
    61
    62
    63
    64
    65
    66

signature.asc
Description: PGP signature