[bug #24049] [mm] signature (SG) under MT 5 (external letter) puts spurious hyphens in output

[G. Branden Robinson]

Update of bug #24049 (project groff):

                  Status:               Confirmed => In Progress            
             Assigned to:                    None => gbranden               
         Planned Release:                    None => 1.23.0                 
                 Summary: [mm] signature (SG) under MT 5 (external letter) =>
[mm] signature (SG) under MT 5 (external letter) puts spurious hyphens in
output

    _______________________________________________________

Follow-up Comment #2:

DWB 3.3 mm and Heirloom Doctools mm don't do this.

They also don't require a title for memorandum type 5 (external letter),
though this is clearly implied by the DWB 3.3 manual, which is worth quoting
for two reasons.

> In the external-letter style (`.MT 5`), only the title (without
> "subject") and the date are printed in the upper left and right
> corners, respectively, on the first page.  It is expected that
> preprinted stationary [sic] will be used providing the author's
> company logo and address.

(The claim about the date being in the upper-right "corner" is either
approximate, or wrong.  What all of these mm macro packages do with the date
more closely resembles a traditional U.S. business letter than a hyper-literal
interpretation of the quote above.)

Example input:


$ cat EXPERIMENTS/letter-with-signature.mm 
.ND 2022-12-17
.TL
Our Automobile Requisition
.AU "S.\& Clown"
.MT 5
Dear Regional Manager,
.P 1
It's circuses all the way down.
.FC
.SG


DWB 3.3 mm output:


$ DWBHOME=. ./bin/nroff -mm
~/src/GIT/groff/EXPERIMENTS/letter-with-signature.mm | ul | cat -n
     1
     2
     3
     4
     5         Our Automobile Requisition
     6
     7
     8
     9
    10
    11
    12
    13
    14
    15                                                      2022-12-17
    16
    17
    18         Dear Regional Manager,
    19
    20              It's circuses all the way down.
    21
    22                                       Yours very truly,
    23
    24
    25
    26                                       S. Clown
    27
    28
    29
    30
    31
    32
    33
    34
    35
    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


There is a minor difference in vertical spacing between DWB mm and Heirloom
Doctools mm.


--- /tmp/dwbmm.txt      2022-12-17 14:33:01.286672580 -0600
+++ /tmp/heirmm.txt     2022-12-17 14:33:28.854556501 -0600
@@ -10,20 +10,20 @@
     10
     11
     12
-    13
+    13                                                     2022-12-17
     14
-    15                                                     2022-12-17
-    16
+    15
+    16        Dear Regional Manager,
     17
-    18        Dear Regional Manager,
+    18             It's circuses all the way down.
     19
-    20             It's circuses all the way down.
+    20                                      Yours very truly,
     21
-    22                                      Yours very truly,
+    22
     23
-    24
+    24                                      S. Clown
     25
-    26                                      S. Clown
+    26
     27
     28
     29


And here is groff 1.22.4 output.


$ nroff -mm ~/src/GIT/groff/EXPERIMENTS/letter-with-signature.mm | cat -n 
     1
     2
     3
     4
     5         Our Automobile Requisition
     6
     7
     8
     9                                                          
2022‐12‐17
    10
    11
    12
    13         Dear Regional Manager,
    14
    15              It’s circuses all the way down.
    16
    17                                       Yours very truly,
    18
    19
    20
    21         ‐‐                            S. Clown
    22
    23
[blank lines to 66]


There is a more substantial difference in vertical spacing here.  My guess is
that this is because, as the DWB mm manual says, the user of preprinted
stationery is expected.

Backward compatibility _and_ reasonable accommodation of users who don't have
such stationery both seem reasonable, particularly since these days it is more
likely that users will employ Encapsulated PostScript or something like that
instead.

What we could do is add an extension option to reduce the space reserved for
preprinted stationery, decide on a reasonable default, and document this.

But that's an issue for another day.

The bigger problem here is line 21 in the last example.

This turned out to be easy.  All one had to do was turn on GNU troff warnings
and one was led directly to the problems.

Jörgen Hägg interpolated strings that might be undefined.  Specifically, if
the document author didn't lard `AU` calls with at least 4 parameters.

So here's a patch.


diff --git a/contrib/mm/m.tmac b/contrib/mm/m.tmac
index e31a8acb7..cf9d9da53 100644
--- a/contrib/mm/m.tmac
+++ b/contrib/mm/m.tmac
@@ -3333,11 +3333,16 @@ in=\\n[.i] fi=\\n[.u] .d=\\n[.d] nl=\\n[nl] pg=\\n[%]
 .ne \\n[cov*au]u*4v
 .ie \\n[.$]>1 .nr let*k 1
 .el .nr let*k \\n[cov*au]
-.ds let*tmp \\*[cov*au!\\n[let*k]!3]-\\*[cov*au!\\n[let*k]!4]-
+.ds let*tmp \" empty
+.if d cov*au!\\n[let*k]!3 \{\
+.      as let*tmp \\*[cov*au!\\n[let*k]!3]\"
+.      if d cov*au!\\n[let*k]!4 \
+.              as let*tmp -\\*[cov*au!\\n[let*k]!4]-\"
+.\}
 .nr let*i 0 1
 .while \\n+[let*i]<=\\n[cov*au] \{\
-.      if \\n[let*i]>1 .as let*tmp /
-.      as let*tmp \\*[cov*au!\\n[let*i]!2]
+.      if \\n[let*i]>1 .as let*tmp /\"
+.      if d cov*au!\\n[let*i]!2 .as let*tmp \\*[cov*au!\\n[let*i]!2]\"
 .\}
 .if !''\\$1' .as let*tmp -\\$1
 .in (u;\\n[.l]/2)


The solution originally proposed does get rid of the excess hyphens (by using
the formatted output comparison operator to test strings derived from `AU`
arguments for emptiness), but it still _dereferences_ them, which correctly
produces warning diagnostics when enabled.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?24049>

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