Re: mdoc(7) prologue regressions

[G. Branden Robinson]

[self-follow-up]

At 2022-07-16T02:21:49-0500, G. Branden Robinson wrote:
> groff man(7) in groff Git behaves pretty badly if `TH` is omitted,
> whereas groff 1.22.4 degrades much more gracefully.

I was incorrect about this.  I was comparing 1.22.4's "groff -mandoc" to
Git HEAD's "groff -ww -man".  We might recall (I didn't, at first) that
"an.tmac" is no longer a wrapper for andoc.tmac, but the actual groff
man(7) implementation.  And of course in one case I asked for more
warnings.

A pair of fair comparisons would be:

1.22.4 groff -ww -mandoc
HEAD   groff -ww -mandoc

and

1.22.4 groff -man-old
HEAD   groff -man

And these indeed have comparable output both in content and diagnostic
messages.  Skipping andoc, in either version, elicits noticeable
distress from the formatter when terminal output devices are used.

I had been preparing a "normal" Savannah ticket about this, but this
finding demotes it to "wish" severity since this is in no sense a
regression.

Regards,
Branden

signature.asc
Description: PGP signature