Update of bug #67902 (group groff):
Severity: 3 - Normal => 1 - Wish
Status: Need Info => None
Assigned to: barx => None
Summary: [troff] sometimes a blank page is emitted after the
end of a document => [troff] want implicit page trap behavior changed to not
eject page unless/until more formatting work remains
_______________________________________________________
Follow-up Comment #14:
[comment #13 comment #13:]
> I feel like this ground has been covered.
I do too.
> "It's documented" doesn't really address that it is undesired by most users,
You keep moving the goal posts here. Stop.
Earlier, we briefly found common ground:
DK> Default behavior should reflect the common case.
GBR> The common case _isn't_ having vertical page margins of zero.
DK> Sure, maybe only 5% of users do that.
...but just as quickly, your acknowledgement of the "niche-ness" of this issue
slipped away.
DK> But of that subset of users, if 95% of them want the last page to have
content, and 5% of them want it to be blank, which should be the default?
We're still dealing with at most 4.75% of users, using your statistics.
I'd say two things.
(a) This 4.75% of users can learn and use the `\c` escape sequence. (Maybe
invoking `.vpt 0` before the input that populates the final output line would
work, too.)
(b) I think it likely that 5% was an overstatement--GNU _troff_ doesn't record
statistics of its invocations, but if it did, I'll bet that not only would
`-mandoc`, `-man`, or `-mdoc` be present more than 95% of all its command
invocations, but that the other full-service macro package selections `-me`,
`-mm`, `-mom`, and `-ms` would occur in a very large proportion of the
remainder. My guess that is the premier application of GNU _troff_ without
loading of a full-service macro package is the execution of its automated test
suite.
Wait-- **three** kinds of aftershave!
(c) At no point do you consider the costs or tradeoffs involved in making the
change you desire. I believe doing so will involve deep surgery to GNU
_troff_'s parser, which I assert based on some experience to be costly.
However I welcome the attention of others to this code--possibly the cost is
lower than I suspect, and in the meantime we might "sucker" other people into
building some expertise with the formatter, which would be good for our Bus
Factor.
> nor that people who encounter the behavior (including at least one who was an
> original *roff developer) consider it a bug. You could loop Doug in for his
> take if you want.
As can you.
> (Fun aside: about four hours ago, I received a PDF from my insurance company
> that consisted of one page of content, labeled "Page 1 of 1" at the bottom,
> followed by a blank page. Either they're *roff users, or the bug exists in
> other PDF-rendering software as well.)
Obviously they have a page footer: "Page _m_ of _n_". It sounds like their
software has a bug that _groff_ lacks.
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?67902>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
