#192144 groff-base: eqn formats complex matrices badly on nroff devices

Package:
groff-base
Source:
groff
Description:
GNU troff text-formatting system (base system components)
Submitter:
David Schmitt
Date:
2023-07-21 13:45:04 UTC
Severity:
normal
Tags:
#192144#5
Date:
2003-05-06 11:36:24 UTC
From:
To:
Dear X-Team!

The matrices and formulae in the gl* manpages are broken to the point of
being unusable. Perhaps this is a groff problem, but the gl* manpages
are the only broken ones I know of.

For examples look at gluPerspective(3) and gluProject(3).


Thank you for your time and work.

Regards, David
#192144#10
Date:
2003-05-21 12:55:21 UTC
From:
To:
I don't see anything wrong with gluPerspective(3x), but gluProject(3x)
throws "eqn:/tmp/zmanA7ACTF:80:fatal error: nested .EQ", which won't
help. In one of the equations in the description section, there's a .EN
that's been accidentally joined to the end of the previous line rather
than being on a line by itself; I don't have time to work up a patch
right now but it'll be obvious at a glance.

Maybe you could elaborate on what other problems you're seeing?
#192144#15
Date:
2003-05-21 14:18:44 UTC
From:
To:

Sorry for the meager explanation, but I thought it was obvious when
viewing the manpages. Matrices are displayed with _huge_ vertical space:

From gluPerspective(3x):

       Given f defined as follows:

                                                            f=cotangent(fo2y)




























       The generated matrix is


                                                    aspect 0      0           0



















                                                       0   f      0           0



.... and so on.

For the records, I am running unstable with groff-base 1.18.1-9.


Regards, David
#192144#20
Date:
2003-05-21 14:29:00 UTC
From:
To:
Bizarrely, I can't reproduce this with man but can when I run the exact
same commands executed by man by hand. I'm now confused and will try to
figure out what on earth's going on.

Thanks,
#192144#25
Date:
2003-05-21 18:58:55 UTC
From:
To:
By default, man passes the -s option to pager, which (normally)
squeezes multiple blank lines into one.  This is (generally) the Right
Thing(tm).  David: you might want to look into what's wrong with your
pager (check the $PAGER env variable and the /etc/alternatives/pager
link), and stick to something that supports -s.  Colin: you might want
to try postscript output (man -Tps) and see how the page looks that
way; as far as I'm concerned, that's the real test of a man page.

cheers
#192144#30
Date:
2003-05-21 19:10:13 UTC
From:
To:
Good call, I'd forgotten about that.

There's almost certainly nothing wrong with his pager. man only sets -s
for its default pager, not for random stuff in the $PAGER variable
(which I had unset). I should add -s to the default $LESS set by man, I
think.

However, there is still a clear bug in eqn's output, forgetting about
the pager issues. I don't yet know whether it's a bug in the input or in
the output.

It's not quite as bad, but the matrix could easily be squeezed onto the
first page and isn't, so there's probably a spacing issue there, and the
brackets around the matrix are set too high.

Cheers,
#192144#35
Date:
2003-05-22 17:23:40 UTC
From:
To:
reassign 192144 groff
thanks

Per ongoing discussions, I'm reassigning this bug to groff.

If there really is something hideously wrong with the Mesa manpages,
I'll welcome a reassignment back to me once that problem has been
identified.  A patch would be most appreciated.
#192144#40
Date:
2003-05-22 21:48:59 UTC
From:
To:
Indeed, I had set PAGER=less and unsetting it sqeezed the matrices into
readable dimensions.

That would at least solve the thing for people who have $PAGER set to
less.


Thank you all for your time and work again!



Regards, David
#192144#45
Date:
2003-05-24 06:42:38 UTC
From:
To:
Hi,

The attached file, extracted from XFree86's gluPerspective(3x) man page,
formats very poorly with 'groff -e -Tascii'; the rows of the matrix are
far too widely spaced. I know very little about eqn, though. Can anyone
tell me whether eqn, troff, or the man page is at fault?

Thanks,
#192144#46
Date:
2003-05-24 11:51:59 UTC
From:
To:
Hi Colin,

Isn't groff's eqn meant to cope poorly with ASCII output?  If it helps
decide, here's AIX 3.2.5's output, i.e. troff not groff.

    $ eqn <k | nroff | col -b | cat -S
               (                                        )
               |                 0            0         |
               |  ______  0                   0         |
               |  aspect  f      0       ____________   |
               |    0     0  __________   zNear-zFar    |
               |    0        zNear-zFar       0         |
               |    0     0      -1                     |
               (                                        )

So it seems like it isn't the man page that's the problem.

Cheers,


Ralph.
#192144#47
Date:
2003-05-24 22:26:50 UTC
From:
To:
Right.

Well, this isn't correct either since the top of each fraction is
missing.  I strongly suggest that the author of gluPerspective(3x)
uses `.ie t \{...\} .el \{...\}' to provide a hand-tuned TTY version
of the matrix if it is essential to the man page.


    Werner
#192144#48
Date:
2003-05-25 08:48:15 UTC
From:
To:
.. and the rows are broken. eg the second row should read ( 0 f 0 0 ).

Even as a user, I wouldn't say, that it is essential for the man page.
One can use the gluPerspective() function without knowing the
mathematical background. But there are some special use cases, where one
needs deep matrix magic and where is the information then?

And why provide a facility which is broken to the point of being
not only unusable, but also producing dangerously misleading output?

Regards, David
#192144#53
Date:
2014-05-24 12:03:18 UTC
From:
To:
Sehr geehrte/r Arbeitsuchender,

folgendes Arbeitsangebot ist für alle geeignet, da diese Arbeit ohne besondere Anforderungen auch von zu Hause zu bewerkstelligen ist. Der Arbeitnehmer hat keine Ausgaben und muss keine besonderen Kenntnisse mitbringen. Die benötigte technische Ausrüstung wird von uns kostenlos zur Verfügung gestellt. Zu Ihrem Arbeitsfeld gehört die Betreuung von Onlinebibliotheken, das Digitalisieren von Buchwerken, die Koordinierung, das Erstellen von Mediatheken und die Vorbereitung der Dokumentenbearbeitung. Wir bieten eine attraktive Vergütung in Höhe von 20€ die Stunde Brutto.

Unser Betrieb verfügt über internationale Nebenstellen auf der ganzen Welt und wir arbeiten im Bereich der Mediendigitalisierung. Im Moment sind wir auf der Suche nach neuen Mitarbeitern.

Was Sie mitbringen sollten: Sie verfügen über ein Paar Stunden Zeit am Tag, Sie besitzen Flexibilität, Sie arbeiten gern von zu Hause, Sie besitzen eine teamorientierte Arbeitsweise, Grundkenntnisse von MS-Office sind von Vorteil und sie verfügen über ein Paar Stunden Zeit am Tag.

Haben wir Ihr Interesse geweckt? Dann freuen wir uns über Ihre Bewerbung! Senden Sie Ihre kompletten Bewerbungsbögen an: LacyCroucht@gmx.com
Wir freuen uns auf Ihre Bewerbung.

Beste Grüße

Topl GmbH
#192144#62
Date:
2022-04-12 02:21:36 UTC
From:
To:
package groff groff-base
forwarded 192144 https://savannah.gnu.org/bugs/?62298
thanks

I've opened a ticket for this upstream.  In reproducing it, I found
diagnostics that no one ever commented on before which suggest that eqn
could be improved, even if it doesn't ultimately make the input
useful in this case (that remains to be seen).

Thanks for Colin for the simple reproducer.

Regards,
Branden