On Mon, Nov 24, 2014 at 10:49:11PM +0100, Ingo Schwarze wrote:
> Hi Theo,
>
> Theo Buehler wrote on Mon, Nov 24, 2014 at 04:56:00PM +0100:
>
> > This is a groff(1)-related mail, so I'm not sure if CC'ing schwarze@
> > is appropriate, if not, my apologies.
>
> I'm maintaining the groff port, so if you suspect a problem with
> groff, it is.
>
> > I noticed that the djview4(1) man page contains lots of artefacts
> > stemming from an incorrect rendering of tables, as in the paragraph
> > below:
> >
> > ----- 8< -----
> > center,box; lfI l lfB l lfB l lfB l lfB l
> > number Magnification factor in range 10% to 999%.
> > one2one Select the "one-to-one" mode. width Select the
> > "fit width" mode. page Select the "fit page" mode.
> > stretch Stretch the image to the plugin window size.
> > ----- >8 -----
> >
> > which is the result of the source
> >
> > .TS
> > center,box;
> > lfI l
> > lfB l
> > lfB l
> > lfB l
> > lfB l
> > .
> > number Magnification factor in range 10% to 999%.
> > one2one Select the "one-to-one" mode.
> > width Select the "fit width" mode.
> > page Select the "fit page" mode.
> > stretch Stretch the image to the plugin window size.
> > .TE
>
> That happens because the page doesn't contain the standard
> annotation telling groff(1) that it needs tbl(1).
>
> > when one lets it compile with `groff -Tascii' without the `-t' option
> > for preliminary tbl(7) rendering.
>
> Yes, that is what pkg_create(1) ends up doing, and i can't blame it.
>
> > The full manual is available here:
> > http://sourceforge.net/p/djvu/djview-git/ci/master/tree/src/djview.1
> >
> > The artefacts are irritating, but I'm not sure if the result of
> > `groff -t -Tascii' is preferable since it confuses less(1) and more(1)
> > at least in uxterm(1) in that the resulting terminal escapes for bold
> > and italic fonts aren't rendered correctly either and produce other
> > irritating artefacts.
>
> Probably you forgot to use the -P-c option to groff.
> Alternatively, you can run less(1) with the -R option.
> This is just a guess, of course, because you are not
> showing the actual commands you ran, nor hexdump(1) -C
> output of the result.
>
> > I'm not sure whether this is a common problem and whether there is a
> > standard way of dealing with this kind of situation.
>
> I'd suggest pushing the following diff upstream, it fixes the
> issue.
>
> Do you want me to commit it, too?
>
ok shadchin@. Thanks
> Yours,
> Ingo
>
>
> Index: Makefile
> ===================================================================
> RCS file: /cvs/ports/graphics/djview4/Makefile,v
> retrieving revision 1.20
> diff -u -p -r1.20 Makefile
> --- Makefile 21 Mar 2013 08:45:18 -0000 1.20
> +++ Makefile 24 Nov 2014 21:47:21 -0000
> @@ -3,6 +3,7 @@
> COMMENT= portable DjVu viewer and browser plugin
>
> DISTNAME= djview-4.9
> +REVISION= 0
> PKGNAME= ${DISTNAME:S/djview/djview4/}
> CATEGORIES= graphics print
>
> Index: patches/patch-nsdejavu_nsdejavu_1_in
> ===================================================================
> RCS file: patches/patch-nsdejavu_nsdejavu_1_in
> diff -N patches/patch-nsdejavu_nsdejavu_1_in
> --- /dev/null 1 Jan 1970 00:00:00 -0000
> +++ patches/patch-nsdejavu_nsdejavu_1_in 24 Nov 2014 21:47:21 -0000
> @@ -0,0 +1,9 @@
> +$OpenBSD$
> +Requires tbl(1).
> +--- nsdejavu/nsdejavu.1.in.orig Tue Apr 10 05:56:55 2012
> ++++ nsdejavu/nsdejavu.1.in Mon Nov 24 22:39:46 2014
> +@@ -1,3 +1,4 @@
> ++'\" t
> + .\" -*- nroff -*-
> + .\"
> + .\" Copyright (c) 2001 Leon Bottou, Yann Le Cun, Patrick Haffner,
> Index: patches/patch-src_djview_1
> ===================================================================
> RCS file: patches/patch-src_djview_1
> diff -N patches/patch-src_djview_1
> --- /dev/null 1 Jan 1970 00:00:00 -0000
> +++ patches/patch-src_djview_1 24 Nov 2014 21:47:21 -0000
> @@ -0,0 +1,9 @@
> +$OpenBSD$
> +Requires tbl(1).
> +--- src/djview.1.orig Tue Apr 10 05:56:55 2012
> ++++ src/djview.1 Mon Nov 24 22:31:15 2014
> +@@ -1,3 +1,4 @@
> ++'\" t
> + .\" Copyright (c) 2007- Leon Bottou
> + .\"
> + .\" This is free documentation; you can redistribute it and/or
>
--
Alexandr Shadchin
