Re: [Groff] odd interaction between .bl and .(c in -me macros

[Dave Kemper]

On Mon, Jan 23, 2012 at 4:45 PM, Tadziu Hoffmann
<address@hidden> wrote:
> By the way, I've also figured out why the centering of wide
> blocks in the "me" macros does not work correctly: the macros
> change the line width to 6.0 inches, but the indent for blocks
> is computed in environment 1 which retains the default line
> width of 6.5 inches.

I also discovered a pair of -me macro bugs that involve line lengths
and environments.  I suspect these bugs are related to each other; I
don't yet know enough to guess whether they're related to the problem
you describe above, but they cover the same ground at least.  I reported
them to bug-groff several months ago, and have not yet received a reply.

1. The following input file:

The quick brown fox
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
.ev foo
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
.ev
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which
jumped over a quick brown fox which...

produces the expected output (with all lines of equal length) when using
basic groff, but changes line length in the foo environment when using
the -me macro set.  There is nothing in the meref.me documentation that
leads me to believe this is expected behavior, so I presume it's a bug.

2. meref.me documents that the .ll macro sets the line length across all
environments, but it does not work this way.  In particular, if you add
".ll 4i" to the beginning of the sample file above, the line length
still changes when switching environments.

It's possible, of course, that the first bug is a side effect of the
second.  There is a ".ll 6.0i" line near the end of e.tmac which, were
it working as documented, might fix bug 1.  (The -me output in bug 1
looks to me like the line length is going back to the groff default of
6.5i when entering foo.)