Re: Another yodl 3.0 oddity?

[Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx>]

On Jun 7,  4:19pm, Jun T. wrote:
} Subject: Re: Another yodl 3.0 oddity?
}
} On 2013/06/04, at 15:01, Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx> wrote:
} >          number the directories in the stack when printing.
} > 
} > []
} > 
} > disable [ -afmprs ] NAME ...
} 
} 
} Please try using WIDER terminal window so that info browser can
} display whole of the status line (the bottom line of the info browser);
} yes, this may sound nonsense, but just give it a try.

I still get different results from yodl 1.31.18 / texinfo 4.7 than I do
from yodl 3.00.0 / texinfo 4.13.  However, I do see that the blank lines
are the same in both, so my first guess at a -k effect was wrong.

} I guess the index now brings you to one line (not two lines) above the
} 'disable ...' line (this may be a bug of the info browser).

Nope, still two lines above.  However, if I make it wide enough to fit
the whole "File: ..., Node: ..., Next: ..., Prev: ..., Up: ..." display
(from the top of the node) without wrapping, then it's only off by one.

However, yes, the misplaced cursor is what caused me to notice the extra
whitespace.

} Next try deleting an extra blank line (line 289) in builtin.yo:

That does help, but it doesn't change the line number in the index for
"disable" (it changes the line numbers for a whole lot of other stuff
that follows later).  So that line number must have been wrong (but it
is off by one in the older yodl output, too).

I think I know what's going on, and it's a change in the info browser,
not in yodl output.  In 4.7, the browser jumps to the node name read
from the index and then does a forward search for the keyword, which
can put you in very much the wrong place (using the index to look for
typeset lands you in the middle of the declare command).

4.13, it appears, reads both the node and line number from the index,
jumps to the node and then moves down that many lines.  That puts you
a lot closer to the right spot a lot more of the time, and explains
why a line at the top of the node that wraps would leave the cursor
in the wrong place farther down.

} > This in turn seems to be happening because yodl has been
} > invoked with the -k flag,
} 
} zsh.texi produced by yodl with and without the '-k' option are identical:

That's nice to know, maybe we've fixed up the source enough to remove the
configure case.  Or maybe it's still needed for yodl 2, which I haven't
had a chance to test.

-- 
Barton E. Schaefer