Re: texinfo-format-buffer text.texi

[Robert J. Chassell]

[None of my previous messages got through to
address@hidden  Let's try these other addresses.  
Please tell me which address succeeds and then please have Emacs set
the `mail-default-reply-to' variable so I don't try sending them to
address@hidden   Thanks!]

As for your latest message, 

    makeinfo-buffer doesn't work at all with

    GNU Emacs 22.0.50.2 (i686-pc-linux-gnu, X toolkit, Xaw3d scroll bars) of 
    2006-06-18

That was about the time my motherboard was zapped by lightning.

However, both makeinfo-buffer and texinfo-format-buffer succeeded with
this morning's GNU Emacs CVS snapshot, Wed, 2006 Jul  5  11:47 UTC
GNU Emacs 22.0.50.20 (i686-pc-linux-gnu, GTK+ Version 2.8.18) 
although I saw

    Warning (initialization): Building Emacs overflowed pure space.
    (See the node Pure Storage in the Lisp manual for details.)

I used regular but old commands in the first Texinfo file on which I
tested them.

In addition, I tried both commands with a second file.  In that file,
the makeinfo-buffer succeeded but texinfo-format-buffer died at the
@slanted command.  The texinfo-format-buffer command succeeded after I
change @slanted to @emph.

So the texinfo-format-buffer command does not know all the @-commands
that have been introduced since makeinfo superceded it.

Also, I was able to run

    makeinfo --force --fill-column=70 --no-split --paragraph-indent=0 \
    --verbose foo.texi

    makeinfo --fill-column=70 --no-split --paragraph-indent=0 \
    --verbose --no-headers --output=foo.txt  \
    foo.texi

    makeinfo --no-split --html foo.texi

in a shell inside that instance of Emacs with no reported errors where
foo.texi is the second file with @slant replaced with @emph.


Regarding, the other messages:

       AFAIU the var `fill-column' is commonly known as
       line-break.

I wrote:

    No, not at all.  When automatic filling is set to a specified
    value, the variable `fill-column' is the value *up to which* (and
    including) a carriage return or other such line-break may occur.
    Often, the variable `fill-column' is larger than that line's value
    of its line break.  At least, that is how I learned the difference
    years ago.

    A line break may occur anywhere [in what is defined as inter-word
    space in the syntax table],  as in this sentence.
    (In the previous sentence, the line break occurred at column 53,
    not at the value of the variable `fill-column' which is 70 in this
    instance of Emacs.)

       Is there a way to refer the var `fill-column'?

    Yes.  I use the term `fill-column'; that works for me.  Sometimes,
    I have to explain that `automatic filling' refers to the same
    concept as `automatic wrapping' in other programs.

You wrote,

       Calling `texinfo-format-buffer' at text.texi
       produces a line

            @anchor-yes-refill{Definition of sentence-end-double-space}

       which is probably not correct. (Line 1408 now)

    RMS sent me a note saying `if you are interested' and I responded

        Hah!  I see that @anchor-yes-refill is attributed to you on 2
        July 1998.  I have never used @anchor and do not know anything
        about it.  I see what you did:  take into account that some
        paragraphs should be filled and others not.  That makes a
        great deal of sense.

    As far as I know, the @anchor-yes-refill command is correct [for
    an intermediate command].

You also wrote that there was a problem formatting the Emacs Lisp
Reference manual.  It did not format for me either but stopped
formatting sooner than you reported:  at Node: Syntax Table Internals.
(Possibly you did not run the command so early.)  As I wrote,

    And as far as I know the `texinfo-format-region' and the
    `texinfo-format-buffer' commands do not understand all the
    @-commands introduced in the past 10 years.  I know for sure that
    it worked with the Emacs Lisp Reference Manual when it was first
    written -- makeinfo had not yet been invented -- but not since.

    Interestingly, I am still listed as the maintainer of texinfmt.el.
    I guess no one else wants it.  You are the first person to send me
    a bug report in a long time.  My hunch is that
    `texinfo-format-buffer' and similar commands will be troubled by
    the `modern' @-commands, that is to say, those of the last decade
    or more, few or perhaps none of which I know.

That hunch looks as if it was confirmed this morning, when I was able
to run `texinfo-format-buffer' successfully on a file in which @emph
replaced @slanted, on which the command failed.

    (I use Texinfo frequently; after all, you can produce seven direct
    output formats from a single source file [for printing in hard
    copy or for fast online listening or viewing or slow online
    listening or viewing].  Also, you can create PostScript and RTF in
    two steps and LaTeX in three steps.  Strangely enough, Info is
    still the most efficient online format.  But I hardly ever use
    @-commands invented since 1989, even for fiction.  The exceptions
    are @image and @footnote.)

    I am not quite sure what to do, since I do not know of anyone who
    uses the commands in the texinfmt.el library any more, even for
    short segments.  They use `makeinfo' and `texi2dvi'.

The relevant library is not in `obsolete'; I thought it was.

If you send me a list of the @-commands which are not handled by
`texinfo-format-buffer' I can at least tell texinfmt.el not to do
anything with them but carry on.

-- 
    Robert J. Chassell                         
    address@hidden                         GnuPG Key ID: 004B4AC8
    http://www.rattlesnake.com                  http://www.teak.cc