unofficial mirror of emacs-devel@gnu.org 
 help / color / mirror / code / Atom feed
* Re: pxref
       [not found] <17238.44739.179203.166568@kahikatea.snap.net.nz>
@ 2005-10-20  9:29 ` Eli Zaretskii
  2005-10-20 11:16   ` pxref Juri Linkov
  0 siblings, 1 reply; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-20  9:29 UTC (permalink / raw)
  Cc: bug-texinfo, emacs-devel

> From: Nick Roberts <nickrob@snap.net.nz>
> Date: Thu, 20 Oct 2005 09:38:27 +1300
> 
> pxref seems to put in a full stop in the generated info page, when the
> reference includes the manual name:
> 
> (@pxref{Registers,,, gdb, The GNU debugger}) gives:  (see Registers(gdb).)
> 
> which seems wrong since
> 
> (@pxref{Registers}) gives:  (see Registers)

Wrong mailing list ;-)  It's not a Texinfo problem, it's a problem with
Emacs's reformatting of cross-references.

makeinfo _always_ produces either a "::" or a "." at the end of the
@pxref cross-reference, before the closing paren, because those are
the rules of the Info files.  If you look in the emacs-* files with a
pager, you will see that a @pxref with more than one argument always
ends in a period, no matter if it's the in the same or another manual.


_______________________________________________
Texinfo home page: http://www.gnu.org/software/texinfo/
bug-texinfo@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-texinfo


^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-20  9:29 ` pxref Eli Zaretskii
@ 2005-10-20 11:16   ` Juri Linkov
  2005-10-20 12:41     ` pxref Eli Zaretskii
  0 siblings, 1 reply; 39+ messages in thread
From: Juri Linkov @ 2005-10-20 11:16 UTC (permalink / raw)
  Cc: nickrob, bug-texinfo, emacs-devel

>> pxref seems to put in a full stop in the generated info page, when the
>> reference includes the manual name:
>> 
>> (@pxref{Registers,,, gdb, The GNU debugger}) gives:  (see Registers(gdb).)
>> 
>> which seems wrong since
>> 
>> (@pxref{Registers}) gives:  (see Registers)
>
> Wrong mailing list ;-)  It's not a Texinfo problem, it's a problem with
> Emacs's reformatting of cross-references.
>
> makeinfo _always_ produces either a "::" or a "." at the end of the
> @pxref cross-reference, before the closing paren, because those are
> the rules of the Info files.  If you look in the emacs-* files with a
> pager, you will see that a @pxref with more than one argument always
> ends in a period, no matter if it's the in the same or another manual.

The reason not to hide a period is because it may end the sentence.

-- 
Juri Linkov
http://www.jurta.org/emacs/

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-20 11:16   ` pxref Juri Linkov
@ 2005-10-20 12:41     ` Eli Zaretskii
  2005-10-20 16:46       ` pxref Juri Linkov
  0 siblings, 1 reply; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-20 12:41 UTC (permalink / raw)
  Cc: nickrob, emacs-devel

> From: Juri Linkov <juri@jurta.org>
> Cc: nickrob@snap.net.nz, bug-texinfo@gnu.org, emacs-devel@gnu.org
> Date: Thu, 20 Oct 2005 14:16:22 +0300
> 
> > makeinfo _always_ produces either a "::" or a "." at the end of the
> > @pxref cross-reference, before the closing paren, because those are
> > the rules of the Info files.  If you look in the emacs-* files with a
> > pager, you will see that a @pxref with more than one argument always
> > ends in a period, no matter if it's the in the same or another manual.
> 
> The reason not to hide a period is because it may end the sentence.

The problem is that the period is hidden inconsistently: it is hidden
when the @pxref references a node in the same manual, but not if it
references a different manual.  Are you saying that the former can
never end a sentence, while the latter can?

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-20 12:41     ` pxref Eli Zaretskii
@ 2005-10-20 16:46       ` Juri Linkov
  2005-10-20 19:35         ` pxref Eli Zaretskii
  0 siblings, 1 reply; 39+ messages in thread
From: Juri Linkov @ 2005-10-20 16:46 UTC (permalink / raw)
  Cc: nickrob, emacs-devel

> The problem is that the period is hidden inconsistently: it is hidden
> when the @pxref references a node in the same manual, but not if it
> references a different manual.

The period is never hidden in the Emacs Info reader, whereas :: is hidden.

The difference is in Info files produced by makeinfo.  When a Texinfo
source file has

    (@pxref{Registers,,, gdb, The GNU debugger})

then makeinfo produces in the Info output file:

    (*note Registers: (gdb)Registers.)

where the point is not hidden by the Emacs Info reader, but

    (@pxref{Registers})

produces in the Info output file:

    (*note Registers::)

where :: is hidden.

-- 
Juri Linkov
http://www.jurta.org/emacs/

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-20 16:46       ` pxref Juri Linkov
@ 2005-10-20 19:35         ` Eli Zaretskii
  2005-10-23  4:42           ` pxref Richard M. Stallman
  0 siblings, 1 reply; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-20 19:35 UTC (permalink / raw)
  Cc: nickrob, emacs-devel

> From: Juri Linkov <juri@jurta.org>
> Cc: nickrob@snap.net.nz, emacs-devel@gnu.org
> Date: Thu, 20 Oct 2005 19:46:23 +0300
> 
> The period is never hidden in the Emacs Info reader, whereas :: is hidden.
> 
> The difference is in Info files produced by makeinfo.  When a Texinfo
> source file has
> 
>     (@pxref{Registers,,, gdb, The GNU debugger})
> 
> then makeinfo produces in the Info output file:
> 
>     (*note Registers: (gdb)Registers.)

This happens with @pxref's that refer to the same manual as well.  You
can find a few examples in glossary.texi; here's one:

    @pxref{Matching,,Matching Parens}

> where the point is not hidden by the Emacs Info reader, but
> 
>     (@pxref{Registers})
> 
> produces in the Info output file:
> 
>     (*note Registers::)
> 
> where :: is hidden.

This is IMHO a bug in Emacs: the :: has the same semantic role as the
period before the closing paren.  Morover, it is _never_ correct
Texinfo usage to put a full sentence inside @pxref, so this particular
period can _never_ end a valid sentence.

So I think Emacs should always hide this period.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-20 19:35         ` pxref Eli Zaretskii
@ 2005-10-23  4:42           ` Richard M. Stallman
  2005-10-23  7:27             ` pxref Eli Zaretskii
  0 siblings, 1 reply; 39+ messages in thread
From: Richard M. Stallman @ 2005-10-23  4:42 UTC (permalink / raw)
  Cc: juri, nickrob, emacs-devel

      Morover, it is _never_ correct
    Texinfo usage to put a full sentence inside @pxref, so this particular
    period can _never_ end a valid sentence.

That is true.  However, the reason it is not correct Texinfo usage
to use @pxref for an entire parenthetical sentence
is that the correct way to do that is with `(@xref{args}.)'.
That will also generate a period before the closeparen
resulting in similar Info output.

I don't think Emacs can distinguish one case from the other.

If it would give good results to hide that period in both cases, then
it can be done.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-23  4:42           ` pxref Richard M. Stallman
@ 2005-10-23  7:27             ` Eli Zaretskii
  2005-10-23 13:43               ` pxref Robert J. Chassell
  2005-10-24  1:01               ` pxref Richard M. Stallman
  0 siblings, 2 replies; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-23  7:27 UTC (permalink / raw)
  Cc: juri, nickrob, emacs-devel

> From: "Richard M. Stallman" <rms@gnu.org>
> CC: juri@jurta.org, nickrob@snap.net.nz, emacs-devel@gnu.org
> Date: Sun, 23 Oct 2005 00:42:49 -0400
> 
> the correct way to do that is with `(@xref{args}.)'.
> That will also generate a period before the closeparen
> resulting in similar Info output.
> 
> I don't think Emacs can distinguish one case from the other.

Thanks, now I understand the problem.

However, I think a cross-reference in parens which is a complete
sentence is a very rare situation.  I found only 45 such cases in all
the Emacs manuals (in man/, lispref/ and lispintro/), and only 11 in
the Texinfo manual.  Almost all of these look like a @pxref in
disguise, as the sentence in parens is very short, mostly just the
@xref itself.

So I think it is safe for us to hide the period in this case.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-23  7:27             ` pxref Eli Zaretskii
@ 2005-10-23 13:43               ` Robert J. Chassell
  2005-10-23 16:14                 ` pxref Eli Zaretskii
  2005-10-24  1:01               ` pxref Richard M. Stallman
  1 sibling, 1 reply; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-23 13:43 UTC (permalink / raw)


In the default Info, @pxref succeeds, but @xref fails.

This is a bug in Emacs display, not makeinfo.

This morning's  CVS snapshot, Sun, 2005 Oct 23  12:42 UTC
GNU Emacs 22.0.50.1 (i686-pc-linux-gnu, GTK+ Version 2.6.10)
built using only

 (compile
        "time make -C lisp autoloads && cd lisp && \
         time make recompile \
             EMACS=/usr/local/src/emacs/src/emacs && \
         cd /usr/local/src/emacs/ && \
         time make info html"))

and the (presumably temporary and irrelevant) eval-at-startup bug
handled so I do not get the `Symbol's function definition is void:' 
error, and using GNU Texinfo 4.8.

In the default, Info-hide-note-references is non-nil.
Then @xref lacks a proper capitalized `See' using both 
emacs -Q and emacs -Q -D.

But with emacs -Q and evaluating

    (setq Info-hide-note-references nil)

the cross references appear properly.

With the default, which is t, @xref display wrongly:

    Here is a @pxref (see Musings on fossil fuels) in a sentence.

    Here is a standalone @xref within parentheses:

    (see Musings on fossil fuels.)

    Here is another @xref within parentheses, but with following text:

    (see Musings on fossil fuels, with non-cross-reference text.)



With  the following evaluated,  (setq Info-hide-note-references nil)
the GNU Emacs Info is correct:

    Here is a @pxref (*note Musings on fossil fuels::) in a sentence.

    Here is a standalone @xref within parentheses:

    (*Note Musings on fossil fuels::.)

    Here is another @xref within parentheses, but with following text:

    (*Note Musings on fossil fuels::, with non-cross-reference text.)


Here is the output in DVI:

    Here is an @pxref (see Chapter 2 [Musings on fossil fuels], page
    2) in a sentence.

    Here is a standalone @xref within parentheses:

    (See Chapter 2 [Musings on fossil fuels], page 2.)


    Here is another @xref within parentheses, but with following text:

    (See Chapter 2 [Musings on fossil fuels], page 2, with
    non-cross-reference text.)


Here is the Texinfo source for the two output formats:

    Here is an @@pxref (@pxref{Musings on fossil fuels}) in a sentence.

    Here is a standalone xref within parentheses:

    (@xref{Musings on fossil fuels}.)

    Here is another @@xref within parentheses, but with following text:

    (@xref{Musings on fossil fuels}, with non-cross-reference text.)

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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-23 13:43               ` pxref Robert J. Chassell
@ 2005-10-23 16:14                 ` Eli Zaretskii
  2005-10-23 20:20                   ` pxref Robert J. Chassell
  0 siblings, 1 reply; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-23 16:14 UTC (permalink / raw)
  Cc: emacs-devel

> Date: Sun, 23 Oct 2005 13:43:04 +0000 (UTC)
> From: "Robert J. Chassell" <bob@rattlesnake.com>
> 
> In the default Info, @pxref succeeds, but @xref fails.

``Fails'' in what way?

> This is a bug in Emacs display, not makeinfo.

So this ``failure'' is just a problem with how @xref is displayed?
That is, the cross-reference itself does work when you type `f' or
RET?

> With the default, which is t, @xref display wrongly:
> 
>     Here is a @pxref (see Musings on fossil fuels) in a sentence.
> 
>     Here is a standalone @xref within parentheses:
> 
>     (see Musings on fossil fuels.)
> 
>     Here is another @xref within parentheses, but with following text:
> 
>     (see Musings on fossil fuels, with non-cross-reference text.)

What is wrong with this display?  Please spell that out, since AFAIK
what you show is how the @xref's are supposed to be displayed in
Emacs.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-23 16:14                 ` pxref Eli Zaretskii
@ 2005-10-23 20:20                   ` Robert J. Chassell
  2005-10-23 22:45                     ` pxref Eli Zaretskii
  0 siblings, 1 reply; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-23 20:20 UTC (permalink / raw)


    >     Here is a standalone @xref within parentheses:
    > 
    >     (see Musings on fossil fuels.)
    > 
    >     Here is another @xref within parentheses, but with following text:
    > 
    >     (see Musings on fossil fuels, with non-cross-reference text.)

    What is wrong with this display?  Please spell that out, since
    AFAIK what you show is how the @xref's are supposed to be
    displayed in Emacs.

In English, a standalone sentence is supposed to start with a capital
letter.  This has been a convention for centuries.

Consequently, a standalone @xref should start with a capital S
regardless whether it is inside of parentheses or not.

The parentheses should have no effect and indeed they do not.  In the
default display of today's GNU Emacs CVS snapshot, @xref does wrong in
all circumstances.  When Texinfo 4.8 builds this:

    Here is an @@xref  standing abosolutely alone:

    @xref{Musings on fossil fuels}.

The default for Info started from `emacs -Q -D' is

    Here is an @xref  standing abosolutely alone:

    see Musings on fossil fuels.

That should be a capital S for `See'.  (The reference itself is
underlined and a different color; it is fine.)

In the default GNU Emacs display, @ref is supposed to start with a
lower case letter so the reference can be used within a sentence; and
it does correctly.

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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-23 20:20                   ` pxref Robert J. Chassell
@ 2005-10-23 22:45                     ` Eli Zaretskii
  2005-10-24 13:38                       ` pxref Richard M. Stallman
  0 siblings, 1 reply; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-23 22:45 UTC (permalink / raw)
  Cc: emacs-devel

> Date: Sun, 23 Oct 2005 20:20:00 +0000 (UTC)
> From: "Robert J. Chassell" <bob@rattlesnake.com>
> 
>     What is wrong with this display?  Please spell that out, since
>     AFAIK what you show is how the @xref's are supposed to be
>     displayed in Emacs.
> 
> In English, a standalone sentence is supposed to start with a capital
> letter.

Now I understand what is your complaint.  Yes, I agree that the
sentence produced by an @xref should begin with a capitalized "See".

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-23  7:27             ` pxref Eli Zaretskii
  2005-10-23 13:43               ` pxref Robert J. Chassell
@ 2005-10-24  1:01               ` Richard M. Stallman
  1 sibling, 0 replies; 39+ messages in thread
From: Richard M. Stallman @ 2005-10-24  1:01 UTC (permalink / raw)
  Cc: juri, nickrob, emacs-devel

    However, I think a cross-reference in parens which is a complete
    sentence is a very rare situation.  I found only 45 such cases in all
    the Emacs manuals (in man/, lispref/ and lispintro/), and only 11 in
    the Texinfo manual.  Almost all of these look like a @pxref in
    disguise, as the sentence in parens is very short, mostly just the
    @xref itself.

    So I think it is safe for us to hide the period in this case.

I have no particular opinion about it.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-23 22:45                     ` pxref Eli Zaretskii
@ 2005-10-24 13:38                       ` Richard M. Stallman
  2005-10-24 20:17                         ` pxref Robert J. Chassell
  2005-10-25  8:05                         ` pxref Eli Zaretskii
  0 siblings, 2 replies; 39+ messages in thread
From: Richard M. Stallman @ 2005-10-24 13:38 UTC (permalink / raw)
  Cc: bob, emacs-devel

    Now I understand what is your complaint.  Yes, I agree that the
    sentence produced by an @xref should begin with a capitalized "See".

Isn't that how they actually begin, in the Info file?

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-24 13:38                       ` pxref Richard M. Stallman
@ 2005-10-24 20:17                         ` Robert J. Chassell
  2005-10-26 16:46                           ` pxref Richard M. Stallman
  2005-10-25  8:05                         ` pxref Eli Zaretskii
  1 sibling, 1 reply; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-24 20:17 UTC (permalink / raw)
  Cc: bob

       Now I understand what is your complaint.  Yes, I agree that the
       sentence produced by an @xref should begin with a capitalized "See".

   Isn't that how they actually begin, in the Info file?

That is the bug:  sentences stopped beginning that way in the default.

(Using today's CVS snapshot.)  There are two reasons that I discovered
in emacs/lisp/info.el in `Info-fontify-node':

  * First, the previous line of text, before a blank line, ended with a
    colon.  (You can see that here.)  A colon followed by a blank line
    should not force lower case when a blank line intervenes.

    To fix, move the colon from the conditional that always produces a
    lower case `see' and to the conditional that produces an upper
    case `See' and add a colon to the conditional involving a blank
    line.

  * Second, the `Info-fontify-node' function used `looking-at' rather than
    `looking-back' which meant never checking backwards.

    To fix, replace `looking-at' with `looking-back'.

I think these two changes fix all the problems; but I do not use the
default except for tests and would appreciate your trying this
yourself.  Thank you.

Here is the diff:


diff -c2p /usr/local/src/emacs/lisp/info.el.\~1.453.\~ /usr/local/src/emacs/lisp/info.el
*** /usr/local/src/emacs/lisp/info.el.~1.453.~	Mon Oct 24 19:22:40 2005
--- /usr/local/src/emacs/lisp/info.el	Mon Oct 24 20:10:24 2005
*************** the variable `Info-file-list-for-emacs'.
*** 3738,3744 ****
                    (skip-syntax-backward " ")
                    (setq other-tag
!                         (cond ((memq (char-before) '(nil ?\. ?! ??))
                                 "See ")
!                               ((memq (char-before) '(?\, ?\; ?\: ?-))
                                 "see ")
                                ((memq (char-before) '(?\( ?\[ ?\{))
--- 3738,3744 ----
                    (skip-syntax-backward " ")
                    (setq other-tag
!                         (cond ((memq (char-before) '(nil ?\. ?! ?\: ??))
                                 "See ")
!                               ((memq (char-before) '(?\, ?\; ?-))
                                 "see ")
                                ((memq (char-before) '(?\( ?\[ ?\{))
*************** the variable `Info-file-list-for-emacs'.
*** 3746,3753 ****
                                 ;; an end of sentence
                                 (skip-syntax-backward " (")
!                                (if (memq (char-before) '(nil ?\. ?! ??))
                                     "See "
                                   "see "))
!                               ((save-match-data (looking-at "\n\n"))
                                 "See "))))
                  (goto-char next)
--- 3746,3753 ----
                                 ;; an end of sentence
                                 (skip-syntax-backward " (")
!                                (if (memq (char-before) '(nil ?\. ?! ?\: ??))
                                     "See "
                                   "see "))
!                               ((save-match-data (looking-back "\n\n"))
                                 "See "))))
                  (goto-char next)



Here is a test Texinfo file that you can convert to Info:

\input texinfo.tex                             @c -*-texinfo-*-
@comment %**start of header
@setfilename foo.info
@settitle Texinfo Test
@smallbook
@comment %**end of header

@ignore

 ## Summary of shell commands to create various output formats:

    pushd /u/texinfo/

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

    ## ;; (kill-buffer "*info*")
    ## ;; (info "/u/texinfo/foo.info" nil)

    ## DVI output
    texi2dvi foo.texi

    ## View DVI output
    ##     xdvi foo.dvi &

    ## HTML output
    makeinfo --no-split --html foo.texi

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


    ## DocBook output
    makeinfo --docbook --no-split --paragraph-indent=0 \
    --verbose foo.texi

    ## XML output
    makeinfo --xml --no-split --paragraph-indent=0 \
    --verbose foo.texi

    popd

@end ignore

@titlepage
@sp 6
@re@center @titlefont{Test document}
@sp 4
@center by Robert J. Chassell

@page
@vskip 0pt plus 1filll
@end titlepage

@contents

@ifnottex
@node Top, Chapter One, (dir), (dir)
@top Test Top
@end ifnottex

@menu
* Chapter One::
* Musings on fossil fuels::
@end menu

@node Chapter One, Musings on fossil fuels, Top, Top
@chapter Chapter One

Contents of chapter 1.

See
@url{http://www.rattlesnake.com/notions/Choice-and-Constraint.html},
which is intended as a description of `what is'
@dots{} not the `what is' that most often concern citizens,
politicians, and political scientists, but a `what is' that reflects
the current world and of `what can be done'.


    Here is an @@pxref (@pxref{Musings on fossil fuels}) in a sentence.

    Here is a standalone xref within parentheses:

    (@xref{Musings on fossil fuels}.)

    Here is another @@xref, no parentheses; this ends with a colon and
    is followed by a blank line:

    @xref{Musings on fossil fuels}.

    Here is another @@xref, no parentheses; this ends with a period and
    is followed by a blank line.

    @xref{Musings on fossil fuels}.

More contents of chapter 1.


@node Musings on fossil fuels,  , Chapter One, Top
@chapter Musings on Fossil Fuels

More than three decades ago, some modern people I knew tried to
replace fossil fuels with alternate sources of energy.

@bye



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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-24 13:38                       ` pxref Richard M. Stallman
  2005-10-24 20:17                         ` pxref Robert J. Chassell
@ 2005-10-25  8:05                         ` Eli Zaretskii
  2005-10-25 20:29                           ` pxref Richard M. Stallman
  1 sibling, 1 reply; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-25  8:05 UTC (permalink / raw)
  Cc: bob, emacs-devel

> From: "Richard M. Stallman" <rms@gnu.org>
> CC: bob@rattlesnake.com, emacs-devel@gnu.org
> Date: Mon, 24 Oct 2005 09:38:08 -0400
> 
>     Now I understand what is your complaint.  Yes, I agree that the
>     sentence produced by an @xref should begin with a capitalized "See".
> 
> Isn't that how they actually begin, in the Info file?

No, in the Info file they begin with "*Note".  Emacs hides that and
replaces it with "see".

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-25  8:05                         ` pxref Eli Zaretskii
@ 2005-10-25 20:29                           ` Richard M. Stallman
  0 siblings, 0 replies; 39+ messages in thread
From: Richard M. Stallman @ 2005-10-25 20:29 UTC (permalink / raw)
  Cc: bob, emacs-devel

    No, in the Info file they begin with "*Note".  Emacs hides that and
    replaces it with "see".

Please forgive me.  My memory is really going.

One possible solution is to check whether the open-paren around the
xref is a sentence-begin.  If so, it could use "See".

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-24 20:17                         ` pxref Robert J. Chassell
@ 2005-10-26 16:46                           ` Richard M. Stallman
  2005-10-26 18:29                             ` pxref Robert J. Chassell
                                               ` (2 more replies)
  0 siblings, 3 replies; 39+ messages in thread
From: Richard M. Stallman @ 2005-10-26 16:46 UTC (permalink / raw)
  Cc: bob, emacs-devel

      * First, the previous line of text, before a blank line, ended with a
	colon.  (You can see that here.)  A colon followed by a blank line
	should not force lower case when a blank line intervenes.

	To fix, move the colon from the conditional that always produces a
	lower case `see' and to the conditional that produces an upper
	case `See' and add a colon to the conditional involving a blank
	line.

That would make colon always act like the end of a sentence.
That isn't correct.

I installed a fix I think is correct.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-26 16:46                           ` pxref Richard M. Stallman
@ 2005-10-26 18:29                             ` Robert J. Chassell
  2005-10-26 18:34                             ` pxref Robert J. Chassell
  2005-10-26 19:02                             ` pxref Robert J. Chassell
  2 siblings, 0 replies; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-26 18:29 UTC (permalink / raw)
  Cc: emacs-devel

Today's GNU Emacs CVS snapshot, Wed, 2005 Oct 26  17:49 UTC
GNU Emacs 22.0.50.6 (i686-pc-linux-gnu, GTK+ Version 2.6.10)
started with `emacs -Q -D'

   I installed a fix I think is correct.

The change for Info looks right to me.  I added more test lines to
foo.texi and they all look right in the default Emacs Info.

However, I also checked the DVI version, created by `texi2dvi'.
Unfortunately, that makes a capitalized `See' when the preceding text
with a colon and is not followed by a blank line.

In DVI as viewed with xdvi:

    Here is another @xref, no parentheses; this ends with a colon and
    is not followed by a blank line: See Musings on fossil fuels.

But in the default Info display:

    Here is another @xref, no parentheses; this ends with a colon and
    is not followed by a blank line:  see Musings on fossil fuels.


Here is the test file I used:

\input texinfo.tex  @c -*-texinfo-*-
@comment %**start of header
@setfilename foo.info
@settitle Texinfo Test
@smallbook
@comment %**end of header

@ignore

 ## Summary of shell commands to create various output formats:

    pushd /u/texinfo/

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

    ## ;; (kill-buffer "*info*")
    ## ;; (info "/u/texinfo/foo.info" nil)

    ## DVI output
    texi2dvi foo.texi

    ## View DVI output
    ##     xdvi foo.dvi &

    ## HTML output
    makeinfo --no-split --html foo.texi

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


    ## DocBook output
    makeinfo --docbook --no-split --paragraph-indent=0 \
    --verbose foo.texi

    ## XML output
    makeinfo --xml --no-split --paragraph-indent=0 \
    --verbose foo.texi

    popd

@end ignore

@titlepage
@sp 6
@center @titlefont{Test document}
@sp 4
@center by Robert J. Chassell

@page
@vskip 0pt plus 1filll
@end titlepage

@contents

@ifnottex
@node Top, Chapter One, (dir), (dir)
@top Test Top
@end ifnottex

@menu
* Chapter One::
* Musings on fossil fuels::
@end menu

@node Chapter One, Musings on fossil fuels, Top, Top
@chapter Chapter One

Contents of chapter 1.

See

@url{http://www.rattlesnake.com/notions/Choice-and-Constraint.html},

which is intended as a description of `what is'
@dots{} not the `what is' that most often concern citizens,
politicians, and political scientists, but a `what is' that reflects
the current world and of `what can be done'.


    Here is an @@pxref (@pxref{Musings on fossil fuels}) in a sentence.

    Here is a standalone xref within parentheses:

    (@xref{Musings on fossil fuels}.)

    Here is another @@xref, no parentheses; this ends with a colon and
    is followed by a blank line:

    @xref{Musings on fossil fuels}.

    Here is another @@xref, no parentheses; this ends with a period and
    is followed by a blank line.

    @xref{Musings on fossil fuels}.


    Here is another @@xref, no parentheses; this ends with a colon and
    is not followed by a blank line:  @xref{Musings on fossil fuels}.

    Here is another @@xref, no parentheses; this ends with a period and
    is not followed by a blank line.  @xref{Musings on fossil fuels}.

    Here is a standalone xref within parentheses, no blank line:
    (@xref{Musings on fossil fuels}.)



More contents of chapter 1.


@node Musings on fossil fuels,  , Chapter One, Top
@chapter Musings on Fossil Fuels

More than three decades ago, some modern people I knew tried to
replace fossil fuels with alternate sources of energy.

@bye

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-26 16:46                           ` pxref Richard M. Stallman
  2005-10-26 18:29                             ` pxref Robert J. Chassell
@ 2005-10-26 18:34                             ` Robert J. Chassell
  2005-10-26 19:02                             ` pxref Robert J. Chassell
  2 siblings, 0 replies; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-26 18:34 UTC (permalink / raw)
  Cc: emacs-devel

Today's GNU Emacs CVS snapshot, Wed, 2005 Oct 26  17:49 UTC
GNU Emacs 22.0.50.6 (i686-pc-linux-gnu, GTK+ Version 2.6.10)
started with `emacs -Q -D'

   I installed a fix I think is correct.

The change for Info looks right to me.  I added more test lines to
foo.texi and they all look right in the default Emacs Info.

However, I also checked the DVI version, created by `texi2dvi'.
Unfortunately, that makes a capitalized `See' when the preceding text
with a colon and is not followed by a blank line.

In DVI as viewed with xdvi:

    Here is another @xref, no parentheses; this ends with a colon and
    is not followed by a blank line: See Musings on fossil fuels.

But in the default Info display:

    Here is another @xref, no parentheses; this ends with a colon and
    is not followed by a blank line:  see Musings on fossil fuels.


Here is the test file I used:

% \input /usr/local/src/texinfo-4.8/doc/texinfo.tex  @c -*-texinfo-*-
\input texinfo.tex  @c -*-texinfo-*-
@comment %**start of header
@setfilename foo.info
@settitle Texinfo Test
@smallbook
@comment %**end of header

@ignore

 ## Summary of shell commands to create various output formats:

    pushd /u/texinfo/

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

    ## ;; (kill-buffer "*info*")
    ## ;; (info "/u/texinfo/foo.info" nil)

    ## DVI output
    texi2dvi foo.texi

    ## View DVI output
    ##     xdvi foo.dvi &

    ## HTML output
    makeinfo --no-split --html foo.texi

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


    ## DocBook output
    makeinfo --docbook --no-split --paragraph-indent=0 \
    --verbose foo.texi

    ## XML output
    makeinfo --xml --no-split --paragraph-indent=0 \
    --verbose foo.texi

    popd

@end ignore

@titlepage
@sp 6
@center @titlefont{Test document}
@sp 4
@center by Robert J. Chassell

@page
@vskip 0pt plus 1filll
@end titlepage

@contents

@ifnottex
@node Top, Chapter One, (dir), (dir)
@top Test Top
@end ifnottex

@menu
* Chapter One::
* Musings on fossil fuels::
@end menu

@node Chapter One, Musings on fossil fuels, Top, Top
@chapter Chapter One

Contents of chapter 1.
@sp 1

@ifnotinfo
@noindent
See

@noindent
@url{http://www.rattlesnake.com/notions/Choice-and-Constraint.html},

@noindent
which is intended as a description of `what is'
@dots{} not the `what is' that most often concern citizens,
politicians, and political scientists, but a `what is' that reflects
the current world and of `what can be done'.

@sp 1

    Here is an @@pxref (@pxref{Musings on fossil fuels}) in a sentence.

    Here is a standalone xref within parentheses:

    (@xref{Musings on fossil fuels}.)

    Here is another @@xref, no parentheses; this ends with a colon and
    is followed by a blank line:

    @xref{Musings on fossil fuels}.

    Here is another @@xref, no parentheses; this ends with a period and
    is followed by a blank line.

    @xref{Musings on fossil fuels}.


    Here is another @@xref, no parentheses; this ends with a colon and
    is not followed by a blank line:  @xref{Musings on fossil fuels}.

    Here is another @@xref, no parentheses; this ends with a period and
    is not followed by a blank line.  @xref{Musings on fossil fuels}.

    Here is a standalone xref within parentheses, no blank line:
    (@xref{Musings on fossil fuels}.)



@sp 1
More contents of chapter 1.


@node Musings on fossil fuels,  , Chapter One, Top
@chapter Musings on Fossil Fuels

More than three decades ago, some modern people I knew tried to
replace fossil fuels with alternate sources of energy.

@bye

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-26 16:46                           ` pxref Richard M. Stallman
  2005-10-26 18:29                             ` pxref Robert J. Chassell
  2005-10-26 18:34                             ` pxref Robert J. Chassell
@ 2005-10-26 19:02                             ` Robert J. Chassell
  2005-10-26 20:55                               ` pxref Robert J. Chassell
  2005-10-26 20:57                               ` pxref Karl Berry
  2 siblings, 2 replies; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-26 19:02 UTC (permalink / raw)
  Cc: karl, emacs-devel

Today's GNU Emacs CVS snapshot, Wed, 2005 Oct 26  17:49 UTC
GNU Emacs 22.0.50.6 (i686-pc-linux-gnu, GTK+ Version 2.6.10)
started with `emacs -Q -D'

"Richard M. Stallman" <rms@gnu.org> wrote,

   I installed a fix I think is correct.

The change for Info looks correct to me.  I added more test lines to
foo.texi and they all look right in the default Emacs Info.

(When I set `Info-hide-note-references' to nil, as I do customarily, 
the cross references are shown by `*note' and `*Note' in a manner that
is correct.) 

However, using `xdvi', I also checked the DVI version that was created
with `texi2dvi' (GNU Texinfo 4.8) 1.34.  Unfortunately, the typeset
output displays a capitalized `See' when the preceding text both ends
with a colon and is not followed by a blank line.

Karl Berry, what do you see?

In DVI as viewed with xdvi:

    Here is another @xref, no parentheses; this ends with a colon and
    is not followed by a blank line: See Musings on fossil fuels.

(Note also the one space before the `See'; this is correct for
typesetting according to my copy of the Chicago Manual of Style,
although when I type, I tend to put in two spaces.  The Texinfo source
has two spaces.)

But in the default Info display:

    Here is another @xref, no parentheses; this ends with a colon and
    is not followed by a blank line:  see Musings on fossil fuels.

(Note the two spaces before the `see', which looks right to me.)

Here is the test file I used:


\input texinfo.tex  @c -*-texinfo-*-
@comment %**start of header
@setfilename foo.info
@settitle Texinfo Test
@smallbook
@comment %**end of header

@ignore

 ## Summary of shell commands to create various output formats:

    pushd /u/texinfo/

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

    ## ;; (kill-buffer "*info*")
    ## ;; (info "/u/texinfo/foo.info" nil)

    ## DVI output
    texi2dvi foo.texi

    ## View DVI output
    ##     xdvi foo.dvi &

    ## HTML output
    makeinfo --no-split --html foo.texi

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


    ## DocBook output
    makeinfo --docbook --no-split --paragraph-indent=0 \
    --verbose foo.texi

    ## XML output
    makeinfo --xml --no-split --paragraph-indent=0 \
    --verbose foo.texi

    popd

@end ignore

@titlepage
@sp 6
@center @titlefont{Test document}
@sp 4
@center by Robert J. Chassell

@page
@vskip 0pt plus 1filll
@end titlepage

@contents

@ifnottex
@node Top, Chapter One, (dir), (dir)
@top Test Top
@end ifnottex

@menu
* Chapter One::
* Musings on fossil fuels::
@end menu

@node Chapter One, Musings on fossil fuels, Top, Top
@chapter Chapter One

Contents of chapter 1.
@sp 1

@ifnotinfo
@noindent
@end ifnotinfo
See
@ifnotinfo

@noindent
@end ifnotinfo
@url{http://www.rattlesnake.com/notions/Choice-and-Constraint.html},
@ifnotinfo

@noindent
@end ifnotinfo
which is intended as a description of `what is'
@dots{} not the `what is' that most often concern citizens,
politicians, and political scientists, but a `what is' that reflects
the current world and of `what can be done'.

@sp 1

    Here is an @@pxref (@pxref{Musings on fossil fuels}) in a sentence.

    Here is a standalone xref within parentheses:

    (@xref{Musings on fossil fuels}.)

    Here is another @@xref, no parentheses; this ends with a colon and
    is followed by a blank line:

    @xref{Musings on fossil fuels}.

    Here is another @@xref, no parentheses; this ends with a period and
    is followed by a blank line.

    @xref{Musings on fossil fuels}.


    Here is another @@xref, no parentheses; this ends with a colon and
    is not followed by a blank line:  @xref{Musings on fossil fuels}.

    Here is another @@xref, no parentheses; this ends with a period and
    is not followed by a blank line.  @xref{Musings on fossil fuels}.

    Here is a standalone xref within parentheses, no blank line:
    (@xref{Musings on fossil fuels}.)


@sp 1
More contents of chapter 1.


@node Musings on fossil fuels,  , Chapter One, Top
@chapter Musings on Fossil Fuels

More than three decades ago, some modern people I knew tried to
replace fossil fuels with alternate sources of energy.

@bye


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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-26 19:02                             ` pxref Robert J. Chassell
@ 2005-10-26 20:55                               ` Robert J. Chassell
  2005-10-26 20:57                               ` pxref Karl Berry
  1 sibling, 0 replies; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-26 20:55 UTC (permalink / raw)


Oops!  I did not mean to send so many messages on pxref and am not
sure what happened.  Please refer to the one for 19:02:50 UTC, the
last one, which has a CC to Karl Berry.

Quick summary:  the Info fix looks good
                I discovered a problem with texi2dvi

My apologies.

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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-26 19:02                             ` pxref Robert J. Chassell
  2005-10-26 20:55                               ` pxref Robert J. Chassell
@ 2005-10-26 20:57                               ` Karl Berry
  2005-10-27 12:20                                 ` pxref Robert J. Chassell
  1 sibling, 1 reply; 39+ messages in thread
From: Karl Berry @ 2005-10-26 20:57 UTC (permalink / raw)
  Cc: rms, emacs-devel

    output displays a capitalized `See' when the preceding text both ends
    with a colon and is not followed by a blank line.

I don't understand why you expect anything else.  @xref always produces
a capitalized "See" in TeX output.  It's documented that way and been
that way forever.  I'm sure you know this perfectly well, so I'm
confused as to what you have in mind.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-26 20:57                               ` pxref Karl Berry
@ 2005-10-27 12:20                                 ` Robert J. Chassell
  2005-10-27 14:03                                   ` pxref Karl Berry
  2005-10-28  3:47                                   ` pxref Richard M. Stallman
  0 siblings, 2 replies; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-27 12:20 UTC (permalink / raw)
  Cc: rms, emacs-devel

       output displays a capitalized `See' when the preceding text both ends
       with a colon and is not followed by a blank line.

   I don't understand why you expect anything else.  @xref always produces
   a capitalized "See" in TeX output.  It's documented that way and been
   that way forever.  I'm sure you know this perfectly well, so I'm
   confused as to what you have in mind.

The new default Info and TeX sometimes produce different outputs,
uncapitalized and capitalized.

Specifically, in the new Info default, @xref does not always produce a
capitalized "See".  This means a change in documentation.

(Personally, I prefer the older way, but that is neither here nor
there.  The older way means the author makes the decision, not the
computer.  The new default means that the computer makes the decision,
but, currently, only in Info.  Except for tests, I do not see the new
Info default but set Info-hide-note-references to nil.)

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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 12:20                                 ` pxref Robert J. Chassell
@ 2005-10-27 14:03                                   ` Karl Berry
  2005-10-27 15:03                                     ` pxref Robert J. Chassell
  2005-10-27 19:06                                     ` pxref Eli Zaretskii
  2005-10-28  3:47                                   ` pxref Richard M. Stallman
  1 sibling, 2 replies; 39+ messages in thread
From: Karl Berry @ 2005-10-27 14:03 UTC (permalink / raw)
  Cc: rms, emacs-devel

    Specifically, in the new Info default, @xref does not always produce a
    capitalized "See".  This means a change in documentation.

Sorry, I have no background for and don't understand any of this.  How
can such a fundamental thing about Texinfo be changed?  Who did it and
why?  Why should "Info" (do you mean info.el?) be downcasing/upcasing
"See" (last I knew it was *[nN]ote" in Info, anyway, not "[sS]ee"])?

As far as I know neither makeinfo nor texinfo.tex have changed in this
regard.  Also, it is not feasible to change texinfo.tex in any such way,
if I'm understanding right.  (Messing with stuff based on surrounding
text.)

karl

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 14:03                                   ` pxref Karl Berry
@ 2005-10-27 15:03                                     ` Robert J. Chassell
  2005-10-27 15:29                                       ` pxref Karl Berry
  2005-10-27 19:06                                     ` pxref Eli Zaretskii
  1 sibling, 1 reply; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-27 15:03 UTC (permalink / raw)
  Cc: rms, emacs-devel

     ... (last I knew it was *[nN]ote" in Info, anyway, not "[sS]ee"])?

This has changed in the default Info display.  

Now the default display (which as I said, I do not see normally) is
"[sS]ee"].  As far as I know, this has been the case for more than a
year.  (The earlist vc-annotate change date that seems relevant is
08-Apr-04.)

    As far as I know neither makeinfo nor texinfo.tex have changed in
    this regard.

True.  (I have been using Texinfo version 4.8 for both makeinfo and
texi2dvi with the texinfo.tex saying version 2003-07-16.18.)  The
change is to the online display, as shown in a current default Emacs
from CVS.

Thus, using today's GNU Emacs CVS snapshot, Thu, 2005 Oct 27 12:31 UTC
GNU Emacs 22.0.50.7 (i686-pc-linux-gnu, GTK+ Version 2.6.10) started
with

    emacs/src/emacs -q

we see this in Info:

    Here is another @xref, no parentheses; this ends with a colon and
    is not followed by a blank line:  see Musings on fossil fuels.

not `*Note'.

When you set

    (setq Info-hide-note-references nil)

then you see `*Note' in your display.  `*Note' is also what is copied
between different instances of Emacs, even when the display of the
source shows `see'.

A copy of Emacspeak with Emacspeak code from 24 May 2005 speaks
`*Note' but visually shows `See' with a capital `S'.

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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 15:03                                     ` pxref Robert J. Chassell
@ 2005-10-27 15:29                                       ` Karl Berry
  2005-10-27 17:32                                         ` pxref Robert J. Chassell
  0 siblings, 1 reply; 39+ messages in thread
From: Karl Berry @ 2005-10-27 15:29 UTC (permalink / raw)
  Cc: rms, emacs-devel

    this has been the case for more than a year.

Not in any released version, which is why I've never seen it.

        Here is another @xref, no parentheses; this ends with a colon and
        is not followed by a blank line:  see Musings on fossil fuels.

    not `*Note'.

Anyway, the original source is wrong if it uses @xref in that context.
@xref is specified precisely to output a capital "See", always.
If/Since Emacs Info now changes it to "see" sometimes, then yes, that
will certainly cause incorrect output and should be changed.  I hope
whoever made the change in info.el can deal with it.  I see nothing to
change in Texinfo itself in this regard.  I certainly don't think we
should change the specification of @xref.

Karl

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 15:29                                       ` pxref Karl Berry
@ 2005-10-27 17:32                                         ` Robert J. Chassell
  2005-10-27 23:07                                           ` pxref Karl Berry
  2005-10-29 20:34                                           ` pxref Richard M. Stallman
  0 siblings, 2 replies; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-27 17:32 UTC (permalink / raw)
  Cc: rms, emacs-devel

The problem for @xref and @ref is that the Info output format reduces
information.

`texi2dvi' works from the Texinfo source.
For display, the current CVS Emacs works from the Info output format.

If `*Note' is to replaced by `[Ss]ee', then the display must do
something in Emacs Lisp to choose whether or not to capitalize.  This
can only be done by emulating in code what a human does when choosing
@xref and @ref.

Karl, I am becoming more and more convinced that you should do
nothing.  But those who think `Info-hide-note-references' should be
set to a non-nil value (not me; I set the value to nil) should figure
out a good display algorithm.

Or else they should re-write Emacs Info to use Texinfo files when they
are available rather than Info files, being sure the new Info looks
good on slow connections, on text based terminals, and the various
window systems, such as X.

Or else they should follow my suggestion mentioned below for single
input/multiple output.

To see the issue, add the following to the Texinfo source test file.
(This text is indented in the source as shown.)

    Here is an @@xref as a sentence.  @xref{Musings on fossil fuels}.

    Here is an @@ref in a sentence, see @ref{Musings on fossil fuels}.

    Here is an @@pxref parenthetically (@ref{Musings on fossil fuels})
    in a sentence.

The Texinfo source produces this in the Info output format.  (This
output is not indented four spaces; I did that for this message.)

    Here is an @xref as a sentence.  *Note Musings on fossil fuels::.

    Here is an @ref in a sentence, see *Note Musings on fossil fuels::.

    Here is an @pxref parenthetically (*Note Musings on fossil fuels::)
     in a sentence.

Note the `see' from the source for the @ref and the one space
indentation of the `in' for the @pxref.  (Perhaps the source should
not be written with `see'.  I have left it because many make this
mistake although the Texinfo manual counsels against it.)

Here is the default Info display using today's CVS Emacs snapshot:

    Here is an @xref as a sentence.  See Musings on fossil fuels::.

    Here is an @ref in a sentence, see see Musings on fossil fuels::.

    Here is an @pxref parenthetically (see Musings on fossil fuels::)
     in a sentence.

Note the doubled `see' and the one space indentation of the `in'.

Here is the DVI output from texi2dvi 4.8 as seen using xdvi version 22.40v:

    Here is an @xref as a sentence. See Chapter 2 [Musings on fossil
    fuels], page 2.

    Here is an @ref in a sentence, see Chapter 2 [Musings on fossil
    fuels], page 2.

    Here is an @pxref parenthetically (Chapter 2 [Musings on fossil
    fuels], page 2) in a sentence.

W3M mode in GNU Emacs displays this:

    Here is an @xref as a sentence. See Musings on fossil fuels.

    Here is an @ref in a sentence, see Musings on fossil fuels.

    Here is an @pxref parenthetically (Musings on fossil fuels) in a sentence.

Both the DVI and the HTML display look correct to me (except for the
single space between sentences).

As far as I can see, the best result comes from writing Texinfo source
so the output looks good in Info (both with and without
`Info-hide-note-references' set to nil), in DVI (or PostScript or PDF;
I think they all display the same), in HTML, in plain text, and
listened to.  Texinfo is, after all, a single input format that
provides multiple outputs.

    http://www.rattlesnake.com/notions/single-multiple.html

But this depends on the virtue of the author ...

Or else, we should make it quick to update Info displays in Emacs;
then tell people to use an Emacs/Emacspeak that shows the high
resolution printed, the Web, and the two Info displays, and also
provides input for text-to-speech generation, which comes from Info.
Several years ago I wrote a prototype but it did not update Info
quickly and was a failure.

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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 14:03                                   ` pxref Karl Berry
  2005-10-27 15:03                                     ` pxref Robert J. Chassell
@ 2005-10-27 19:06                                     ` Eli Zaretskii
  1 sibling, 0 replies; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-27 19:06 UTC (permalink / raw)
  Cc: bob, rms, emacs-devel

> Date: Thu, 27 Oct 2005 09:03:40 -0500
> From: karl@freefriends.org (Karl Berry)
> Cc: rms@gnu.org, emacs-devel@gnu.org
> 
>     Specifically, in the new Info default, @xref does not always produce a
>     capitalized "See".  This means a change in documentation.
> 
> Sorry, I have no background for and don't understand any of this.  How
> can such a fundamental thing about Texinfo be changed?  Who did it and
> why?  Why should "Info" (do you mean info.el?) be downcasing/upcasing
> "See" (last I knew it was *[nN]ote" in Info, anyway, not "[sS]ee"])?
> 
> As far as I know neither makeinfo nor texinfo.tex have changed in this
> regard.  Also, it is not feasible to change texinfo.tex in any such way,
> if I'm understanding right.  (Messing with stuff based on surrounding
> text.)

Karl, you are absolutely right: there's no need to change anything in
Texinfo.  There's a huge misunderstanding here, mainly due to the fact
that Bob insists on saying that some ``new Info default'' changed what
@xref produces.

The reality is that @xref still produces the same output as it did
_in_the_Info_file, except that the Emacs Info reader displays the
leading "*Note" as "[Ss]ee", and hides the explicit reference to the
Info node and file, leaving just the reference name.  This was done to
make the cross-references look more like hyperlinks in other GUI
applications.

This is only a display trickery; info.el does not change the text read
from the Info file, it just uses the Emacs display features to display
certain text as a something different.  Thus, if one copies from the
buffer to another one, they will still see the same "*Note" and the
full cross-reference as makeinfo produces.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 17:32                                         ` pxref Robert J. Chassell
@ 2005-10-27 23:07                                           ` Karl Berry
  2005-10-28  8:28                                             ` pxref Eli Zaretskii
  2005-10-28 17:32                                             ` pxref Robert J. Chassell
  2005-10-29 20:34                                           ` pxref Richard M. Stallman
  1 sibling, 2 replies; 39+ messages in thread
From: Karl Berry @ 2005-10-27 23:07 UTC (permalink / raw)
  Cc: rms, emacs-devel

    If `*Note' is to replaced by `[Ss]ee', 

A dubious desire in the first place, IMHO.

    Karl, I am becoming more and more convinced that you should do nothing.

Good.  That is my plan :).

    Or else they should re-write Emacs Info to use Texinfo files when they

And not call it Info.  Info should remain as it is.

Doing such a thing has been suggested in other contexts.  It amounts to
a new Texinfo parser, so that any changes to the language have to be
propagated there.

A better alternative would be to use the Texinfo XML output, which is
not dependent on the exact Texinfo language, but (is intended to)
preserve(s) all the semantics, so that it can be displayed as desired.
There are some projects doing this now, amazingly enough.

    Here is an @@pxref parenthetically (@ref{Musings on fossil fuels})

I think you used @ref where you meant @pxref.  FWIW, my recollection
(not checked) is that @pxref produces *note, not *Note.

Best,
k

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 12:20                                 ` pxref Robert J. Chassell
  2005-10-27 14:03                                   ` pxref Karl Berry
@ 2005-10-28  3:47                                   ` Richard M. Stallman
  1 sibling, 0 replies; 39+ messages in thread
From: Richard M. Stallman @ 2005-10-28  3:47 UTC (permalink / raw)
  Cc: emacs-devel, karl

    Specifically, in the new Info default, @xref does not always produce a
    capitalized "See".  This means a change in documentation.

It should always produce "See" _when used properly_.
That is, at the beginning of a sentence.

If you find an example when it does not produce "See"
when used properly, that is a bug, so please report it.

As for what happens when it is used improperly, that is
user error, not a bug.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 23:07                                           ` pxref Karl Berry
@ 2005-10-28  8:28                                             ` Eli Zaretskii
  2005-10-29  2:24                                               ` pxref Luc Teirlinck
  2005-10-28 17:32                                             ` pxref Robert J. Chassell
  1 sibling, 1 reply; 39+ messages in thread
From: Eli Zaretskii @ 2005-10-28  8:28 UTC (permalink / raw)
  Cc: bob, emacs-devel

> Date: Thu, 27 Oct 2005 18:07:50 -0500
> From: karl@freefriends.org (Karl Berry)
> Cc: rms@gnu.org, emacs-devel@gnu.org
> 
> A better alternative would be to use the Texinfo XML output

That has the disadvantage that one needs to install 2 sets of manuals:
one in XML for Emacs, the other in Info for the stand-alone reader and
other Info readers out there (since on a typical multi-user machine,
the admins cannot make sure only one reader is used).

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 23:07                                           ` pxref Karl Berry
  2005-10-28  8:28                                             ` pxref Eli Zaretskii
@ 2005-10-28 17:32                                             ` Robert J. Chassell
  1 sibling, 0 replies; 39+ messages in thread
From: Robert J. Chassell @ 2005-10-28 17:32 UTC (permalink / raw)
  Cc: rms, emacs-devel

Karl Berry wrote

    A better alternative would be to use the Texinfo XML output, which is
    not dependent on the exact Texinfo language, but (is intended to)
    preserve(s) all the semantics, so that it can be displayed as desired.
    There are some projects doing this now, amazingly enough.

Yes, that is an excellent idea.  It moves a great deal forward,
including DebianDoc.  Which, if any, of the projects are in Emacs?

However, Eli Zaretskii has a point: without educating all users,
including guests, this means that system adminstrators will need to
install 2 sets of manuals.

        Here is an @@pxref parenthetically (@ref{Musings on fossil fuels})

    I think you used @ref where you meant @pxref.  

Yes, my mistake.

    FWIW, my recollection (not checked) is that @pxref produces *note,
    not *Note.

Yes, you are right.  I just checked.  Also, in the default Info
display in CVS GNU Emacs, it produces `see', as does `texi2dvi'.
@pxref displays correctly in the two versions of the Info display and
with xdvi (and presumably others).

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

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-28  8:28                                             ` pxref Eli Zaretskii
@ 2005-10-29  2:24                                               ` Luc Teirlinck
  2005-10-29 19:47                                                 ` pxref Juri Linkov
  0 siblings, 1 reply; 39+ messages in thread
From: Luc Teirlinck @ 2005-10-29  2:24 UTC (permalink / raw)
  Cc: bob, emacs-devel, karl

Eli Zaretskii wrote:

   That has the disadvantage that one needs to install 2 sets of manuals:
   one in XML for Emacs, the other in Info for the stand-alone reader and
   other Info readers out there (since on a typical multi-user machine,
   the admins cannot make sure only one reader is used).

I guess that if, some time in the future, Emacs were able to read the
XML format, info.el would need to remain in the Emacs distribution to
read Info files for which no XML version exists.  Since the XML format
would be available for those who find a modern browser look so very
important, info.el could then once again become free of unfixable bugs
and could once again be a true Info reader instead of a buggy would-be
formatter that attempts to do the impossible: format plain text.

Disk space is getting cheaper and cheaper so I guess most admins would
elect to install both.  But if some would only install the Info files,
users would still have access to exactly the same documentation.  The
only problem (if they consider it one) is that they get presented with
something that looks less "modern", less like a browser and less colorful.

To me it seems obvious that making Emacs able to read the Info files
in XML format would be very desirable.  The main problem right now is
simple: Emacs currently quite simply can not read the XML output.
Another problem could be that many Emacs maintainers do not know XML
or do not know it well enough.  (I myself do not know XML, although I
could always sooner or later learn it.)

Sincerely,

Luc.

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-29  2:24                                               ` pxref Luc Teirlinck
@ 2005-10-29 19:47                                                 ` Juri Linkov
  2005-11-08 20:59                                                   ` pxref Kevin Rodgers
  0 siblings, 1 reply; 39+ messages in thread
From: Juri Linkov @ 2005-10-29 19:47 UTC (permalink / raw)
  Cc: bob, eliz, karl, emacs-devel

> To me it seems obvious that making Emacs able to read the Info files
> in XML format would be very desirable.  The main problem right now is
> simple: Emacs currently quite simply can not read the XML output.

Emacs already can read the XML output.  Just try, for example, after:

  makeinfo --xml -o info.xml info.texi

to evaluate:

  (xml-parse-file "info.xml")

`xml-parse-file' is a top-level autoloaded function, so you will see
its output, which is just a simple Lisp structure.

So the main problem is not reading the XML output, but rendering such
Lisp structure into the final representation.  This could be achieved
most easily with using DSSSL-like rules.

-- 
Juri Linkov
http://www.jurta.org/emacs/

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-27 17:32                                         ` pxref Robert J. Chassell
  2005-10-27 23:07                                           ` pxref Karl Berry
@ 2005-10-29 20:34                                           ` Richard M. Stallman
  1 sibling, 0 replies; 39+ messages in thread
From: Richard M. Stallman @ 2005-10-29 20:34 UTC (permalink / raw)
  Cc: emacs-devel, karl

    The Texinfo source produces this in the Info output format.  (This
    output is not indented four spaces; I did that for this message.)

	Here is an @xref as a sentence.  *Note Musings on fossil fuels::.

	Here is an @ref in a sentence, see *Note Musings on fossil fuels::.

	Here is an @pxref parenthetically (*Note Musings on fossil fuels::)
	 in a sentence.

    Note the `see' from the source for the @ref and the one space
    indentation of the `in' for the @pxref.

That last seems to be a bug in Texinfo.  Why does it indent that space?

      (Perhaps the source should
    not be written with `see'.  I have left it because many make this
    mistake although the Texinfo manual counsels against it.)

It is correct to write "see" before @ref; that is necessary to get
the right output in DVI.

This change in Emacs should prevent Emacs from displaying a second "see".
Does it work right for you?

*** info.el	26 Oct 2005 12:35:51 -0400	1.454
--- info.el	29 Oct 2005 16:05:29 -0400	
***************
*** 3741,3747 ****
  		    ;; an end of sentence
  		    (skip-syntax-backward " ("))
                    (setq other-tag
!                         (cond ((memq (char-before) '(nil ?\. ?! ??))
                                 "See ")
  			      ((save-match-data
  				 (save-excursion
--- 3741,3749 ----
  		    ;; an end of sentence
  		    (skip-syntax-backward " ("))
                    (setq other-tag
! 			(cond ((save-match-data (looking-back "see "))
! 			       "")
! 			      ((memq (char-before) '(nil ?\. ?! ??))
                                 "See ")
  			      ((save-match-data
  				 (save-excursion

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-10-29 19:47                                                 ` pxref Juri Linkov
@ 2005-11-08 20:59                                                   ` Kevin Rodgers
  2005-11-09  9:34                                                     ` pxref Juri Linkov
  0 siblings, 1 reply; 39+ messages in thread
From: Kevin Rodgers @ 2005-11-08 20:59 UTC (permalink / raw)


Juri Linkov wrote:
 > Emacs already can read the XML output.  Just try, for example, after:
 >
 >   makeinfo --xml -o info.xml info.texi
 >
 > to evaluate:
 >
 >   (xml-parse-file "info.xml")
 >
 > `xml-parse-file' is a top-level autoloaded function, so you will see
 > its output, which is just a simple Lisp structure.
 >
 > So the main problem is not reading the XML output, but rendering such
 > Lisp structure into the final representation.  This could be achieved
 > most easily with using DSSSL-like rules.

The alternative would be to use XSLT directives.  But you've then got to
implement a DSSSL or XSLT processor in Emacs Lisp, for a reasonable
subset of the chosen stylesheet language.

-- 
Kevin Rodgers

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-11-08 20:59                                                   ` pxref Kevin Rodgers
@ 2005-11-09  9:34                                                     ` Juri Linkov
  2005-11-09 19:27                                                       ` pxref Kevin Rodgers
  0 siblings, 1 reply; 39+ messages in thread
From: Juri Linkov @ 2005-11-09  9:34 UTC (permalink / raw)
  Cc: emacs-devel

>> So the main problem is not reading the XML output, but rendering such
>> Lisp structure into the final representation.  This could be achieved
>> most easily with using DSSSL-like rules.
>
> The alternative would be to use XSLT directives.  But you've then got to
> implement a DSSSL or XSLT processor in Emacs Lisp, for a reasonable
> subset of the chosen stylesheet language.

XSLT is a poor man's Lisp, and DSSSL uses Scheme syntax.

I meant rather using their semantics, but to write rules in Emacs Lisp.

-- 
Juri Linkov
http://www.jurta.org/emacs/

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-11-09  9:34                                                     ` pxref Juri Linkov
@ 2005-11-09 19:27                                                       ` Kevin Rodgers
  2005-11-10  1:39                                                         ` pxref Miles Bader
  0 siblings, 1 reply; 39+ messages in thread
From: Kevin Rodgers @ 2005-11-09 19:27 UTC (permalink / raw)


Juri Linkov wrote:
 >>>So the main problem is not reading the XML output, but rendering such
 >>>Lisp structure into the final representation.  This could be achieved
 >>>most easily with using DSSSL-like rules.
 >>
 >>The alternative would be to use XSLT directives.  But you've then got to
 >>implement a DSSSL or XSLT processor in Emacs Lisp, for a reasonable
 >>subset of the chosen stylesheet language.
 >
 > XSLT is a poor man's Lisp, and DSSSL uses Scheme syntax.

XSLT is a rich man's CSS, and it is actively evolving.  DSSSL is dead.
And given xml-parse-file, Scheme syntax has little advantage over XML
syntax.

 > I meant rather using their semantics, but to write rules in Emacs Lisp.

DSSSL and XSLT are not just stylesheet languages but transformation
languages, with complex semantics that are difficult to implement.

-- 
Kevin Rodgers

^ permalink raw reply	[flat|nested] 39+ messages in thread

* Re: pxref
  2005-11-09 19:27                                                       ` pxref Kevin Rodgers
@ 2005-11-10  1:39                                                         ` Miles Bader
  0 siblings, 0 replies; 39+ messages in thread
From: Miles Bader @ 2005-11-10  1:39 UTC (permalink / raw)
  Cc: emacs-devel

2005/11/10, Kevin Rodgers <ihs_4664@yahoo.com>:
> And given xml-parse-file, Scheme syntax has little advantage over XML
> syntax.

Er, well it does have a few advantages: (1) It's less ugly than XML,
(2) It can be edited more Easily in emacs (because Emacs does a very
good job with lisp-like syntaxes), and (3) It's not a
buzzword-of-the-week.

-Miles
--
Do not taunt Happy Fun Ball.

^ permalink raw reply	[flat|nested] 39+ messages in thread

end of thread, other threads:[~2005-11-10  1:39 UTC | newest]

Thread overview: 39+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <17238.44739.179203.166568@kahikatea.snap.net.nz>
2005-10-20  9:29 ` pxref Eli Zaretskii
2005-10-20 11:16   ` pxref Juri Linkov
2005-10-20 12:41     ` pxref Eli Zaretskii
2005-10-20 16:46       ` pxref Juri Linkov
2005-10-20 19:35         ` pxref Eli Zaretskii
2005-10-23  4:42           ` pxref Richard M. Stallman
2005-10-23  7:27             ` pxref Eli Zaretskii
2005-10-23 13:43               ` pxref Robert J. Chassell
2005-10-23 16:14                 ` pxref Eli Zaretskii
2005-10-23 20:20                   ` pxref Robert J. Chassell
2005-10-23 22:45                     ` pxref Eli Zaretskii
2005-10-24 13:38                       ` pxref Richard M. Stallman
2005-10-24 20:17                         ` pxref Robert J. Chassell
2005-10-26 16:46                           ` pxref Richard M. Stallman
2005-10-26 18:29                             ` pxref Robert J. Chassell
2005-10-26 18:34                             ` pxref Robert J. Chassell
2005-10-26 19:02                             ` pxref Robert J. Chassell
2005-10-26 20:55                               ` pxref Robert J. Chassell
2005-10-26 20:57                               ` pxref Karl Berry
2005-10-27 12:20                                 ` pxref Robert J. Chassell
2005-10-27 14:03                                   ` pxref Karl Berry
2005-10-27 15:03                                     ` pxref Robert J. Chassell
2005-10-27 15:29                                       ` pxref Karl Berry
2005-10-27 17:32                                         ` pxref Robert J. Chassell
2005-10-27 23:07                                           ` pxref Karl Berry
2005-10-28  8:28                                             ` pxref Eli Zaretskii
2005-10-29  2:24                                               ` pxref Luc Teirlinck
2005-10-29 19:47                                                 ` pxref Juri Linkov
2005-11-08 20:59                                                   ` pxref Kevin Rodgers
2005-11-09  9:34                                                     ` pxref Juri Linkov
2005-11-09 19:27                                                       ` pxref Kevin Rodgers
2005-11-10  1:39                                                         ` pxref Miles Bader
2005-10-28 17:32                                             ` pxref Robert J. Chassell
2005-10-29 20:34                                           ` pxref Richard M. Stallman
2005-10-27 19:06                                     ` pxref Eli Zaretskii
2005-10-28  3:47                                   ` pxref Richard M. Stallman
2005-10-25  8:05                         ` pxref Eli Zaretskii
2005-10-25 20:29                           ` pxref Richard M. Stallman
2005-10-24  1:01               ` pxref Richard M. Stallman

Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/emacs.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).