Re: [gnu.org #181558] format error in html mono web page

[Juri Linkov]

Richard Stallman <address@hidden> writes:
> I looked at this, and I think that the problem occurred when Dave
> checked the formatter HTML file into CVS.  That is because the file
> contains something that looks like an RCS keyword.  It should contain
> $Log$.

This is an interesting problem when the output of a file with
description of some feature is processed according to its own
description on literal strings it contains.

The aforementioned HTML file contains also other expanded RCS keywords:
the node "(emacs)Version Headers" has $Id$ expanded in several places.
For example, in the following paragraph:

  It is often necessary to use "superfluous" backslashes when
writing the strings that you put in this variable.  For instance, you
might write <CODE>"$Id\$"</CODE> rather than <CODE>"$Id: emacs.html,v 1.1 
2003/10/30 01:27:33 sinuhe Exp $"</CODE>.  The extra
backslash prevents the string constant from being interpreted as a
header, if the Emacs Lisp file containing it is maintained with
version control.

the $Id$ keyword is expanded while $Id\$ is not.

One possibility is to take the advice from this paragraph and to quote
all RCS keywords in the Emacs manual, with explanations for the reader
of the manual how s?he could get the right value from a quoted string.

> The long-term fix, perhaps, is to change makeinfo to output some
> HTML no-op there which will prevent CVS from recognizing the "keyword".

Changing makeinfo to add some HTML no-op, or e.g. to replace a dollar
sign by the HTML numerical character reference in RCS keywords will
definitely work for HTML output.  But one can imagine a situation with
the need to put other Texinfo output file types under version control
where it will not work.  For example, I don't see how RCS keywords
could be prevented from expansion in Info output files.

> Or could you turn off RCS keyword substitution for the HTML files
> of the Emacs manual, and regenerate these and check them in?

This problem exists not only in the HTML files like

http://www.gnu.org/software/emacs/manual/html_node/emacs_141.html
http://www.gnu.org/software/emacs/manual/html_node/emacs_158.html
http://www.gnu.org/software/emacs/manual/html_chapter/emacs_16.html

but also in other files, e.g.

http://www.gnu.org/software/emacs/manual/text/emacs.txt

and some Emacs Lisp files under webcvs/emacs.

I don't know exactly about workflow for putting files on the GNU Web
server, but generally I'd rather suggest not to put generated files
under version control.

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