[GRASS-dev] doc bug #411: quoting <> in html pages

[Glynn Clements]

Hamish wrote:

> re bug 411
> http://wald.intevation.org/tracker/?func=detail&atid=207&aid=411&group_id=21
> 
> 
> 
> Any <> chars used in module option descriptions are broken when
> converted into html docs.
> 
> Daniel pointed out that parser.c has print_escaped_for_xml() for this.
> 
> but....
> 
> lib/gis/parser.c  G_usage_html()
> 
> 
> if (opt->description) {
> 	fprintf(stdout, "<DD>");
> 	newbuf = G_str_replace(opt->description, "\n","<br>");
> 	if (newbuf) {
> 	print_escaped_for_xml(stdout, newbuf);
> 		G_free(newbuf);
> 	} else
> 	print_escaped_for_xml(stdout, opt->description);
> 	fprintf(stdout, "</DD>\n");
> }
> 
> 
> print_escaped_for_xml() reformats the <BR>s which were inserted by
> G_str_replace() into "&lt;BR&gt;". :(
> 
> any ideas? Use G_tokenize() with '\n' as the field sep?

Add e.g. "char *escape_xml(const char *)" which returns either its
argument or a dynamically-allocated copy with &<> replaced. Apply that
first, then apply G_str_replace() to the result, then print that.

I wouldn't worry about free()ing the result, as this will only be
applied to a relatively limited amount of data.

-- 
Glynn Clements <glynn at gclements.plus.com>