Recent changes to bugs

#6 Some links with parentheses not parsed correctly

Paul Worrall — Tue, 27 Dec 2022 12:13:48 -0000

Some links with parentheses not parsed correctly

Paul Worrall — Tue, 27 Dec 2022 12:10:05 -0000

e.g. the link '(libc)Parsing of Floats' in Coreutils: 2.4 Floating point numbers fails although the page is accessible directly via:

&DirnameCheck requires '' in @INFODIR

Christopher Yeleighton — Sat, 13 Nov 2010 00:19:20 -0000

The most obvious invocation

./info2html '(binutils)Top' >/dev/null

causes an error message 'could not be opened for reading'. In other words, it requires an explicit path.

To fix this, add '' to @INFODIR in info2html.conf.

pages invalid

Christopher Yeleighton — Wed, 19 May 2010 20:56:10 -0000

The pages produced by info2html are invalid. First of all, they have no DOCTYPE, which makes them render in quirks mode. Moreover, the ALT attribute for images is missing.
See e.g. <URL:http://validator.w3.org/check?uri=http://info2html.sourceforge.net/cgi-bin/info2html-demo/info2html?(texinfo)Overview&charset=(detect+automatically)&doctype=Inline&group=0>

$* no longer available

Michael James — Wed, 17 Sep 2008 03:08:21 -0000

The subroutine "ParsCrossRefs" uses $* to get multi-line matching. This has been deprecated for ages and in no longer available in the current perl 5.10.

Without any understanding of the workings of "sub ParsCrossRefs" I took out "$*" and added the "m" parameter to all matches. It seems to work, I've attached the new info2html file.

bad links

Hans — Thu, 19 Apr 2007 10:02:56 -0000

One of my directory links has the form:

http://localhost/cgi-bin/info2html?\(StumpWM)Top

Clicking on this link gives an error: could not open the info file. However, the following link is valid:

http://localhost/cgi-bin/info2html?\(stumpwm)Top

Escaping/Deescaping is broken

Anonymous — Wed, 10 Jan 2007 23:48:14 -0000

In MAIN the query string gets decoded twice. This is incorrect and will e.g. replace any '+' characters by spaces, effectively making the node string invalid. This can easily be seen when a link with a '+' in it is selected. The corresponding node will not be found.

To fix this in MAIN simply replace:
...
$BaseInfoFile = &DeEscape($1);
$BaseInfoFile =~ s#\.\./##g; # jonh 5/20/97 -- sanitize up-references
$NodeName = &DeEscape($2);
...
with
...
$BaseInfoFile = $1;
$BaseInfoFile =~ s#\.\./##g; # jonh 5/20/97 -- sanitize up-references
$NodeName = $2;
...

to avoid De-Escaping the QUERY_STRING twice.

Also, Escape() and DeEscape() can be stripped down to be simple wrappers for the corresponding CGI functions.
Modify:
...
sub Escape{
local($Tag) = @_;
#-- escaping is not needed anymore KG/28.6.94
$Tag =~ s/ /%20/g; # space
$Tag =~ s/\+/%AB/g; # +
#-- oh yes it is -- jonh 5/16/1997
#$Tag;
return CGI::escape($Tag);
}
...
to
...
sub Escape{
local($Tag) = @_;
return CGI::escape($Tag);
}
...

and
...
sub DeEscape{
local($Tag) = @_;
#-- deescaping is not needed anymore. KG/28.6.94
$Tag =~ s/%AB/+/g;
$Tag =~ s/%20/ /g;
#-- oh yes it is -- jonh 5/16/1997
#$Tag;
return CGI::unescape($Tag);
}
...
to
...
sub DeEscape{
local($Tag) = @_;
return CGI::unescape($Tag);
}
...

Note that using a %AB to encode/decode a '+' is wrong in the first place. It should be %2B.
Furthermore, it does not make much sense to pre-encode the string in Escape() and then encode it. This effectively encodes the '+' and the ' ' twice.