phpxpath-users

Re: special char at pos 0 problem

[J. C. <jan...@im...>]

Found a simple solution:

Add these special chars to preg_match expression on line 2672:

Old:
    elseif (preg_match('/^[a-zA-Z0-9\-_]+/', $nodeTest)) {

New:
    elseif (preg_match('/^[a-zA-Z0-9\-_ÄÖÜäöü]+/', $nodeTest)) {

Maybe there are more chars to add?

Cheers
 Jan

> -----Ursprüngliche Nachricht-----
> Von: php...@li...
> [mailto:php...@li...]Im Auftrag von J.
> Carmanns
> Gesendet: Donnerstag, 16. Mai 2002 17:19
> An: php...@li...
> Betreff: special char at pos 0 problem
>
>
>
> > >    * It makes use of the get_html_translation_table(HTML_ENTITIES) php
> > > library
> > >    * call, so is limited in the same ways.  At the time of
> writing this
> > > seemed
> > >    * be restricted to iso-8859-1
>
>
> Hmmm - iso-8859-1 should allow chars like "ä" "ü" "ö" ("ae",
> "ue", "oe") and
> my XML is valid - but when I use tags like <über/> and call
> evaluate() then
> I get an error:
>
> XPath error in XPath.class_cvs.php:2680 While parsing the XPath query
> "/Daten[1]/Visite[1]/über[1]/*" an empty and therefore invalid
> node-test has
> been found.
>
> When I change the first letter it is fine. It only happens when
> the special
> char is at position 0 in the nodeName.
>
> Help?

Re: special char at pos 0 problem

[Peter R. <php...@pe...>]

On Friday 17 May 2002 9:03, J. Carmanns wrote:
> Found a simple solution:
>
> Add these special chars to preg_match expression on line 2672:
>
> Old:
>     elseif (preg_match('/^[a-zA-Z0-9\-_]+/', $nodeTest)) {
>
> New:
>     elseif (preg_match('/^[a-zA-Z0-9\-_ÄÖÜäöü]+/', $nodeTest)) {
>
> Maybe there are more chars to add?

Looks more like a quick fix than a solution - why only German accents and not 
French or anything else?!

raises the question in my mind: why is this limitation on node names in there 
anyway? Surely there is no such limitation in the xml spec?

Re: special char at pos 0 problem

[Nigel S. <nig...@us...>]

> On Friday 17 May 2002 9:03, J. Carmanns wrote:
> > Found a simple solution:
> >
> > Add these special chars to preg_match expression on line 2672:
> >
> > Old:
> >     elseif (preg_match('/^[a-zA-Z0-9\-_]+/', $nodeTest)) {
> >
> > New:
> >     elseif (preg_match('/^[a-zA-Z0-9\-_ÄÖÜäöü]+/', $nodeTest)) {
> >
> > Maybe there are more chars to add?
>
> Looks more like a quick fix than a solution - why only German accents and
not
> French or anything else?!
>
> raises the question in my mind: why is this limitation on node names in
there
> anyway? Surely there is no such limitation in the xml spec?

There is no such limitation.  In that line we are purely trying to work out
if the $nodeTest is JUST a node name, or whether it is an XPath Expression.

My suggestion is that we change to:

elseif (preg_match('/^[\w\-]+$/', $nodeTest)) {

The PHP manual says \w is a word character where a word character is defined
as:

=============
A "word" character is any letter or digit or the underscore character, that
is, any character which can be part of a Perl "word". The definition of
letters and digits is controlled by PCRE's character tables, and may vary if
locale-specific matching is taking place (see "Locale support" above). For
example, in the "fr" (French) locale, some char- acter codes greater than
128 are used for accented letters, and these are matched by \w.
=============

Cheers,

Nigel

Re: special char at pos 0 problem

[Peter R. <php...@pe...>]

what context/baseXPath does this new feature of being able to use an 
XPathquery in say getData use? Does it always use the evaluate default, i.e. 
the root node? If so, you should be able to check on presence of / or [
If present, then it's a query, if not, it's a nodename. (er, I think, or am I 
getting confused? :-)

I think we need to move away from the concept of locale/charset. If you're 
using ucs, a file, or for that matter node, can contain characters from any 
number of different languages/locales, and has nothing to do with what locale 
the machine/operating system works under.

On Tuesday 21 May 2002 11:06 pm, Nigel Swinson wrote:
>
> There is no such limitation.  In that line we are purely trying to work out
> if the $nodeTest is JUST a node name, or whether it is an XPath Expression.
>
> My suggestion is that we change to:
>
> elseif (preg_match('/^[\w\-]+$/', $nodeTest)) {
>
> The PHP manual says \w is a word character where a word character is
> defined as:
>
> =============
> A "word" character is any letter or digit or the underscore character, that
> is, any character which can be part of a Perl "word". The definition of
> letters and digits is controlled by PCRE's character tables, and may vary
> if locale-specific matching is taking place (see "Locale support" above).
> For example, in the "fr" (French) locale, some char- acter codes greater
> than 128 are used for accented letters, and these are matched by \w.

Re: special char at pos 0 problem

[Nigel S. <nig...@us...>]

> what context/baseXPath does this new feature of being able to use an
> XPathquery in say getData use? Does it always use the evaluate default,
i.e.
> the root node? If so, you should be able to check on presence of / or [
> If present, then it's a query, if not, it's a nodename. (er, I think, or
am I
> getting confused? :-)

The first thing we do is to check to see if the $xPathQuery is actually an
absoluteXPath.  If it is not then we call evaluate/match treating it as a
query.  See setModMatch() and _resolveXPathQuery() for more detailed
documentation of how the class operates.

> I think we need to move away from the concept of locale/charset. If you're
> using ucs, a file, or for that matter node, can contain characters from
any
> number of different languages/locales, and has nothing to do with what
locale
> the machine/operating system works under.

I agree.  But using the built in \w is a whole lot better than explicitly
listing european characters in the [] regex.  Suggestions for what to do
with that test would be welcomed.  Bascially the test needs to capture a
string that could feasibly be an XML element name.  Here's the references to
the specs:

http://www.w3.org/TR/xpath#NT-NameTest

http://www.w3.org/TR/REC-xml-names#NT-NCName

http://www.w3.org/TR/REC-xml#NT-Letter

Cheers

Nigel