Checking that a string is a valid attribute name ...

[emma_middlebrook]

Hi

I've found the rules that cover what is a valid attribute name here:

http://www.w3.org/TR/2006/REC-xml-20060816/
2.3 Common Syntactic Constructs

Is there anything already in the dotnet framework which I can use to
test whether a string can be a valid attribute name or do I have to
roll my own implementation?

Thanks for the help!

Emma

[Martin Honnen]

em************* *@fastmail.fm wrote:

I've found the rules that cover what is a valid attribute name here:

http://www.w3.org/TR/2006/REC-xml-20060816/
2.3 Common Syntactic Constructs

Is there anything already in the dotnet framework which I can use to
test whether a string can be a valid attribute name or do I have to
roll my own implementation?

The name rule is the same for element and attribute names.

If you use XmlTextWriter (in .NET 1.x) or XmlWriter (in .NET 2.0) to
produce your XML then you have methods that ensure that a valid name is
used and otherwise throw an exception, see the WriteName method
<http://msdn.microsoft. com/library/default.asp?url =/library/en-us/cpref/html/frlrfSystemXmlX mlTextWriterCla ssWriteNameTopi c.asp>

And with .NET 2.0 if you use any of the methods to for instance write
attributes e.g. WriteAttributeS tring then name arguments passed in are
checked too.

So producing XML with correct names should not be a problem. And parsing
XML should report any name violations any way.

Does that help? Or where/when exactly do you need to check a name?

--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[emma_middlebrook]

Martin Honnen wrote:

em************* *@fastmail.fm wrote:

I've found the rules that cover what is a valid attribute name here:

http://www.w3.org/TR/2006/REC-xml-20060816/
2.3 Common Syntactic Constructs

Is there anything already in the dotnet framework which I can use to
test whether a string can be a valid attribute name or do I have to
roll my own implementation?

The name rule is the same for element and attribute names.

If you use XmlTextWriter (in .NET 1.x) or XmlWriter (in .NET 2.0) to
produce your XML then you have methods that ensure that a valid name is
used and otherwise throw an exception, see the WriteName method
<http://msdn.microsoft. com/library/default.asp?url =/library/en-us/cpref/html/frlrfSystemXmlX mlTextWriterCla ssWriteNameTopi c.asp>

And with .NET 2.0 if you use any of the methods to for instance write
attributes e.g. WriteAttributeS tring then name arguments passed in are
checked too.

I'm using XmlWriter and WriteAttributeS tring which flagged up the
problem!

It was letting through "text!" as an attribute name and only once the
resulting XMLDocument was loaded via the written-to Stream and then
validated against my schema did the error occur.

So producing XML with correct names should not be a problem. And parsing
XML should report any name violations any way.

Does that help? Or where/when exactly do you need to check a name?

We are allowing users to define their own attributename value pairs
ultimately to be put into an XML document as attributes i.e.
attributename = "value". I want to be able to validate the string they
enter as the attributename once they enter it and if there's any of the
disallowed characters just replace them with _.

See what I mean? Thanks for responding so quickly - any ideas? I was
thinking I'd have to bundle the whole definition of what was valid into
a regular expression or something.

Emma

[emma_middlebrook]

em************* *@fastmail.fm wrote:

Martin Honnen wrote:

em************* *@fastmail.fm wrote:

I've found the rules that cover what is a valid attribute name here:
>
http://www.w3.org/TR/2006/REC-xml-20060816/
2.3 Common Syntactic Constructs
>
Is there anything already in the dotnet framework which I can use to
test whether a string can be a valid attribute name or do I have to
roll my own implementation?

The name rule is the same for element and attribute names.

If you use XmlTextWriter (in .NET 1.x) or XmlWriter (in .NET 2.0) to
produce your XML then you have methods that ensure that a valid name is
used and otherwise throw an exception, see the WriteName method
<http://msdn.microsoft. com/library/default.asp?url =/library/en-us/cpref/html/frlrfSystemXmlX mlTextWriterCla ssWriteNameTopi c.asp>

And with .NET 2.0 if you use any of the methods to for instance write
attributes e.g. WriteAttributeS tring then name arguments passed in are
checked too.

I'm using XmlWriter and WriteAttributeS tring which flagged up the
problem!

It was letting through "text!" as an attribute name and only once the
resulting XMLDocument was loaded via the written-to Stream and then
validated against my schema did the error occur.

This is wrong - I've just checked. It does catch the !.

So the main problem is that I need to find some way of validating
user-inputted attribute name's and replace any illegal characters with
underscore.

Any ideas for something simple otherwise I'm going to have to implement
a regular expression according to the w3.org rules - and that's going
to take me a few days to understand!

:-(

Emma

[Martin Honnen]

em************* *@fastmail.fm wrote:

I'm using XmlWriter and WriteAttributeS tring which flagged up the
problem!

It was letting through "text!" as an attribute name and only once the
resulting XMLDocument was loaded via the written-to Stream and then
validated against my schema did the error occur.

Is that .NET 2.0? It does report a "System.Argumen tException: Invalid
name character" for me.
Doesn't seem to do that with .NET 1.x.

>>So producing XML with correct names should not be a problem. And parsing
XML should report any name violations any way.

Does that help? Or where/when exactly do you need to check a name?

We are allowing users to define their own attributename value pairs
ultimately to be put into an XML document as attributes i.e.
attributename = "value". I want to be able to validate the string they
enter as the attributename once they enter it

That could be one in .NET 1.x with WriteName e.g. alike

static public bool IsXmlName (string nameToCheck) {
XmlTextWriter xmlWriter = new XmlTextWriter(n ew StringWriter()) ;
bool isXmlName = true;
try {
xmlWriter.Write Name(nameToChec k);
}
catch (ArgumentExcept ion e) {
isXmlName = false;
}
finally {
xmlWriter.Close ();
}
return isXmlName;
}

and if there's any of the
disallowed characters just replace them with _.

See what I mean? Thanks for responding so quickly - any ideas? I was
thinking I'd have to bundle the whole definition of what was valid into
a regular expression or something.

I don't think there is any API in .NET 1.x or .NET 2.0 that flags the
individual disallowed characters in an XML name and then offers to
replace them with some allowed character. In that case you will need to
write that character replacement routine yourself.

Regular expressions in the .NET framework support Unicode character
groups/blocks with e.g. \p{name} so that might help to write the regular
expression.
--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[emma_middlebrook]

Martin Honnen wrote:

em************* *@fastmail.fm wrote:

I'm using XmlWriter and WriteAttributeS tring which flagged up the
problem!

It was letting through "text!" as an attribute name and only once the
resulting XMLDocument was loaded via the written-to Stream and then
validated against my schema did the error occur.

Is that .NET 2.0? It does report a "System.Argumen tException: Invalid
name character" for me.
Doesn't seem to do that with .NET 1.x.

>So producing XML with correct names should not be a problem. And parsing
XML should report any name violations any way.

Does that help? Or where/when exactly do you need to check a name?

We are allowing users to define their own attributename value pairs
ultimately to be put into an XML document as attributes i.e.
attributename = "value". I want to be able to validate the string they
enter as the attributename once they enter it

That could be one in .NET 1.x with WriteName e.g. alike

static public bool IsXmlName (string nameToCheck) {
XmlTextWriter xmlWriter = new XmlTextWriter(n ew StringWriter()) ;
bool isXmlName = true;
try {
xmlWriter.Write Name(nameToChec k);
}
catch (ArgumentExcept ion e) {
isXmlName = false;
}
finally {
xmlWriter.Close ();
}
return isXmlName;
}

and if there's any of the
disallowed characters just replace them with _.

See what I mean? Thanks for responding so quickly - any ideas? I was
thinking I'd have to bundle the whole definition of what was valid into
a regular expression or something.

I don't think there is any API in .NET 1.x or .NET 2.0 that flags the
individual disallowed characters in an XML name and then offers to
replace them with some allowed character. In that case you will need to
write that character replacement routine yourself.

Regular expressions in the .NET framework support Unicode character
groups/blocks with e.g. \p{name} so that might help to write the regular
expression.

Hi

Found XmlConvert.Veri fyName and XmlConvert.Enco deName!

EncodeName just replaces anything invalid with _xNNNN_ where NNNN is
its unicode value. I can then use a simple regular expression and
Regex.Replace to replace that encoding with the underscore!

Done it - well at least enough so I can cope with anything the user
puts in without having to inform them, and with the _ rather than the
_xNNNN_ the XML doesn't look too bad if a human has a peek.

Emma