fabianlauer/tsxml

A small XML parser and serializer. No dependencies, no hassle.

tsxml's Issues

Newlines around text content

Original issue: FabianLauer/vs-code-xml-format#4

Excess attribute with empty string as attribute name in processing instructions

When parsing PIs, the parsed ast nodes always get an attribute with undefined value and an empty string as attribute name. When serialized again, something similar to this appears:

<?alpha ="undefined" ?>

Add XPath query support

Lock-up on unterminated comments

It appears that parsing input which contains an unterminated XML comment causes tsxml to begin traversing beyond the end of the string searching for the closing token. I’m hoping to work on reproducing this and coming up with a fix soon!

Remove return type declarations on chainable methods

Nodes that were self-closing in source are serialized as non-self-closing

For example, when parsing this:
<alpha />
the serialisation will output:
<alpha></alpha>

This is because the parser/AST do not store whether nodes in source code were self closing or not.

Attribute ending in backslash fails formatting

From FabianLauer/vs-code-xml-format#6:

An attribute like:

<absPath url="D:\Users\kgunders\Documents\MR\Projects\HyperLinkSheets"

Will cause the following error:

XML formatting failed: at line 2, column 694: namespace prefix must be followed by a tag name.

Here is more context from the problematic XML:

<Choice Requires="x15">
    <absPath url="D:\Users\kgunders\Documents\MR\Projects\HyperLinkSheets\" xmlns:x15ac="http://schemas.microsoft.com/office/spreadsheetml/2010/11/ac" />
</Choice>

A work around is to temporarily add a character after the backslash, format, and remove the extra character.

Namespace prefixes in closing tags produce a syntax error

They throw a syntax error although clearly they shouldn't. See da85440.

Support periods in tag names

See FabianLauer/vs-code-xml-format#7:

A node like

<system.webServer>
<!-- .... -->
</system.webServer>

Will cause an error and will not format the document.

XML formatting failed: at line #, column #; expected whitespace or end of opening tag.

Provide document information on `DocumentNode` instances

Instances of DocumentNode should provide information about DOCTYPEs and XML versions to user code so the user doesn't have to process literals and attributes in MDOs and PIs themselves.

Empty attribute names in MDOs

In cases where MDOs have both literals and attributes, an attribute with an empty name is parsed.

格式后换行问题

格式前

格式后

如何才能不换行呢       

Max Line Length

There should be a way to limit the line length and insert a line break after a certain number of characters, eg:

<a>
    <values alpha="1" beta="2" delta="3"
            gamma="4" />
</a>

Strips of XML Namespaces when formatting

This makes the XML invalid - Tried on SOAP request.

Order of attributes and literals not preserved in serialised XML

Support void elements as in HTML5 <meta>, <link>, etc

To be able to parse HTML, support for "void elements" is necessary. These elements do not require a closing tag or closing slash, for example:
<meta charset="utf-8">

To support these elements, the parser will need a list of tag names where this is allowed (or even enforced). Without such a list, the parser can not determine whether a void element is one on purpose or whether the missing closing tag or slash is actually syntax error.

The list of void tags for HTML5 can be found here:
https://www.w3.org/TR/html-markup/syntax.html#syntax-elements

Error on format code in a Ternary Operator

I have tried to use Ternary Operators with expression using < and I got the error above, but not when using something like that !=:

XML formatting failed: at line 13, column 76: expected exclamation mark, question mark or alphabetic tag name.

<StackLayout>
    <Label text="label using < " visibility="{{ p.cond < 1 ? 'visible' : 'collapsed' }}"  />
    <Label text="label using != " visibility="{{ p.cond != 1 ? 'visible' : 'collapsed' }}" />
</StackLayout>