magmalake xml

xml · API

xml.reader

xml.reader — a pull parser over an in-memory XML document.

XmlReader.next() hands back one XmlEvent at a time: a start tag (with its attributes), an end tag, a run of text, or end of input. Self-closing tags produce a start and an end event, so a consumer never has to special-case <a/>. Comments, processing instructions, the XML declaration and DOCTYPE are skipped; CDATA sections come back as text, verbatim.

Entity references are decoded in text and attribute values: the five predefined ones (&lt; &gt; &amp; &quot; &apos;) and numeric character references (&#65;, &#x41;). Anything else is an error rather than being passed through, because an API response carrying an undeclared entity is malformed and silently keeping &foo; would hide that.

What this is not: a validating parser. It checks that tags nest and close with matching names — the property every consumer relies on — and nothing about DTDs or schemas. Namespaces are left as written (ec2:item stays ec2:item); local_name strips a prefix when a consumer wants to ignore it.

Structs

Attr

struct Attr

One name="value" pair, value already entity-decoded.

Implements Copyable, Movable

Fields
name: String
value: String

XmlEvent

struct XmlEvent

Implements Copyable, Movable

Fields
kind: Int: One of EVENT_START, EVENT_END, EVENT_TEXT, EVENT_EOF.
name: String: The tag name for start and end events; empty otherwise.
text: String: The decoded text for text events; empty otherwise.
attrs: List[Attr]: The attributes of a start tag, in document order.

XmlReader

struct XmlReader

Pull parser over a document held in memory.

The reader owns a copy of the bytes, so the source string may go away. Call next() until it returns an EVENT_EOF event; a malformed document raises with the byte offset where parsing stopped.

Implements Movable

__init__

def __init__(out self, doc: String)

depth

def depth(self) -> Int

How many elements are open at the current position.

next

def next(mut self) -> XmlEvent

Functions

is_blank

def is_blank(s: String) -> Bool

True when s is only XML whitespace — the indentation between tags.

local_name

def local_name(name: String) -> String

ec2:item → item; a name without a prefix is returned unchanged.

decode_entities

def decode_entities(raw: Span[UInt8]) -> String

Decodes the predefined and numeric entity references in raw.

Aliases

comptime EVENT_START = 1
comptime EVENT_END = 2
comptime EVENT_TEXT = 3
comptime EVENT_EOF = 4