text_scan

Tiny text-extraction helpers shared by the credential providers and the runtime’s error parser. Both shapes — XML <Tag>X</Tag> and JSON "key": "value" — appear in places where dragging in a full XML / JSON parser would be over-investment.

xml_tag_text and json_string_after_key each scan once with two split_once calls; neither validates the surrounding document. Use them only on inputs whose structure you trust (STS / SSO responses, error envelopes), never on arbitrary user-supplied text.

Values

json_string_after_key

</>

pub fn json_string_after_key(
  text: String,
  key: String,
) -> Result(String, Nil)

Pull the string value that follows a quoted JSON "key". Skips the first quote after the key (i.e. tolerates "key": "value", "key":"value", "key" : "value"), then reads the next two quotes as the value’s delimiters. Returns Error(Nil) if the key is absent or the value isn’t a string-typed JSON field.

The function deliberately doesn’t unescape backslash sequences in the value — callers either don’t expect them (STS tokens, error codes) or do their own decoding.

xml_tag_text

</>

pub fn xml_tag_text(
  text: String,
  tag: String,
) -> Result(String, Nil)

Pull the text content between the first <tag> and its matching </tag>. Returns Error(Nil) if either tag is absent. Used by the STS providers (<AccessKeyId>...</AccessKeyId>) and the runtime’s restXml error path (<Code>NoSuchBucket</Code>).