shashank created CAMEL-26087:
--------------------------------
Summary: camel-stax - xtokenize fails on valid XML that starts
with a comment or doctype, and mode 'w' fails or gives unbalanced wrappers when
the matches are at different depths
Key: CAMEL-26087
URL: https://issues.apache.org/jira/browse/CAMEL-26087
Project: Camel
Issue Type: Bug
Components: camel-stax
Reporter: shashank
Assignee: shashank
Two cases where {{xtokenize}} ({{XMLTokenExpressionIterator}}, main
6a2f66ddec3e) fails on valid XML.
# A document whose prolog has a comment or a doctype and no XML declaration
fails in every mode with {{XMLStreamException: XMLStreamReader does not
supporting getting location}} ({{<!-- generated by the order export
-->\n<orders><order>1</order>...}}). The second compliance check (CAMEL-8164)
treats text before the first start element that starts with {{<}} but not
{{<?}} as proof of a reader without usable locations; with a compliant reader
(Woodstox) that text is the prolog. Nothing is split and the message points at
the StAX implementation.
# Mode {{w}} with a descendant path whose matches are at different depths:
{{<orders><order>1</order><box><order>2</order></box></orders>}} with
{{xtokenize("//order", 'w')}} routes the first token and then fails with
{{IndexOutOfBoundsException: Index -1 out of bounds for length 0}}; with a
container inside a container of the same name a later token has an unbalanced
wrapper ({{<orders><box><order>2</order></orders>}}). The segment log is popped
by element name, but an element that starts right after a match has no entry of
its own, and an inner element of the same name stops the pop too early.
*Fix*: the compliance check looks at what the reader read after that text: a
reader reporting the start of the start tag has the tag still ahead ({{<name}}
followed by {{>}}, {{/}} or white space); Woodstox 7.3.0 does, the JDK reader
reports the end of the tag and is still refused. Mode {{w}} records the depth
of each segment and the end of an element removes the segments deeper than it;
the preceding siblings kept in the wrapper (CAMEL-17466) are unchanged.
*Tests*: new {{XMLTokenizePrologAndWrapDepthTest}}, 5 tests, all fail on main
(the {{XMLStreamException}} above for a comment, a doctype and a comment
containing the root's start tag; the {{IndexOutOfBoundsException}}; {{Expected:
<<orders><order>2</order></orders>> but was:
<<orders><box><order>2</order></orders>>}}). Unit tests of camel-stax with the
fix: 116 run, 0 failures, 2 skipped. A differential run against a DOM parser on
20,000 generated documents: on main 3,065 fail in every mode and mode {{w}}
with {{//*:order}} fails on 633 more and gives 72 tokens that are not
well-formed; with the fix none.
*Related*: mode {{t}} does not decode entities and modes {{i}}/{{u}} cut the
start tag at a {{>}} inside an attribute value: separate issue.
*How it was found*: a reference differential of {{xtokenize}} against a DOM
parser.
*Affected*: main, camel-4.23.x, camel-4.22.x, camel-4.18.x and camel-4.14.x
(same code). Read, not run.
_Filed with Claude Code on behalf of allthingssecurity._
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
