`DOMWrappingReader.getDepth()` is one too low at `START_ELEMENT`, and typed array reads lose a level

オープン
#62 コメント 0 件 リアクション 0 件 担当者 0 名 GitHub で見る

まだ誰も着手していません。

評価

難易度
3/5
見積もり時間
1〜2日
初心者へのやさしさ
58/100
issue の種類
バグ
明瞭さ
明確に書かれている
活発さ
活発
技術スタック
java
領域
api

調査の方向性

Start in DOMWrappingReader, where getDepth() returns _depth as is and readElementAsArray() moves into children or the virtual END_ELEMENT without the ++_depth that next() applies. Compare with Stax2ReaderAdapter.getDepth(), which is the reference for the expected values, and reproduce the two tables in the issue with a DOMSource. Done when getDepth() matches the expected values at START_ELEMENT, END_ELEMENT and after array reads, and the existing tests still pass. The compatibility impact on WstxDOMWrappingReader users needs a maintainer decision.

索引モデルが issue の本文から書いたものです。

説明

Bug found when fixing #60, I'm afraid Claude is right here, too.
I can provide the PR if you wish.

XMLStreamReader2.getDepth() is documented as 1 at the root element itself, and the same for a start element and its matching end element. DOMWrappingReader breaks this in two ways:

  1. At every START_ELEMENT, getDepth() returns one less than it should. _depth is only incremented when next() moves past the start element, and getDepth() returns it as is. Values at END_ELEMENT and inside elements are correct.
  2. readElementAsArray() (and so readElementAsIntArray() etc.) permanently lowers the depth by one. Starting at START_ELEMENT, it moves into the children (or to the virtual END_ELEMENT of an empty element) without the ++_depth that next() does there, but leaving the END_ELEMENT still does --_depth. This happens for every array read, whether partial, complete or of an empty element. Each one lowers all later depths by one more.

This affects Woodstox too: WstxDOMWrappingReader, returned for any DOMSource, inherits both methods without overriding them. So its depths differ from Woodstox's own stream reader for the same document (TestXMLStreamReader2 in Woodstox checks the stream reader only; no DOM test checks getDepth()). Stax2ReaderAdapter reports the correct values in all cases below.

Repro (depth after each event):

Document, action Expected DOMWrappingReader
<root><a><b/></a>x<c/></root>, plain next() loop <root>1 <a>2 <b>3 </b>3 </a>2 x:1 <c>2 </c>2 </root>1 <root>0 <a>1 <b>2 </b>3 </a>2 x:1 <c>1 </c>2 </root>1
<root><a>1 2 3</a><b/></root>, <a> read with readElementAsIntArray() to the end <root>1 <a>2 </a>2 <b>2 </b>2 </root>1 <root>0 <a>1 </a>1 <b>0 </b>1 </root>0

Coalescing or not makes no difference. Same results for a partial read followed by next(), and for an empty <a/>.

Fix:

  1. getDepth(): return _depth+1 when the current event is START_ELEMENT. This mirrors the END_ELEMENT adjustment in Stax2ReaderAdapter.getDepth().
  2. readElementAsArray(): ++_depth when starting at START_ELEMENT, before either branch (children or empty element).

Checked with a temporary patch: all cases above then match Stax2ReaderAdapter, and the existing tests pass.

Compatibility: fix 1 changes the value callers get at every START_ELEMENT from a DOM reader, including Woodstox users with a DOMSource. Callers that adjusted for the old value will be off by one in the other direction. Fix 2 only changes behaviour after array reads, where the old values were wrong in any case.

主要言語
Java
スター
42
フォーク
21
平均マージ
3時間 50分
マージ済み PR(30日)
2

環境構築

このプロジェクトには開発コンテナ、Dockerfile、コントリビューションガイドがありません。まず README を読み、一般的な手順ははじめてのコントリビューションガイドを参照してください。

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

FasterXML/stax2-api のほかの issue

FasterXML/stax2-api の issue をすべて見る

似ている issue

Java の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。