assertj / assertj/assertj-generator
The generator is a little too greedy with Iterable of Path and make tests fails
- 主要言語
- Java
- スター
- 72
- フォーク
- 47
- 平均マージ
- 1日 16時間
- マージ済み PR(30日)
- 2
説明
With 2.1.0 version, the generator is too greedy when encountering a class implementing an `java.lang.Iterable`. It should ignore it unless the type extends `java.util.Collection` (where that make sense) or explicitly a `java.lang.Iterable` (where it is harder to tell it makes sense).
Sample classes: [assertj-foo.zip](https://github.com/joel-costigliola/assertj-assertions-generator/files/1377297/assertj-foo.zip)
The test (`FooTest`) will fail because the usage of hasPath is NOT the same than using an iterable: we want to test an equals rather than a collection predicate. Here the exception: _java.lang.AssertionError: Expecting: to contain: <[foobar\a]>but could not find: <[foobar\a]>_
A working alternative is to convert the Path into a List extracting the Iterable content, but that is wrong by all means:
- first, the default semantic is not good (the default = search for subsequence in any order)
- second, as an user, I want to test the path using `equals` (and that is what the generator produce if I have a class not implementing `java.lang.Iterable`):
```
final List paths = new ArrayList<>();
test.forEach(paths::add); // says ["foobar", "A"]
assertThat(foo).hasPath(paths);
```
コントリビューションガイド
調査の方向性
Start by unpacking the linked assertj-foo.zip sample and running FooTest to reproduce the failing hasPath assertion. Trace how the generator handles types implementing java.lang.Iterable, including the distinction from java.util.Collection. Done means Path-like Iterable types use equality semantics while collection types retain collection predicates, with the test passing.
索引モデルが issue の本文から書いたものです。
評価
- 技術スタック
- java
- 領域
- tooling
- issue の種類
- バグ
- 難易度
- 3/5
- 見積もり時間
- 1〜2日
- 活発さ
- 停滞
- 明瞭さ
- おおむね明確
- 初心者へのやさしさ
- 42/100