RFR: 7009069: ZipFile.getEntry(String name) does NOT respect the "language encoding flag" [v2]

[Alan Bateman]

On Sun, 31 Dec 2023 23:25:51 GMT, Eirik Bjørsnøs <eirbjo at openjdk.org> wrote:

>> Please review this test-only PR which adds test coverage for `ZipFile.getEntry` under certain charset conditions. 
>> 
>> When `ZipFile.getEntry` is called for an entry which has the `Language encoding flag` general purpose bit flag set,  then `ZipCoder.UTF8` is used unconditionally, even when a different charset was supplied to the `ZipFile` constructor.
>> 
>> It turns out we do not have any testing for this particular case, as can be verified by commenting out the following line of code in `ZipFile.Source.getEntryPos`:
>> 
>> 
>> //ZipCoder zc = zipCoderForPos(pos);
>> ``` 
>> 
>> and then running `make test TEST="test/jdk/java/util/zip"`
>> 
>> The current test verifies that the correct ZipCoder is used by `ZipFile.entries()`, but does not exercise `ZipFile.getEntry` the same way.
>> 
>> Seeing that [JDK-7009069](https://bugs.openjdk.org/browse/JDK-7009069) was (accidentally?) fixed by [JDK-8243469](https://bugs.openjdk.org/browse/JDK-8243469), I think it is worthwhile to add explicit testing for this case to avoid regressions.
>> 
>> While visiting `ZipCoding.java`, I took the opportunity to convert it to JUnit 5. The conversion and modernization of the code is done in the first commit 1384850ed51ec845af06dd6d13616f20f8bbaa6a in this PR, while the second commit 1776b258b0fe8383709ae0c095f2631a4e6237f6 actually adds the code required to verify the `Language encoding flag` condition for `ZipFile.getEntry`.
>> 
>> Testing: Verified that the test indeed fails when `ZipFile.Source.getEntryPos` is updated to use the ZipFile's ZipCoder as suggested above.
>
> Eirik Bjørsnøs has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Introduce separate tests for ZipInputStream and ZipFile, use parameterized tests with providers for write charset, read charset, name and comment arguments.

It's great to improve the test coverage in this area but I'm sure about re-purposing JDK-7009069 to do this. Consider a future archaeologist looking through the commits where it initially looks this might be a fix to ZipFile to respect the language encoding flag. So I think create a new issue for the test work or else change the title on this one.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/17207#issuecomment-1873225058