Refactor stream handling for Cocoa and GTK in ImageLoader class

[Michael5601]

This PR refactors the stream handling by using try-with-resource for opening streams in the ImageLoader class for Cocoa and GTK.

[Michael5601]

@HannesWell can you please review and merge this PR?

[github-actions]

Test Results

   494 files  ±0     494 suites  ±0   9m 0s ⏱️ - 1m 25s
 4 331 tests ±0   4 318 ✅ ±0   13 💤 ±0  0 ❌ ±0 
16 572 runs  ±0  16 464 ✅ ±0  108 💤 ±0  0 ❌ ±0 

Results for commit 6776734. ± Comparison against base commit 5c2611d.

♻️ This comment has been updated with latest results.

[HeikoKlare]

In general, using try-with-resources for these streams makes sense and the change looks sound.

But this also changes the behavior of the methods, as now an SWTError (ERROR_IO) will be thrown in case an IOException happens when closing the stream while currently that exception is swallowed.
I am still in favor of doing it, as it has already been changed that in Windows more than a year ago and (1) no one complained so far and (2) this change will make the implementations consistent across the operating systems again.

We might update the Javadoc as well to clarify that an error will also be thrown when closing the stream fails.

[laeubi]

If closing fails we are probably all lost anyways...

[Michael5601]

We might update the Javadoc as well to clarify that an error will also be thrown when closing the stream fails.

I agree and updated the javadoc accordingly.

[laeubi]

I think we should not complicate this.... reading includes closing, and that a stream is used is and implementation detail, so simply say it do so when I/O error occurs when reading the file is enough.

[HeikoKlare]

I think we should not complicate this.... reading includes closing, and that a stream is used is and implementation detail, so simply say it do so when I/O error occurs when reading the file is enough.

Agreed 👍

[Michael5601]

Now it should be ready.

[HannesWell]

I think we should not complicate this.... reading includes closing, and that a stream is used is and implementation detail, so simply say it do so when I/O error occurs when reading the file is enough.

Fully agree, even though Heiko is technically correct, this is a potential change in behavior, I think this is neglectable. In other clean-ups (from myself and others) in other parts of Eclipse the same simplification was already applied multiple times. So I also think this is fine.

[HannesWell]

The failure of browser-tests is unrelated and known, submitting.
Thanks for this code enhancement.