eclipse.platform.ui icon indicating copy to clipboard operation
eclipse.platform.ui copied to clipboard

Mark several unused deprecated methods/constants/classes in org.eclipse.jface package for removal

Open deepika-u opened this issue 2 months ago • 10 comments

deepika-u avatar Oct 29 '25 06:10 deepika-u

Test Results

 3 018 files  ±0   3 018 suites  ±0   2h 8m 52s ⏱️ - 3m 48s  8 242 tests ±0   7 994 ✅ +1  248 💤 ±0  0 ❌  - 1  23 646 runs  ±0  22 855 ✅ +1  791 💤 ±0  0 ❌  - 1 

Results for commit ecf7a1e0. ± Comparison against base commit 26ddd0da.

:recycle: This comment has been updated with latest results.

github-actions[bot] avatar Oct 29 '25 07:10 github-actions[bot]

@merks Could you take a look at this when you get some time please.

deepika-u avatar Nov 03 '25 07:11 deepika-u

@merks Can you take a look at this when you get some time please.

deepika-u avatar Nov 07 '25 05:11 deepika-u

@merks can you see if the change is not good for you? @deepika-u can you re-trigger the build?

vogella avatar Nov 10 '25 07:11 vogella

I wanted to make a general point here as well. If you mark something for removal, please ensure you have the full SDK IDE setup and that you look at the call hierarchy of each and every field, method, or class you plan to mark for deletion. If you find even one use of it, you can assume that you will find 1000 uses of it out there in the ecosystem such that if you delete that thing, there will be things that simply stop working. Then you should ask yourself, is there significant value in the removal that offsets the broken functionality downstream, keeping in mind that any broken functionality in any IDE for any component will generally create a bad impression overall; one that generally comes back to roost in the Platform. Is it worth it?

merks avatar Nov 11 '25 10:11 merks

Especially at least Platform itself must be free from any references first, as otherwise builds in the sub repositories will start to fail!

laeubi avatar Nov 11 '25 10:11 laeubi

If i run this QuickAccessComputerTest locally, with and without pr - in both cases test is passing.

deepika-u avatar Nov 11 '25 12:11 deepika-u

If i run locally the ProgressContantsTest, all the tests are passing locally with and without the pr. image

deepika-u avatar Nov 12 '25 06:11 deepika-u

1 out of 3 runs failed: testBinaryContentTypeWithDescriberParallel (AllSearchTests AllFileSearchTests FileSearchTests) is known to be flaky. I have a look.

vogella avatar Nov 12 '25 08:11 vogella

1 out of 3 runs failed: testBinaryContentTypeWithDescriberParallel (AllSearchTests AllFileSearchTests FileSearchTests) is known to be flaky. I have a look.

Thanks for your input.

deepika-u avatar Nov 12 '25 11:11 deepika-u

@merks : can you take a look at this once you get some time please?

deepika-u avatar Dec 16 '25 07:12 deepika-u