[Win32] Enable monitor-specific scaling by default

[HeikoKlare]

The monitor-specific UI scaling on Windows is disabled by default and can be activated via a preference.

This change inverts the default enablement: it enables the feature by default and allows to disable it via preference. The existing experimental preference is renamed to reset the configuration for every consumer. All information about the feature being experimental is removed.

[github-actions]

Test Results

 1 824 files  ±0   1 824 suites  ±0   1h 45m 35s ⏱️ + 11m 10s
 7 918 tests ±0   7 688 ✅  - 2  228 💤 ±0  2 ❌ +2 
23 841 runs  ±0  23 091 ✅  - 2  748 💤 ±0  2 ❌ +2 

For more details on these failures, see this check.

Results for commit b45c4a0. ± Comparison against base commit 88e83a8.

♻️ This comment has been updated with latest results.

[fedejeanne]

I tested this and I'd like to propose some minor changes.

Also, shouldn't the option be then enabled when starting on a new workspace? I started a brand new Eclipse Application with a clean WS and it wasn't

[HeikoKlare]

Also, shouldn't the option be then enabled when starting on a new workspace? I started a brand new Eclipse Application with a clean WS and it wasn't

@fedejeanne The preference is not stored in the workspace but in the configuration. Did you clean that or create a new one as well? At least I cannot reproduce this issue.

[fedejeanne]

Also, shouldn't the option be then enabled when starting on a new workspace? I started a brand new Eclipse Application with a clean WS and it wasn't

@fedejeanne The preference is not stored in the workspace but in the configuration. Did you clean that or create a new one as well? At least I cannot reproduce this issue.

Thank you for the pointer. I must have screwed up somewhere else when checking out your PR because I tried it again and I cannot reproduce it either. One pointer (for my future self too) is that when creating a new Run Configuration, this option is enabled by default:

This makes testing this kind of PRs impossible: I was getting always the default value (true) for the property and it was therefore always enabled after a restart. I disabled that option and was able to test this PR just fine.

I have one final remark: the initial value of true should also be applied in line 375 of the ViewsPreferencePage, otherwise changing the preference from true to false for the very 1st time doesn't not ask for a workbench restart (subsequent changes will ask for a restart though)

[merks]

There is a checkbox in each launch configuration whether to clear the configuration before launching.

[fedejeanne]

There is a checkbox in each launch configuration whether to clear the configuration before launching.

Do you mean this one? #2868 (review)

[fedejeanne]

From my point of view, this PR is now good to go ✔️

@HeikoKlare fyi I made 2 changes (2 commits), they are already squashed.

[merks]

There is a checkbox in each launch configuration whether to clear the configuration before launching.

Do you mean this one? #2868 (review)

Yes. Reading on phone so harder to see all details. Sorry.

[fedejeanne]

There is a checkbox in each launch configuration whether to clear the configuration before launching.

Do you mean this one? #2868 (review)

Yes. Reading on phone so harder to see all details. Sorry.

No worries, I was just making sure I got it right :-) . Thank you, Ed!

[fedejeanne]

Test failures are unrelated:

• Some ResourceInitialSelectionTest tests fail intermittently #294
• Restore equality for all kinds of Point/Rectangle eclipse.platform.swt#1952 fixes the test failure in Windows. Need to wait until tonight's I-Build so the test stops failing in the PRs

[HeikoKlare]

Thank you! Changes look good to me.

fwiw: when introducing the preference, we had to move it to the configuration area because it has to be read before the workspace is available.

[iloveeclipse]

We have a regression: eclipse-platform/eclipse.platform.swt#2003