Replace usages of new Image(device, width, height)

[ShahzaibIbrahim]

Replacing it with imageGcDrawer

[github-actions]

Test Results

   545 files  + 6     545 suites  +6   28m 38s ⏱️ - 4m 46s
 4 367 tests +37   4 351 ✅ +35   16 💤 +3  0 ❌  - 1 
16 616 runs  +37  16 502 ✅ +35  114 💤 +3  0 ❌  - 1 

Results for commit 7214710. ± Comparison against base commit e8ed06e.

♻️ This comment has been updated with latest results.

[HeikoKlare]

Can you please share some information on how to test the change to StyledText (for regression)? Where can those affected carets be seen?

[ShahzaibIbrahim]

Can you please share some information on how to test the change to StyledText (for regression)? Where can those affected carets be seen?

I have tested it with snippet332 replacing the styleText string with "English متن بالعربية עברית " and forcing isBidiCaret() method in StyledText class to return true (cos it checks from OS if it is Bidirectional Platform) and with my changes the caret looks same compared to previous versions. I am not sure I could programmatically write any test for it.

[HeikoKlare]

I have tested it with snippet332 replacing the styleText string with "English متن بالعربية עברית " and forcing isBidiCaret() method in StyledText class to return true (cos it checks from OS if it is Bidirectional Platform)

Thank you. I have just tested with that as well and found the adapted code to be working fine. I actually expected the carets to become "better" because of proper scaling, but they are regenerated on a zoom change for the StyledText anyway, which makes the results look exactly the same, even in monitor-specific scaling scenarios.

Could you please inline at least the single-line lamda as proposed by Hannes? For the others I (personally) would be in favor of having it assigned to a variable to not make the code overly complex, but having a better name like leftCaretDrawer or the like would be good.