[win] Generic TextEditor: Some unicode characters cause text to be rendered incorrectly

[nlisker]

Describe the bug Redirected from Eclipse JDT: https://github.com/eclipse-jdt/eclipse.jdt.ui/issues/1403

Adding some unicode characters to a text editor can cause either a change in the font of the line in which the character is added, or the character to not be displayed correctly.

To Reproduce Open any text file in a text editor. My demonstration is with a .java file, but in the link above a .txt file is also used. Type text into that file, for example: (" abc ").

Copy the character ⭳ (\u2B73).

1. Paste it at the start of the string, the character is shown correctly, but the string changes its font:
2. Paste it at the end of the string, the character is shown incorrectly, but the string does not change font:
3. Copy a non-problematic character, like ← (\u2190). a. Paste it between ⭳ and abc, the characters are shown correctly and the font does not change: , b. Paste it at the edge, and the problem resurfaces:

The same issue occurs with the character ⮠ (\u2BA0).

Expected behavior The characters will show correctly and the string will not change its font regardless of the positions those characters appear at.

Environment:

1. Select the platform(s) on which the behavior is seen: Only tested on Windows.

   • • [ ] All OS
   • • [x] Windows
   • • [ ] Linux
   • • [ ] macOS

2. Additional OS info (e.g. OS version, Linux Desktop, etc) Windows 10

3. JRE/JDK version 21 and 22

Version since Eclipse 2024-03, haven't tested older versions. Text file encoding doesn't seem to matter. Tested UTF-8, UTF-16, UTF-16BE, UTF-16LE and ISO-8859-1.

[fedejeanne]

@nlisker thank you for reporting this issue!

Is this something you would like to work on or were you just reporting it? :-)

[nlisker]

Just reporting. I don't know SWT and the Eclipse inner workings well enough to be able to work on this in a meaningful amount of time.

[jukzi]

not reproducible with a plain org.eclipse.swt.custom.StyledText

[jakub-suliga]

I was able to reproduce this issue with plain org.eclipse.swt.custom.StyledText. I used org.eclipse.swt.snippets.Snippet126 for testing and was able to replicate the bug. Here’s what I found:

• The width of the string unexpectedly changes when a Unicode character like ⭳ (\u2B73) is inserted after a single char. This leads incorrect display of the normal character (e.g. " ), either the normal char gets smaller or bigger.
• However, when the Unicode character is inserted after two or more characters, the width of the string remains constant and the character is displayed correctly.

I tested various positions and chars -> see image The width of the text is set in the Method org.eclipse.swt.graphics.TextLayout.shape(long, StyleItem) on line 3897. It gets the width from the method OS.ScriptPlace. There is probably somewhere a rounding error in the code. I will test if the same problem occurs on Linux. When it works normally on Linux, it could be a Windows problem.

Image Explanation: For Example: D⭳ should be 8 + 5 = 13 and not 14.