Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 331453

Summary: unicode nonprintable / zero width space completely bricks editors
Product: [Eclipse Project] Platform Reporter: Tobi Vollebregt <tobivollebregt>
Component: SWTAssignee: Felipe Heidrich <eclipse.felipe>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: P3 CC: eclipse.felipe, eclipse, mehmet.gurevin, remy.suen
Version: 3.6.1   
Target Milestone: 3.7 M4   
Hardware: PC   
OS: Linux   
Whiteboard:
Attachments:
Description Flags
ArrayIndexOutOfBoundsException that appears in the error log none

Description Tobi Vollebregt CLA 2010-11-30 11:35:57 EST 
Build Identifier: M20100909-0800

Eclipse is completely bricked (rendering artifacts and/or hang and many errors in error log) by pasting the unicode character U+200B (zero width space) in a Java editor.

I will attach the stacktrace of the exception that is repeatedly put in the error log right after pasting this character in an Eclipse Java editor.

Curiously, with a plain text file only the exception is spammed in the error log, but Eclipse seems to continue running normally. Major problems occur in the Java editor (and other language-specific editors).

Reproducible: Always

Steps to Reproduce:
1. Open eclipse & open a Java file
2. Paste the unicode character U+200B (zero width space) in the editor. Assuming bugzilla and the webserver support unicode properly I have included this invisible characters between the two x's: x&#8203;x
3. Observe that the error log is spammed full of ArrayIndexOutOfBoundsException; rendering of the editor window breaks and often (when pasting multiple times) rendering of other parts of Eclipse break too and/or it hangs indefinitely.
Comment 1 Tobi Vollebregt CLA 2010-11-30 11:38:35 EST 
Created attachment 184146 [details]
ArrayIndexOutOfBoundsException that appears in the error log
Comment 2 Tobi Vollebregt CLA 2010-11-30 11:42:11 EST 
Details about the character copied from KCharSelect

Character: Non-printable U+200B
Name: ZERO WIDTH SPACE
Annotations and Cross References
Notes:
commonly abbreviated ZWSP
this character is intended for invisible word separation and for line break control; it has no width, but its presence between two characters does not prevent increased letter spacing in justification
General Character Properties
Block: General Punctuation
Unicode category: Other, Format
Various Useful Representations
UTF-8: 0xE2 0x80 0x8B
UTF-16: 0x200B
C octal escaped UTF-8: \342\200\213
XML decimal entity: &#8203;
Comment 3 Felipe Heidrich CLA 2010-12-02 15:25:39 EST 
I suspect this was fixed in Bug 308329/Bug 295513
Can you please verify if this problem happens on HEAD ?
Comment 4 Felipe Heidrich CLA 2010-12-02 15:46:18 EST 
I see the problem:
	public static void main(String[] args) {
		Display display = new Display();
		TextLayout layout = new TextLayout(display);
		layout.setText("a\u200B");
		layout.setAscent(20);
		layout.setDescent(6);
		layout.setSegments(new int[] {0, layout.getText().length()});
		layout.getBounds();
	}
Comment 5 Tobi Vollebregt CLA 2010-12-03 07:08:39 EST 
I tested with the nightly eclipse-SDK-N20101202-2000-linux-gtk; issue still occurs with it.

Build id: N20101202-2000
Comment 6 Felipe Heidrich CLA 2010-12-06 10:54:02 EST 
Fixed in HEAD 
Please verify the fix later this week using M4. Thank you.
Comment 7 Tobi Vollebregt CLA 2010-12-20 05:15:56 EST 
Verified; works fine in M4.
Comment 8 Felipe Heidrich CLA 2011-07-05 10:45:32 EDT 
*** Bug 350988 has been marked as a duplicate of this bug. ***
Comment 9 Eric Williams CLA 2016-12-19 16:10:57 EST 
*** Bug 358970 has been marked as a duplicate of this bug. ***