781038 - java.lang.IndexOutOfBoundsException: setSpan (<a> ... <b>) ends beyond length 0

Status: RESOLVED WONTFIX

(Firefox for Android Graveyard :: Keyboards and IME, defect, P3)

Description

[Chris Peterson [:cpeterson]]

This bug was filed from the Socorro interface and is 
report bp-6cab57f1-6c4f-4652-a831-2bb3b2120804 .
============================================================= 

We've seen at least two crash reports where Firefox tries to select a zero-length string. I suspect our call to mEditable.replace(content) is not succeeding, leaving mEditable with a zero-length string.

Comment 1

[Chris Peterson [:cpeterson]]

Attachment: part-1-select-editable-length.patch

Part 1: Span and select the Editable's actual length (not the length we hope it will be).

Comment 2

[Chris Peterson [:cpeterson]]

Attachment: part-2-extract-setSpan-method.patch

Part 2: Extract setSpan() and setSelection() boilerplate into a convenience method.

Also collapse extraneous initEditable() method into GeckoInputConnection's constructor.

Comment 4

[Chris Peterson [:cpeterson]]

Green tryserver run:
https://tbpl.mozilla.org/?tree=Try&rev=6db4694cc2bf

Comment 5

[Chris Peterson [:cpeterson]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/3422d26da179
https://hg.mozilla.org/integration/mozilla-inbound/rev/8fbaea0f82e8

Comment 6

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/3422d26da179
https://hg.mozilla.org/mozilla-central/rev/8fbaea0f82e8

Comment 7

[Chris Peterson [:cpeterson]]

Comment on attachment 649868 [details] [diff] [review]
part-1-select-editable-length.patch

[Approval Request Comment]
Bug caused by (feature/regressing bug #): N/A
User impact if declined: Socorro reports just a few crash reports from Beta 15, Aurora 16, and Nightly 17 users.
Testing completed (on m-c, etc.): m-c
Risk to taking this patch (and alternatives if risky): Low risk because the patch is just double-checking the string length before copying.
String or UUID changes made by this patch: N/A

Comment 8

[Chris Peterson [:cpeterson]]

Patch 2 does not need to be uplifted because it is just a minor refactoring.

Comment 9

[Chris Peterson [:cpeterson]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/5ca4c8fe973c
https://hg.mozilla.org/releases/mozilla-aurora/rev/a4bdc645127c

Comment 10

[Cristian Nicolae (:xti)]

This crash was not reproduced at all on any branch past 4 weeks. I cannot reproduce it either on the latest Nightly. Closing bug as verified fixed.

--
Firefox 18.0a1 (2012-09-17)
Device: Galaxy Note
OS: Android 4.0.4

Comment 11

[Scoobidiver (away)]

(In reply to Cristian Nicolae (:xti) from comment #10)
> This crash was not reproduced at all on any branch past 4 weeks. I cannot
> reproduce it either on the latest Nightly. Closing bug as verified fixed.
I don't understand how you can be sure it's not reproducible without STR and it no longer happens with so many crashes: https://crash-stats.mozilla.com/query/query?product=FennecAndroid&version=ALL%3AALL&query_search=signature&query_type=contains&query=android.text.SpannableStringBuilder.checkRange&do_query=1
Can you clarify that?

Comment 12

[Chris Peterson [:cpeterson]]

This bug lives again because I backed out the fix (to deal with a regression).

Comment 13

[Cristian Nicolae (:xti)]

(In reply to Scoobidiver from comment #11)
> Can you clarify that?

I checked the crash reports, on what build number and dates were reported to make sure that everything is fine on m-a and m-c. Since Chris has backed out the fix, it means that's opened again, as already mentioned.

Comment 15

[Sylvestre Ledru [:Sylvestre]]

Closing because no crash reported since 12 weeks.

Comment 16

[Sylvestre Ledru [:Sylvestre]]

Closing because no crash reported since 12 weeks.