389651 - ​ (zero width space) should be ignored while searching/finding text

Status: RESOLVED FIXED

(Core :: Find Backend, defect)

Description

[Martijn Wargers (dead)]

This is a follow-up from bug 294615, comment 5.
"
We _may_ also want to ignore ​ (zero width space), which can be used as a
line-breaking character without visible hyphen (for example in URLs).
"
That seems like a good idea to me.
And maybe there are more unicode characters that need to ignored.
It might be useful to see what other browsers are doing.

Comment 1

[Jonathan Haas]

Most other browsers don't even ignore ­ when searching. Opera ignores it only when the hyphen is invisible (=not at the end of a line). 
No browser I tested (Opera, IE6, Safari) ignored ​.

Comment 2

[Martijn Wargers (dead)]

http://www.unicode.org/Public/UNIDATA/PropList.txt might be an interesting list (from bug 389056, comment 45).

Comment 3

[Mardeg]

Added a data: entry splitting the word "testing" up with various zero width space characters in the URL field for testing with any proposed patches.

Comment 4

[j. 'mach' wust]

Bug 647805 is pretty much about the same problem, but with regard to ZWJ and ZWNJ. Bug 640856 more generally addresses all Unicode normalization issues in the "Find" operation. If it is patched, this bug will be fixed, too.

Comment 6

[Eelis]

Tested in Chromium 50 just now, and it handles this correctly.

Comment 7

[Chris]

We recently hit this problem with our documentation pages when viewed by Firefox.
The raw HTML source gets rendered as:

<code>database.history&#8203;.kafka.bootstrap.servers</code>

If a user attempts to search for database.history.kafka.bootstrap.servers, the user gets no hits because of the inclusion of the non-breaking space unicode character; however if a user performs the same search on the same page in Chrome, it is able to find this reference.

Comment 8

[Emilio Cobos Álvarez (:emilio)]

Should be easy enough to fix... :)

Comment 9

[Emilio Cobos Álvarez (:emilio)]

Attachment: Bug 389651 - Make find-in-page ignore non-matching default-ignorable characters. r=jfkthame

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/f84ef8062ad9
Make find-in-page ignore non-matching default-ignorable characters. r=jfkthame

Comment 11

[Bogdan Tara[:bogdan_tara | bogdant]]

https://hg.mozilla.org/mozilla-central/rev/f84ef8062ad9