Allow injecting additional encoding detector

It would be nice to allow injection of a better encoding detector then the ICU or simple UTF8 detection.

[zhoresh, 2009-11-19 02:47:16]

Fix error in creation of patch file

[zhoresh, 2009-11-19 02:51:02]

Fix path issue in patch

[johnfargo, 2009-11-24 00:03:11]

A few factoring comments. thanks Ziv!

http://codereview.appspot.com/157054/diff/1006/6
File
java/gadgets/src/main/java/org/apache/shindig/gadgets/encoding/EncodingDetector.java
(right):

http://codereview.appspot.com/157054/diff/1006/6#newcode60
java/gadgets/src/main/java/org/apache/shindig/gadgets/encoding/EncodingDetector.java:60:
return alternateDecoder.detectEncoding(input);
IMO we should return ISO_8859_1 here to keep the API intact. We'd then rename
CustomEncodingDetector to FallbackEncodingDetector, documenting that it only
applies when !assume88591IfNotUtf8, putting the slow ICU impl in the fallback.

http://codereview.appspot.com/157054/diff/1006/8
File
java/gadgets/src/test/java/org/apache/shindig/gadgets/encoding/EncodingDetectorTest.java
(right):

http://codereview.appspot.com/157054/diff/1006/8#newcode88
java/gadgets/src/test/java/org/apache/shindig/gadgets/encoding/EncodingDetectorTest.java:88:
}
rather than passing in the custom detector from EncodingDetector, we should mock
out a detector. Then in the "normal" tests, just ensure that the detector was or
wasn't called (the mock will record this state), as appropriate.

[zhoresh, 2009-11-24 03:01:06]

Updated according to comments

[zhoresh, 2009-11-24 03:02:54]

Updated according to comments

[zhoresh, 2009-11-24 03:04:55]

I fixed the code according to the comments

http://codereview.appspot.com/157054/diff/1006/6
File
java/gadgets/src/main/java/org/apache/shindig/gadgets/encoding/EncodingDetector.java
(right):

http://codereview.appspot.com/157054/diff/1006/6#newcode60
java/gadgets/src/main/java/org/apache/shindig/gadgets/encoding/EncodingDetector.java:60:
return alternateDecoder.detectEncoding(input);
On 2009/11/24 00:03:11, johnfargo wrote:
> IMO we should return ISO_8859_1 here to keep the API intact. We'd then rename
> CustomEncodingDetector to FallbackEncodingDetector, documenting that it only
> applies when !assume88591IfNotUtf8, putting the slow ICU impl in the fallback.

Done.

http://codereview.appspot.com/157054/diff/1006/8
File
java/gadgets/src/test/java/org/apache/shindig/gadgets/encoding/EncodingDetectorTest.java
(right):

http://codereview.appspot.com/157054/diff/1006/8#newcode88
java/gadgets/src/test/java/org/apache/shindig/gadgets/encoding/EncodingDetectorTest.java:88:
}
On 2009/11/24 00:03:11, johnfargo wrote:
> rather than passing in the custom detector from EncodingDetector, we should
mock
> out a detector. Then in the "normal" tests, just ensure that the detector was
or
> wasn't called (the mock will record this state), as appropriate.

Done.

[johnfargo, 2009-11-24 06:55:36]

Looks great, patch committed, with only slight formatting updates (lines <100
chars, Falbback -> Fallback).