[PATCH] KEncodingProber fixes

[Peter Oberndorfer]

Actually return the values detected by unicodeTest().
Make the old deprecated encodingName() method still return values.
The new encoding method is preferred since it does not leak memory.

Fix a memory leak in setProberType
Rename encodingNameByteArray() to encoding() as suggested
Remove a unused #define
Add some unit tests

thanks to Jeff Mitchell for parts of the patch
---

This is my work in progress for today.
This makes the unicodeTest() function useful again.
I plan to continue tomorrow, and to add some more testcases,
test if things still work properly without a BOM.
And finally if everything is OK, and everybody agrees, commit the patch.
I might need some help backporing it.

 kate/document/katebuffer.cpp             |    2 +-
 kdecore/localization/kencodingprober.cpp |   30 ++++---
 kdecore/localization/kencodingprober.h   |    4 +-
 kdecore/tests/CMakeLists.txt             |    1 +
 kdecore/tests/kencodingprobertest.cpp    |  126 ++++++++++++++++++++++++++++++
 kdecore/tests/kencodingprobertest.h      |   36 +++++++++
 6 files changed, 184 insertions(+), 15 deletions(-)
 create mode 100644 kdecore/tests/kencodingprobertest.cpp
 create mode 100644 kdecore/tests/kencodingprobertest.h

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 82f2e582de7e0a21632a14421330a18f670874bc.diff
Type: text/x-diff
Size: 10710 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20090321/773a0954/attachment.diff>