r1065913 | winterz | 2009-12-24 20:34:31 +0100 (Thu, 24 Dec 2009) | 4 lines

updated, from Juliano
BUG: 219758
MERGE: 4.3

r1065939 | winterz | 2009-12-25 00:09:48 +0100 (Fri, 25 Dec 2009) | 3 lines

minor coding style fixes
SVN_SILENT:

r1066123 | tmcguire | 2009-12-26 00:13:43 +0100 (Sat, 26 Dec 2009) | 2 lines

Print the actual input instead of "aStr"

r1066128 | tmcguire | 2009-12-26 00:29:26 +0100 (Sat, 26 Dec 2009) | 1 line

Port to qtestlib.

r1066129 | tmcguire | 2009-12-26 00:31:21 +0100 (Sat, 26 Dec 2009) | 1 line

Simplify

r1066131 | tmcguire | 2009-12-26 00:39:13 +0100 (Sat, 26 Dec 2009) | 1 line

Simplifiy

r1066158 | ahartmetz | 2009-12-26 07:01:16 +0100 (Sat, 26 Dec 2009) | 1 line

link with gold

r1066349 | tmcguire | 2009-12-26 21:08:57 +0100 (Sat, 26 Dec 2009) | 1 line

SVN_SILENT typo

r1066358 | tmcguire | 2009-12-26 22:56:21 +0100 (Sat, 26 Dec 2009) | 13 lines

I admit defeat: Disable the test.

I found no way to fix the test. The problem was that the test manually included ../kmime_headerfactory.cpp, which
was already included by linking to kmime. However, that cpp file had non-exported private symbols, which
is the reason it was included manually again.

Including the same compilation unit twice introduces interesting problems though, such as two instances of 
the same K_GLOBAL_STATIC being created, which caused the tests to fail.

So there is no sane way I found to test private functions. But maybe private functions ought not to be tested.

Anybody an idea how to make the test work again?

r1066360 | tmcguire | 2009-12-26 23:00:21 +0100 (Sat, 26 Dec 2009) | 12 lines

Prevent spoofing of the From address by using Unicode direction control chars
in the display name.

Make sure to balance the Bidi control characters so that the direction changes
can not flow over to the addrspec.

We balance the characters when parsing the mailbox or addresslist.

Add some tests for that.

Note that this only fixes KMime-based clients, a fix for KMail will follow soon.

r1066375 | tmcguire | 2009-12-26 23:50:29 +0100 (Sat, 26 Dec 2009) | 5 lines

Grmbl, KHTML seems to ignore the PDF char, so all my work was moot...

Instead of balancing the bidi control characters, simply remove them now,
that works even with KHTML.

r1066378 | tmcguire | 2009-12-26 23:52:01 +0100 (Sat, 26 Dec 2009) | 4 lines

Remove bidi control characters here as well.

Not strictly necessary when using KMime, but KMail doesn't use KMime.

r1066585 | tmcguire | 2009-12-27 15:10:05 +0100 (Sun, 27 Dec 2009) | 5 lines

Disable these annoying qdebugs.

Since there is no KDE dependency here, we can't use kDebug which would have the ability to
turn them off.

r1066588 | tmcguire | 2009-12-27 15:16:59 +0100 (Sun, 27 Dec 2009) | 5 lines

Comment out debug output with std::cerr, which is annoying since it
can't be turned off.

If you want debug output, use kDebug.

r1066589 | tmcguire | 2009-12-27 15:17:46 +0100 (Sun, 27 Dec 2009) | 2 lines

Fix warning about unused function.

r1066604 | djarvie | 2009-12-27 15:48:07 +0100 (Sun, 27 Dec 2009) | 2 lines

apidox corrections

r1066611 | djarvie | 2009-12-27 15:53:06 +0100 (Sun, 27 Dec 2009) | 2 lines

Add info supplied by Volker

r1066806 | djarvie | 2009-12-28 00:03:52 +0100 (Mon, 28 Dec 2009) | 2 lines

Change apidox references to MimeTypeFilterProxyModel -> EntityMimeTypeFilterModel.

r1066807 | djarvie | 2009-12-28 00:06:13 +0100 (Mon, 28 Dec 2009) | 2 lines

Revert unintentional change to entityData()

r1066905 | scripty | 2009-12-28 11:29:17 +0100 (Mon, 28 Dec 2009) | 1 line

SVN_SILENT made messages (.desktop file)

r1066936 | cgiboudeaux | 2009-12-28 12:57:27 +0100 (Mon, 28 Dec 2009) | 1 line

SVN_SILENT remove an old comment.

r1066941 | tokoe | 2009-12-28 13:17:57 +0100 (Mon, 28 Dec 2009) | 2 lines

Hide private API

r1066942 | tokoe | 2009-12-28 13:31:51 +0100 (Mon, 28 Dec 2009) | 3 lines

Replace our QSharedDataPointer::detach() hack with the
proper solution that is available since Qt 4.5

r1067280 | tmcguire | 2009-12-29 12:17:33 +0100 (Tue, 29 Dec 2009) | 1 line

Add comment for API change in KDE5.

r1067336 | tmcguire | 2009-12-29 12:59:48 +0100 (Tue, 29 Dec 2009) | 1 line

When no charset is given, use the name of the codec, instead of none.

r1067339 | tmcguire | 2009-12-29 13:00:53 +0100 (Tue, 29 Dec 2009) | 4 lines

Add a test for UTF16 mails.
While UTF16 works fine, this uncovered the bug fixed with the previous commit
(no charset in RFC2047-string) and also a not yet fixed newline bug.

r1067378 | tmcguire | 2009-12-29 13:54:55 +0100 (Tue, 29 Dec 2009) | 3 lines

Fix removing trailing newlines in decodedText() when not trimming other whitespace.
Add a quick test.