Error in encoded-words (RFC 2047)

[A. Brandt]

The word encoder does not encode all characters:

Subject: =?UTF-8?Q?"=C3=9Cml=C3=A4ute"?=

should be:

Subject: =?UTF-8?Q?=22=C3=9Cml=C3=A4ute=22?=

The RFC (http://tools.ietf.org/html/rfc2047#section-5) says:

    In this case the set of characters that may be used in a "Q"-encoded
    'encoded-word' is restricted to: <upper and lower case ASCII
    letters, decimal digits, "!", "*", "+", "-", "/", "=", and "_"
    (underscore, ASCII 95.)>.  An 'encoded-word' that appears within a
    'phrase' MUST be separated from any adjacent 'word', 'text' or
    'special' by 'linear-white-space'.

The "-character is not allowed in an 'encoded-word'.

[Vincent Richard]

Hello,

Thank you for your report, I just fixed it.

Here is the patch:
http://www.vmime.org/download/patches/vmime-20101012-rfc2047quot.patch

Regards,

Vincent

[A. Brandt]

Thanks for the fast patch.

But I think there is more to do.:
- The encode() function in qpEncoder.cpp has to encode more special characters like '§$%&'.
- The "outputCount" in function getNextChunk() in wordEncode.cpp is wrong for this special characters.

Best Regards,
Achim

[Vincent Richard]

You are right.

I have rewritten RFC-2047 encoding, here is a better patch:
http://www.vmime.org/download/patches/vmime-20101012-rfc2047better.patch

Please let me know whether it works for you.

Regards,

Vincent

[A. Brandt]

Thanks for the fast and better patch.

Best Regards,
Achim

[A. Brandt]

There is an other text encoder problem;

    vmime::messageBuilder mb;
    mb.setExpeditor(vmime::mailbox(vmime::text("\"Achim Brändt\"", vmime::charsets::UTF_8), "me@vmime.org"));

encodes to:

From: =?us-ascii?Q?=22Achim?= =?utf-8?Q?Br=C3=A4ndt=22?= <me@vmime.org>

and this decodes to:

"AchimBrändt" me@vmime.org

I think it should encode to ("_" for space)
From: =?us-ascii?Q?=22Achim?= =?utf-8?Q?_Br=C3=A4ndt=22?= <me@vmime.org>
or
From: =?utf-8?Q?=22Achim_Br=C3=A4ndt=22?= <me@vmime.org>

[Vincent Richard]

Hello,

There was a bug in text::createFromString(): whitespaces were simply used as word separator, then ignored. Now, whitespace is added to the previously parsed word.

Try with this patch:
http://www.vmime.org/download/patches/vmime-20101012-word-whitespace.patch

Vincent

[A. Brandt]

Thanks,
Encoding works for me now.

Best Regards
Achim

[Esteban Lombeyda]

Hello,

I am trying to generate mails with vmime. Using the function
vmime::text::newFromString for assigning values with no ascii
characters to headerField, I have observed that the new strings contain
additional white-spaces.
Sample: Input String 'ü e)' . The Output of vmime::text::newFromString
is 'ü  e)'. It contains an additional white space. I could reproduce
this behaviour without  vmime::text::newFromString too, with following code

    vmime::string gen_str;
    vmime::utility::outputStreamStringAdapter os1(gen_str);
    vmime::text gen_text("ü e)");
    gen_text.generate(os1);
    std::cout << "["<< gen_str << "]"<< std::endl;

I get the string (value of gen_str):

=?UTF-8?B?w7wg?= e)

In my opinion there is an additional white space (which is too much)

Using the method decodeAndUnfold one gets the converted string

ü  e)

Resume:

Input string

'ü e)'

Converted string

'ü  e)'

This happens  in following constellations:
"non ascii character" white-space ascii-character !"$%&/()=?,;.:<>|@

I would appreciate, if you could have a look into this issue

Kind regards
Esteban Lombeyda