Fix GH-20674 mb_decode_mimeheader does not handle separator

[youkidearitai]

Closes: #20674

In RFC2047

An 'encoded-word' may not be more than 75 characters long, including
'charset', 'encoding', 'encoded-text', and delimiters. If it is
desirable to encode more text than will fit in an 'encoded-word' of
75 characters, multiple 'encoded-word's (separated by CRLF SPACE) may
be used.

?= =? is skipped if long term, so skip space character.

cc: @alexdowad

[alexdowad]

@youkidearitai, thanks very much for working on this issue!

My understanding of RFC 2047 was that when performing encoding of a non-ASCII string into "encoded words", the encoder should put CR + LF + space (in that order) between each encoded word.

However, the reporter of GH 20674 says that iconv decodes MIME encoded words in the way which he is requesting.

I also just checked what Python's email.header library does, and it also does the same thing which is requested in GH 20674.

[alexdowad]

Looks great. Many thanks to @youkidearitai

[youkidearitai]

I found a bug, so fix it.

var_dump(mb_decode_mimeheader("(=?ISO-8859-1?Q?a?=  =?ISO-8859-1?Q?b?=)"));

Last result is (a b ), RFC 2047 example is (a b).

[alexdowad]

I found a bug, so fix it.

var_dump(mb_decode_mimeheader("(=?ISO-8859-1?Q?a?=  =?ISO-8859-1?Q?b?=)"));

Last result is (a b ), RFC 2047 example is (a b).

Awesome! Thank you!!

I noticed that RFC 2047 contains 7 test cases in section 8, and you brought 6 of those test cases into the PHP test suite... could we use all 7 of them?

[youkidearitai]

@alexdowad Thank you! I got it. I deleted duplicate test cases.

[alexdowad]

😅 Sorry that my comment maybe wasn't very clear.
I was suggesting that since RFC 2047 has provided 7 "official" test cases, it would be good to use all of those official test cases in the PHP test suite.
（RFCがせっかくに七つのテストケースを提供してくれたので、それを全部活用した方が良いではないかという提案でした。でも、浜田さんがここまで頑張ってくださったので、今のままにPRをマージしたいと思われたら、僕が別個のPRで追加させていただいても構いません。🙇🏻‍♂️）

[youkidearitai]

Thank you. I see. Added😊.

(ありがとうございます、承知しました。追加しなおしますね😊)