The Trunk: Collections-pre.747.mcz

Patrick Rein uploaded a new version of Collections to project The Trunk:
http://source.squeak.org/trunk/Collections-pre.747.mcz

==================== Summary ====================

Name: Collections-pre.747
Author: pre
Time: 20 April 2017, 11:37:10.168994 am
UUID: 6fe02314-e789-064a-b678-0caec3506aac
Ancestors: Collections-eem.746

changes decoding of mime headers to decode directly to squeak encoding

=============== Diff against Collections-eem.746 ===============

Item was changed:
  ----- Method: String>>decodeMimeHeader (in category 'internet') -----
  decodeMimeHeader
  "See RFC 2047, MIME Part Three: Message Header Extension for Non-ASCII  
+ Text and RFC 1342. Text containing non-ASCII characters is encoded by the sequence  
- Text. Text containing non-ASCII characters is encoded by the sequence  
  =?character-set?encoding?encoded-text?=  
  Encoding is Q (quoted printable) or B (Base64), handled by  
  Base64MimeConverter / RFC2047MimeConverter.
 
  Thanks to Yokokawa-san, it works in m17n package.  Try the following:
 
  '=?ISO-2022-JP?B?U1dJS0lQT1AvGyRCPUJDKyVpJXMlQRsoQi8=?= =?ISO-2022-JP?B?GyRCJVElRiUjJSobKEIoUGF0aW8p?=' decodeMimeHeader.
  "
  | input output temp charset decoder encodedStream encoding pos |
  input := ReadStream on: self.
  output := WriteStream on: String new.
  [output
  nextPutAll: (input upTo: $=).
  "ASCII Text"
  input atEnd]
  whileFalse: [(temp := input next) = $?
  ifTrue: [charset := input upTo: $?.
  encoding := (input upTo: $?) asUppercase.
  temp := input upTo: $?.
  input next.
  "Skip final ="
  (charset isNil or: [charset size = 0]) ifTrue: [charset := 'LATIN-1'].
+ encodedStream := WriteStream on: String new.
- encodedStream := MultiByteBinaryOrTextStream on: String new encoding: charset.
  decoder := encoding = 'B'
  ifTrue: [Base64MimeConverter new]
  ifFalse: [RFC2047MimeConverter new].
  decoder
  mimeStream: (ReadStream on: temp);
  dataStream: encodedStream;
  mimeDecode.
+
+ output nextPutAll: (MultiByteBinaryOrTextStream with: encodedStream contents encoding: charset) contents.
- output nextPutAll: encodedStream reset contents.
  pos := input position.
  input skipSeparators.
  "Delete spaces if followed by ="
  input peek = $=
  ifFalse: [input position: pos]]
  ifFalse: [output nextPut: $=;
  nextPut: temp]].
  ^ output contents!

On Thu, 20 Apr 2017, [hidden email] wrote:

Creating a WriteStream on an empty string is just a waste, since it'll
have to allocate another string on the first write. #on: will always
create an empty stream, the argument will be the initial buffer.

The above conversion should be done without creating a
MultiByteBinaryOrTextStream.

Levente

P.S.: The whole method could use some refactoring.

> - output nextPutAll: encodedStream reset contents.
>   pos := input position.
>   input skipSeparators.
>   "Delete spaces if followed by ="
>   input peek = $=
>   ifFalse: [input position: pos]]
>   ifFalse: [output nextPut: $=;
>   nextPut: temp]].
>   ^ output contents!