Schrievkrom

Report4PDF and PDFTalkLibrary are the base of all reports I create within Gemstone/S in all our products/projects.

But I noticed problems in Report4PDF when using specific Unicode oriented strings – then R4PString crashes.

The problem comes up, if you have an Unicode string and it somehow gets an #asString (so it tries to convert itself) and the result gets its way into Report4PDF.

Gemstone/S tries to convert it into a single-byte string (if that works for all characters in the Unicode string) … if not possible, then you get a DoubleByteString instance.

And with instances of DoubleByteString Report4PDF does not work.

The problem is in the method 
R4PString instance 
calculateNextLineAt: aPosition lineEnd: aLineEndIndex

| ... |

originalString := self string.
breakIndex := originalString
		indexOfLastByte: Character space codePoint
		startingAt: aLineEndIndex.
...

The problem here is the method indexOfLastByte:*, which is only defined for one-byte strings.

So, you should define a similiar method in the MultiByteString hierarchy … and then the stuff is working.

I defined a new method

breakIndex := 
  originalString
    mskIndexOfLastCharacter: Character space codePoint
    startingAt: aLineEndIndex.

And here is the implementation for located in MultiByteString class:

mskIndexOfLastCharacter: codepoint startingAt: aLineEndIndex
  | lastPosition  |

  lastPosition := 0.
  self isEmpty ifTrue:[ ^lastPosition ].
  [ lastPosition < aLineEndIndex ]	whileTrue:[ 
    | tmp |
    "nichts gefunden - letzte Position zurückgeben"
    (tmp := self 
              indexOf: codepoint 
              startingAt: lastPosition + 1) = 0 
	ifTrue:[ ^lastPosition ].
    tmp > aLineEndIndex ifTrue:[ ^lastPosition ].
    (tmp > 0 and:[ tmp <= aLineEndIndex ]) 
      ifTrue:[ lastPosition := tmp ]
   ].
   ^lastPosition