On Jun 27, 2008, at 00:25, [EMAIL PROTECTED] wrote:
--- Comment #4 from Adrian Cumiskey <[EMAIL PROTECTED]> 2008-06-26
15:25:09 PST ---
Thanks Andreas, that does seem to reproduce the problem! Boy thats
bug/limitation that has been lurking there since the beginning of
253 char text block limit...
The fact that it has remained undetected for so long does say
something about the average length of text-nodes in most general use
Not really new for me, as I have some statistics-gathering
stylesheets that I run from time to time.
The last similar one I remember was the 32K char-limit for
TextLayoutLayoutManager due to the use of shorts as indexes into the
FOText's char array. Also a low priority issue. Who on earth would
need /that/ much characters? The bug-report was killer. When Jeremias
asked to provide a test-file reproducing the issue, the reporter
replied with a FO that simply contained TextLayoutManager.java, put
inside a single fo:block... Great sense of humor! :-)
I haven't looked at the related code here yet, but purely from that
number, I'd guess a byte is the source of the issue. 253 is pretty
strange, though... Something like an array with Byte.MAX_VALUE
elements, but the first and the last are unused (very wild guess?).