DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG 
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://nagoya.apache.org/bugzilla/show_bug.cgi?id=4546>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND 
INSERTED IN THE BUG DATABASE.

http://nagoya.apache.org/bugzilla/show_bug.cgi?id=4546

normalize-space does not properly handle text data





------- Additional Comments From [EMAIL PROTECTED]  2002-04-22 19:41 -------
Fix checked in.

Chunk size range is set when the FastStringBuffer is created. Each FSB instance 
may set this differently, depending on which constructor was called and what 
parameters were supplied. Generally, the DTM implementation is responsible for 
setting these allocation parameters. SAX2DTM sets the initial chunk size at 10 
bits, which comes out to 1024 characters per chunk. Other DTMs, such as the SQL 
extension, may be using other values.

Due to some issues that arose in our support for variable chunk sizes -- 
specifically, the fact that we weren't responding correctly when appending 
content from a FastStringBuffer back into that same FSB -- we are currently 
forcing the size to a specific value. The code for this workaround is enabled by 
DEBUG_FORCE_FIXED_CHUNKSIZE. I suppose we should make it permanant at some 
point; the varying size was a nice idea, but if we aren't using it we probably 
shouldn't spend cycles on it.

(We should also consider whether we want to update SAX2DTM to increase the 
minimum FSB chunk size, now that we've got the shared DTM code running for 
Result Tree Fragments and hence will tend to use these buffers more 
efficiently.)

Reply via email to