Hi to everybody!
I have just switched to JServ now that there is a stable release
and I got it running on a RedHat 6.0, JDK1.2, JSDK2.0, Apache 1.3.6
configuration.
When I tried the ApacheJSSI extension, I trapped obviously a bug when
using the init argument buffered=yes
Its somewhere arround the following part:
if (pipeStream != null) {
// close the stream to force the compression algorithm
// to finish.
pipeStream.close();
//
// The size of generated pages is dynamic and thus, subject of
// change. So calculate the heuristic size as a weighted
// mean value. Weights are 1/2 + 1/4 + 1/8 + ...
if (compressLargerThan != null) { // workaround conditional
Long oldSize = (Long) heuristicSize.get(file);
if (oldSize != null) {
heuristicSize.put (file,
new Long ((pipeStream.size() +
oldSize.longValue()) / 2));
}
else heuristicSize.put (file, new Long(pipeStream.size()));
} //workaround conditional end
}
I included a workaround, but well I skimmed through the code,
and the problem is obviously the Hashtable heuristicSize which is
only initialized if the init Argument compresslarge is set.
Maybe the two cases (compresslarge set or not set) could be seperated
better...?
Furthermore the profiling should maybe really be commented for the
release so that it is not carried out (or only with a switch):
//long StartTime = System.currentTimeMillis();
^^
.....
.....
.....
/*
* to be removed for release
*/
//long duration = System.currentTimeMillis() - StartTime;
^^
//log ("processing '" + file + "' took " + duration + "ms");
^^
Ah I almost forgot. I am talking about the latest stable release
that I retrieved today from the java.apache.org website (1.1).
Regards
Wimpi
---
[EMAIL PROTECTED]
PS: Thanks for your great work!
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Problems?: [EMAIL PROTECTED]
