> > this patch fixes a problem when posting with more than about > 8K of data. it > seems that when the java side of the ajp connection has to > ask the webserver > (jk) side for more data, it goes into an infinite loop > because the next > chunk of data isn't actually sent. > > i'm 97% sure about this patch, but since i don't know this > code as well as > some, i'd like someone else to review it. > > attached is the patch and a servlet (FileUpload.java) and a client > (Upload2.java) that demonstrate the problem. > > the client uses HTTPClient, which can be downloaded from > http://www.innovation.ch/java/HTTPClient/. incidentally, i > would have used > the http client in jakarta-commons, if i could have figured > out how to post > using an output stream :) > > this same test fails on the java side of ajp using the ajp > stuff on the head > in j-t-c/jk. again, an infinite loop. i haven't been able > to figure it out > yet... if somebody wants extra points ... ;) >
well, i'm pretty sure i figured out this one. RequestHandler.refillReadBuffer() is missing the all important check for the empty packet that signals the end of the stream :) -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
