David, Thanks for the reply - that helps. I'm still curious to know if this should be configurable. Let's say an application server changes its behavior and no one remembers to change the Java code. Wouldn't it help if those settings were in a configuration file?
-Adrian --- On Sun, 6/6/10, David E Jones <[email protected]> wrote: > From: David E Jones <[email protected]> > Subject: Re: Question About UtilJ2eeCompat.java > To: [email protected] > Date: Sunday, June 6, 2010, 1:27 AM > > The simple answer is that in the past not all app servers > supported repeated writes to one or the other of these two > options. In other words, you could write once to either one, > but if you wanted to "stream" stuff through the writer or > the stream you had to use the one the app server supported, > and there was nothing in the spec that said which one that > should be... hence the need for a compatibility layer to > handle the differences in different servlet containers. > > -David > > > On Jun 5, 2010, at 10:44 PM, Adrian Crum wrote: > > > I might have found the answer to this. Some > application servers perform better with an output stream > instead of a print writer (or the other way around). If that > is the reason for this code, shouldn't that be something > that is configurable? Right now the choices are hard-coded > in Java. Why not allow OFBiz users choose the settings most > appropriate for their installation? > > > > -Adrian > > > > --- On Sat, 6/5/10, Adrian Crum <[email protected]> > wrote: > > > >> From: Adrian Crum <[email protected]> > >> Subject: Question About UtilJ2eeCompat.java > >> To: [email protected] > >> Date: Saturday, June 5, 2010, 12:11 PM > >> I noticed that the screen renderers > >> call the UtilJ2eeCompat.useOutputStreamNotWriter > method to > >> determine whether to use the response's print > writer or > >> output stream. Why is that check necessary? From > my > >> understanding of the javax.servlet spec, you use a > writer to > >> send text and you use the output stream to send > binary > >> data. > >> > >> -Adrian > >> > >> > >> > >> > >> > > > > > > > >
