Tomcat 3.3 Release Date?
Can anyone shed any light as to when Tomcat 3.3 will actually be released (i.e. as a non-beta release)? What are the open/known issues? [I checked CVS, dev-mailing list, and Bugzilla to no avail.] -- Jess Holle
RE: SAX 2.0, sealing, Tomcat 3.2.3
A better question is why Tomcat 3.3 which has the critical classloader separation feature as well as performance improvements over 3.2.1 has not yet been released. As the previous message said, many of us have to use released software *period*. Jess Holle -Original Message- From: Craig R. McClanahan [mailto:[EMAIL PROTECTED]] Sent: Tuesday, July 31, 2001 4:35 PM To: '[EMAIL PROTECTED]' Subject: RE: SAX 2.0, sealing, Tomcat 3.2.3 On Tue, 31 Jul 2001, Bryan Rood wrote: Pier, I am interested in using a 4.0.something tomcat build. Are there official sources at apache that can confirm the VERY_stable declaration that you have made? My company won't let me use anything but a stable production release. Is the 4.0 the same as prod quality? thanks so much, I'm one of the primary authors of Tomcat 4.0, so I've got some thoughts about this (and undoubtedly some bias as well :-). The only reason that 4.0 has not been declared final yet is that the underlying specifications it is based on (Servlet 2.3 and JSP 1.2) are not yet final. In fact, small changes and clarifications are still going on, and it would be pretty silly to declare 4.0 final and then have to go change it because the specs changed underneath. There will shortly be a beta 7 release, to pick up the most recent specification-related changes. It should be considered a release candidate, and development efforts between now and release day will be focused on bug fixes (at the moment, there are very few bugs recorded against Tomcat 4 in the bug tracking system at http://nagoya.apache.org/bugzilla/) and improved documentation. Craig McClanahan Bryan -Original Message- From: Pier P. Fumagalli [mailto:[EMAIL PROTECTED]] Sent: Tuesday, July 31, 2001 9:33 AM To: [EMAIL PROTECTED] Subject: Re: SAX 2.0, sealing, Tomcat 3.2.3 Andrew Cooke at [EMAIL PROTECTED] wrote: I don't want to use 4.0 as it's beta. Don't worry about being a beta... It's stable, _VERY_ stable. Pier
RE: Beyond Tomcat 4..?
I personally second call for use of the new I/O calls and performance efforts in general. As for IIS -- until one can easily give per servlet and per-JSP authentication settings (e.g. anonymous for servlet A, authenticated via LDAP for servlet B, authenticated via password file for servlet C, anonymous for /anon/*.jsp, but authenticated for /auth/*.jsp) with the ease one can in Apache I can only say no thanks. Win2K is well and good, but I'd sooner see Apache 2.0 finished up then further work on IIS -- unless IIS is suddenly and *dramatically* improved. [For instance IIS should have a purely text-based configuration capability, and that doesn't mean 'regedit', like Apache to simplify automated configuration.] Jess Holle -Original Message- From: Curtis Dougherty [mailto:[EMAIL PROTECTED]] Sent: Wednesday, August 01, 2001 1:31 PM To: '[EMAIL PROTECTED]'; '[EMAIL PROTECTED]' Subject: RE: Beyond Tomcat 4..? YES... integration with IIS5 and Win2K... Documentation for same... :P -Original Message- From: Craig R. McClanahan [mailto:[EMAIL PROTECTED]] Sent: Wednesday, August 01, 2001 1:18 PM To: [EMAIL PROTECTED] Subject: Re: Beyond Tomcat 4..? On Wed, 1 Aug 2001, chris brown wrote: Just wondering what's being planned for the future of Tomcat beyond the 4.0 release (if anyone knows...). For example, are there plans to make a release that uses JDK 1.4's scalable I/O features? (as both products are currently in beta, I imagine that could be some way off!) Experimenting with the new I/O calls is certainly something I'm interested in exploring (along with performance tuning in general). But the reality is that what actually gets added post-4.0 is based on what features people suggest, combined with what features people actually write code for. Are there particular things you're interested in seeing? Craig McClanahan - Original Message - From: Craig R. McClanahan [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, July 31, 2001 11:35 PM Subject: RE: SAX 2.0, sealing, Tomcat 3.2.3 On Tue, 31 Jul 2001, Bryan Rood wrote: Pier, I am interested in using a 4.0.something tomcat build. Are there official sources at apache that can confirm the VERY_stable declaration that you have made? My company won't let me use anything but a stable production release. Is the 4.0 the same as prod quality? thanks so much, I'm one of the primary authors of Tomcat 4.0, so I've got some thoughts about this (and undoubtedly some bias as well :-). The only reason that 4.0 has not been declared final yet is that the underlying specifications it is based on (Servlet 2.3 and JSP 1.2) are not yet final. In fact, small changes and clarifications are still going on, and it would be pretty silly to declare 4.0 final and then have to go change it because the specs changed underneath. There will shortly be a beta 7 release, to pick up the most recent specification-related changes. It should be considered a release candidate, and development efforts between now and release day will be focused on bug fixes (at the moment, there are very few bugs recorded against Tomcat 4 in the bug tracking system at http://nagoya.apache.org/bugzilla/) and improved documentation. Craig McClanahan Bryan -Original Message- From: Pier P. Fumagalli [mailto:[EMAIL PROTECTED]] Sent: Tuesday, July 31, 2001 9:33 AM To: [EMAIL PROTECTED] Subject: Re: SAX 2.0, sealing, Tomcat 3.2.3 Andrew Cooke at [EMAIL PROTECTED] wrote: I don't want to use 4.0 as it's beta. Don't worry about being a beta... It's stable, _VERY_ stable. Pier
ETA for 3.3???
What is the expected time of arrival / release of Tomcat 3.3??? It's been at the same milestone seemingly for aeons while Tomcat 4.0 continues to show more obvious (to those not doing lots of CVS snooping) signs of progress. Problem is we need 3.3 ASAP whereas 4.0 is "out there". -- Jess Holle
RE: Xalan2 in server environment (was Xalan2 Stree Module spans sec ondthread to do transform?)
Thanks, Scott.
I doubt it does the trick generally for EJB, but it does the trick for our
app server :-)
EJB is unkind to threaded algorithms to the point that I see this as a major
hole in EJB, though I suppose you could hide this in a Connector and
potentially skirt some of the restrictions
Still, there *are* algorithms which should be threaded and EJB shouldn't
make it a supreme pain to do this!
--
Jess Holle
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, January 09, 2001 2:51 PM
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Subject: Re: Xalan2 in server environment (was Xalan2 Stree Module spans
sec ondthread to do transform?)
In response to this I've created the following method in
org.apache.xalan.TransformerImpl:
/**
* Create a thread for the transform. This can be overridden by derived
* implementations to provide their own thread, for thread pooling and
the
* like.
*
* @return thread suitable to use for the transformation.
*/
public Thread createTransformThread()
{
return new Thread(this);
}
Someone who is using XalanJ2 in an EJB environment or the like can set the
javax.xml.transform.TransformerFactory system property to provide a
derivation from TransformerImpl, and provide a thread suitable for the
environment.
I don't know if this will do the trick. Please advise.
-scott
"Roytman,
Alex" To:
"'[EMAIL PROTECTED]'" [EMAIL PROTECTED]
roytmana@peaccc:
"'[EMAIL PROTECTED]'"
etech.com[EMAIL PROTECTED],
(bcc: Scott Boag/CAM/Lotus)
Subject: Xalan2 in server
environment (was Xalan2 Stree Module spans
01/07/2001sec ond thread to do transform?)
12:14 AM
Please respond
to xalan-dev
Scott,
Thank you very much for your reply.
One part which is sensitive to thread origin in my opinion is Extensions
In extensions people can do all sorts of things. I am not very experienced
with EJB implementation but I can imagine that many things (transactional
context, bean environment, etc.) depend on threads and some static
variables
which helps with setting /switching context for beans being managed by the
server.
Example: in J2EE it is recommended to create JNDI initial context using
default constructor InitialContext(). The result of the instantiation
depends on the context where it was executed and the context is set by the
server for the thread on which your component is running. So if you call
new
InitialContext() in your extension (in sql extension to get JDBC DataSource
for example ) it might fail.
Do extension run on the second (created by Xalan) thread?
I will forward your message to Tomcat news group lets see what Tomcat
developers think.
Alex
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Saturday, January 06, 2001 11:39 PM
To: [EMAIL PROTECTED]
Subject: Re: Xalan2 Stree Module spans second thread to do transform?
In many
cases component should not attempt to create its own threads.
For example thread's ContextClassLoader or ThreadLocal variables might
need
to be initialized by the server.
Also it defeats thread pooling done by the server etc.
Could you please comment on this issue
Yeah, I've been a bit worried about this. However, I've not yet heard of
any problems that have been caused by it, and XalanJ1 has long had a two
threaded system (though not as effective as in XalanJ2). Ultimately, I
would rather use a pull model for this, and only have one thread, but a)
there is no standard "pull" API for XML parsers, and b) this doesn't work
anyway when SAX events are used, for whatever reason.
I'm open to any input to how Xalan might request the thread from the
servlet environment, though it has to be able to be run outside a servlet
environment too. Also, it would be good if someone with deep knowledge of
EJB's and the like could comment. I talked this over in a hallway
conversation with someone who is fairly familiar with EJB's, and he didn't
think there was a problem, though I forget why. It seems insane/crazy to
me that a component can't use a thread in it's internal modules.
-scott
"Roytman,
Alex" To:
"'[EMAIL PROTECTED]'" [EMAIL PROTECTED]
roytmana@peaccc: (bcc: Scott
Boag/CAM/Lotus)
etech.comSubject: Xalan2 Stree Module
spans second thread to do transform?
01/06/2001
10:48 PM
Please respond
to xalan-dev
Dear Xalan developers,
I have a question about using upcoming Xalan2 in server env.
In Xalan2 design specs it is said:
"The Stree module implements the default Source Tree
