The patch from QFJ-285 by Oliver seems to solve this issue (I also applied it and successfully ran all the unittests). We'll be using the patched version in a UAT environment and I'll follow up with any observed problems.

In reply to comment #1:
> The patch from QFJ-285 by Oliver seems to solve this issue (I also applied it
> and successfully ran all the unittests). We'll be using the patched version in a
> UAT environment and I'll follow up with any observed problems.

Sounds good. Please provide feedback as soon as you have some.

sorry, late delay... i will rediff on the latest tree and post the patch with all cleanups. the problem is, that i have to split out the proxy hack stuff und the mina stuff. i hope i can work tomorrow in train on the diff, so i can repost the patch tomorrow or on wednesday.

In reply to comment #3:
> sorry, late delay... i will rediff on the latest tree and post the patch with
> all cleanups. the problem is, that i have to split out the proxy hack stuff und
> the mina stuff. i hope i can work tomorrow in train on the diff, so i can repost
> the patch tomorrow or on wednesday.

That's fine, Oliver. I am also short of time due to my work.

latest patch and last one.... please test this patch in your UAT environment. there is only a small diff to the v01 patch. only some System.out and System.err and Log.info are removed. rest ist stable and the ant test was successfull. don't forget to replace the mina-1.1.7 with the mina-2.0.4 or higher libs...

Unfortunately it seems that we will have to wait until at least 1.6.0 to incorporate this change since we need to discontinue the JDK1.4 support of QF/J when we want to use MINA2. MINA2 is only available for JDK5+ and the QFJ code needs to be changed to work with MINA2.

This is an updated version of the previous patch, done against quickfix/j 1.5.3.

I would like to mention that at this point Java 1.4 is ancient history and not even Java 1.6 is publicly supported anymore. Along this line of reasoning you should run java 1.7 because (other than common sense - which is not all that common):

• you get patches for free
• you avoid possibly violating regulations like PCI for example by not running unsupported software

Thanks for the patch. I agree that we should definitely use a recent JDK. I will send a message to the quickfixj-users group.

One more reason to remove 1.5 and lower: the JDK itself will remove support for them: http://openjdk.java.net/jeps/182

Applied patch quickfix-1.5.3-mina2.path and built without issues. Can someone point me how to make a change to quickfix/j and let it pick up proxy settings and establish connection via socks v4 proxy server?

Tried returning a ProxyConnector with ProxyIoSession set from ProtocolFactory.createIoConnector() but got no luck. Can't even pass through jSocks tests. Very frustrated.

Any help is highly appreciated.

Aaron

[email protected]

i did only tests with http proxies. socks i dont know

if mina support it, perhaps passing socks option
to mina would work

but as you can see, the patch is old and never
accepted to mainstream, because jdk discussions.

We have a dependency on both quickfixj & apache qpid - but qpid uses mina 2.0 which clashes with quickfixj's dependency on mina 1.0. So would be really good if quickfixj could move to mina 2.0

Committed as http://sourceforge.net/p/quickfixj/code/1131/ and http://sourceforge.net/p/quickfixj/code/1132/.
Thanks to Oliver Jehle and Attila-Mihaly Balazs for the submitted patches.

With the new MINA libraries some deadlock situations on disconnect are now more prominent than before. This will be handled in QFJ-738.