Well it felt like it took me forever to get this all put together but I finally have a release candidate for ActiveMQ-CPP 3.0. The first candidate RC0 can be downloaded from the Apache ActiveMQ-CPP website here.
I've tested the release on Fedora 10 and Windows XP both in 32bit and 64bit compiles. Now I need some users to grab it and start testing it out to find the bugs I didn't.
Monday, April 20, 2009
Thursday, April 16, 2009
ActiveMQ-CPP Release Candidte
I've been trying all week to get the CPP client to build cleanly on Windows but still haven't been able to resolve all the DLL build issues, and don't even get me started on 64bit builds. Why is building on Windows always such a pain in the arse?
In the meantime I've resolved a couple of issues that were submitted last week that fix some pretty important things. We had previously written some code to hand sending the C++ strings to the Java broker in the native Modified UTF-8 format that Java uses but we made a mistake and didn't properly encode NULLs or any ASCII character over 127, well that now all fixed. As a bonus I realized that our code was also not taking advantage of the fact that the Openwire processing on the Java side allows for strings to be encoded with a size larger than 65535 bytes for string values in the Message properties not just in the TextMessage payload, so the long standing limit on string length in Message properties is now gone as well.
One other small fix adds the ability to set the max inactivity duration value on the URI and to properly negotiate it with the broker.
So all in all not a bad week as far as improvements to the client goes, now if only this Windows build would finish...
In the meantime I've resolved a couple of issues that were submitted last week that fix some pretty important things. We had previously written some code to hand sending the C++ strings to the Java broker in the native Modified UTF-8 format that Java uses but we made a mistake and didn't properly encode NULLs or any ASCII character over 127, well that now all fixed. As a bonus I realized that our code was also not taking advantage of the fact that the Openwire processing on the Java side allows for strings to be encoded with a size larger than 65535 bytes for string values in the Message properties not just in the TextMessage payload, so the long standing limit on string length in Message properties is now gone as well.
One other small fix adds the ability to set the max inactivity duration value on the URI and to properly negotiate it with the broker.
So all in all not a bad week as far as improvements to the client goes, now if only this Windows build would finish...
Thursday, April 2, 2009
ActiveMQ-CPP 3.0 Right around the corner
There's been a lot of work going on to revamp ActiveMQ-CPP over the past couple of months. The major new feature of course if Failver support, but besides that there's also been a lot of work done to make the code faster and more reliable. The end is in sight, Failover is working now, and most of the open issues are closed. There are a few small tasks remaining as well as a lot of testing to do but the code is now at a point where folks can start using it. I just finished making several fixes to the Transactions handling code and also cleaned up the way Message's are ack'd. If all goes well in my testing I'm hoping to have a release candidate out in another week or so.
Subscribe to:
Posts (Atom)