I have done a stab at a "Fair" StoreQueueCursor.
It would be great if we could discuss whether something like this is
desireable, and feasible to get into ActiveMQ.
I am not sure of the semantics wrt. the store - whether this could leave a
message stranded in case of powerloss or somesuch. I hope
Hmm..
You asked for code. As you hopefully saw, there was already code in my
initial message - the minor suggested very simple change actually makes the
situation better, by consuming "fairly" between the two buckets.
And here's some more code - my research-repo for this, with a set of tests,
and
Hi Endre-
> On Jun 19, 2023, at 2:49 PM, Endre Stølsvik wrote:
>
> I have explained why. I have suggested a half-way fix, which for most users
> probably would be better than the current situation. I have also suggested
> the gist of a complete fix.
The ActiveMQ v5.18.2 release is in progress,
Hi!
I must clearly have miscommunicatied or undercommunicated the issue here.
Or you have not read my email fully.
This is the gist: There is a bug in ActiveMQ. ActiveMQ's default queue
handling mechanism is completely broken if you mix persistent and
non-persistent messages on the same queue.
P
Hi Endre-
Contributions are always welcome =). Sample unit test scenarios and/or PRs
that demonstrate the issue and an approach to solve for it.
That being said, keep in mind that any system that deals with the big three—
network I/O, disk I/O & CPU (like messaging and databases) always have t
