Thx, the update stop button is back
Regarding BrokerAsset: so when the data for the asset is available, but the asset itself is currently not tradeable, then the plugin should return 1 anyway?
I think I will add additional checks for tradeability for all order related commands(BrokerBuy, BrokerStop, BrokerSell).
This is beneficial in giving back a more detailed failure description.

While implementing the BrokerAsset function, I can see that only the price and spread pointers are valid.
I wonder under which conditions the other pointer outputs are ready to be written.
I opened a position, but still the pointers are not used by Zorro.
Are they obsolete or do I need to enable something?

I don't have my IDE open, so let me ask you:

1) Which pointers are valid during the asset subscription phase?
2) Which pointers are valid during the live trading phase?

I have not checked it during the subscription phase, but I know for sure that the price pointer is NULL as I can subscribe the assets.
After subscription is done, I see that only price and spread pointers are valid.
All others are NULL.
After opening a trade, the situation does not change.
Maybe everything is OK, since I check for valid pointers anyway.
I'm just curious under which conditions the other pointers(besides price and spread) become valid.
It also makes sense to test these values when developing a plugin, so I need a way to tell Zorro to use these pointers.

Or it could be that the TradeTest.c script is the wrong tool for that.

Ok, now I see in the log that the other pointers are valid only sometimes.
When looking at timestamps, I would guess that Zorro queries the other values only once per minute.
Now I can see that all values are set correct and everything works

Next questions arise for BrokerBuy2:
The manual says for the return value 0 that:
when the market or limit order was not filled within 30 seconds. Not or partially filled orders must be cancelled.
I'm not sure what that means.
Does Zorro support partially filled orders at all?
If you look at the output parameter pFill, then it seems that partial fills are allowed.
On the other hand, when the order was not filled within 30sec(does this mean fully filled?), partially filled ones must be closed.
So, partial fills are not allowed?!
The only scenario where pFill makes sense to me, is when after 30sec the order is still only partially filled.

This depends. Does the broker API use a high-level "trades" abstraction or does it only process "orders"?

It's Dukascopy and I think they work only with orders.
But "that depends" should be clearly stated, such that a Broker plugin is fully compliant to the specification.
BrokerBuy2 is in general very unclear to me.
For example, if there is no connection to the broker for 30sec then there is also no order fill.
Should the plugin then return 0 or -2, because both conditions would be satisfied?
And what about the NFA closing way, when the order was rejected.
Then 0 and 1 would be candidates for the result.

In general it would be helpful if the documentation is very strict here, i.e. a boolean matrix with possible return codes, like this:

Order fully filled | Order partially filled(30sec) | Order not filled(30s) | No BrokerConnecition(30s) | NFA-Account =
1 0 0 0 0 Expected Result: TradeId
1 0 0 0 1 Expected Result: 1
0 1 0 0 0 Expected Result: ????
0 1 0 0 1 Expected Result: ????(maybe 1)
0 0 1 0 0 Expected Result: 0
0 0 1 0 1 Expected Result: ????(maybe 1)
0 0 1 1 0 Expected Result: 0 or -2
0 0 1 1 1 Expected Result: 0,-2 or 1

I agree with you, that the API is not very well defined. For example, see the link below:

https://opserver.de/ubb7/ubbthreads.php?ubb=showflat&Number=473867#Post473867

But in case with Dukascopy you can port your strategy to Java and execute it directly inside the JForex environment.