Overall, the thing which upset me the most is that the software crashes. C++ library.
Would it be because I am running so many bots? Is there a limit to the maximum number to run simultaneously?
It has nothing to do with number of bots.
I do see it more often when Polo is fucking up, though. I think the bot isn't handling something properly when Polo is wonky. I don't know if it messes up on other exchanges.
It did this error for me on Bittrex a number of times, up to the point that I couldn't open the Bittrex market at all. It was clearly a weird API response from Bittrex (I could see in the logs that it wasn't JSON at all).
So it seems to me that some improvement of error handling would be required here. Normally, leonArdo should be able to continue past the unintelligible API response, and not simply crash wioth that C++ Library error instead.
We agree. Seg fault crashes should not happen.
If you can re-produce one - we'd love to hear how you can do it.
If you can produce a crash report please send it on.
Any specific information helps us to improve leonArdo. This is complex stuff, but we are working hard on it.
Thanks for all your support guys!
I am under the impression that I wrote it a few times already: Make orders, then cancel before the order is placed / confirmed on the exchange, or move an existing order then cancel it before it has finished to replace it. That's a near-guaranteed crash, at least in my experience.
If you want to test and reproduce that, just place many bots (with 0.001 BTC value, or something like twice the minimal value), and have fun moving them around in an aggressive way, like if you want to push leonArdo to the limit. You'll see it crash quite easily, I believe.
Not a day passes by without leonArdo crashing.
As for a "crash report", would any of the logs be helpful for you? I've offered by email to send you that a few times, but I didn't get a clear reply on it.
Also, I believe it would be important that you setup PGP keys, so that we can communicate securely with you by email.