Arena moveerror?

[Gary]

Hey guys!

I recently implemented pondering in my engine.

Arena seems to think that my engine is outputting illegal pv lines on ponder. However, the line it has marked seems legal to me. Here is a log excerpt:

Code: Select all

2015-03-19 21:49:41.744-->1:position startpos moves e2e3 e7e5 d2d4
2015-03-19 21:49:41.807-->1:go ponder wtime 126000 btime 123755 winc 6000 binc 6000
2015-03-19 21&#58;49&#58;41.807<--1&#58;info score cp 4 depth 1 pv d7d6 
2015-03-19 21&#58;49&#58;41.807<--1&#58;info score cp 3 depth 2 pv d7d5 d4e5 
2015-03-19 21&#58;49&#58;41.807<--1&#58;info score cp 4 depth 3 pv d7d5 d4e5 c8f5 
2015-03-19 21&#58;49&#58;41.823<--1&#58;info score cp -7 depth 4 pv d7d5 d4e5 c8f5 f2f3 
2015-03-19 21&#58;49&#58;41.885<--1&#58;info score cp -6 depth 5 pv e5d4 e3d4 d7d5 f1b5 c7c6 
2015-03-19 21&#58;49&#58;41.885*moveerror* start piece<0 and target piece<0 !e3xd4!

That doesn't make any sense to me since e2e3 e7e5 d2d4 + e5d4 e3d4 is valid.

Any idea on what Arena is trying to tell me?

Thanks for any help,

Gary

[Daniel Anulliero]

May be it try to re play e3xd4 after c7c6 (in pv) but after c6 the e3 square is empty ?

[cdani]

I think is a bug on Arena. I see the same behavior on 3.5, but not on Arena 3. I saw another post somewhere talking about the same. Can someone confirm this?
Thanks.

[Joost Buijs]

I think is a bug on Arena. I see the same behavior on 3.5, but not on Arena 3. I saw another post somewhere talking about the same. Can someone confirm this?
Thanks.

I also had this problem with Arena on several occasions when running test-suites with my engine. My engine is perfectly reliable with every other GUI. Arena seems to have some annoying bugs.
My guess is that sometimes the communication between the engine and Arena gets messed up in some way.

[Gary]

Yeah I thought it might be an Arena bug.

Thanks for the confirmations.