En passant

[cyanfish]

Unless I am mistaken, I couldn't capture en passant.
http://chesstempo.com/chess-problems/37964
I can't imagine that this would be a big deal, but nonetheless... Not that I'm complaining, of course - it probably saved me a few rating points.


P.S. I just joined a couple days ago, and this site is great!

[richard]

Hi Cyanfish,

Thanks for reporting that, it does appear the client side validity checker doesn't see that as a valid move. I just had a quick check of problems that do require en-passant, and en-passant moves that are part of the solution do appear to be accepted, so the worst thing that can happen here is you get saved from losing some points :-)  I'll have a look at the validity checker and see if I can fix this (it is a fairly low impact issue for the tactics viewer but might have more severe impact on the PGN Viewer which uses the same validity checking code during PGN parsing).

Glad you're enjoying the site.

Regards,
Richard.

[revenant]

Hmm, problem 8298 does the same thing.  In blitz mode, I tried 1... hxg3 but the interface threw my pawn back to g4 and let me try again, same effect as if I had made an illegal move when in check (which happens a fair amount in blitz).

An unexpected blessing, as it was the wrong answer.  Kinda like in real life when your chess teacher shakes his head, reaches out and undoes your bad move.  :-)

White has just played 1. h4 as the first move and we have a black pawn on g4, so en passant capture should be legal in that position.

Is it Chess Tempo, or is it maybe my browser?  (Firefox on SUSE Linux.)

[drahacikfm]

I have solved a problem by taking en passant: 36898.  But the en passant was on move 2 of the solution, not on move 1. 

[richard]

The bug only impacts situations where you try and play en-passants that are not the best moves.  The UI always accepts en-passants that are the 'best move'.  What happens is that first the UI checks if a move is the best move, if it is not then the UI makes sure it was a valid move and if it was not valid no further processing occurs and the piece is put back in it's place.  The move validity tester is returning invalid for at least some en-passant attempts. I've not looked further into this yet as the impact is low at the moment (people don't seem to mind so much when they play a bad move and don't get punished :-)  ).

I've added 8298 to the tests cases for this issue.

Regards,
Richard.

[mindbreaker]

Another one for testing: 22043...not that I am happy about my blunder;)

[richard]

Thanks mindbreaker,

Added to list.

Richard.