[bug #27666] "misc shogi" on FICS

[Sylvain Beucler-4]

URL:
  <http://savannah.gnu.org/bugs/?27666>

                 Summary: "misc shogi" on FICS
                 Project: XBoard
            Submitted by: None
            Submitted on: Sun 11 Oct 2009 11:54:32 AM UTC
                Category: None
                Severity: 3 - Normal
              Item Group: None
                  Status: None
                 Privacy: Public
             Assigned to: None
             Open/Closed: Open
         Discussion Lock: Any

    _______________________________________________________

Details:

Not that I will ever need this, but:

Logon to FICS and type:

ex b misc shogi

Wrong board appears (compared to xboard 4.2.7).
Move a white piece (eg. a3-a4) -> Segmentation fault.

A similar thing happens, when you try to examine a gothic game on "the
Variant Chess Server" (playing the game works, though).




    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Follow-up Comment #1, bug #27666 (project xboard):

Well, one thing is sure: this is not Shogi. Shogi is played on a 9x9 board.

There is no way to prevent ICS operators picking a name for their boards that
is identical to the name of an XBoard variant, while it is in fact a board for
an unrelated variant. It can run a suicide game under the name crazyhouse, or
call nocastle frc, etc. This is a good way to wreck proper operation of the
client. In general this should be considered an error of the ICS.

In this particular case it could be solved by beefing up the
variant-recognition heuristic, to give priority to recognizing 'misc' over
'shogi', and recognize it as 'normal', so that 'misc shogi' would be
recognized as 'normal'. Would this be worth it?

H.G. Muller

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Update of bug #27666 (project xboard):

                  Status:                    None => Ready For Test        

    _______________________________________________________

Follow-up Comment #2:

commited patch by HGM. Does this fix this bug?

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Follow-up Comment #3, bug #27666 (project xboard):

For me it did, except the Gothic case, which I think is a server problem, not
an XBoard problem. (OK, admitted, a segfault is not the best kind of error
message...)

H.G. Muller

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Follow-up Comment #4, bug #27666 (project xboard):

OK, I fixed the Gothic examining case on the ICS. Unfortunately the ICS
handles it (like any variant switch, e.g. normal -> crazyhouse) in such a way
that you get the board first, without knowing what the variant is, and only
sends you the variant name when you request a move list. XBoard could not
handle this without patching up the code to switch variant during an (ICS)
game, to recognize more variants.

Now  patched that as well, examining of all variants should now be possible.
I had never used examine mode on an ICS, so I probably did not fully test it;
If there is still something wrong, let me know.

H.G. Muller

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Follow-up Comment #5, bug #27666 (project xboard):

Yes, I already noticed, that "Get move list" is now necessary for proper
variant recognition. Otherwise observing a suicide game (with xboard in normal
mode) shows some effects, like move animation disabled, when kings are not on
the board.

"misc shogi" is now working.

In the gothic case, I still get wrong piece placement when typing the
following after a fresh login:

ex admin 36

A "refresh" command (or making a move) repairs the positions. No segfaults
any more.

Other observations on "VICS" (probably server issues):

1.
command "bsetup" opens xboard error window with "Failed to parse board
string..."

2.
ex admin 31
forward 999999
backward 999999

gives this output:

Game 0: GuestIAEP backs up 999999 moves.
Game 0: Beginning of game.
Disconnecting you from game number 1.

This disconnection only happens with some games.




    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Follow-up Comment #6, bug #27666 (project xboard):

Get Move List has always been necessary for variant recognition, as ICS
simply don't send any variant indication with their boards. When al variants
were played on 8x8 boards with the same pieces this was a whole lot less
critical, though. In ICS mode the ICS send a new board on every move, so it
doesn't matter a whole lot what WinBoard thinks of the legality of the move,
as long as it displays the new position it received. (Which it always does.)

I already fixed the problem you notice in Gothic and e-mailed the patch to
Arun yesterday, but I see he did not commit it to git yet.

You seem to have stumbled on nother bug, though. That animation stops when
you have no King in suicide suggests that it considers the moves illegal, and
just displays the new board. This is not supposed to happen, so I looked at
the CheckTest code, and it turns out I introduced a bug there: when it finds
no King I return a check code, for the benifit of -variant atomic, to make it
consider moves that destroy its own King illegal (so MateTest would work). But
stupidly I did not reckon that this could also happen in suicide / giveaway.
So I should limit this no-king = in-check patch to -variant atomic only.

The other things indeed sound like server issues. The "Disconnecting you from
game..." message typically occurs when the ICS finds no white and black Kings
on the board. (Note that the open-source version of the Chessd server does not
support suicide or atomic.) I should be able to see in the error log here what
board the ICS actually was analysing when it came to this verdict.

H.G. Muller

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Follow-up Comment #7, bug #27666 (project xboard):

With the lastest git my points are fixed.
I cannot decide for:
1. the bug mentioned in comment #6.
2. "bsetup" on VICS (xboard or server issue?)

Thanks!

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Follow-up Comment #8, bug #27666 (project xboard):

I can confirm that the bsetup and disconnect problem are server errors. The
hack I made supports variable board size, and I initialize it in th boad-setup
routin that starts a new game. Apparently bsetup uses another routine that
does not set nr of ranks and files, so you get sent a 0x0 board!

The disconnect is caused by a board that is the initial position minus Ke1,
Ke8, Rh1 and Rh8. Presumably the backward command does not know how to take
back O-O...

I still have to try reproducing the FICS animation problem in Suicide.

H.G. Muller

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[Sylvain Beucler-4]

Update of bug #27666 (project xboard):

                  Status:          Ready For Test => Fixed                  
             Open/Closed:                    Open => Closed                

    _______________________________________________________

Follow-up Comment #9:

ok. I'll close this bug then. If the others are still valid we should file a
new bugreport, so that we can keep better track of things.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?27666>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/



_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard

[h.g. muller]

The problems with bsetup and castling takeback in examine mode
should now have been solved in the server. The bsetup case
required some additions to XBoard, to make it switch variant
during the setup when the ICS changes board size.

When you enter "bsetup" from idle mode you now get an empty 8x8
board (in stead of 0x0). Once in setup mode you can enter
"bsetup category board", or "bsetup category", and this will set up
the initial position of the variant loaded from the file category/board
or category/0. This means that commands like "bsetup shogi",
"bsetup gothic" or "bsetup capablanca bird" work without problems,
and XBoard adapts variant and board size accordingly.

The problem of taking back castlings in examine mode should be
fixed now; it was due to a new internal representation of castling
in the ICS I had implemented to unambiguously specify Chess960
castling, but which was not recognized by the take-back routine
yet.

More bug reports on this are welcome, as I don't really know how
setting up boards is supposd to work. Is one supposed to use
edit-position mode in XBoard to setup the position?


_______________________________________________
Bug-XBoard mailing list
Bug-XBoard@...
http://lists.gnu.org/mailman/listinfo/bug-xboard