Ratings & Comments
Seems interesting, I played a game and got annihilated in like 5 moves by the computer though.
Which computer was that?
Thanks, Jörg! These corrections have been incorporated.
It is not clear to me what happens when a Pawn is on 9th rank, and nothing of that color is captured yet. Is the move to 10th rank now forbidden? Or would the Pawn just stay a Pawn? (This could make sense to capture something.) If the move to 10th rank is forbidden, would it still check a King on 10th rank?
I think we should follow the Grand Chess rules for promote-by-replacement. A pawn cannot move to the 10th rank if there is nothing to promote to, but it can still give check. This would never happen in a real game anyway.
There is something very fishy with this Man & Beast page: I cannot get the page source! When I request it through my browser, I get the page source of a page announcing a 520 error by Cloudflare. This doesn't happen for any other page on CVP.
I cannot reproduce that problem, pressing CTRL-U in Firefox gives me a clean and readable source page.
Must be on your end, it works for me: https://pastebin.com/ex2CW0Hj
There must be something very wrong with my FireFox. Anyway, I managed to work around it by logging on to a remote Linux system, and using wget there to fetch the page. I wanted to see what it uses to generate the faulty images. This appears to be the following:
<IMG SRC="/play/pbm/drawdiagram.php?code=9-18-3.1.6.1.3-11.3.2-1.5.5N4-4J6.3.2-1.5.4.1.3-18-3.1.12-9&cols=19&nocoordinates=on&set=alfaerie-many">
Apparently it relies on the diagram designer to display a period as some marker. And for some reason it doesn't.
Thanks Greg. Can you update the Emperor's Game as well?
The Emperor's Game
Apparently it relies on the diagram designer to display a period as some marker. And for some reason it doesn't.
I've noted this before. We used to allow a period for a movement marker.
https://www.chessvariants.com/index/listcomments.php?id=37824
I started to look into a fix last night, but was having problems connecting to the server. I'll try again tonight. If I can't find a clever enough search-and-replace method, I might consider a spinoff (legacy version) DiagramDesigner for these pages.
Why are there no pieces displayed? Even when I try customizing the preset.
Can you provide more information about what you are doing? All I can tell you right now is that there is no system-wide problem.
Is here a compelling reason to not simply fix the diagram designer so that it shows a marker for a period? At the moment it seems to completely ignore the periods in the FEN code. So it doesn't seem to be in use for something else.
Just a few days ago I created some diagrams with dots meaning plain and empty squares. I usually use numbers for fully empty lines, but dots for spaces in lines with pieces or decorations.
I suspect that there are more diagrams of this kind.
I only notice now that these variants have weird castling rules, where the Rook doesn't end up next to the King. At the moment the diagram doesn't support that. Also for lack of a Betza notation for this. The On castling implies the castling partner ends on the opposit adjacent square.
[Edit] For now I solved this problem by allowing additional specification of the castling type through diagram parameters, rather than introducing new XBetza notation. For this I implemented a new parameter castleGap, default value 1, which corresponds to normal castling. The idea is that for castleGap > 0 the destination of the Rook will be that of the King, minus castleGap times the basic King step. For castleGap=1 the Rook would thus end up next to the King on the inside, while castleGap=2 would leave an empty square between R and K, etc. If castleGap < 1, however, the reference point is the King's square of origin, instead of the destination. That means that castleGap=-1 gives the type of castling we want here: the Rook ends next to the square where the King started (just as the King ends next to where the Rook started, so that they move by an equal amount.)
The new setup is mistaken, the Knights are between the Rooks and the Bishops, and the Bishops are on different colours, compare p. 77 in the book by L. Tressan here: https://books.google.de/books?id=n64UAAAAYAAJ&printsec=frontcover&hl=de&source=gbs_ge_summary_r&cad=0#v=onepage&q&f=false
Oops, sorry. I just cloned the diagram for The Sultan's Game and removed the Marshall, without realizing the Sultan's game swaps the a-side N and B compared to their orthodox position. I fixed this now. For the castling I still have to figure out a solution.
Sorry Fergus I meant to comment on this preset. All I get is a blank board.
I wish I could use the Alfaerie icon for the Sergeant for the Maasai on ranks 3 instead of the one I used instead but I don't know how to do it.
@H.G., The change was to conform better to FFEN notation; together with that change occuring back in 2018, I think fixing the handful of pages in the comment I linked to last time is the better solution.
Here's the comments recording the change in behavior:
https://www.chessvariants.com/index/listcomments.php?itemid=DiagramDesigner&order=DESC&first=36719&last=36723
On this page, I copied out Pieces section to Notepad++, and used this regex:
find: (drawdiagram.php\?code[^&{"]*)\.
repl: $1$2$3{.}
This finds periods after drawdiagram.php?code
and any number of characters that are not among &
, {
, "
. The exclusions prevent searching beyond the ffen code from the drawdiagram
script. The match always gets the furthest .
, and so we replace from the rightmost side; repeating this (and the excluded {
) eventually replaces everything we want.
This means I have to mash "replace all" until all the replacements are done, but that's fine.
There's a bigger issue with generated images, like the ZEBU (.ZRH
) and BISON (.JZ
). When copying from the webUI editor, the surrounding squiggly brackets got url-encoded, so that they were missed by the regex, and the leading .
got caught and changed. I manually fixed these two.
I'll ponder ways to improve this process, but thought I'd describe it here so that (1) I'll remember, and (2) anybody (handier with regex?) can provide feedback.
Finally, the diagram containing the forward-only versions had suffered a newline in its url, which together with some added indentation for some reason, interpreted the additional whitespaces as requesting stones added to the board. I fixed that as well.
And on a possibly totally unrelated note: some of the math expressions are weird, seemingly including extraneous characters just before every exponent character?
@H.G., The change was to conform better to FFEN notation; together with that change occuring back in 2018, I think fixing the handful of pages in the comment I linked to last time is the better solution.
Here's the comments recording the change in behavior:
https://www.chessvariants.com/index/listcomments.php?itemid=DiagramDesigner&order=DESC&first=36719&last=36723
Yep, that is how you wreck a website.
a Royal and his Pet
Why are there no pieces displayed? Even when I try customizing the preset.
I recently changed all sets to use relative URLs for the pieces. While the URLs were correct, the pieces all had a height and width of 0. This is because you were rescaling a board rendered as a table, and it needed to read information about each image file to do this properly, but it no longer had the correct path to the image to do this. I corrected this, and now the pieces will show up when you rescale your preset.
25 comments displayed
Permalink to the exact comments currently displayed.
Same for me, none of the diagrams work in these MAB articles. These articles seem really useful but are really dense with information and hard to follow, so without functioning diagrams I can't really make heads or tails out of them.