1Possible improvements **=somewhat higher priority 2--------------------------------------------------- 3 420. Some way to back up, explore variations, then return to the main 5line and restore moves we backed out. Should also support exploring 6side branches while loading a game file. Push/pop buttons? Button to 7fork another xboard in same state as current one? 8 920a. Allow side branches to be remembered in game files (and 10understood on LoadGame). Use PGN variation notation. 11 12**62a. Support series of time controls as in later gnuchess 4.0.x 13versions (and real tournaments!). Needs a protocol extension to 14"level"; this is now planned for protover 3. 15 1662b. Perhaps we could also support the kind of clock where you get K 17seconds per move, plus a reserve of J seconds per game that is used 18when a move takes more than K seconds. The degenerate form of this 19with J=0 is desirable in itself. Needs a protocol extension. 20 2168. WinBoard has some features not in xboard. Among others: 2268a. LoadGameOptions dialog (good to add) 2368b. SaveGameOptions dialog (good to add) 2468c. TimeControl dialog (good to add) 2568g. Font dialog. (leave out) 2668h. Colors dialog. (leave out) 2768i. Board size on Options menu (good to add) 2868j. Resize with mouse 2968k. Highlight dragging 3068m. Move typein from keyboard. (good to add) 31 3269. xboard has some features not in WinBoard. Among others: 3369c. Font size selection by pattern fill-in. (leave out) 3469d. Alternate bitmaps specified at runtime. (good to add, often requested) 3569e. Flashing. 36 3771. Allow restart after ICS logout instead of exiting? Many requests 38for this. 39 4071a. Would be useful to drop into NCP mode after ICS logout and allow 41for reconnect later. For instance, this would let you save your last 42partial game if you didn't have autosave on. 43 4472. Allow switching among ICS, NCP, GNU modes?? 45 4674. Observe/play multiple games, each with its own window. Hard. 47 4880. Clone command. Saves current game to a file foo, then starts up 49another copy of xboard with -lgf foo. Should probably note current 50state of some options and give command line flags to set them the same 51way in child. What about -ncp mode, though? Might want that to be 52different. Would be nice to have a -moveNumber flag to position the 53child to the same move number in the current game. 54 5591. Would be nice to have an option to overwrite save file instead of 56appending. 57 5895. "Have the clock indicate how long the computer/player has been 59thinking for when the -st option is used. This should be how long for 60the current move only." (Dino Dini) 61 6297. "Have the ability to change the level of the gnuchess during a 63game. For analysis purposes, sometimes it is useful to have the 64computer think really hard about certain positions." (Dino Dini) 65 6698. Reorganize the source and split it into more files. 67 68107. Changing time control during a game could be made to work. 69Current (WinBoard only) way of changing at start is suboptimal. 70 71**108. Could add search depth and time per move to (WinBoard) 72TimeControl dialog. 73 74140. LoadPosition needs better checking for invalid data. 75 76142. xboard needs a real file chooser dialog. Jeff Woolsey submitted 77code to merge in the one from Ghostview, but said it works only on 78X11R5 or later. I have not yet tried his code or looked to see if it 79could be made to work on X11R4. 80 81147. Improve display of held pieces in bughouse mode, putting them 82closer to the player's side of the board, and perhaps using icons (and 83allowing dragging from the icons to drop a piece). Actually, moving 84the clocks would be beneficial too, so that it's easier to visually 85associate the right clock with the right player when looking at two 86boards during a bughouse match (with two copies of xboard running). 87Maybe multiple window layouts should be user-selectable. [4.3 supports 88drag-drops from holdings] 89 90150. Keep track of castling availability, en-passant availability, and 91FEN half-move clock (for 50-move rule), pass the availability 92information to the move generator through the existing interface, and 93load/save it in the FEN code. [Fixed in 4.3] 94 95150a. Note that not keeping track of ep creates a bug in mate 96detection. In certain positions, moving a pawn from the 3rd to 4th 97rank gives mate, but it is not detected because if the pawn had moved 98to the same position from the 2nd rank, ep capture would have been 99legal and there would be no mate. [problem disappeared in 4.3] 100 101150b. Note that although we try to put in e.p. availability when 102writing FEN, this works only if we have the move that got us to this 103position in the moveList. If we load a position from FEN, the 104e.p. availability field of the FEN is lost. [disappeared in 4.3] 105 106152. Generate the info file, man page, perhaps html, and perhaps even 107the WinBoard help file from the same sources. Right now, 108substantially the same text is duplicated in three places. This is 109a huge pain in the neck. 110 111157. Idea: if Zippy flags some number of times in a row and keeps 112getting told the opponent is not out of time (presumably due to 113timestamp/timeseal), he should issue an adjourn or abort. 114 115160. The xboard CTRL key flashback feature is bad in Analysis mode. 116Flashing back throws away the analysis. 117 118161. SIGINT doesn't work for local programs on Win32. There is an 119equivalent, GenerateConsoleCtrlEvent(CTRL_C_EVENT), but I haven't 120gotten it to work thus far. It seems that WinBoard and the chess 121engine don't "have the same console"; indeed, neither has a console. 122 123162. Supply a non-X Zippy for Unix. Perhaps use the code from Klaus 124Knopper. 125 126167. Alternative bitmaps loaded at runtime for WinBoard. [font-based 127piece rendering provides this functionality since WinBoard_x] 128 129169. Upload Game on file menu? This would upload the current game 130moves into a scratch examined game on ICS. Could do Upload Position 131too. Some code to do Upload Game was contributed, but it didn't 132handle setup positions, variants, checking for proper gameMode, etc., 133so I couldn't use it. 134 135170. Auto step every k seconds should be a general feature you can 136turn on at any time, not just in LoadGame. What should GUI be? 137 138171. Displaying thinking in a separate window should be an option 139available at all times when Show Thinking makes sense, not just in 140analysis mode. It should perhaps be orthogonal to analysis mode. 141[added in WinBoard_x] 142 143172. Should analysis window scroll instead of showing only the last 144line? There could be an extra checkbox to show all chess engine 145output in it. [fixed in WinBoard_x] 146 147178. Need reliable kill for chess engine on Win32. See also 161. 148 149181. Some people want -sgf to save only played games, not observed and 150examined games. Others like it to save them all. Make this an option? 151 152183. Search feature in Game List dialog [Igor Eged]. 153 154187. "Log session" feature for ICS window. 155 156193. Some users have reported xboard redraw bugs, but I have not 157been able to reproduce them on the Digital Unix, Linux, or eXcursion X 158servers (xboard). 159 160c. From Hawk: 161Dragging in xboard does not work 100%. 162* The dragged piece may disappear (try some fast e2-d3 movements - 163 this is easy to reproduce). 164* The dragged piece may be placed on an illegal position (this 165 happens less frequently, but it _can_ be reproduced by doing the same 166 as above). 167* In both of the above cases, the board is correct after forcing a 168 redraw (ie. minimizing/restoring the window) 169I'm using eXodus - a win32 X server, perhaps it's buggy. [Seems that way, 170or just possibly, these could be bugs that are fixed in 3.6.11beta or 1713.6.12beta.] 172 173d. One user reports: "If you are 'holding' a piece over the square you 174want to go to and that move becomes illegal but you make it anyway, 175the display sometimes corrupts and a partial piece appears..sometimes 176a whole piece." This was under RedHat 6.0. 177 178194. A few users have reported drawing problems with WinBoard. There 179may be some color map problem on certain displays. I might possibly 180have fixed this already; I cannot reproduce it. The common thing 181seems to be black pieces changing to white. This may depend on having 182customized the board colors. I suspect many of these problems are 183buggy Windows drivers for particular display cards. (May be worked 184around in 4.2.0 with reprocessed .bmp files.) 185 186a. From mulesmurf: if i log in and then move the board, it will happen 187that the black [pieces] change to white. 188 189c. People have sent pictures where the grid didn't get redrawn. This 190produces random bits in the grid area from uninitialized memory. 191Could this be a resource leak? 192 193d. From James Van Buskirk: I did notice one problem tonight playing 194against Comet under FX!32 1.2 (Comet, not winboard ran under FX!32): 195for a couple of moves winboard didn't turn off the highlighting of the 196square I dropped the piece on until Comet had moved. Later, the 197problem fixed itself. Have other users noticed this? [No.] 198 199e. One user reports: Lastly, there is some sort of slight bug in the 200piece dragging code, but it comes up only very occasionally (once in 201500 moves?)...basically, I'm dragging, change my mind, throw the piece 202of the side of the board, and instead it gets stuck not inside a 203square. It's really disconcerting, this piece just sitting across 204square boundaries. When you make another move, it gets cleared up and 205goes back where it's supposed to. [Might be fixed in 4.0.6, but I 206couldn't reproduce this, so I can't be sure.] 207 208f. DAV says: "When I resized the board on win 95 and made it smaller 209some of the pieces turned into junk (I'm not able to give a screen 210shot)." Later: "Seems to happen on the three sizes below small. 211Win98 and NT seem to be ok, trying to get more win95 users to test 212it." Works fine for me on Win 95... [Might have been fixed by the 213cleanup of the Windows .bmp's that happened sometime before version 2144.2.3] 215 216195. Flash Moves works at the end of every change to the board, using 217some rather gross code in DrawPosition. This code should perhaps be 218ripped out and the flashing done from inside AnimateMove instead. 219 220197. More color distinctions. Maybe allow several user-selected channels 221to have their own colors. 222 223198. More sounds. Some possiblities: distinguish White vs. Black move 224or your move vs. opponents, distinguish different kinds of 225win/loss/draw/unfinished (checkmate, resignation, stalemate, other 226draws, opponent disconnected, etc.), game ends when you aren't playing 227and/or aren't in ICS mode, differentiate amongst requests (draw offer, 228etc.), **check, got new offboard piece (by piece type?), opponent got 229new offboard piece (ditto), arrival notification, "* who has an 230adjourned game with you," draw declined, moretime received. 231 232199. Maybe put user names instead of White and Black on clocks, or 233user names above or below clocks in a smaller font. 234 235200. I don't like these ideas, but some people do: 236a. Highlight all of the squares that the piece you are moving can go to. 237b. Clicking on the piece moves it immediately if it has only one legal move. 238c. Allow clicking on the destination, then the source. 239d. I guess highlight the pieces that can go to that destination. 240e. Clicking on a destination that only one piece can go to moves it there. 241 242201. mulesmurf: piece counter so we know exactly the score 243 244202. mulesmurf: ziics pieces, board along with color intensity 245 246204. Hawk at Fri May 15, 12:58 CDT 1998: a minor thing with xboard: 247the coords disappear while dragging from a square with coords on it. 248[They come back when the drag is finished.] 249 250208. WinBoard sometimes comes up with one of the menu bar items grayed 251out (one that should not be). I've seen this on different entries at 252different times. 253 254213. Maybe add Silence to sound menu -- parallel to Monochrome on 255color menus. 256 257214. Maybe easier startup dialog customize: 258a. Buttons to add/delete/move(?) current string to/from/within current list. 259b. Or maybe auto-add newly typed/edited strings up to some limit. 260c. Also a drop-down for Additional Options. 261 262216. Maybe shorten menu names in WinBoard to one char if and only if 263the menu bar would wrap with longer names. 264 265**217. Easier way to add engines to the WinBoard startup dialog. 266Ideally, an engine author should be able to supply a simple 267configuration file or script that installs his engine. "*.wbe" idea. 268 269218. Some folks would like Highlight Last Move to apply only to 270opponent's moves. 271 272219. From Jorge Monasterio: Support automatic game starts when in 273tournament. In other words, do that "match bigolfish mann 5 0 black" 274automatically based on the info sent from mamer. [I would need to 275play in some tournaments before I could evaluate this idea or fill in 276the details.] 277 278222. Could support time handicaps. [Added in 4.3] 279 280223. Bug: very long comments cause buffer overflow. Need to revamp 281the way comments are parsed and returned from parser.l to fix this. 282 283224. Larry Parrish requests: In an automated multi-games match I would 284like to load the n-th game from the file for the n-th game of the 285match. [That is, when -mg and -lgf or -lpf are used together without 286-lgi or -lpi.] [4.3.15 provides auto-increment og game and pos index] 287 288225. From pulp: another bug (maybe with the Microsoft RichEdit 289control) I've observed is when you flood the edit control it loses the 290bottom. Eg. "games" will scroll along then start slipping until by 291the end of the list you are really looking at the middle of the list 292and have to manually scroll back to the bottom. [May be fixed?] 293 294225a. Resizing the WinBoard ICS interaction window can cause the cursor 295to be off the screen, which stops scrolling. 296 297226. From pulp: Clipboard persistence. If I copy text, then the text 298disappears when Winboard is closed. Eg.: Log in, highlight and copy a 299web page address, close, go to browser, paste *nothing*. [How does 300one do this in Windows?] 301 302227a. In engine-engine matches, an engine that crashes should lose and 303perhaps get restarted for the next game. [In 4.3 a crashing engine 304forfeits the game, but still terminates a match.] 305 306227b. Maybe when an engine playing in Zippy mode exhibits a bug, we 307should do a little more too. Perhaps "say Sorry, the engine seems to 308have a bug", resign or abort, and maybe force an engine restart for the 309next game. 310 311**227c. We should also detect and stop the game on false mate or draw 312claims, rejecting a legal move, or (maybe) missed claims of mate or 313draw. Note that draw claims require checking for repetition, 50 move 314rule, and insufficient material as well as stalemate, so this is 315nontrivial. [Added in 4.3] 316 317228. RichEdit bug: sometimes a line gets indented as if it were 318wrapped when it should be flush left. 319 320**229. Support for sending moves from the partner's game to a bughouse 321engine. Ben Dean-Kanemura submitted support that works without adding 322general multi-game support, by relaying the moves without processing 323them. Should also send holding updates and clock times; not sure if 324Ben's code does that. Needs to be a new "feature" command option, and 325to be documented in engine-intf.html, removing the note that this may 326be done in the future. 327 328231. See Ed Collins's WinBoard bug list on his pages: 329 330231b. If you save an ICS examined game to disk as a PGN file, it 331doesn't report the end-of-game result in both the header and at the 332end of the notation. [It doesn't get this info from the server, but 333maybe there is a way it could.] 334 335232. From Ed Collins's WinBoard wish list on his pages: 336 337232b. Multiple winboard.ini files for different moods, loadable from 338inside the program. 339 340232c. Fix bug where fics% prompt is not suppressed when using 341timeseal. This has something to do with a timeseal deficiency whose 342details I've forgotten, but there must be some way to fix it. 343 344232e. When using the server's "examine" mode to examine a previously 345played game, have the game clocks display the time that each player 346currently had after each move was made. 347 348232f. [Not from Ed] Deal with ICC game relays where an "examined" game 349is supposed to have running clocks. Not sure if this is still an issue. 350 351232g. Customizable on-screen buttons [i.e., toolbar or toolbox]. 352 353232h. Customizable function keys. Currently xboard can bind a key to 354any built-in xboard function. Should extend this to WinBoard, and 355also (in both programs) support sending an arbitrary ICS command when 356button is pressed. 357 358233. See if anything from LightningRulez should be added. Premove is 359now included (from Winboard Plus), but maybe there are other useful 360things too. 361 362241. There might still be bugs in xboard and/or winboard when moves 363come in from ICS during drag animation, though I can't reproduce them. 364DAV saw the opponent's piece fail to be erased from the source square, if 365I understood his message right. [DAV says this still exists in 4.2.5. 366I think bug 426 is a duplicate. The description is clearer there.] 367 368242. User-defined "looking_at" patterns and actions. This is hard to 369bombproof. 370 371**243. "Rematch" shortcut in /cp mode. 372 373246. Maybe auto-close comment window when user goes to the next move. 374Whether this is desirable really depends on the content of the 375comment, but maybe it is rare to have a long comment that discusses 376the plan several moves ahead and should be left up. 377 378248. Would be nice to support retract move after checkmate or draw (etc.). 379 380254. Different tray icons for interaction window vs. board (suggestion 381from FICS). 382 383256. One user asked for "delete" option in game lists. This requires 384rewriting the file to close up the hole, so it's pretty heavyweight. 385Next people will be asking for move, sort, search, and other game 386database functionality; ugh. 387 388**257. Would be nice to find a way to safely implement "retract move" 389when the engine is still on move. (using ping/pong) 390 391**260. General mechanism where engine can declare engine properties that 392the user can frob and xboard/winboard will put up a simple dialog to 393allow user to view and set them. See email message "special engine 394commands" from Steffen Jakob: 395 396 From: "Steffen A. Jakob" <saj@xxx> 397 To: mann@xxx 398 Cc: chess-engines@xxx 399 Subject: special engine commands 400 Date: Fri, 26 Feb 1999 09:12:26 +0100 401 402 Hi Tim! 403 404 Here comes a summary of our ICC chat (I cc this mail to the 405 list... maybe other engine programmers can give some feedback if 406 they find this feature useful): 407 408 Please add a protocol command which enables the engine to tell 409 xboard its special commands. Maybe a set of (label, command string) 410 pairs would be useful. Then create a menu for these commands. It 411 could look like this (that's what I would like to see for Hossa): 412 413 engine->interface: 414 ------------------ 415 command "Reset the log file" "reset log" 416 command "Save Diagram as PNG" "png dia.png" 417 command "Disable Book" "book off" 418 command "Show Evaluation details" "eval" 419 command "Show Search statistics" "stat" 420 421 etc. 422 423 While typing this list I realized that an additional parameter like 424 this would be needed/nice (of course you can think of many more 425 improvements): 426 427 command "Save Diagram as PNG" "png %s" 1 428 429 "1" tells xboard to ask for an additional parameter (here: filename) 430 which will replace the "%s" in the command string. 431[XBoard 4.3.16 has an experimental implementation of this ide through 432the 'feature option="..."' command.] 433 434261. Check that "quit" is sent to engine properly if ICS connection 435fails to open due to error from timestamp/timeseal. "wzrdking" had a 436case where it didn't. Timetamp had failed with "Host name not found 437by name server" (because in fact there was no internet connection, arrgh). 438 439273. Displaying the names of the players from a PGN file in the header 440of the window does not work if there is only one game in the file. 441The filename is shown instead. (This was intentional, but maybe a 442poor idea.) 443 444274. Command to go to move number N, with entry box for N. 445 446276. Make grid optional in winboard? Note that the highlight options 447don't work without it. [The Winboard_x -overrideLineGap option provides 448this.] 449 450276a. Make color of grid an option, instead of always black. 451 452277. Flooding xboard -zt + timestamp with tells hangs timestamp after 453a while. xboard is blocked writing to the timestamp pipe. timestamp 454is in an unknown state, not reading from the pipe nor responding to 455pings. (Confirmed with the software running on milwaukee, Red Hat 4565.2) 457 458278. Flooding winboard -zt + timestamp with tells hangs winboard after 459a while. timestamp still responds to pings. This reported by 460nudnick; I haven't tried to reproduce it yet. He also says that 461windows explorer then crashed when he tried to close winboard. 462 463279. Idea from an anonymous ICSer: piece drop menu could have images, 464not words. [Or maybe both?] Also might be easier to select 465accurately from it if horizontally arranged, or in a 3x2 grid. 466[Now 4.3 has drag & drop, piece-drop menu should be deprecated] 467 468284. dxterm hangs when it receives the escape sequence that sets the 469title bar on xterm. I think some other xterm substitutes do too. 470 471**287. WinBoard can look funny if you are using themes because it 472unconditionally uses white for certain colors instead of the default 473background. 474 475289. Merge WinBoard support for the Henter-Joyce JAWS screen reader. 476Source code (zip file containing modified WinBoard 4.0.2 source) is at 477ftp://ftp.henter-joyce.com/users/hj/winboard/WinboardSource.exe. It 478would be interesting to make this work with festival (free Unix speech 479synth) or other speech synthesizers too. 480 481290. Could have limited voice output without a speech synthesizer by playing 482wav files for piece names and square numbers. Idea suggested by Hannu 483Wegner. 484 485291. Would be great to support DGT, Novag Universal chessboard, etc. 486I'm not likely to fork out the cash to buy either for testing, however. 487Note users have done various things along these lines. 488 489292. When an engine is analyzing, it would be cool to have a way to 490automatically play out its current PV (or even older ones) on a board. 491These PVs could be saved as variations when we have variation support. 492 493297. Windows users expect all programs to have a Print function, even 494though there are other ways to print. Useful items to make printable: 495PGN game scores (which are plain text, of course), maybe FEN position 496diagrams, maybe graphical position diagrams (i.e., same as pressing 497the "Print Screen" function key and printing the clipboard). 498 499298. One user suggests that automatic move stepping should not start 500until you press OK in the tags window. This could be a good idea, but 501I think I'd rather have automatic stepping never start until you press 502a "play" button on the main window, and have that button always 503available, not just when loading a game from a file. 504 505299. One user suggested being able to jump to a specific move by 506typing its number (followed by ... if it's to be black's move). 507 508300. Can I put a hyperlink (that will invoke a browser) directly onto 509the WinBoard Help menu or in the Help file? Useful to take users to 510my chess page. 511 512302. On NT, check if the board is redrawn properly when exiting from 513the screen saver. 514 515308. There is some tricky way to maximize the WinBoard board window, 516although there isn't supposed to be. Maybe prevent this, or else 517legalize it and make sure something reasonable looking happens. 518Note: the trick is to select one of the "tile" options on the Windows 519task bar. Not clear what should happen in this case. 520 521309. Avoid endlessly flagging every second when opponent is lagged out 522and autoflag is on. Automatically courtesy-adjourn (or abort if 523unrated) instead, at least if in zippy mode. 524 525310. Weird bug: the last character of a FICS output can appear on the 526next line, *after* the fics% prompt, according to one user. I might 527have seen something like this myself once too. 528 529314. One person wanted his engine to see tells when running in Zippy 530mode. I doubt this is generally useful, so I haven't added the code I 531wrote for him. (I can query for "tellfrom" in my mail file to find 532it.) 533 534314a. Another person wants his computer account to relay tells to his 535personal account. 536 537319. One user would like to be able to reload a PGN file and resume a 538game against an engine as if after an adjournment, with the clocks 539(and moves to next time control) restored as they were when the game 540was saved. 541 542320. Display ratings of ICS players in the title bar (if there is room, 543presumably), or somewhere, anyway. 544 545321. Maybe refresh the EditTags window if the gameInfo changes and the 546window is up. (WinBoard now does this when entering MachineWhite, 547MachineBlack, or TwoMachines mode -- not very general. From Winboard Plus.) 548 549322. One user requested that Zippy games use the engine name instead 550of the ICS handle in the PGN tags, as he uses multiple engines on his 551account. (My response was "I'd say it should give both"; not sure just 552how to do that, though.) 553 554324. I'm told that Blitzin supports animated dragging without holding 555down the mouse button and that this is "much better" and I should try 556it too. As yet I haven't. 557 558326. Full support for variants and variant-capable engines in all 559modes, not just ICS mode. Here are some items still needed. Some if 560not all should probably wait for a rewrite, not be added to the 561current code base. [4.3 supports many variants, even those not supported 562by ICS. Not bughouse, though] 563 564326a. Menu to let you see and change the current variant type, not 565just a command-line argument as at present. [Added in 4.3] 566 567326b. crazyhouse: Keep track of holdings and display them, keep track 568of what onboard pieces are promoted so they will revert if captured, 569understand offboard interposition to stop mate. [Added in 4.3, (not 570sure about the interposition)] 571 572326bb. One user asked for a Czech variant of crazyhouse where pawns are 573given to the opponent when they reach the 8th rank (after capturing, if 574it's a capture) instead of promoting. 575 576326c. bughouse: Features in 326a would be helpful for editing one board 577of a bughouse game, and getting mate indication right even in ICS mode. 578In the longer term, could understand BPGN, support editing two-board games. 579 580326d. wildcastle: Allow wildcastle castling only when playing this variant! 581Maybe support automatic shuffling for local play. [4.3 adds shuffling 582for any variant played locally] 583 584326e. nocastle: Maybe support automatic shuffling for local play. 585Note that several types of ICS shuffle chess fall under this variant. 586Note that this isn't really a variant, since if we kept proper track 587of castling rights and used "setboard FEN" or the ChessBase extension 588of "edit" to set them, it could be treated as normal chess with a 589different starting position. 590 591326f. fischerandom: Maybe support automatic shuffling for local play. 592Support castling. Currently, you can probably play Fischer Random 593successfully on ICS by typing castling moves into the ICS Interaction 594window, but they will not be animated correctly, and saved games will 595not be loaded correctly if castling occurs. FR castling is 596problematic because the move cannot be specified by the starting and 597ending coordinates of the king; in some starting positions there is a 598legal non-castling move that moves the king the same way. This 599violates a deeply ingrained assumption in the implementation. It also 600breaks the GUI -- you need more than just dragging the king two 601squares to indicate castling. (Note: In FR notation, O-O is h-side 602castling, O-O-O is a-side. See "help fr_rules" and "help fr" on FICS 603for rules.) [All fixed in 4.3] 604 605326g. kriegspiel: Could give some thought to how to support a 606kriegspiel engine. I doubt anyone will ever write one, though. 607 608326h. losers: Know that captures are mandatory. Understand the win 609condition. 610 611326i. suicide: Know that captures are mandatory. Understand the win/draw 612condition. 613 614326j. giveaway: Know that captures are mandatory. Understand the win 615condition. 616 617326k. atomic: Understand the win condition. 618 619326l. 3check: Understand the win condition -- ugh, have to keep a 620check counter! 621 622327. Option to set the coord font color (or something) to make sure 623there is always enough contrast with the square colors. 624 625329. Bernhard Maerz would like a command line option to choose the PGN 626player name for an engine, overriding the default of generating the 627name from the engine filename. This is for his ChessCall program. 628 629331. Maybe the local autoflag menu option should *also* turn 630server-side autoflag on/off by sending an appropriate set command. 631 632**332. Bug: Autoflag for local engine/engine or engine/human matches 633does not recognize that having your flag fall when your opponent has 634insufficient mating material is a draw. It counts the flag fall as a 635loss. [Partly fixed in 4.3, which corrects any kind of losses against 636a bare King to draws.] 637 638336. Of course lots of people ask for a graphical seek window. 639 640339. There are probably some incompatibilities with chess.net, as they 641have made a few gratuitous changes to message formats from the FICS 642code base they started with, and perhaps they've added features since 643without duplicating the syntax closely enough. In particular it's 644been reported that Zippy does not correctly send "partner" and 645"variant" commands to bughouse engines on chess.net. Also, I doubt 646colorization works. As a matter of personal choice I do not spend any 647of my time supporting chess.net, but if others want to donate patches 648that allow chess.net to be supported without breaking the support for 649ICC and FICS, I am willing to install them. 650 651342. Analysis mode automatically turns on the Show Thinking option; 652the user must then turn it off manually after leaving analysis mode if 653he doesn't want it. 654 655343. Would be nice to change xboard keyboard shortcuts to not be plain 656letters, make the new shortcuts work also in the ICS input box, and 657make typing plain letters to the board redirect to the ICS input box 658or the xterm. Might want to change/remove the Ctrl binding at that time. 659 660344. One user suggested having a way to save a position as a character 661string that would produce a board diagram when rendered in MS Word in 662a chess font. Perhaps there is already a nice standalone tool for 663this, though? 664 665346. Could use a better way to pick default size. Now we choose the 666largest that fits on the screen, which is typically too big. Not too 667bad for WinBoard, but painful for xboard since it can't be resized 668after starting. 669 670349. Check that "quit" definitely makes it to the engine before exit, 671if you exit via the File menu while a match is running. Uli Tuerke 672reported that maybe it sometimes doesn't (in WinBoard 4.0.5). 673 674352. Perhaps recognize the ANSI escape sequence that FICS sends out 675for the cls command and clear the ICS interaction window. 676 677**353. WinBoard can crash when the engine outputs very long PV lines 678(or debug output that looks like a PV). In particular, lines that 679start with 4 or more blanks following a "thinking" output line are 680treated as continuation lines, and get concatenated into a 512-byte 681buffer with no checking for overflow. Generally we need a lot more 682care to avoid buffer overflows inside both xboard and WinBoard. 683[Note: changed the buffers to be 5120 bytes as a band-aid.] 684 685Guido Flohr expands on this bug: 686 687 The problem is easy to reproduce, simply write a program/script that 688 spits out lines of more than 8192 bytes (INPUT_SOURCE_BUF_SIZE) and 689 connect it as a chess engine to xboard. No surprise, xboard cannot 690 handle such long lines, but the resulting behavior is really weird: 691 692 The program enters DoInputCallback() in xboard.c, reads the first 693 8192 bytes, searches for a linefeed in vain, then closes its eyes, 694 sets the member "unused" of the InputSource to is->buf + 8192 and 695 waits for more input although the buffer is already full. 696 697 The next time it calls the same function, is->unused is exactly 698 these 8192 bytes away from is->buf, xboard computes the amount of 699 bytes to read in this round, which sums up to 0 nothing. It then 700 calls read(2) with a bufsize argument of 0, and then assumes end 701 of file when the kernel call returns 0 (which is the documented 702 behavior). 703 704 For local engines, we now jump into backend.c, ReceiveFromProgram(), 705 and this function reports that the engine has exited unexpectedly. 706 In fact the engine is still very vivid, which is good, because in 707 spite of xboard's assumption that its child has died, it still 708 sends "force", "ping N", and "quit" to the false dead. If the child 709 was really dead, the alert boxes "error writing to ... engine" 710 would hide the first alert box "... has exited unexpectedly". 711 712 From xboard's side, it is not exactly clean to call read(2) with 713 a zero bufsize argument. In my opinion, a sane solution would be 714 to eat the INPUT_SOURCE_BUF_SIZE bytes, and when they contain no 715 linefeed, print the partial line to standard error (in debug mode) 716 and report something like �Maximum line size exceeded� (and then 717 kill the engine). 718 719**354. We currently let the user go directly from TwoMachines mode to 720various other modes like MachineWhite or MachineBlack. This may 721confuse the engine that was on move, since we don't have a way to 722force it to stop before it makes the move it was thinking about. 723(use ping/pong to know when the engine has stopped) 724 725355. Idea from LiveAndLetDie: have Zippy specially log games where it 726loses to a substantially lower-rated player. Such a game could 727indicate an xboard bug, engine bug, or perhaps a computer-using 728cheater. 729 730356. Colorization for FICS tourney tells? They look like this (ugh): 731:DChessmaster [tourney #4]: 732:hi 733 734357. A few reports of WinBoard pieces being drawn incorrectly at 735certain sizes. I haven't reproduced this. [Maybe worked around in 7364.2.0. Or is/was this the same as 194f? I don't recall the details.] 737 738358. One report of WinBoard colorization not working, even after 739restart, until the computer was rebooted. Weird. 740 741362. Running a very long automated match (/mg option) seems to crash 742WinBoard eventually, usually after 300 to 500 games, sometimes later. 743Reported by Josh Haglund. Gets a "WINBOARD caused an invalid page 744fault in module KERNEL32.DLL" dialog box from Windows. See messages 745from Josh for examples of the registers and stack dump. 746 747362a. A few people have reported WinBoard crashing upon exit. 748 749363. Maybe xboard should not set the window geometry for new popups, 750or at least not the position. The default positions aren't very good, 751and setting them programmatically keeps knowledgeable X users from 752customizing the positions with X resources. 753 754366. Feature to let an engine observe an ICS game and run analysis 755mode on the moves, perhaps whispering or channel-telling the results. 756It's a little too easy to cheat with this feature or I might have 757added it long ago. [Added in 4.2.8 and 4.4] 758 759369. Optionally send a periodic keepalive command to ICS. Can be 760used to defeat autologout, so should probably not be on by default. 761Also should be useful to help Zippy notice more reliably when its 762connection has hung, so that it can exit and a shell script (batch 763file) can restart it. 764 765370. One user requested a local noplay list for his computer player 766because the FICS limit of 50 is too small. He wants several hundred so 767that he can noplay everyone who disconnects in a lost position. 768 769374. Would be good to have a way to change engines without restarting 770the interface. 771 772376. WinBoard is said to crash or "lock up" when loading enormous PGN 773files, like a 98 MB file that comes with (??) Crafty. Win32 (at least 774in the Win95 lineage) has a limit of 32,767 items in a list box 775widget. That might explain this problem. 776 777376a. xboard exceeds limits in X and/or Xaw if a PGN file has too many 778games. (The virtual window that the list box can scroll through 779becomes too many pixels high.) This has odd effects; the window 780height of the game list seems to get reduced mod 2^16 or something. 781Maybe detect this and do something nicer about it, like stop loading 782when the window is full and pop up a warning. 783 784378. One Solaris user didn't have the "strip" program, so the 785install-sh script used by xboard "make install" failed. Check for 786this? Yuck. 787 788381. "Illegal move: post" is apparently not handled. Should it be? 789 790382. The UI for entering Training mode is a bit obscure: you must 791first select Load Game, then Training becomes enabled in the upper (!) 792part of the Mode menu and you can select it. You can then select it 793again to toggle it off. This mode is something like Analyze File 794(which has a problematic UI too, but somewhat less so); both should 795probably use a similar UI style. I consider disabling the 796step-forward button to be overkill, and I think you need to manually 797turn off auto-step when loading a game (?). One user expected it to 798work with .epd files that give a position and a recommended move. 799That seems reasonable, but doing it would require quite different code 800from the current pgn-game training mode. 801 802383. Premove piece drop for bughouse/crazyhouse. 803 804384. Pre-first-move for xboard. I don't much like the fill-in field 805method used in winboard. Maybe allow a move to be made for white 806before the game starts (in both xboard and winboard). If you get 807black, it's cancelled. Black can already premove after the game 808starts, during white's first move think time, so I don't see that a 809black pre-first-move is needed. 810 811**387. Have winboard draw opaque details on black pieces too. 812Hmm, this should probably be optional, and the detail colors should 813be selectable instead of hardwired to the opposite piece color. 814[The 4.3 -allWhite option goes in this direction. Details always 815in black, though.] 816 817**387a. Add "no piece border" option to WinBoard. See code from 818Robert Jurjevic sent 4/9/2001. 819 820391. I have a bug report from Stefan Zipproth about engines sometimes 821failing to make their first move in an engine-engine match. So far I 822can't reproduce it. This might be fixed by the Sleep(500) voodoo in 4.2.4. 823 824394. See about tightening matching for the "request" event. Right now 825it matches on help files a bit too often. (e.g. FICS "help sr_info") 826**Also clarify documentation on what the request event is. 827 828396. xboard's -size defaulting code for bitmaps unfortunately always 829picks a size from the built-in list, even if you gave the 830-bitmapDirectory option. It would be more user-friendly to have some 831extra code that lists the directory and considers exactly the sizes 832that are found there, as the pixmap code does. 833 834397. It would be nice to be able to configure xboard with bitmap 835pieces as the default without having to disable xpm support entirely. 836This should be a pretty simple mod to configure.in (and xboard.c). 837 838397a. It would be very nice to be able to compile in both the bitmap 839and pixmap pieces. This would allow us to automatically switch to bitmap 840pieces in -mono mode, including when -mono mode was automatically selected 841because too few colors were available. 842 843398. It would be nice to have the new .xpm pieces in .xim format too, 844since I restrained myself from ripping out the .xim support. 845 846399. It would be nice to have a configure option to change the 847directory where the default piece set is taken from. Slightly nicer 848than having to rename the desired directory to "bitmaps" or "pixmaps" 849as at present. 850 851400. TypeinMove should be grayed in ICS mode, or should raise the 852console window, or something. 853 854402. Internationalization. It probably isn't practical to retrofit 855this into the current code base. [4.2.8 and 4.4 have international- 856ization of xboard.] 857 858409. One user requested a way to put a time delay in the middle of 859sending zippyGameEnd commands. This would (for example) let the 860computer do a seek only if its opponent doesn't accept a rematch 861within 10 seconds. Nontrivial because Zippy has to continue to listen 862to input during the delay; it can't simply call sleep(). Maybe 863ScheduleDelayedEvent is OK, but I think it may already be in use in 864this situation, and it only supports one event at a time. 865 866411. Add zippyKibitzThinking and zippyWhisperThinking features? 867[4.3 has an -autoKibitz option] 868 869412. Maybe optimize setting the move list in the case where only the 870current move number changes; avoid slow XawListChange. 871 872**413. Add movelist window to WinBoard too. Wilkin Ng submitted a 873patch that I believe includes this; need to integrate and test it. 874 875**415. It was suggested that a few common options that many chess 876engines need be promoted to xboard/WinBoard options so that the user 877can set them from the GUI in a uniform way. In particular: hash size, 878directory name for endgame tables. [4.3 has extended WB protocol 879with a memory, egtpath and cores command.] 880 881**416. Can we avoid dying when xboard is in 8-bit mode, pixmap pieces 882are chosen (by default), and we can't get enough colors? Ugh. 883 884**419. It was suggested that stat01 be generated/accepted during normal 885play with "post" on, not just during analysis mode. This would need 886to be a "feature" command option. 887 888422. Track FICS change to details of how timeseal works; is this 889implemented at FICS yet? 890 891423. When locally stepping through an ICS game that is over, it would 892be nice to display the historical clock times at each move. 893 894424. DAV at Thu Nov 9, 01:20 CST 2000: any chance on winboard/xboard 895of a key press and a move executing draw <move>? that way you at least 896have a chance of claiming a draw in 1 0. 897 898**426. DAV says, in xboard: If you are not on clock and you hover a 899piece over a square where your opponent then moves to and then move 900the piece away the piece the opp moved or part of it has not been 901drawn. [Duplicate of bug 241?] 902 903429a. Check whether more support is needed for WinBoard to work with 904scroll wheel mice in the absence of special mouse drivers that generate 905scroll bar events directly. 906 907431. People find Analyze File confusing; they think it should do 908Crafty's relatively new Annotate function. Also, it's somewhat 909superfluous now that you can enter Analysis mode and then load the 910file, or vice versa. 911 912**436. WinBoard is said to crash if started while WinBoard.debug is 913being tail -f'ed. 914 915**441. Can I prevent an undo after new in the odd case where it can 916happen now? See mail from Bob Hyatt. 917 918450. Suggested: an option to let an engine keep pondering while an ICC 919game is adjourned. This seems a bit difficult. 920 921451. Suggested: support SuperMamer messages on FICS, aka qtells. See 922http://www.freechess.org/SuperMamer/parsing.html. 923 924452. Suggested: a way to say "prompt for the password" in a login.ini file. 925See 510 instead. 926 927**454. Add a way to claim a draw together with making a move, as can 928be needed for draw by repetition (creating the 3rd repetition with the 929move) and by the 50 move rule (the move is the 50th). Add to 930protocol, and implement using the individual methods on FICS and ICC. 931Need to be sure it's really a FICS that has the feature (somehow). 932[Fixed in 4.3 by usng 'offer draw' for this purpose] 933 934455. Suggested: Add the epdkit commands such as epdpfga to the 935protocol, since Crafty and "many" (which?) other engines include the 936epdkit. 937 938**456. Bob Hyatt suggests: "feature match=1" which says "send match 939requests to the engine and let it decide on who to play next". 940I.e., maybe send an abbreviated but fixed format message like this: 941match <opponent_name> time inc R/U. The engine would use "tellics 942accept <opponent_name>" to accept. Add as feature to protover 3. 943 944**457. Standardize on the meaning of scores (e.g. in stat01 responses). 945Is positive good for white or good for the engine? Do certain scores 946mean forced (tablebase) draw, or forced mate in N? This should align 947with the EPD spec. Maybe a feature command should indicate that the 948engine matches this part of the spec, but then again, it's not clear 949what we can do about it if the engine doesn't give that command, so 950maybe there is no point. See discussion on chess-engines list. 951[Winboard_x adds an argument to tell WinBoard an egine uses absolute 952scores] 953 954458. Bug: xboard will try to use an engine to play normal chess even 955if the engine has given the feature variant command and did not 956include "normal" on the list. Crafty 18.1 inadvertently relies on 957this bug, so it probably can't be fixed right away. 958 959**468. Need to look at Wilkin Ng's contributions! 960 961**469. Look at Zippy's behavior if the operator observes, follows, 962examines, etc. Some bugs in this area have been fixed, but it's 963received very little testing, as Zippy was never meant to be used that 964way. Noted later: Examining in -zp mode definitely has problems. If 965you go forward one move at a time, moves are sent to the engine 966correctly. If you go backward one move, "undo" is not sent to engine 967as it needs to be. Going forward or back by more than one move at a 968time needs to be looked at too. 969 970**471. Look at 3/17/2001 message from Laszlo Vecsey about crashes when 971exiting on Win ME while data is coming in from the server. 972 973**473. Visual low-time indicator. See patch in Brian Mays message of 9743/27/2001. 975 976**474. Crash reported when sending certain 8-bit characters through 977WinBoard (180, 167, others). I'm almost certain this is a FICS 978timeseal issue and nothing to do with WinBoard itself. 979 980**475. Look at Laszlo Vecsey's mousewheel patch of 4/10/2001. 981 982476. For some reason (I guess some GUI supports them), most WinBoard 983engines now have a graphical logo, 100x50 pixels in Windows .bmp 984format. It's been requested that WinBoard have a place to show the 985logo. [4.3.16 supports logos] 986 987477. Implement the "book engine" idea I proposed in the protover 3 988discussions. Some folks have offered to write them (e.g., Bruce 989Morland). [4.3.15 supports a GUI book, which is another way to 990acheive the same, as the only thing that would distinguish book 991engines is their book, not what they do.] 992 993478. Recognize URLs in the WinBoard ics interaction window and make 994them work as hyperlinks (firing up your browser). 995 996**479. Possibly use Scott Gasch's 7 May 2001 patch to delay accepting 997challenges when ping > pong instead of declining them. Bug: if you 998get more than one challenge, Scott's patch forgets about all but the 999last one. If that guy then withdraws his challenge, the others are 1000left in limbo, neither accepted nor declined. I suppose we could 1001decline challenges when there is one waiting. However, if we make the 1002change to optionally pass match requests through to the engine, Bruce 1003Moreland would like us to queue up challenges and pass them all on. 1004 1005481. 8 May 2001, Bob Hyatt "matchscore" feature suggestion could go in 1006protover 3: "matchscore A B C D E says we are in a match, this is game 1007A of B games. The current match score is I have won C, drawn D, and 1008lost E games in this match. This would give the engine a chance to 1009adjust the draw score as needed to try to win the match rather than 1010just trying to win a single game." 1011 1012**482. Suggestion from Bruce Moreland: if the user goes into ICS 1013examine mode while -zp is active, and the engine supports analyze, 1014then put it into analyze mode and let it work on the examined game. I 1015guess this could work for observed games too. Should it be optional 1016or always done? [4.2.8 and 4.4 support analysis in zippy mode, but 1017as yet only while observing games.] 1018 1019483. It would be nice if the KingShine 3D board could be integrated 1020as an alternative xboard GUI rendering. 1021 1022**484. Add -feature1 and -feature2 options to pretend the engine sent 1023us those feature commands. Useful kludge for certain buggy engines. 1024 1025485. Add "reset" to protover 3. (Clear the hash tables and other such 1026info; used to put the engine into a known state before running a 1027test/benchmark position.) [the WB 'memory' command could be defined 1028to have this side effect? Normally we would have to clear the hash 1029table when changing its size.] 1030 1031486. Other protover 3 features listed in my 6/21/2001 message to 1032chess-engines list. Write a full protover 3 draft for comment. 1033[Note that 4.3 already uses protocol extensions that differ from 1034the v3 specs.] 1035 1036487. Adjudicate non-ICS games as a draw when there is insufficient 1037mating material on both sides, especially engine/engine games. Well, 1038maybe. I argued against this in a 6/26/2001 message to chess-engines. 1039[Added in 4.3] 1040 1041487a. Probably also adjudicate when we detect checkmate or stalemate, 1042before bothering to notice whether the engine detected this. [Added 1043in 4.3] 1044 1045487b. Note: Do not adjudicate 50 move rule or 3x repetition since those 1046must be claimed by a player -- at least not by default. It would be 1047OK to have a flag to do this someday when we are smart enough to 1048notice it. Would definitely be good to be smart enough to notice an 1049incorrect draw claim by an engine. [In 4.3 you can set the number of 1050moves or repeats after which draw is adjudicated] 1051 1052487c. Another request: allow the operator to adjudicate engine/engine games. 1053[Winboard_x provided menu items for this.] 1054 1055488. Option to suppress tags popup on game load? This has been 1056requested a few times, and I received one partial implementation (not 1057enough to be useful). Needs to be on the options menu, etc. Note 1058conflict with 298 above. 1059 1060488a. Also option to suppress comment popups. 1061 1062489. Maybe have Zippy use xtell for tells to avoid redefining "." if 1063the operator is also doing tells. Do all ICSes support xtell, though? 1064 1065490. Bruce Moreland request: a way to turn Zippy auto-accept on and 1066off dynamically. 1067 1068491. Bruce Moreland request: a defined way to turn on/off PV (etc.) 1069kibitzing from the engine dynamically. I had thought it would be 1070better to make the GUI able to relay thinking output as kibitzing on 1071its own (and be able to turn this on/off dynamically), but Bruce 1072points out that what engine authors want to send as kibitzing output 1073usually includes more than the PV thinking output. [Note that engines 1074can put whatever info they want in the PV field of the thinking 1075output]. 1076 1077492. One user reports that GNU Chess 5 crashes if you start a game, 1078play far enough to get GNU Chess out of book, select Reset, and play 1079e3 in the new game. This is probably a GNU Chess 5.00 bug, unrelated 1080to WinBoard. Can't reproduce on Linux; didn't try yet on Windows. 1081Need to see if 5.02 fixes it. 1082 1083493. DAV suggests stripping off first arguments that are obviously 1084not the name of the engine when trying to parse out the engine name 1085from its command line -- things like nice, nohup, sh, csh, bash, etc. 1086Ugh, but I suppose it would be helpful sometimes. 1087 1088494. Internationalization. A couple of people have done (or told me 1089they were doing) partial conversions of WinBoard to German or Spanish 1090by manually translating the messages. I think it would be much better 1091(but a big job) to use GNU gettext. I haven't looked into the details 1092of that at all. [4.2.8 and 4.4 have gettext support.] 1093 1094495. JamesBaud points out that FEN is not sufficient to correctly 1095represent a crazyhouse position (or bughouse, for that matter), 1096because it does not distinguish promoted pawns that will revert when 1097captured from real pieces. ICS style 12 also does not give us this 1098information. I really don't know what I could do about this -- take 1099the lead in proposing an extension to FEN and to style 12 with new 1100piece letters, I suppose. I don't really have time or motivation for 1101that. If you always use PGN, not FEN, to save Crazyhouse, that works 1102around the problem. [4.3 uses bFENs which fix this, but not in ICS 1103mode] 1104 1105**496. DAV sent some code to save a screen shot in xboard. Look at 1106merging this in. [4.3 has an option to save the board display (but 1107not the entire window) as bitmap.] 1108 1109**497. Igor Eged reports: I make a move on opponents time using 1110premove. Right in the time I am holding another piece in case of 1111unexpected check that I can have other possibility. But then the piece 1112hangs in the void and its invalid move, what means often loss in 1 0 1113lightning. 1114 1115498. It would be nice for Zippy to be able to play ICC wild 20 and 21 1116(the types where loadgame is allowed). Presumably there would have to 1117be an option to either send the loadgame command with a specified 1118position, or to deal with an opponent sending it. I'm not sure how 1119this should work exactly, or how to prevent opponents from cheating in 1120various ways. 1121 1122499. It would be nice for Zippy to be able to switch amongst several 1123different engines depending on which variant type is being played. 1124 1125500. It would be nice if Zippy could have an option to accept 1126takebacks. Hmmm, also, the operator might want to manually offer a 1127takeback; it would be good if that worked. 1128 1129501. People would like to be able to add more bot functionality to 1130Zippy without C programming. I don't have any ideas on how to do that. 1131 1132502. One user requests the ability to have a game list file open for 1133more than one PGN file at a time. 1134 1135503. We don't do protover 2 feature negotiation properly when 1136restarting an engine due to -xreuse mode. The timeout and done=0/1 1137features don't work on the restart; instead we basically assume that 1138all the features will be the same as last time and that the engine 1139will be ready to accept commands immediately. This probably doesn't 1140cause a lot of problems, but if you try to replace the engine with one 1141that sets different features, the changes might not be noticed before 1142the first move, and the defaults will be the values from the old 1143engine, not the documented defaults. Also, if the engine takes time 1144to initialize and wants us to wait to send commands until it says 1145done=1, we fail to wait. However, we do send commands in the right 1146order and use ping, so most engines should work fine. 1147 1148506. One user wants a board size even smaller than Tiny. 1149 1150507. Suggested: have the cursor change to a gripping hand while moving 1151pieces. Currently in WinBoard it's always an arrow, while in xboard 1152it's always a hand with pointing finger while over the board and an 1153arrow elsewhere. 1154 1155508. It would be nice to make -xclock work with -ics, just for 1156orthogonality, though it's not terribly useful. 1157 1158509. It would be nice to make -clock work with -st. Perhaps each 1159player's clock should start with the specified time on each move and 1160count down until he moves. 1161 1162**510. A FICS service rep tells me that many users expect an autologin 1163feature in winboard (other than the non-user-friendly login.ini file). 1164Seems easy enough to add such a thing. 1165 1166512. Some users dislike the clocks counting in tenths when below 10 1167seconds, as it's distracting, and would like this to be configurable. 1168(From the Debian bug tracking system.) 1169 1170513. Perhaps have Alt+P and Alt+C as synonyms for Pause in WinBoard, 1171since the P or C on the button makes it look like that should be the 1172shortcut. I suppose Alt+< and Alt+>, and the same with Shift for << 1173and >>, would also be a good idea. Oops, but Alt+C is already 1174CopyGame, so maybe this is a bad idea. 1175 1176514. Gian-Carlo Pascutto reports: "when pasting a FEN/EPD position 1177into WinBoard, and selecting analysis mode, the engine analyses the 1178starting position instead of the pasted one. Re-pasting fixes this." 1179 1180515. Troy Edwards suggests: "It would be nice to have the option to 1181'wrap text' in the comment window." 1182 1183516. DAV says: "FICS now supports the setting and the making clocks 1184tick using unpause/pause in examine mode. It seems either the clocks 1185don't tick or are not present when using xboard. Is it possible to 1186change the policy and base things upon the clock is paused flag in 1187style 12?" 1188 1189517. Clarify in engine-intf.html that "sd" is an extra dumbing-down 1190feature, not a time control. Some text: 1191 1192In the protocol, the "sd" command isn't a time control. It doesn't 1193say that your engine has unlimited time but must search to exactly the 1194given depth. It says that you should pay attention to the time 1195control as normal, but cut off the search at the specified depth even 1196if you have time to search deeper. If you don't have time to search 1197to the specified depth, given your normal time management algorithm, 1198then you will want to stop sooner than the given depth. 1199 1200The "new" command should set the search depth back to unlimited. This 1201is already stated in the spec. The "level" command should not affect 1202the search depth. As it happens, xboard/WinBoard currently always 1203sends sd (if needed) right after level, but that isn't part of the 1204spec. [Added to the specs that go with 4.4] 1205 1206518. Warn in doc that -initString is a dangerous feature. Also move 1207doc of dalogs for sending commands directly to the engine out of the 1208protocol spec and into the man page and help file, but warn they are 1209dangerous too. 1210 1211519. In engine-intf.html, change "st" description to not use the word 1212"exact", which is misleading. Make sure it says this is the maximum 1213time per move (and does not accumulate). 1214 1215520. Check whether Alt+C (CopyGame) works in XP. One user says it 1216gives an error message. 1217 1218522. engine-intf.html should really be rewritten to describe the 1219protocol as a state machine (on each end), with every command 1220precisely defined as to what it does in each state (including perhaps 1221being illegal). 1222 1223523. Figure out what, if anything, to do with the FICS "iset singleboard". 1224I lean toward not bothering to set it, as it only disables functionality, 1225but maybe I'm confused. 1226 1227524. theblob at Mon Mar 4, 23:23 PST 2002: it'd be nice in the xboard 1228interaction window if control-w and control-u respectively erased the 1229line and word, even in a Windows environment. At least, I think it'd 1230be nice. 1231 1232525. Support a 64-bit node counter in analysis mode when compiled with 1233something that has 64-bit integer support. Both gcc and msvc++ have 1234this, though under different names. [Done in 4.2.8 and 4.4] 1235 1236526. "-initialMode analysis -lgf -" doesn't get the engine into 1237analysis mode properly. "-initialMode analyzeFile" is a workaround, 1238but bug should be fixed. 1239 1240527. Clarify ICS.ini more in the doc, if possible. Some sample words: 1241> You should create this file in 1242> WinBoard's directory (usually C:\Program Files\WinBoard) using a plain 1243> text editor such as Notepad. Make sure that Notepad doesn't decide to 1244> call the file ICS.ini.txt for you -- you may have to select "All 1245> Files" instead of "Text Files" in the Notepad save dialog. 1246> 1247> The file is simply sent to the chess server, just as if you had typed it. 1248> So if your username is foo and password is bar, the file would look 1249> like what is between ==== below, without the ===='s: 1250> 1251> ==== 1252> foo 1253> bar 1254> iset movecase 1 1255> ==== 1256 1257528. Check on updates/revisions to PGN from Steven Edwards and others. 1258Any support needed? 1259 1260529. Protover 3 needs a prefix character that says "this is debug 1261output" and prevents the GUI from parsing that line (say ";"). 1262[4.3 uses '#' for this purpose, when the engine sets feature debug=1] 1263 1264530. The undocumented (?) checking for rsh error messages tripped up 1265one engine author who had a "not found" in a message. Tighten this 1266up. 1267 1268531. Do we always wait for pongs to catch up with pings before 1269starting the engine's clock? I suppose we should. 1270 1271532. On FICS the clock does not start until after the first move has 1272been made. So perhaps we don't need the "not ready, try again soon" 1273response from Zippy there. What if Zippy is black, does black's clock 1274also not start until he makes his first move? 1275 1276533. theblob at Mon Jun 24, 20:05 PDT 2002: here's a suggestion for 1277xboard -- give priority to fitting in amounts of material in the title 1278bar, and truncate names instead, e.g., use "tyrionsAp...(50) vs. woden 1279(51)" rather than "TyrionsApprentice (50) vs. woden..." -- it's much 1280more important to see the material than the full names, especially 1281when you're playing. 1282 1283534. Steffen Jakob reports: "I use xboard 4.2.6. Yesterday I played 1284some bullet games at ICC. In one game I lost on time. In the last move 1285I already clicked on my queen and wanted it to move to its destination 1286square when I forfeited on time. After that my mouse pointer was the 1287queen piece. I couldnt drop it to any square." 1288 1289535. For protover 3, the "st" command in the protocol could be 1290extended to allow for smaller units, not just seconds. 1291 1292537. Clarify in the protocol that "force" while searching means you 1293should stop the search as soon as you can. (Isn't this clear 1294already?) 1295 1296538. Need a way for the user to adjust the clocks. This is important 1297for WCCC and other computer competitions, where xboard/winboard's 1298clock is not the master. I was going to do it in time for the July 12992002 WCCC. Oops. [4.3 allows adjusting the clocks by clicking them 1300in 'edit game' mode, in steps of 1 minute] 1301 1302541. Related to #327, John Iverson says, "It would be nice to have the 1303coordinates shown outside of the board along the edges, rather than in 1304the squares where they are hard to see and give a cluttered 1305appearance, especially on smaller board sizes. (The coordinates can 1306be handy in certain situations, such as posting a screenshot of xboard 1307showing an email game in progress, with the view flipped.)" 1308 1309542. If you start xboard with the -geometry NxM option, it will 1310dutifully make the window exactly that size, but it doesn't choose a 1311-size option that fits into such a window size, so the board is cut 1312off or has a lot of whitespace around it. In fact, the layout is 1313generally messed up because the code doesn't expect the window to be 1314the wrong size for the layout. We should at least ignore the size 1315part of the -geometry option if we aren't going to do something 1316sensible with it. How is that done with Xt? 1317 1318543. See the patches and extensions on 1319http://www.tim-mann.org/extensions.html#patches and integrate those 1320that are worthwhile into the main code base. (Some are also mentioned 1321elsewhere in this ToDo list.) They are pretty much all over 15 lines, 1322so copyright assignment from their authors is needed. Unfortunately 1323some of the authors may be hard to reach. 1324 1325544. Possibly rework looking_at pattern matching to have a "match at 1326beginning of line" character (perhaps ^). It looks like this may 1327actually be useful after all. The historical cases where you could 1328get a board or other output that needs parsing on the same line as a 1329prompt *may* have been fixed. Confirming this would need testing on 1330all the ICSes that are in use. 1331 1332546. Odd bug reported by travelGirl. The last line of a multi-line 1333macro from the ICS context menu sometimes "sticks" at the bottom of 1334the ICS interaction window. New text gets inserted before instead of 1335after it. This started when she upgraded to win2k. See mail. 1336 1337548. Guido Flohr reports: 1338 1339Another "not really a bug" problem: When xboard forks in order to 1340start a second chess engine, it is only concerned about duplicating 1341the descriptors for the standard channels in the child. This is 1342very interesting for the second chess engine, since it inherits the 1343pipes to its opponent, and can use them for sending commands like 1344"sd 2", "easy", or even for feeding a bogus FEN string into the 1345first engine (exchanging the opponent's queen with a bishop looks 1346really attractive to me). A very interesting feature, but not exactly 1347what you intended? ;-) 1348 1349Closing the other descriptors after the fork will not completely 1350avoid the problem, since - at least under Solaris and Linux - it 1351is still possible to peek and write into xboard's open descriptors via the 1352/proc pseudo file system (unless you make xboard setuid root), but 1353leaving them open, makes this cheat very obvious. 1354 1355549. "Ron" (no last name given) reports that when compiling WinBoard 1356under VC++ 6.0, there are many warnings. These ought to be looked at. 1357 1358551. "Miguel" (no last name given) reports "When you make a premove in 1359Winboard 4.2.6, and then opponent move and premove is illegal, then 1360and you try to make a legal move but you cannot, because piece are 1361erased and you have to replace your move again, move twice... in 1362lightining this is annoying." I'm not really sure what that means or 1363if it is the same bug as 497. 1364 1365552. Get rid of the need for proprietary tools to build WinBoard. 1366It's already possible to compile winboard.exe using the Cygwin 1367toolchain. You can even compile it with -mno-cygwin to avoid the need 1368for cygwin1.dll. However: 1369 1370552a. The Help file needs Microsoft Word to edit the source file 1371winboard.rtf and the Microsoft Help Compiler to compile it into 1372winboard.hlp. I tried using OpenOffice.org on the RTF file, but the 1373result was missing some formatting that the help compiler needs. The 1374help compiler itself is a free-as-in-beer download, but it's still 1375proprietary software. Perhaps the help could be converted to vanilla 1376HTML, at the cost of losing the index and search capability. 1377 1378552b. The WinBoard installer package is built with InstallShield. 1379 Possible replacements: 1380 Inno Setup (http://www.jrsoftware.org/isinfo.php) - looks best 1381 NSIS (http://nsis.sourceforge.net/site/index.php) 1382 Cygwin setup (http://www.cygwin.com) - a long shot 1383 1384552c. Although the cygwin resource compiler works on winboard.rc, it's 1385a pain to create new dialogs by editing the text file. The original 1386ones were all created using the graphical dialog editor in MSVC 5. 1387 1388553. It would be nice for xboard's ICS Input box to have a history 1389feature similar to WinBoard's ICS Interaction window. Actually it 1390would be nice for xboard to have its own ICS Interaction window with 1391all the bells and whistles that WinBoard's has instead of limping 1392along using an xterm for interaction. 1393 1394**554. Port evaluation-graph window to XBoard. 1395 1396555. Split x/wengineoutput.c in a front-end and common backend file. 1397 1398556. Fix -stickyWindows option under XP and Vista, and port to XBoard. 1399 1400557. Port display of engine logos to XBoard. 1401 1402558. Device a smart way for -autoLogo to select logos in other modes 1403than Two Machines. 1404 1405 1406 1407* * * 1408 1409Before each release: 1410- make sure both xboard and winboard build and are up to date with each other 1411- run zippy with new version as a shakedown 1412- be sure gnu5 binary is up to date in winboard, and gnu5 patch (if needed) 1413 is included in both xboard and winboard kits 1414- recheck cygwin compiling winboard 1415- make sure ChangeLog is complete 1416- test changes 1417 1418*** up to date with mail received through 5 Nov 2003, except 1419for the "protover 3" discussions on the chess-engines mailing list *** 1420