Problems angband winxp

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • smokie007
    Rookie
    • Nov 2007
    • 9

    #16
    Hmm, but why does the 305, 306 & the most variants runs without any Problem when 307s-309 & o don't like to over the whole day? Was it just a pure, tragically and randomly event that every try for 309 hangs when on the other side every 305 start up perfectly?
    I tried it with process killing and restarting...

    Are there significant changes in the source for displaying the main window from <=306 to >=307 and what about o 110? I think it came out before 307 vanilla...

    Mystical greetings from
    smokie

    Comment

    • zaimoni
      Knight
      • Apr 2007
      • 590

      #17
      Yes, there was a moderate rewrite of the main-win.c between V3.0.6 and V3.0.7. Part of it was ditching support for Win16 (Windows 3.0/3.1), but there were other changes to the graphical display code.
      Zaiband: end the "I shouldn't have survived that" experience. V3.0.6 fork on Hg.
      Zaiband 3.0.10 ETA Mar. 7 2011 (Yes, schedule slipped. Latest testing indicates not enough assert() calls to allow release.)
      Z.C++: pre-alpha C/C++ compiler system (usable preprocessor). Also on Hg. Z.C++ 0.0.10 ETA December 31 2011

      Comment

      • Pete Mack
        Prophet
        • Apr 2007
        • 6883

        #18
        XP launch problems..

        Is this a problem with launching, or a problem with running? Can you run the thing from the command line, either via cygwin or DOS?

        I always run from the command line in WinXP, using cygwin. I've never seen this problem, only bad fonts-file problems.

        Comment

        • zaimoni
          Knight
          • Apr 2007
          • 590

          #19
          Technically, Smokie and I are talking about different issues, as I'm on W2K.

          For W2K, it's a launch issue; it's not getting the OS term windows it needs to continue launching. Sometimes once the "bad processes" are all killed the others will come up before I can finish.
          Zaiband: end the "I shouldn't have survived that" experience. V3.0.6 fork on Hg.
          Zaiband 3.0.10 ETA Mar. 7 2011 (Yes, schedule slipped. Latest testing indicates not enough assert() calls to allow release.)
          Z.C++: pre-alpha C/C++ compiler system (usable preprocessor). Also on Hg. Z.C++ 0.0.10 ETA December 31 2011

          Comment

          • Nick
            Vanilla maintainer
            • Apr 2007
            • 9647

            #20
            Originally posted by Pete Mack
            Is this a problem with launching, or a problem with running? Can you run the thing from the command line, either via cygwin or DOS?
            Dunno; after months of not working, O is running perfectly for me now
            One for the Dark Lord on his dark throne
            In the Land of Mordor where the Shadows lie.

            Comment

            • Pete Mack
              Prophet
              • Apr 2007
              • 6883

              #21
              Originally posted by zaimoni
              Technically, Smokie and I are talking about different issues, as I'm on W2K.

              For W2K, it's a launch issue; it's not getting the OS term windows it needs to continue launching. Sometimes once the "bad processes" are all killed the others will come up before I can finish.

              Sorry, I wasn't clear on what I meant. I understand that it doesn't launch properly from the file explorer.
              But can you run angband.exe from a DOS window?
              (Start up cmd.exe or cygwin, cd to the Angband root directory, and run angband.exe from there.)

              Comment

              • zaimoni
                Knight
                • Apr 2007
                • 590

                #22
                No, DOS command line does not work for V3.0.9/W2K once W2K starts refusing to hand out OS windows to Angband instances. The first instance to hang did manage to get more memory than the following instances, but is consuming 99&#37; CPU. Killing the 99% CPU one enabled opening more instances, including those queued after the initial failure.

                (For my system:
                * opened savegame: 8,192K
                * before opening savegame: 5,912K
                * first hang, 99% CPU: 4912K
                * later hangs, 0% CPU: 2324K)
                Last edited by zaimoni; November 9, 2007, 23:48.
                Zaiband: end the "I shouldn't have survived that" experience. V3.0.6 fork on Hg.
                Zaiband 3.0.10 ETA Mar. 7 2011 (Yes, schedule slipped. Latest testing indicates not enough assert() calls to allow release.)
                Z.C++: pre-alpha C/C++ compiler system (usable preprocessor). Also on Hg. Z.C++ 0.0.10 ETA December 31 2011

                Comment

                • Pete Mack
                  Prophet
                  • Apr 2007
                  • 6883

                  #23
                  Originally posted by zaimoni
                  No, DOS command line does not work for V3.0.9/W2K once W2K starts refusing to hand out OS windows to Angband instances. The first instance to hang did manage to get more memory than the following instances, but is consuming 99% CPU. Killing the 99% CPU one enabled opening more instances, including those queued after the initial failure.
                  The only time I've seen something like this was when I selected fontsize 200 (or something silly) on OS X. It opened a monster canvas, and pinned all resources attempting to render it. I don't see how this can happen with a clean Windows install, since Windows uses the installation directory for storing preferences, rather than ~/Library/Preferences. Until I see it happen on my (WinXP) box, I am stumped.

                  If you can get a stack trace (via gdb or--better--Visual Studio) of where it is spinning, that might help. You need a debug build for that to work.

                  Comment

                  • zaimoni
                    Knight
                    • Apr 2007
                    • 590

                    #24
                    I'm thinking that something getting a Windows resource isn't reality-checking the return value. Either that, or there's some system resource that's unusually limited that most programs don't use. (I checked that I could launch other programs fine even when Angband instance 64 went into 99&#37; CPU mode rather than open. This limits my empirical testing of the RNG.)

                    I don't have a clear idea of when I'd allocate time for an "irrationally cautious" version of main-win.c.
                    Zaiband: end the "I shouldn't have survived that" experience. V3.0.6 fork on Hg.
                    Zaiband 3.0.10 ETA Mar. 7 2011 (Yes, schedule slipped. Latest testing indicates not enough assert() calls to allow release.)
                    Z.C++: pre-alpha C/C++ compiler system (usable preprocessor). Also on Hg. Z.C++ 0.0.10 ETA December 31 2011

                    Comment

                    • smokie007
                      Rookie
                      • Nov 2007
                      • 9

                      #25
                      It's me again
                      and what can it be... angband don't run again.
                      Today i start everything as usual and after checking mail and a little web-browsing and starting all the programs i need for the day, i give a try for angband, just to be secure. And it just start the process (0% CPU and 2.764 K Memory) and nothing else.
                      At this time i think it must be a program i start just before. For now there are two questions: which program & why it runs on Thursday and Friday (on these days i use the same programs as today and every other day). On the next days i try to find that #*.~#,: program.

                      Can it be that angband is bound to any (whatever) resource and have this problem if some other program was faster to occupy this special resource...? The other way seems to run fine for all programs. Maybe i just have to put angband in the autostart group

                      In search for any clue i wish a good start in the new week (but today for me no angband ;( )
                      smokie

                      Comment

                      Working...
                      😀
                      😂
                      🥰
                      😘
                      🤢
                      😎
                      😞
                      😡
                      👍
                      👎