Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

Faster, no problems

Faster, problematic

Same speed, no problems

Same speed, problematic

Slower, no other problems

Slower, problematic

Doesn't work at all

[Baughn]

Malicus: GLView's database is not just out of date, it's basically useless. It claims older cards support extensions they just plain don't do.

[illiterate]

Yay!  Got it working.  Getlibs solved my problem with the first lib, but couldn't find libjpeg8.  So I had to link that to 6.2. 

[bombcar]

Movies now fixed. Text-mode movies too. Not sure about the running out of color pairs issue.. how did that happen?

I was able to make it happen anytime I wanted by increasing the window size while the movie was playing. I'll try again with HEAD.

[bombcar]

The movie handled resizing correctly, and worked great! I did not get the color warning this time.

I did get "Display initialization failed, DF will crash if asked to use the screen." - but I was using ncurses and it worked fine (except for the unselected menu options still being black on black).

[bombcar]

Checking the "Use bright colors for bold text" option in Preferences - Settings - Text in Terminal.app made the unselected menu items appear.

The cursor is appearing; I think it probably should be hidden - I assume there's a way to turn it off. Some claim "Escape-[?25l" might work.

As a side note, Andale Mono 14 pt with Character Spacing and Line Spacing set to 1.0 looks pretty good.

[Baughn]

I've got the cursor hidden in -head, so that's sorted. There's an ncurses call.

And yeah, bold text is supposed to be bright. I should save that information.. somewhere...

bombcar: You get that warning if it fails to initialize GTK, which happens if DISPLAY isn't set. As you've discovered, it's normally harmless.

[bombcar]

Over SSH to a P4, I'm getting FPS: 100 (20) on embark. Is that as expected?

[Baughn]

Well, sure. That's at your FPS cap.

[illiterate]

interesting.  looks ok locally in terminal, but poo in putty via ssh 

[illiterate]

Over SSH to a P4, I'm getting FPS: 100 (20) on embark. Is that as expected?

100 is your default FPS, 20 is the default graphical FPS.  The new version seems to display both, or at least it does in curses. 

[rdwulfe]

I've got a new 40d18-head up now that displays the old FPS counter as well, for your comparison porpoises. It's a lot more jumpy, but I'm not sure it's all that much less accurate overall; hopefully someone will tell me otherwise.

Anyway, it goes "FPS (G_FPS) / old_FPS".

(Download! Please, think of the porpoises!)

Admittedly, I haven't looked... But is there a HEAD version for Windows?

[Baughn]

No, -head is linux only.

[shadow_slicer]

I managed to get the shift-arrow keys working on my machine by hacking on graphics lib. It looks like shift-left and shift-right are defined in curses so, I just used the existing define. Unfortunately shift-up and shift-down aren't recognized by curses in any of the terminals I tried (and don't seem to be specified by terminfo or termcap!), so I had to hard code the escape sequence for them to the one used by my terminal (but it works!). I suppose they could be made configurable if this sequence isn't portable, but I hope that's not necessary  .

Spoiler: code modifications (click to show/hide)

enabler_input.h:

Code: [Select]

//We add the shifted up and down keys so we use some of the unused key codes for these characters
#define KEY_SUP 0700
#define KEY_SDOWN 0701
enabler_sdl.cpp:

Code: [Select]

#ifdef CURSES
  // If we're using ncurses output, we completely short-circuit SDL windowing.
  if (init.display.flag.has_flag(INIT_DISPLAY_FLAG_TEXT)) {
    initscr();
    raw();
    noecho();
    // add our special keys that aren't part of terminfo =(
    assert(define_key("\E[1;2A",KEY_SUP)!=ERR);
    assert(define_key("\E[1;2B",KEY_SDOWN)!=ERR);
    keypad(stdscr, true);
    nodelay(stdscr, true);
    start_color();
    curs_set(0);
    return true;
  }
#endifenabler_input.cpp:

Code: [Select]

    case KEY_NPAGE: sdl.key = SDLK_PAGEDOWN; break;
    case KEY_PPAGE: sdl.key = SDLK_PAGEUP; break;
    case KEY_ENTER: sdl.key = SDLK_RETURN; break;
    case KEY_SRIGHT: sdl.key = SDLK_RIGHT; sdl.mod |= DFMOD_SHIFT; break;
    case KEY_SLEFT: sdl.key = SDLK_LEFT; sdl.mod |= DFMOD_SHIFT; break;
    case KEY_SUP: sdl.key = SDLK_UP; sdl.mod |= DFMOD_SHIFT; break;
    case KEY_SDOWN: sdl.key = SDLK_DOWN; sdl.mod |= DFMOD_SHIFT; break;
    }
I also modified the input code to use get_wch() instead of using getch() to eliminate the UTF8 decoding:

Code: [Select]

static int getch_utf8() {
  wint_t ch;
  int ret = get_wch(&ch);
  switch (ret)
  {
    case ERR:
      return 0; //error
    case OK:
       return -ch; //unicode
    case KEY_CODE_YES:
       return ch; //key code
    default:
       return 0;
  }
}

I ran this a bit and the shift-arrow keys seem to work. I'm not sure I really trust the rest of it though, because I had to really hack the files to get the other parts to compile on my system since I don't have std::bimap.

[koitsu]

Does the source code repo have a browsable interface (something like cvsweb or svnweb)?  I might be able to narrow it down by looking at the code.

http://github.com/Baughn/Dwarf-Fortress--libgraphics-

There's no way this is the entire source code.  There's binary blobs stuffed into the repo all over the place; me tracking this bug down is basically impossible given that.  :-)

[bombcar]

Does the source code repo have a browsable interface (something like cvsweb or svnweb)?  I might be able to narrow it down by looking at the code.

http://github.com/Baughn/Dwarf-Fortress--libgraphics-

There's no way this is the entire source code.  There's binary blobs stuffed into the repo all over the place; me tracking this bug down is basically impossible given that.  :-)

It's the source for the graphics engine; DF itself, of course, is Toady's and a binary blob. All of Baughn's work is in g_src.