Skip to content
Commits on May 25, 2009
Commits on Nov 8, 2008
  1. Mention experimantal branch

    committed Nov 8, 2008
Commits on Oct 13, 2008
  1. Add comment in big iter form

    committed Oct 13, 2008
Commits on Aug 25, 2008
  1. Untabify and whitespace-cleanup

    committed Aug 25, 2008
Commits on Jul 18, 2008
Commits on Jul 17, 2008
  1. @dtorop

    display status line, control chars modify options

    dtorop committed with Jul 17, 2008
    status line shows current emsquare & kerning setting
    it uses a separate gl-font so as not to pollute the test gl-font's texture
    sadly lost the trivial-garbage finalizer, though, should bring that back
      once understand how to use it safely
    use control keys to modify settings, freeing up non-control keys to
      enter text
    added new control keys to change kerning and emsquare
    keys now are:
      tab: insert a random char
      backspace: delete last char
      control-r (and many other keys too!): refresh screen
      control-f: reload font (and in process flush the texture)
      control-m: double emsquare
      control-n: half emsquare
      control-k: toggle kerning on/off
      any non-control key: add itself to the end of the test string
  2. @dtorop

    Store actual-slice in glyph object

    dtorop committed with Jul 17, 2008
  3. Cache compute-actual-slice.

    committed Jul 17, 2008
  4. @dtorop

    Fix for squat chars being too wide at low emsquares

    dtorop committed with Jul 17, 2008
    Remove extra addition in compute-actual-slice, which seemed to have no
    use?  Before this patch, glyphs such as hyphens and periods were
    distorted (too tall) at smaller emsquares (say < 128 in test.lisp).
  5. @dtorop

    handle drawing of 0-length strings

    dtorop committed with Jul 16, 2008
  6. @dtorop

    chars-recoordinate doesn't need to create a new hash

    dtorop committed with Jul 16, 2008 it just modifies the existing glyph object's tex-coord slot.
    I think this is valid, as chars-recoordinate will only be called when
    expanding a texture when all existing glyphs should be kept, but forgive
    me if I'm missing something!
  7. @dtorop

    fix chars-recoordinate

    dtorop committed with Jul 16, 2008
    make call transform-cell with proper first argument
    actually iterate through the hash
    fixes bugs introduced in feb157b
  8. @dtorop

    test.lisp: keys change test string, nearest tex display

    dtorop committed with Jul 16, 2008
    new keys in test program:
      - #\? appends random character to string
      - #\- removes last character from string
      - any unassigned key appends itself to string
    display the view of the whole texture nearest-filtered rather than
      linear, so can see clearly the individual texture pixels
Commits on Jul 16, 2008
  1. @dtorop

    Rely on glyph cell #'s when copying chars around

    dtorop committed with Jul 16, 2008
    May be over-compulsive for safety...
  2. @dtorop

    factor out texture creation/resizing

    dtorop committed with Jul 16, 2008
    Created size-texture which when called will ensure that the current
    texture is large enough to hold the requested # of characters, resizing
    it if necessary.
    Fixes bug in ensure-characters upon texture size increase, when not
    calling chars-recoordinate resulted in garbled characters.
  3. @dtorop

    glyph class encapsulates char tex-coord & cell

    dtorop committed with Jul 16, 2008
    This eliminates the character-cells in the opengl-text class.
    Ideally CLOS is fast enough to allow for its use in something so
    commonly referenced as a glyph.
    Also eliminated the sorting of chars in flush-texture, as this seems to
    no longer be necessary.
  4. @dtorop

    tweak sample string

    dtorop committed with Jul 15, 2008
Commits on Jul 15, 2008
  1. @dtorop

    use gl:color including alpha

    dtorop committed with Jul 15, 2008
    Previously used :texture-env-color and ignored its alpha.
    This allows removing a dimension from the texture array and only passing
    in luminance.
    Updated test and documentation accordingly.
    In test string also demonstrate kerning a bit more.
  2. Drop affi dependency

    committed Jul 15, 2008
    Did not need general affine transformations anyway, and all array sections have been rewritten into specific iteration.
  3. Allocate transient buffers if string is too long.

    committed Jul 15, 2008
    Other option would be to permanently expand them...
  4. Update documentation a bit.

    committed Jul 15, 2008
  5. @dtorop

    draw-char with alpha only as luminance is constant

    dtorop committed with Jul 15, 2008
    Assign luminance of 255 later when copy-character.
    I'm a bit suspicious of the &optional third in destructuring-bind of
    cell-range, but seems ok for now?
  6. @dtorop

    set color on texture draw rather than creation

    dtorop committed with Jul 15, 2008
    This also creates the texture using a luminance-alpha rather than RGBA
    source array.  I have no idea whether this is better/worse for
    performance on the graphics card side, but at least it eliminates a bit
    of Lisp code?
  7. @dtorop

    Test with GLUT rather than SDL

    dtorop committed with Jul 15, 2008
    Also adds two sorts of refresh, one which creates entirely new gl-text
    object, one which just redraws the screen.
Commits on Jul 14, 2008
  1. Unroll loops, was pointlessly allocating arrays before, also map-suba…

    committed Jul 14, 2008
    …rray was slow in this case.
  2. Preserve order on flushing the texture.

    committed Jul 14, 2008
    There is no particular reason to do that.
  3. Transpose coordinates, add one to maximum coordinate.

    committed Jul 14, 2008
    I don't really understand from where the transpose comes from, but it works this way. Changing it in
    transform-cell and send-texture alone does not.
    Add one to maximum coordinate to get 1.0 on maximum width.
Something went wrong with that request. Please try again.