added thick lines and gradient fill support

[tobozo]

Features in this PR:

• Antialiased thick lines, filled with plain color or gradient
• Antialiased circles, filled with plain color or gradient
• Rectangular fill with linear or radial gradient

LGFX Methods:

NOTE: COLOR can be any of uint8_t, uint16_t, uint32_t, RGBColor, rgb888_t, rgb565_t, rgb888_t, etc

• Draw thick+smooth lines:
  void drawWideLine(int32_t x0, int32_t y0, int32_t x1, int32_t y1, float thick, COLOR color)

• Draw thick+smooth lines (gradient fill):
  void drawWideLine(int32_t x0, int32_t y0, int32_t x1, int32_t y1, float thick, const colors_t gradient)

• Draw wedge line:
  void drawWedgeLine(int32_t x0, int32_t y0, int32_t x1, int32_t y1, float thick0, float thick1, COLOR color)

• Draw wedge line (gradient fill):
  void drawWedgeLine(int32_t x0, int32_t y0, int32_t x1, int32_t y1, float thick0, float thick1, const colors_t gradient)

• Fill smooth circle with single color (redundant with fillSmoothCircle(), kept for benchmark):
  void drawSpot(int32_t x, int32_t y, float radius, COLOR fg_color)

• Fill smooth circle, like fillSmoothCircle() but with gradient fill:
  void drawGradientSpot(int32_t x, int32_t y, float radius, const colors_t gradient)

• Draw antialiased line:
  void drawSmoothLine( LovyanGFX* gfx, int32_t x0, int32_t y0, int32_t x1, int32_t y1, COLOR color )

• Draw gradient line (not smoothed, used for filling):
  void drawGradientLine(int32_t x0, int32_t y0, int32_t x1, int32_t y1, const colors_t gradient )
void drawGradientHLine(int32_t x, int32_t y, uint32_t w, const colors_t gradient )
void drawGradientVLine(int32_t x, int32_t y, uint32_t h, const colors_t gradient )

• Fill rect with radial/linear gradient (fill_style_t can be any of RADIAL, HLINEAR or VLINEAR):
  void fillGradientRect(int32_t x, int32_t y, uint32_t w, uint32_t h, COLOR start, COLOR end, fill_style_t style=RADIAL )
void fillGradientRect(int32_t x, int32_t y, uint32_t w, uint32_t h, const colors_t gradient, fill_style_t style=RADIAL )

• Create a gradient item from a sized array of colors or from a pointer to an array of colors:
  colors_t createGradient( rgb888_t[] colors )
colors_t createGradient( rgb888_t* colors, uint32_t count )

• Example:

    lgfx::rgb888_t colors[] = { {0x00,0xff,0x00}, {0xff,0xff,0x00}, {0xff,0x80,0x00}, {0xff,0x00,0x00} };
    auto gradient = tft.createGradient( colors ); // create a gradient item
    tft.fillGradientRect( 10, 10, 200, 200, gradient, HLINEAR );

[lovyan03]

thanks !
There are places where double type is used for coordinate calculation, but wouldn't float be sufficient?
I don't think such exact precision is required.

[tobozo]

fixed, apparently this is also provides a small performance improvement with radial gradients

[tobozo]

btw @lovyan03 do you want me to submit that to M5GFX too ?