/* wrapper functions that implement the floating point sample version of a
function in terms of the 8-bit sample version
*/
-extern int i_ppixf_fp(i_img *im, int x, int y, const i_fcolor *pix);
-extern int i_gpixf_fp(i_img *im, int x, int y, i_fcolor *pix);
-extern int i_plinf_fp(i_img *im, int l, int r, int y, const i_fcolor *pix);
-extern int i_glinf_fp(i_img *im, int l, int r, int y, i_fcolor *pix);
-extern int i_gsampf_fp(i_img *im, int l, int r, int y, i_fsample_t *samp,
+extern int i_ppixf_fp(i_img *im, i_img_dim x, i_img_dim y, const i_fcolor *pix);
+extern int i_gpixf_fp(i_img *im, i_img_dim x, i_img_dim y, i_fcolor *pix);
+extern i_img_dim i_plinf_fp(i_img *im, i_img_dim l, i_img_dim r, i_img_dim y, const i_fcolor *pix);
+extern i_img_dim i_glinf_fp(i_img *im, i_img_dim l, i_img_dim r, i_img_dim y, i_fcolor *pix);
+extern i_img_dim i_gsampf_fp(i_img *im, i_img_dim l, i_img_dim r, i_img_dim y, i_fsample_t *samp,
int const *chans, int chan_count);
/* wrapper functions that forward palette calls to the underlying image,
extern int i_setcolors_forward(i_img *im, int index, const i_color *colors,
int count);
-#define SampleFTo16(num) ((int)((num) * 65535.0 + 0.01))
+/* fallback handler for gsamp_bits */
+extern i_img_dim i_gsamp_bits_fb(i_img *im, i_img_dim x, i_img_dim r, i_img_dim y, unsigned *samp,
+ const int *chans, int chan_count, int bits);
+
+#define SampleFTo16(num) ((int)((num) * 65535.0 + 0.5))
/* we add that little bit to avoid rounding issues */
#define Sample16ToF(num) ((num) / 65535.0)
-#define SampleFTo8(num) ((int)((num) * 255.0 + 0.01))
+#define SampleFTo8(num) ((int)((num) * 255.0 + 0.5))
#define Sample8ToF(num) ((num) / 255.0)
-#define Sample16To8(num) ((num) / 257)
+#define Sample16To8(num) (((num)+128) / 257)
#define Sample8To16(num) ((num) * 257)
extern void i_get_combine(int combine, i_fill_combine_f *, i_fill_combinef_f *);
extern UTIL_table_t i_UTIL_table;
-/* Ideally this will move into imconfig.h if we ever probe */
-#if defined(_GNU_SOURCE) || __STDC_VERSION__ >= 199901L
-/* snprintf() is part of C99 and provided by Glibc */
-#define HAVE_SNPRINTF
-#endif
-
/* test if all channels are writable */
#define I_ALL_CHANNELS_WRITABLE(im) (((im)->ch_mask & 0xF) == 0xf)
typedef struct i_int_hline_seg_tag {
- int minx, x_limit;
+ i_img_dim minx, x_limit;
} i_int_hline_seg;
typedef struct i_int_hline_entry_tag {
- int count;
- int alloc;
+ i_img_dim count;
+ size_t alloc;
i_int_hline_seg segs[1];
} i_int_hline_entry;
/* represents a set of horizontal line segments to be filled in later */
typedef struct i_int_hlines_tag {
- int start_y, limit_y;
- int start_x, limit_x;
+ i_img_dim start_y, limit_y;
+ i_img_dim start_x, limit_x;
i_int_hline_entry **entries;
} i_int_hlines;
extern void
i_int_init_hlines(
i_int_hlines *hlines,
- int start_y,
- int count_y,
- int start_x,
- int width_x
+ i_img_dim start_y,
+ i_img_dim count_y,
+ i_img_dim start_x,
+ i_img_dim width_x
);
extern void i_int_init_hlines_img(i_int_hlines *hlines, i_img *img);
-extern void i_int_hlines_add(i_int_hlines *hlines, int y, int minx, int width);
+extern void i_int_hlines_add(i_int_hlines *hlines, i_img_dim y, i_img_dim minx, i_img_dim width);
extern void i_int_hlines_destroy(i_int_hlines *hlines);
extern void i_int_hlines_fill_color(i_img *im, i_int_hlines *hlines, const i_color *val);
#define I_LIMIT_8(x) ((x) < 0 ? 0 : (x) > 255 ? 255 : (x))
#define I_LIMIT_DOUBLE(x) ((x) < 0.0 ? 0.0 : (x) > 1.0 ? 1.0 : (x))
+#define IM_STRING(x) #x
+
+/* I considered using assert.h here, but perl does it's own thing with
+ assert() and the NDEBUG test is opposite to the direction I prefer */
+#ifdef IM_ASSERT
+extern void im_assert_fail(char const *, int, char const *);
+#define im_assert(x) ((x) ? (void)(0) : im_assert_fail(__FILE__, __LINE__, IM_STRING(x)))
+#else
+#define im_assert(x) (void)(0)
+#endif
+
+i_img_dim i_minx(i_img_dim a, i_img_dim b);
+i_img_dim i_maxx(i_img_dim x, i_img_dim y);
+i_img_dim i_abs(i_img_dim x);
+
+#define i_min(a, b) i_minx((a), (b))
+#define i_max(a, b) i_maxx((a), (b))
+
+#define color_to_grey(col) ((col)->rgb.r * 0.222 + (col)->rgb.g * 0.707 + (col)->rgb.b * 0.071)
+
#endif