0
This repository has been archived on 2025-04-25. You can view files and clone it. You cannot open issues or pull requests or push a commit.
Files
Minecraft-Overviewer/overviewer_core/src/composite.c
Wunkolo 4298aea569 Fix GCC signed-unsigned and pointer-warnings (#1608)
* Fix GCC signed-unsigned and pointer-warnings

A lot of the signed/unsigned issues are related to the fact that I converted
a lot of indexing values to use unsigned types, little did I know that a lot of
other values used when indexing actually come from the python-end. Python does
not have built-in unsigned types so all integers coming from Python are signed
implicitly so a lot of things like image-size and x, y coordiantes are specially
handling negative-integer cases.

Guess we'll just take our `int32_t i = 0; i < blah; ++i` and like it.

Code now compiles with no warnings or nagging.
2019-07-12 15:26:45 +02:00

654 lines
20 KiB
C

/*
* This file is part of the Minecraft Overviewer.
*
* Minecraft Overviewer is free software: you can redistribute it and/or
* modify it under the terms of the GNU General Public License as published
* by the Free Software Foundation, either version 3 of the License, or (at
* your option) any later version.
*
* Minecraft Overviewer is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
* Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with the Overviewer. If not, see <http://www.gnu.org/licenses/>.
*/
/*
* This file implements a custom alpha_over function for (some) PIL
* images. It's designed to be used through composite.py, which
* includes a proxy alpha_over function that falls back to the default
* PIL paste if this extension is not found.
*/
#include "overviewer.h"
typedef struct {
PyObject_HEAD
Imaging image;
} ImagingObject;
inline Imaging
imaging_python_to_c(PyObject* obj) {
PyObject* im;
Imaging image;
/* first, get the 'im' attribute */
im = PyObject_GetAttrString(obj, "im");
if (!im)
return NULL;
/* make sure 'im' is the right type */
if (strcmp(im->ob_type->tp_name, "ImagingCore") != 0) {
/* it's not -- raise an error and exit */
PyErr_SetString(PyExc_TypeError,
"image attribute 'im' is not a core Imaging type");
return NULL;
}
image = ((ImagingObject*)im)->image;
Py_DECREF(im);
return image;
}
/* helper function to setup s{x,y}, d{x,y}, and {x,y}size variables
in these composite functions -- even handles auto-sizing to src! */
static inline void
setup_source_destination(Imaging src, Imaging dest,
int32_t* sx, int32_t* sy,
int32_t* dx, int32_t* dy,
int32_t* xsize, int32_t* ysize) {
/* handle negative/zero sizes appropriately */
if (*xsize <= 0 || *ysize <= 0) {
*xsize = src->xsize;
*ysize = src->ysize;
}
/* set up the source position, size and destination position */
/* handle negative dest pos */
if (*dx < 0) {
*sx = -(*dx);
*dx = 0;
} else {
*sx = 0;
}
if (*dy < 0) {
*sy = -(*dy);
*dy = 0;
} else {
*sy = 0;
}
/* set up source dimensions */
*xsize -= *sx;
*ysize -= *sy;
/* clip dimensions, if needed */
if (*dx + *xsize > dest->xsize)
*xsize = dest->xsize - *dx;
if (*dy + *ysize > dest->ysize)
*ysize = dest->ysize - *dy;
}
/* convenience alpha_over with 1.0 as overall_alpha */
inline PyObject* alpha_over(PyObject* dest, PyObject* src, PyObject* mask,
int32_t dx, int32_t dy, int32_t xsize, int32_t ysize) {
return alpha_over_full(dest, src, mask, 1.0f, dx, dy, xsize, ysize);
}
/* the full alpha_over function, in a form that can be called from C
* overall_alpha is multiplied with the whole mask, useful for lighting...
* if xsize, ysize are negative, they are instead set to the size of the image in src
* returns NULL on error, dest on success. You do NOT need to decref the return!
*/
inline PyObject*
alpha_over_full(PyObject* dest, PyObject* src, PyObject* mask, float overall_alpha,
int32_t dx, int32_t dy, int32_t xsize, int32_t ysize) {
/* libImaging handles */
Imaging imDest, imSrc, imMask;
/* cached blend properties */
int32_t src_has_alpha, mask_offset, mask_stride;
/* source position */
int32_t sx, sy;
/* iteration variables */
int32_t x, y;
uint32_t i;
/* temporary calculation variables */
int32_t tmp1, tmp2, tmp3;
/* integer [0, 255] version of overall_alpha */
UINT8 overall_alpha_int = 255 * overall_alpha;
/* short-circuit this whole thing if overall_alpha is zero */
if (overall_alpha_int == 0)
return dest;
imDest = imaging_python_to_c(dest);
imSrc = imaging_python_to_c(src);
imMask = imaging_python_to_c(mask);
if (!imDest || !imSrc || !imMask)
return NULL;
/* check the various image modes, make sure they make sense */
if (strcmp(imDest->mode, "RGBA") != 0) {
PyErr_SetString(PyExc_ValueError,
"given destination image does not have mode \"RGBA\"");
return NULL;
}
if (strcmp(imSrc->mode, "RGBA") != 0 && strcmp(imSrc->mode, "RGB") != 0) {
PyErr_SetString(PyExc_ValueError,
"given source image does not have mode \"RGBA\" or \"RGB\"");
return NULL;
}
if (strcmp(imMask->mode, "RGBA") != 0 && strcmp(imMask->mode, "L") != 0) {
PyErr_SetString(PyExc_ValueError,
"given mask image does not have mode \"RGBA\" or \"L\"");
return NULL;
}
/* make sure mask size matches src size */
if (imSrc->xsize != imMask->xsize || imSrc->ysize != imMask->ysize) {
PyErr_SetString(PyExc_ValueError,
"mask and source image sizes do not match");
return NULL;
}
/* set up flags for the src/mask type */
src_has_alpha = (imSrc->pixelsize == 4 ? 1 : 0);
/* how far into image the first alpha byte resides */
mask_offset = (imMask->pixelsize == 4 ? 3 : 0);
/* how many bytes to skip to get to the next alpha byte */
mask_stride = imMask->pixelsize;
/* setup source & destination vars */
setup_source_destination(imSrc, imDest, &sx, &sy, &dx, &dy, &xsize, &ysize);
/* check that there remains any blending to be done */
if (xsize <= 0 || ysize <= 0) {
/* nothing to do, return */
return dest;
}
for (y = 0; y < ysize; y++) {
UINT8* out = (UINT8*)imDest->image[dy + y] + dx * 4;
UINT8* outmask = (UINT8*)imDest->image[dy + y] + dx * 4 + 3;
UINT8* in = (UINT8*)imSrc->image[sy + y] + sx * (imSrc->pixelsize);
UINT8* inmask = (UINT8*)imMask->image[sy + y] + sx * mask_stride + mask_offset;
for (x = 0; x < xsize; x++) {
UINT8 in_alpha;
/* apply overall_alpha */
if (overall_alpha_int != 255 && *inmask != 0) {
in_alpha = OV_MULDIV255(*inmask, overall_alpha_int, tmp1);
} else {
in_alpha = *inmask;
}
/* special cases */
if (in_alpha == 255 || (*outmask == 0 && in_alpha > 0)) {
*outmask = in_alpha;
*out = *in;
out++, in++;
*out = *in;
out++, in++;
*out = *in;
out++, in++;
} else if (in_alpha == 0) {
/* do nothing -- source is fully transparent */
out += 3;
in += 3;
} else {
/* general case */
int32_t alpha = in_alpha + OV_MULDIV255(*outmask, 255 - in_alpha, tmp1);
for (i = 0; i < 3; i++) {
/* general case */
*out = OV_MULDIV255(*in, in_alpha, tmp1) +
OV_MULDIV255(OV_MULDIV255(*out, *outmask, tmp2), 255 - in_alpha, tmp3);
*out = (*out * 255) / alpha;
out++, in++;
}
*outmask = alpha;
}
out++;
if (src_has_alpha)
in++;
outmask += 4;
inmask += mask_stride;
}
}
return dest;
}
/* wraps alpha_over so it can be called directly from python */
/* properly refs the return value when needed: you DO need to decref the return */
PyObject*
alpha_over_wrap(PyObject* self, PyObject* args) {
/* raw input python variables */
PyObject *dest, *src, *pos = NULL, *mask = NULL;
/* destination position and size */
int32_t dx, dy, xsize, ysize;
/* return value: dest image on success */
PyObject* ret;
if (!PyArg_ParseTuple(args, "OO|OO", &dest, &src, &pos, &mask))
return NULL;
if (mask == NULL)
mask = src;
/* destination position read */
if (pos == NULL) {
xsize = 0;
ysize = 0;
dx = 0;
dy = 0;
} else {
if (!PyArg_ParseTuple(pos, "iiii", &dx, &dy, &xsize, &ysize)) {
/* try again, but this time try to read a point */
PyErr_Clear();
xsize = 0;
ysize = 0;
if (!PyArg_ParseTuple(pos, "ii", &dx, &dy)) {
PyErr_SetString(PyExc_TypeError,
"given blend destination rect is not valid");
return NULL;
}
}
}
ret = alpha_over(dest, src, mask, dx, dy, xsize, ysize);
if (ret == dest) {
/* Python needs us to own our return value */
Py_INCREF(dest);
}
return ret;
}
/* like alpha_over, but instead of src image it takes a source color
* also, it multiplies instead of doing an over operation
*/
PyObject*
tint_with_mask(PyObject* dest,
uint8_t sr, uint8_t sg, uint8_t sb, uint8_t sa,
PyObject* mask,
int32_t dx, int32_t dy,
int32_t xsize, int32_t ysize) {
/* libImaging handles */
Imaging imDest, imMask;
/* cached blend properties */
int32_t mask_offset, mask_stride;
/* source position */
int32_t sx, sy;
/* iteration variables */
int32_t x, y;
/* temporary calculation variables */
int32_t tmp1, tmp2;
imDest = imaging_python_to_c(dest);
imMask = imaging_python_to_c(mask);
if (!imDest || !imMask)
return NULL;
/* check the various image modes, make sure they make sense */
if (strcmp(imDest->mode, "RGBA") != 0) {
PyErr_SetString(PyExc_ValueError,
"given destination image does not have mode \"RGBA\"");
return NULL;
}
if (strcmp(imMask->mode, "RGBA") != 0 && strcmp(imMask->mode, "L") != 0) {
PyErr_SetString(PyExc_ValueError,
"given mask image does not have mode \"RGBA\" or \"L\"");
return NULL;
}
/* how far into image the first alpha byte resides */
mask_offset = (imMask->pixelsize == 4 ? 3 : 0);
/* how many bytes to skip to get to the next alpha byte */
mask_stride = imMask->pixelsize;
/* setup source & destination vars */
setup_source_destination(imMask, imDest, &sx, &sy, &dx, &dy, &xsize, &ysize);
/* check that there remains any blending to be done */
if (xsize <= 0 || ysize <= 0) {
/* nothing to do, return */
return dest;
}
for (y = 0; y < ysize; y++) {
UINT8* out = (UINT8*)imDest->image[dy + y] + dx * 4;
UINT8* inmask = (UINT8*)imMask->image[sy + y] + sx * mask_stride + mask_offset;
for (x = 0; x < xsize; x++) {
/* special cases */
if (*inmask == 255) {
*out = OV_MULDIV255(*out, sr, tmp1);
out++;
*out = OV_MULDIV255(*out, sg, tmp1);
out++;
*out = OV_MULDIV255(*out, sb, tmp1);
out++;
*out = OV_MULDIV255(*out, sa, tmp1);
out++;
} else if (*inmask == 0) {
/* do nothing -- source is fully transparent */
out += 4;
} else {
/* general case */
/* TODO work out general case */
*out = OV_MULDIV255(*out, (255 - *inmask) + OV_MULDIV255(sr, *inmask, tmp1), tmp2);
out++;
*out = OV_MULDIV255(*out, (255 - *inmask) + OV_MULDIV255(sg, *inmask, tmp1), tmp2);
out++;
*out = OV_MULDIV255(*out, (255 - *inmask) + OV_MULDIV255(sb, *inmask, tmp1), tmp2);
out++;
*out = OV_MULDIV255(*out, (255 - *inmask) + OV_MULDIV255(sa, *inmask, tmp1), tmp2);
out++;
}
inmask += mask_stride;
}
}
return dest;
}
/* draws a triangle on the destination image, multiplicatively!
* used for smooth lighting
* (excuse the ridiculous number of parameters!)
*
* Algorithm adapted from _Fundamentals_of_Computer_Graphics_
* by Peter Shirley, Michael Ashikhmin
* (or at least, the version poorly reproduced here:
* http://www.gidforums.com/t-20838.html )
*/
PyObject*
draw_triangle(PyObject* dest, int32_t inclusive,
int32_t x0, int32_t y0,
uint8_t r0, uint8_t g0, uint8_t b0,
int32_t x1, int32_t y1,
uint8_t r1, uint8_t g1, uint8_t b1,
int32_t x2, int32_t y2,
uint8_t r2, uint8_t g2, uint8_t b2,
int32_t tux, int32_t tuy,
int32_t* touchups, uint32_t num_touchups) {
/* destination image */
Imaging imDest;
/* ranges of pixels that are affected */
int32_t xmin, xmax, ymin, ymax;
/* constant coefficients for alpha, beta, gamma */
int32_t a12, a20, a01;
int32_t b12, b20, b01;
int32_t c12, c20, c01;
/* constant normalizers for alpha, beta, gamma */
float alpha_norm, beta_norm, gamma_norm;
/* temporary variables */
int32_t tmp;
/* iteration variables */
int32_t x, y;
imDest = imaging_python_to_c(dest);
if (!imDest)
return NULL;
/* check the various image modes, make sure they make sense */
if (strcmp(imDest->mode, "RGBA") != 0) {
PyErr_SetString(PyExc_ValueError,
"given destination image does not have mode \"RGBA\"");
return NULL;
}
/* set up draw ranges */
xmin = OV_MIN(x0, OV_MIN(x1, x2));
ymin = OV_MIN(y0, OV_MIN(y1, y2));
xmax = OV_MAX(x0, OV_MAX(x1, x2)) + 1;
ymax = OV_MAX(y0, OV_MAX(y1, y2)) + 1;
xmin = OV_MAX(xmin, 0);
ymin = OV_MAX(ymin, 0);
xmax = OV_MIN(xmax, imDest->xsize);
ymax = OV_MIN(ymax, imDest->ysize);
/* setup coefficients */
a12 = y1 - y2;
b12 = x2 - x1;
c12 = (x1 * y2) - (x2 * y1);
a20 = y2 - y0;
b20 = x0 - x2;
c20 = (x2 * y0) - (x0 * y2);
a01 = y0 - y1;
b01 = x1 - x0;
c01 = (x0 * y1) - (x1 * y0);
/* setup normalizers */
alpha_norm = 1.0f / ((a12 * x0) + (b12 * y0) + c12);
beta_norm = 1.0f / ((a20 * x1) + (b20 * y1) + c20);
gamma_norm = 1.0f / ((a01 * x2) + (b01 * y2) + c01);
/* iterate over the destination rect */
for (y = ymin; y < ymax; y++) {
UINT8* out = (UINT8*)imDest->image[y] + xmin * 4;
for (x = xmin; x < xmax; x++) {
float alpha, beta, gamma;
alpha = alpha_norm * ((a12 * x) + (b12 * y) + c12);
beta = beta_norm * ((a20 * x) + (b20 * y) + c20);
gamma = gamma_norm * ((a01 * x) + (b01 * y) + c01);
if (alpha >= 0 && beta >= 0 && gamma >= 0 &&
(inclusive || (alpha * beta * gamma > 0))) {
uint32_t r = alpha * r0 + beta * r1 + gamma * r2;
uint32_t g = alpha * g0 + beta * g1 + gamma * g2;
uint32_t b = alpha * b0 + beta * b1 + gamma * b2;
*out = OV_MULDIV255(*out, r, tmp);
out++;
*out = OV_MULDIV255(*out, g, tmp);
out++;
*out = OV_MULDIV255(*out, b, tmp);
out++;
/* keep alpha the same */
out++;
} else {
/* skip */
out += 4;
}
}
}
while (num_touchups > 0) {
float alpha, beta, gamma;
uint32_t r, g, b;
UINT8* out;
x = touchups[0] + tux;
y = touchups[1] + tuy;
touchups += 2;
num_touchups--;
if (x < 0 || x >= imDest->xsize || y < 0 || y >= imDest->ysize)
continue;
out = (UINT8*)imDest->image[y] + x * 4;
alpha = alpha_norm * ((a12 * x) + (b12 * y) + c12);
beta = beta_norm * ((a20 * x) + (b20 * y) + c20);
gamma = gamma_norm * ((a01 * x) + (b01 * y) + c01);
r = alpha * r0 + beta * r1 + gamma * r2;
g = alpha * g0 + beta * g1 + gamma * g2;
b = alpha * b0 + beta * b1 + gamma * b2;
*out = OV_MULDIV255(*out, r, tmp);
out++;
*out = OV_MULDIV255(*out, g, tmp);
out++;
*out = OV_MULDIV255(*out, b, tmp);
out++;
}
return dest;
}
/* scales the image to half size
*/
inline PyObject*
resize_half(PyObject* dest, PyObject* src) {
/* libImaging handles */
Imaging imDest, imSrc;
/* alpha properties */
int32_t src_has_alpha, dest_has_alpha;
/* iteration variables */
uint32_t x, y;
/* temp color variables */
uint32_t r, g, b, a;
/* size values for source and destination */
uint32_t src_width, src_height, dest_width, dest_height;
imDest = imaging_python_to_c(dest);
imSrc = imaging_python_to_c(src);
if (!imDest || !imSrc)
return NULL;
/* check the various image modes, make sure they make sense */
if (strcmp(imDest->mode, "RGBA") != 0) {
PyErr_SetString(PyExc_ValueError,
"given destination image does not have mode \"RGBA\"");
return NULL;
}
if (strcmp(imSrc->mode, "RGBA") != 0 && strcmp(imSrc->mode, "RGB") != 0) {
PyErr_SetString(PyExc_ValueError,
"given source image does not have mode \"RGBA\" or \"RGB\"");
return NULL;
}
src_width = imSrc->xsize;
src_height = imSrc->ysize;
dest_width = imDest->xsize;
dest_height = imDest->ysize;
/* make sure destination size is 1/2 src size */
if (src_width / 2 != dest_width || src_height / 2 != dest_height) {
PyErr_SetString(PyExc_ValueError,
"destination image size is not one-half source image size");
return NULL;
}
/* set up flags for the src/mask type */
src_has_alpha = (imSrc->pixelsize == 4 ? 1 : 0);
dest_has_alpha = (imDest->pixelsize == 4 ? 1 : 0);
/* check that there remains anything to resize */
if (dest_width <= 0 || dest_height <= 0) {
/* nothing to do, return */
return dest;
}
/* set to fully opaque if source has no alpha channel */
if (!src_has_alpha)
a = 0xFF << 2;
for (y = 0; y < dest_height; y++) {
UINT8* out = (UINT8*)imDest->image[y];
UINT8* in_row1 = (UINT8*)imSrc->image[y * 2];
UINT8* in_row2 = (UINT8*)imSrc->image[y * 2 + 1];
for (x = 0; x < dest_width; x++) {
// read first column
r = *in_row1;
r += *in_row2;
in_row1++;
in_row2++;
g = *in_row1;
g += *in_row2;
in_row1++;
in_row2++;
b = *in_row1;
b += *in_row2;
in_row1++;
in_row2++;
if (src_has_alpha) {
a = *in_row1;
a += *in_row2;
in_row1++;
in_row2++;
}
// read second column
r += *in_row1;
r += *in_row2;
in_row1++;
in_row2++;
g += *in_row1;
g += *in_row2;
in_row1++;
in_row2++;
b += *in_row1;
b += *in_row2;
in_row1++;
in_row2++;
if (src_has_alpha) {
a += *in_row1;
a += *in_row2;
in_row1++;
in_row2++;
}
// write blended color
*out = (UINT8)(r >> 2);
out++;
*out = (UINT8)(g >> 2);
out++;
*out = (UINT8)(b >> 2);
out++;
if (dest_has_alpha) {
*out = (UINT8)(a >> 2);
out++;
}
}
}
return dest;
}
/* wraps resize_half so it can be called directly from python */
PyObject*
resize_half_wrap(PyObject* self, PyObject* args) {
/* raw input python variables */
PyObject *dest, *src;
/* return value: dest image on success */
PyObject* ret;
if (!PyArg_ParseTuple(args, "OO", &dest, &src))
return NULL;
ret = resize_half(dest, src);
if (ret == dest) {
/* Python needs us to own our return value */
Py_INCREF(dest);
}
return ret;
}