Lines Matching full:but
17 appreciated but is not required.
44 #pragma warning( disable : 4996 ) /*VS does not like fopen, but fopen_s is not standard C so unusab…
100 /* inline is not available in C90, but use it when supported by the compiler */
107 /* restrict is not available in C90, but use it when supported by the compiler */
739 …/*reverse bits, because the huffman bits are given in MSB first order but the bit reader reads LSB… in HuffmanTree_makeTable()
777 but deflate uses 1 bit instead. In case of 0 symbols, no symbols can in HuffmanTree_makeTable()
778 appear at all, but such huffman tree could still exist (e.g. if distance in HuffmanTree_makeTable()
993 symbol, it'd give it the theoretical 0 bits but in practice zlib wants 1 bit*/ in lodepng_huffman_code_lengths()
1084 /*there are 32 distance codes, but 30-31 are unused*/ in generateFixedDistanceTree()
1491 static const unsigned HASH_BIT_MASK = 65535; /*HASH_NUM_VALUES - 1, but C90 does not like that as i…
1499 /*TODO: do this not only for zeros but for any repeated byte. However for PNG
1501 int* headz; /*similar to head, but for chainz*/
1852 /* could fit on stack, but >1KB is on the larger side so allocate instead */ in deflateDynamic()
1872 …for(i = datapos; i < dataend; ++i) lz77_encoded.data[i - datapos] = data[i]; /*no LZ77, but still … in deflateDynamic()
1957 /*trim zeros at the end (using CLCL_ORDER), but minimum size must be 4 (see HCLEN below)*/ in deflateDynamic()
2057 } else /*no LZ77, but still will be Huffman compressed*/ { in deflateFixed()
2586 /*Sets length and name and allocates the space for data and crc but does not
2608 /* like lodepng_chunk_create but with custom allocsize */
2688 This is an error according to the PNG spec, but common PNG decoders make it black instead. in lodepng_color_mode_alloc_palette()
2813 …/* Ignoring casts, the expression is equal to (w * bpp + 7) / 8 + 1, but avoids overflow of w * bp… in lodepng_get_raw_size_idat()
2821 you can safely compute in a size_t (but not an unsigned):
3004 /* same as set but does not delete */
3110 index for a color. It's like an octree, but because the alpha channel is used too, each
3327 /*Similar to getPixelColorRGBA8, but with all the for loops inside of the color
3426 /*Similar to getPixelColorsRGBA8, but with 3-channel RGB output.*/
3500 given color type, but the given color type must be 16-bit itself.*/
3549 /*if the user specified output palette but did not give the values, assume in lodepng_convert()
3607 any palette index but doesn't have an alpha channel. Idem with ignoring color key. */
3881 all pixels of an image but only for a few additional values. */
3974 The parameters are of type short, but should come from unsigned chars, the shorts
4000 reduced image when without filter bytes but with padded scanlines
4002 image without padding between scanlines, but still padding between the images
4146 …nroll independent paths of the paeth predictor. A 6x and 8x version would also be possible but that in unfilterScanline()
4207 …in and out are allowed to be the same memory address (but aren't the same size since in has the ex… in unfilter()
4232 in: Adam7 interlaced image, with no padding bits between scanlines, but between
4234 out: the same pixels, but re-ordered so that they're now a non-interlaced image with size w*h
4288 in and out are allowed to be the same buffer, in may also be higher but still overlapping; in must in removePaddingBits()
5476 but for "the other case", whatever strategy filter_strategy is set to instead of the minimum sum in filter()
5518 This means filtertype 0 is almost never chosen, but that is justified.*/ in filter()
5601 better result with dynamic tree anyway. Using the fixed tree sometimes gives worse, but in rare in filter()
5662 out: the same pixels, but re-ordered according to PNG's Adam7 interlacing, with
5663 no padding bits between scanlines, but between reduced images so that each
6028 …/* Could use strcmp, but we're not calling or reimplementing this C library function for this use … in lodepng_encode()