* Export CRT parameters * This must also be implemented if CRT is not used, for being able to * write DER encoded RSA keys. The helper function mbed

* output = SHA-1( input buffer ) */

* Small divisors test (X must be positive) * * Return values: * 0: no small factor (possible prime, more tests needed) * 1: certain prime * MBEDT

* Convert an ASCII character to digit value */

* Miller-Rabin pseudo-primality test (HAC 4.24) */

* Decrypt with AES-XXX-CBC, using PBKDF1 for key derivation */

* Read a 16-byte hex string and convert it to binary */

* RSAPublicKey ::= SEQUENCE { * modulus INTEGER, -- n * publicExponent INTEGER -- e * } */

* Parse a PKCS#1 encoded private RSA key */

* Parse an unencrypted PKCS#8 encoded private key * * Notes: * * - This function does not own the key buffer. It is the * responsibility of the

* Parse a SEC1 encoded private EC key */

* Use EC parameters to initialise an EC group * * ECParameters ::= CHOICE { * namedCurve OBJECT IDENTIFIER * specifiedCurve Specif

* Put a finder pattern. * @param frame * @param width * @param ox,oy upper-left coordinate of the pattern */

you can both convert from vector to buffer&size and vica versa. If you use init_buffer to take over a buffer and size, it is not needed to use cleanup

* http://www.faqs.org/rfcs/rfc2781.html */

Since an uncompressed block can be max 65535 in size, it actually adds multible blocks if needed. */

Writes DHT and SOS marker segments to out and fills in DC/AC Huffman tables for each component of the image.

Computes 8x8 FFT of each channel of xyb0 and xyb1 and adds the total squared 3-dimensional xybdiff of the two blocks to diff_xyb_{dc,ac} and the avera

Same as CalculateBlockSymbolSize, but for block size smaller than histogram size. */

Gives the exact size of the tree, in bits, as it will be encoded in DEFLATE. */

Returns 32-bit integer value for RGBA color.

Compares a[0..kBlockSize) and b[0..kBlockSize) vectors, and returns 0 : if they are equal -1 : if a is everywhere <= than b and in at least one coordi

Computes out[x] = sum{kIDCTMatrix[8*x+u]*in[u*stride]; for u in [0..7]}

Computes a horizontal convolution and transposes the result.

convolve horizontally and vertically with 1D kernel

Counts amount of colors in the image, up to 257. If transparent_counts_as_one is enabled, any color with alpha channel 0 is treated as a single color

This function will create a Huffman tree. The catch here is that the tree cannot be arbitrarily deep. Brotli specifies a maximum depth of 15 bits for

Sets *pos to the next stream position where parsing should continue. Returns false if the stream ended too early.

Calculates the bit lengths for the symbols for dynamic blocks. Chooses bit lengths that give the smallest size of tree encoding + encoding of all the

Gets a score of the length given the distance. Typically, the score of the length is the length itself, but if the distance is very long, decrease the

Appends the symbol statistics from the store. */

Mimic libjpeg's heuristics to guess jpeg color space. Requires that the jpg has 3 components.

Returns the DC diff or AC value for extra bits value x and prefix code s. See Tables F.1 and F.2 of the spec.

Second step for the ...makeFromLengths and ...makeFromFrequencies functions. numcodes, lengths and maxbitlen must already be filled in correctly. retu

Does a single run for ZopfliLZ77Optimal. For good compression, repeated runs with updated statistics should be performed. s: the block state in: the i

TODO(user) Use linear->SRGB conversion and a lookup-table.

Changes the population counts in a way that the consequent Huffman tree compression, especially its rle-part, will be more likely to compress this dat

Ensures there are at least 2 distance codes to support buggy decoders. Zlib 1.2.1 and below have a bug where it fails if there isn't at least 1 distan

Do the sharpening to the v channel, but only in areas where it will help channel should be 2 for v sharpening, or 1 for less effective u sharpening