Jonathon Fletcher
Basic gzip/gunzip in memory buffer examples using zlib code.
There are small changes needed to the zconf.h file in the zlib distribution (I used 1.2.7). The zlib license applies to the zlib code - I have only imported a subset of the source.
The MBED has limited memory, so we need the following (near the top of zconf.h) to restrict memory allocation sizes:
#define MAX_MEM_LEVEL 3 #define MAX_WBITS 10
Because MAX_MEM_LEVEL and MAX_WBITS are so much lower than the default, there is a danger that the mbed cannot gunzip data compressed by a 'normal' zlib build. My use-case is to gzip on the mbed more than gunzip on the mbed so I have not given much time to this issue.
I also included this (also near the top of zconf.h) to prefix defines with Z_
#define Z_PREFIX
In zconf.h, in the zlib distribution, the includes for <fcntl.h> and <sys/types.h> need commenting out when using the online compiler. No need when using GCC4MBED.
I also looked at miniz. I chose zlib because I needed the gzip headers and miniz does not implement them.
The sample main.cpp reads source data, compresses it, decompresses it, and finally compares the input data with the output data to confirm they are the same.
unsigned char input_data[2048];
unsigned long input_data_length = 0;
FILE *ifp = fopen("/local/src.txt", "r");
if (ifp) {
int br = fread(input_data, 1, sizeof(input_data), ifp);
fclose(ifp);
input_data_length = br;
}
printf("%s:%d: input_data_length:%lu%s", __FILE__, __LINE__, input_data_length, newline);
unsigned char gzip_data[2048];
unsigned long gzip_data_length = 0;
if (input_data_length > 0) {
gzip_data_length = sizeof(gzip_data);
int rv = gzip(gzip_data, &gzip_data_length, input_data, input_data_length);
if (Z_OK == rv) {
FILE *ofp = fopen("/local/dst.gz", "w");
if (ofp) {
int bw = fwrite(gzip_data, 1, gzip_data_length, ofp);
fclose(ofp);
}
} else {
printf("%s:%d: %d%s", __FILE__, __LINE__, rv, newline);
}
}
printf("%s:%d: gzip_data_length:%lu%s", __FILE__, __LINE__, gzip_data_length, newline);
unsigned char output_data[2048];
unsigned long output_data_length = 0;
if (gzip_data_length > 0) {
output_data_length = sizeof(output_data);
int rv = gunzip(output_data, &output_data_length, gzip_data, gzip_data_length);
if (Z_OK != rv) {
printf("%s:%d: %d%s", __FILE__, __LINE__, rv, newline);
}
}
printf("%s:%d: output_data_length:%lu%s", __FILE__, __LINE__, output_data_length, newline);
if (input_data_length > 0 and input_data_length > 0) {
bool input_matches_output = false;
if (input_data_length == output_data_length) {
input_matches_output = true;
for ( size_t i = 0 ; input_matches_output && i < input_data_length ; i++ ) {
if (input_data[i] != output_data[i]) {
input_matches_output = false;
}
}
}
printf("%s:%d: input (%lu bytes) %s output (%lu bytes)%s", __FILE__, __LINE__, input_data_length, input_matches_output?"matches":"does not match", output_data_length, newline);
} else {
printf("%s:%d: input and/or output length is 0%s", __FILE__, __LINE__, newline);
}
Diff: zlib/adler32.c
- Revision:
- 0:54f5be781526
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/zlib/adler32.c Sun Oct 21 07:46:41 2012 +0000
@@ -0,0 +1,179 @@
+/* adler32.c -- compute the Adler-32 checksum of a data stream
+ * Copyright (C) 1995-2011 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* @(#) $Id$ */
+
+#include "zutil.h"
+
+#define local static
+
+local uLong adler32_combine_ OF((uLong adler1, uLong adler2, z_off64_t len2));
+
+#define BASE 65521 /* largest prime smaller than 65536 */
+#define NMAX 5552
+/* NMAX is the largest n such that 255n(n+1)/2 + (n+1)(BASE-1) <= 2^32-1 */
+
+#define DO1(buf,i) {adler += (buf)[i]; sum2 += adler;}
+#define DO2(buf,i) DO1(buf,i); DO1(buf,i+1);
+#define DO4(buf,i) DO2(buf,i); DO2(buf,i+2);
+#define DO8(buf,i) DO4(buf,i); DO4(buf,i+4);
+#define DO16(buf) DO8(buf,0); DO8(buf,8);
+
+/* use NO_DIVIDE if your processor does not do division in hardware --
+ try it both ways to see which is faster */
+#ifdef NO_DIVIDE
+/* note that this assumes BASE is 65521, where 65536 % 65521 == 15
+ (thank you to John Reiser for pointing this out) */
+# define CHOP(a) \
+ do { \
+ unsigned long tmp = a >> 16; \
+ a &= 0xffffUL; \
+ a += (tmp << 4) - tmp; \
+ } while (0)
+# define MOD28(a) \
+ do { \
+ CHOP(a); \
+ if (a >= BASE) a -= BASE; \
+ } while (0)
+# define MOD(a) \
+ do { \
+ CHOP(a); \
+ MOD28(a); \
+ } while (0)
+# define MOD63(a) \
+ do { /* this assumes a is not negative */ \
+ z_off64_t tmp = a >> 32; \
+ a &= 0xffffffffL; \
+ a += (tmp << 8) - (tmp << 5) + tmp; \
+ tmp = a >> 16; \
+ a &= 0xffffL; \
+ a += (tmp << 4) - tmp; \
+ tmp = a >> 16; \
+ a &= 0xffffL; \
+ a += (tmp << 4) - tmp; \
+ if (a >= BASE) a -= BASE; \
+ } while (0)
+#else
+# define MOD(a) a %= BASE
+# define MOD28(a) a %= BASE
+# define MOD63(a) a %= BASE
+#endif
+
+/* ========================================================================= */
+uLong ZEXPORT adler32(adler, buf, len)
+ uLong adler;
+ const Bytef *buf;
+ uInt len;
+{
+ unsigned long sum2;
+ unsigned n;
+
+ /* split Adler-32 into component sums */
+ sum2 = (adler >> 16) & 0xffff;
+ adler &= 0xffff;
+
+ /* in case user likes doing a byte at a time, keep it fast */
+ if (len == 1) {
+ adler += buf[0];
+ if (adler >= BASE)
+ adler -= BASE;
+ sum2 += adler;
+ if (sum2 >= BASE)
+ sum2 -= BASE;
+ return adler | (sum2 << 16);
+ }
+
+ /* initial Adler-32 value (deferred check for len == 1 speed) */
+ if (buf == Z_NULL)
+ return 1L;
+
+ /* in case short lengths are provided, keep it somewhat fast */
+ if (len < 16) {
+ while (len--) {
+ adler += *buf++;
+ sum2 += adler;
+ }
+ if (adler >= BASE)
+ adler -= BASE;
+ MOD28(sum2); /* only added so many BASE's */
+ return adler | (sum2 << 16);
+ }
+
+ /* do length NMAX blocks -- requires just one modulo operation */
+ while (len >= NMAX) {
+ len -= NMAX;
+ n = NMAX / 16; /* NMAX is divisible by 16 */
+ do {
+ DO16(buf); /* 16 sums unrolled */
+ buf += 16;
+ } while (--n);
+ MOD(adler);
+ MOD(sum2);
+ }
+
+ /* do remaining bytes (less than NMAX, still just one modulo) */
+ if (len) { /* avoid modulos if none remaining */
+ while (len >= 16) {
+ len -= 16;
+ DO16(buf);
+ buf += 16;
+ }
+ while (len--) {
+ adler += *buf++;
+ sum2 += adler;
+ }
+ MOD(adler);
+ MOD(sum2);
+ }
+
+ /* return recombined sums */
+ return adler | (sum2 << 16);
+}
+
+/* ========================================================================= */
+local uLong adler32_combine_(adler1, adler2, len2)
+ uLong adler1;
+ uLong adler2;
+ z_off64_t len2;
+{
+ unsigned long sum1;
+ unsigned long sum2;
+ unsigned rem;
+
+ /* for negative len, return invalid adler32 as a clue for debugging */
+ if (len2 < 0)
+ return 0xffffffffUL;
+
+ /* the derivation of this formula is left as an exercise for the reader */
+ MOD63(len2); /* assumes len2 >= 0 */
+ rem = (unsigned)len2;
+ sum1 = adler1 & 0xffff;
+ sum2 = rem * sum1;
+ MOD(sum2);
+ sum1 += (adler2 & 0xffff) + BASE - 1;
+ sum2 += ((adler1 >> 16) & 0xffff) + ((adler2 >> 16) & 0xffff) + BASE - rem;
+ if (sum1 >= BASE) sum1 -= BASE;
+ if (sum1 >= BASE) sum1 -= BASE;
+ if (sum2 >= (BASE << 1)) sum2 -= (BASE << 1);
+ if (sum2 >= BASE) sum2 -= BASE;
+ return sum1 | (sum2 << 16);
+}
+
+/* ========================================================================= */
+uLong ZEXPORT adler32_combine(adler1, adler2, len2)
+ uLong adler1;
+ uLong adler2;
+ z_off_t len2;
+{
+ return adler32_combine_(adler1, adler2, len2);
+}
+
+uLong ZEXPORT adler32_combine64(adler1, adler2, len2)
+ uLong adler1;
+ uLong adler2;
+ z_off64_t len2;
+{
+ return adler32_combine_(adler1, adler2, len2);
+}