xref: /aosp_15_r20/external/zstd/zlibWrapper/examples/fitblk.c (revision 01826a4963a0d8a59bc3812d29bdf0fb76416722)
1*01826a49SYabin Cui /* fitblk.c contains minimal changes required to be compiled with zlibWrapper:
2*01826a49SYabin Cui  * - #include "zlib.h" was changed to #include "zstd_zlibwrapper.h"
3*01826a49SYabin Cui  * - writing block to stdout was disabled                          */
4*01826a49SYabin Cui 
5*01826a49SYabin Cui /* fitblk.c: example of fitting compressed output to a specified size
6*01826a49SYabin Cui    Not copyrighted -- provided to the public domain
7*01826a49SYabin Cui    Version 1.1  25 November 2004  Mark Adler */
8*01826a49SYabin Cui 
9*01826a49SYabin Cui /* Version history:
10*01826a49SYabin Cui    1.0  24 Nov 2004  First version
11*01826a49SYabin Cui    1.1  25 Nov 2004  Change deflateInit2() to deflateInit()
12*01826a49SYabin Cui                      Use fixed-size, stack-allocated raw buffers
13*01826a49SYabin Cui                      Simplify code moving compression to subroutines
14*01826a49SYabin Cui                      Use assert() for internal errors
15*01826a49SYabin Cui                      Add detailed description of approach
16*01826a49SYabin Cui  */
17*01826a49SYabin Cui 
18*01826a49SYabin Cui /* Approach to just fitting a requested compressed size:
19*01826a49SYabin Cui 
20*01826a49SYabin Cui    fitblk performs three compression passes on a portion of the input
21*01826a49SYabin Cui    data in order to determine how much of that input will compress to
22*01826a49SYabin Cui    nearly the requested output block size.  The first pass generates
23*01826a49SYabin Cui    enough deflate blocks to produce output to fill the requested
24*01826a49SYabin Cui    output size plus a specified excess amount (see the EXCESS define
25*01826a49SYabin Cui    below).  The last deflate block may go quite a bit past that, but
26*01826a49SYabin Cui    is discarded.  The second pass decompresses and recompresses just
27*01826a49SYabin Cui    the compressed data that fit in the requested plus excess sized
28*01826a49SYabin Cui    buffer.  The deflate process is terminated after that amount of
29*01826a49SYabin Cui    input, which is less than the amount consumed on the first pass.
30*01826a49SYabin Cui    The last deflate block of the result will be of a comparable size
31*01826a49SYabin Cui    to the final product, so that the header for that deflate block and
32*01826a49SYabin Cui    the compression ratio for that block will be about the same as in
33*01826a49SYabin Cui    the final product.  The third compression pass decompresses the
34*01826a49SYabin Cui    result of the second step, but only the compressed data up to the
35*01826a49SYabin Cui    requested size minus an amount to allow the compressed stream to
36*01826a49SYabin Cui    complete (see the MARGIN define below).  That will result in a
37*01826a49SYabin Cui    final compressed stream whose length is less than or equal to the
38*01826a49SYabin Cui    requested size.  Assuming sufficient input and a requested size
39*01826a49SYabin Cui    greater than a few hundred bytes, the shortfall will typically be
40*01826a49SYabin Cui    less than ten bytes.
41*01826a49SYabin Cui 
42*01826a49SYabin Cui    If the input is short enough that the first compression completes
43*01826a49SYabin Cui    before filling the requested output size, then that compressed
44*01826a49SYabin Cui    stream is return with no recompression.
45*01826a49SYabin Cui 
46*01826a49SYabin Cui    EXCESS is chosen to be just greater than the shortfall seen in a
47*01826a49SYabin Cui    two pass approach similar to the above.  That shortfall is due to
48*01826a49SYabin Cui    the last deflate block compressing more efficiently with a smaller
49*01826a49SYabin Cui    header on the second pass.  EXCESS is set to be large enough so
50*01826a49SYabin Cui    that there is enough uncompressed data for the second pass to fill
51*01826a49SYabin Cui    out the requested size, and small enough so that the final deflate
52*01826a49SYabin Cui    block of the second pass will be close in size to the final deflate
53*01826a49SYabin Cui    block of the third and final pass.  MARGIN is chosen to be just
54*01826a49SYabin Cui    large enough to assure that the final compression has enough room
55*01826a49SYabin Cui    to complete in all cases.
56*01826a49SYabin Cui  */
57*01826a49SYabin Cui 
58*01826a49SYabin Cui #include <stdio.h>
59*01826a49SYabin Cui #include <stdlib.h>
60*01826a49SYabin Cui #include <assert.h>
61*01826a49SYabin Cui #include "zstd_zlibwrapper.h"
62*01826a49SYabin Cui 
63*01826a49SYabin Cui #define LOG_FITBLK(...)   /*printf(__VA_ARGS__)*/
64*01826a49SYabin Cui #define local static
65*01826a49SYabin Cui 
66*01826a49SYabin Cui /* print nastygram and leave */
quit(char * why)67*01826a49SYabin Cui local void quit(char *why)
68*01826a49SYabin Cui {
69*01826a49SYabin Cui     fprintf(stderr, "fitblk abort: %s\n", why);
70*01826a49SYabin Cui     exit(1);
71*01826a49SYabin Cui }
72*01826a49SYabin Cui 
73*01826a49SYabin Cui #define RAWLEN 4096    /* intermediate uncompressed buffer size */
74*01826a49SYabin Cui 
75*01826a49SYabin Cui /* compress from file to def until provided buffer is full or end of
76*01826a49SYabin Cui    input reached; return last deflate() return value, or Z_ERRNO if
77*01826a49SYabin Cui    there was read error on the file */
partcompress(FILE * in,z_streamp def)78*01826a49SYabin Cui local int partcompress(FILE *in, z_streamp def)
79*01826a49SYabin Cui {
80*01826a49SYabin Cui     int ret, flush;
81*01826a49SYabin Cui     unsigned char raw[RAWLEN];
82*01826a49SYabin Cui 
83*01826a49SYabin Cui     flush = Z_SYNC_FLUSH;
84*01826a49SYabin Cui     do {
85*01826a49SYabin Cui         def->avail_in = (uInt)fread(raw, 1, RAWLEN, in);
86*01826a49SYabin Cui         if (ferror(in))
87*01826a49SYabin Cui             return Z_ERRNO;
88*01826a49SYabin Cui         def->next_in = raw;
89*01826a49SYabin Cui         if (feof(in))
90*01826a49SYabin Cui             flush = Z_FINISH;
91*01826a49SYabin Cui         LOG_FITBLK("partcompress1 avail_in=%d total_in=%d avail_out=%d total_out=%d\n", (int)def->avail_in, (int)def->total_in, (int)def->avail_out, (int)def->total_out);
92*01826a49SYabin Cui         ret = deflate(def, flush);
93*01826a49SYabin Cui         LOG_FITBLK("partcompress2 ret=%d avail_in=%d total_in=%d avail_out=%d total_out=%d\n", ret, (int)def->avail_in, (int)def->total_in, (int)def->avail_out, (int)def->total_out);
94*01826a49SYabin Cui         assert(ret != Z_STREAM_ERROR);
95*01826a49SYabin Cui     } while (def->avail_out != 0 && flush == Z_SYNC_FLUSH);
96*01826a49SYabin Cui     return ret;
97*01826a49SYabin Cui }
98*01826a49SYabin Cui 
99*01826a49SYabin Cui /* recompress from inf's input to def's output; the input for inf and
100*01826a49SYabin Cui    the output for def are set in those structures before calling;
101*01826a49SYabin Cui    return last deflate() return value, or Z_MEM_ERROR if inflate()
102*01826a49SYabin Cui    was not able to allocate enough memory when it needed to */
recompress(z_streamp inf,z_streamp def)103*01826a49SYabin Cui local int recompress(z_streamp inf, z_streamp def)
104*01826a49SYabin Cui {
105*01826a49SYabin Cui     int ret, flush;
106*01826a49SYabin Cui     unsigned char raw[RAWLEN];
107*01826a49SYabin Cui 
108*01826a49SYabin Cui     flush = Z_NO_FLUSH;
109*01826a49SYabin Cui     LOG_FITBLK("recompress start\n");
110*01826a49SYabin Cui     do {
111*01826a49SYabin Cui         /* decompress */
112*01826a49SYabin Cui         inf->avail_out = RAWLEN;
113*01826a49SYabin Cui         inf->next_out = raw;
114*01826a49SYabin Cui         LOG_FITBLK("recompress1inflate avail_in=%d total_in=%d avail_out=%d total_out=%d\n", (int)inf->avail_in, (int)inf->total_in, (int)inf->avail_out, (int)inf->total_out);
115*01826a49SYabin Cui         ret = inflate(inf, Z_NO_FLUSH);
116*01826a49SYabin Cui         LOG_FITBLK("recompress2inflate avail_in=%d total_in=%d avail_out=%d total_out=%d\n", (int)inf->avail_in, (int)inf->total_in, (int)inf->avail_out, (int)inf->total_out);
117*01826a49SYabin Cui         assert(ret != Z_STREAM_ERROR && ret != Z_DATA_ERROR &&
118*01826a49SYabin Cui                ret != Z_NEED_DICT);
119*01826a49SYabin Cui         if (ret == Z_MEM_ERROR)
120*01826a49SYabin Cui             return ret;
121*01826a49SYabin Cui 
122*01826a49SYabin Cui         /* compress what was decompressed until done or no room */
123*01826a49SYabin Cui         def->avail_in = RAWLEN - inf->avail_out;
124*01826a49SYabin Cui         def->next_in = raw;
125*01826a49SYabin Cui         if (inf->avail_out != 0)
126*01826a49SYabin Cui             flush = Z_FINISH;
127*01826a49SYabin Cui         LOG_FITBLK("recompress1deflate avail_in=%d total_in=%d avail_out=%d total_out=%d\n", (int)def->avail_in, (int)def->total_in, (int)def->avail_out, (int)def->total_out);
128*01826a49SYabin Cui         ret = deflate(def, flush);
129*01826a49SYabin Cui         LOG_FITBLK("recompress2deflate ret=%d avail_in=%d total_in=%d avail_out=%d total_out=%d\n", ret, (int)def->avail_in, (int)def->total_in, (int)def->avail_out, (int)def->total_out);
130*01826a49SYabin Cui         assert(ret != Z_STREAM_ERROR);
131*01826a49SYabin Cui     } while (ret != Z_STREAM_END && def->avail_out != 0);
132*01826a49SYabin Cui     return ret;
133*01826a49SYabin Cui }
134*01826a49SYabin Cui 
135*01826a49SYabin Cui #define EXCESS 256      /* empirically determined stream overage */
136*01826a49SYabin Cui #define MARGIN 8        /* amount to back off for completion */
137*01826a49SYabin Cui 
138*01826a49SYabin Cui /* compress from stdin to fixed-size block on stdout */
main(int argc,char ** argv)139*01826a49SYabin Cui int main(int argc, char **argv)
140*01826a49SYabin Cui {
141*01826a49SYabin Cui     int ret;                /* return code */
142*01826a49SYabin Cui     unsigned size;          /* requested fixed output block size */
143*01826a49SYabin Cui     unsigned have;          /* bytes written by deflate() call */
144*01826a49SYabin Cui     unsigned char *blk;     /* intermediate and final stream */
145*01826a49SYabin Cui     unsigned char *tmp;     /* close to desired size stream */
146*01826a49SYabin Cui     z_stream def, inf;      /* zlib deflate and inflate states */
147*01826a49SYabin Cui 
148*01826a49SYabin Cui     /* get requested output size */
149*01826a49SYabin Cui     if (argc != 2)
150*01826a49SYabin Cui         quit("need one argument: size of output block");
151*01826a49SYabin Cui     ret = (int)strtol(argv[1], argv + 1, 10);
152*01826a49SYabin Cui     if (argv[1][0] != 0)
153*01826a49SYabin Cui         quit("argument must be a number");
154*01826a49SYabin Cui     if (ret < 8)            /* 8 is minimum zlib stream size */
155*01826a49SYabin Cui         quit("need positive size of 8 or greater");
156*01826a49SYabin Cui     size = (unsigned)ret;
157*01826a49SYabin Cui 
158*01826a49SYabin Cui     printf("zlib version %s\n", ZLIB_VERSION);
159*01826a49SYabin Cui     if (ZWRAP_isUsingZSTDcompression()) printf("zstd version %s\n", zstdVersion());
160*01826a49SYabin Cui 
161*01826a49SYabin Cui     /* allocate memory for buffers and compression engine */
162*01826a49SYabin Cui     blk = (unsigned char*)malloc(size + EXCESS);
163*01826a49SYabin Cui     def.zalloc = Z_NULL;
164*01826a49SYabin Cui     def.zfree = Z_NULL;
165*01826a49SYabin Cui     def.opaque = Z_NULL;
166*01826a49SYabin Cui     ret = deflateInit(&def, Z_DEFAULT_COMPRESSION);
167*01826a49SYabin Cui     if (ret != Z_OK || blk == NULL)
168*01826a49SYabin Cui         quit("out of memory");
169*01826a49SYabin Cui 
170*01826a49SYabin Cui     /* compress from stdin until output full, or no more input */
171*01826a49SYabin Cui     def.avail_out = size + EXCESS;
172*01826a49SYabin Cui     def.next_out = blk;
173*01826a49SYabin Cui     LOG_FITBLK("partcompress1 total_in=%d total_out=%d\n", (int)def.total_in, (int)def.total_out);
174*01826a49SYabin Cui     ret = partcompress(stdin, &def);
175*01826a49SYabin Cui     printf("partcompress total_in=%d total_out=%d\n", (int)def.total_in, (int)def.total_out);
176*01826a49SYabin Cui     if (ret == Z_ERRNO)
177*01826a49SYabin Cui         quit("error reading input");
178*01826a49SYabin Cui 
179*01826a49SYabin Cui     /* if it all fit, then size was undersubscribed -- done! */
180*01826a49SYabin Cui     if (ret == Z_STREAM_END && def.avail_out >= EXCESS) {
181*01826a49SYabin Cui         /* write block to stdout */
182*01826a49SYabin Cui         have = size + EXCESS - def.avail_out;
183*01826a49SYabin Cui    /*     if (fwrite(blk, 1, have, stdout) != have || ferror(stdout))
184*01826a49SYabin Cui     *         quit("error writing output"); */
185*01826a49SYabin Cui 
186*01826a49SYabin Cui         /* clean up and print results to stderr */
187*01826a49SYabin Cui         ret = deflateEnd(&def);
188*01826a49SYabin Cui         assert(ret != Z_STREAM_ERROR);
189*01826a49SYabin Cui         free(blk);
190*01826a49SYabin Cui         fprintf(stderr,
191*01826a49SYabin Cui                 "%u bytes unused out of %u requested (all input)\n",
192*01826a49SYabin Cui                 size - have, size);
193*01826a49SYabin Cui         return 0;
194*01826a49SYabin Cui     }
195*01826a49SYabin Cui 
196*01826a49SYabin Cui     /* it didn't all fit -- set up for recompression */
197*01826a49SYabin Cui     inf.zalloc = Z_NULL;
198*01826a49SYabin Cui     inf.zfree = Z_NULL;
199*01826a49SYabin Cui     inf.opaque = Z_NULL;
200*01826a49SYabin Cui     inf.avail_in = 0;
201*01826a49SYabin Cui     inf.next_in = Z_NULL;
202*01826a49SYabin Cui     ret = inflateInit(&inf);
203*01826a49SYabin Cui     tmp = (unsigned char*)malloc(size + EXCESS);
204*01826a49SYabin Cui     if (ret != Z_OK || tmp == NULL)
205*01826a49SYabin Cui         quit("out of memory");
206*01826a49SYabin Cui     ret = deflateReset(&def);
207*01826a49SYabin Cui     assert(ret != Z_STREAM_ERROR);
208*01826a49SYabin Cui 
209*01826a49SYabin Cui     /* do first recompression close to the right amount */
210*01826a49SYabin Cui     inf.avail_in = size + EXCESS;
211*01826a49SYabin Cui     inf.next_in = blk;
212*01826a49SYabin Cui     def.avail_out = size + EXCESS;
213*01826a49SYabin Cui     def.next_out = tmp;
214*01826a49SYabin Cui     LOG_FITBLK("recompress1 inf.total_in=%d def.total_out=%d\n", (int)inf.total_in, (int)def.total_out);
215*01826a49SYabin Cui     ret = recompress(&inf, &def);
216*01826a49SYabin Cui     LOG_FITBLK("recompress1 inf.total_in=%d def.total_out=%d\n", (int)inf.total_in, (int)def.total_out);
217*01826a49SYabin Cui     if (ret == Z_MEM_ERROR)
218*01826a49SYabin Cui         quit("out of memory");
219*01826a49SYabin Cui 
220*01826a49SYabin Cui     /* set up for next recompression */
221*01826a49SYabin Cui     ret = inflateReset(&inf);
222*01826a49SYabin Cui     assert(ret != Z_STREAM_ERROR);
223*01826a49SYabin Cui     ret = deflateReset(&def);
224*01826a49SYabin Cui     assert(ret != Z_STREAM_ERROR);
225*01826a49SYabin Cui 
226*01826a49SYabin Cui     /* do second and final recompression (third compression) */
227*01826a49SYabin Cui     inf.avail_in = size - MARGIN;   /* assure stream will complete */
228*01826a49SYabin Cui     inf.next_in = tmp;
229*01826a49SYabin Cui     def.avail_out = size;
230*01826a49SYabin Cui     def.next_out = blk;
231*01826a49SYabin Cui     LOG_FITBLK("recompress2 inf.total_in=%d def.total_out=%d\n", (int)inf.total_in, (int)def.total_out);
232*01826a49SYabin Cui     ret = recompress(&inf, &def);
233*01826a49SYabin Cui     LOG_FITBLK("recompress2 inf.total_in=%d def.total_out=%d\n", (int)inf.total_in, (int)def.total_out);
234*01826a49SYabin Cui     if (ret == Z_MEM_ERROR)
235*01826a49SYabin Cui         quit("out of memory");
236*01826a49SYabin Cui     assert(ret == Z_STREAM_END);    /* otherwise MARGIN too small */
237*01826a49SYabin Cui 
238*01826a49SYabin Cui     /* done -- write block to stdout */
239*01826a49SYabin Cui     have = size - def.avail_out;
240*01826a49SYabin Cui     /* if (fwrite(blk, 1, have, stdout) != have || ferror(stdout))
241*01826a49SYabin Cui      *     quit("error writing output"); */
242*01826a49SYabin Cui 
243*01826a49SYabin Cui     /* clean up and print results to stderr */
244*01826a49SYabin Cui     free(tmp);
245*01826a49SYabin Cui     ret = inflateEnd(&inf);
246*01826a49SYabin Cui     assert(ret != Z_STREAM_ERROR);
247*01826a49SYabin Cui     ret = deflateEnd(&def);
248*01826a49SYabin Cui     assert(ret != Z_STREAM_ERROR);
249*01826a49SYabin Cui     free(blk);
250*01826a49SYabin Cui     fprintf(stderr,
251*01826a49SYabin Cui             "%u bytes unused out of %u requested (%lu input)\n",
252*01826a49SYabin Cui             size - have, size, def.total_in);
253*01826a49SYabin Cui     return 0;
254*01826a49SYabin Cui }
255