1 |
/* gzjoin -- command to join gzip files into one gzip file |
2 |
|
3 |
Copyright (C) 2004, 2005, 2012 Mark Adler, all rights reserved |
4 |
version 1.2, 14 Aug 2012 |
5 |
|
6 |
This software is provided 'as-is', without any express or implied |
7 |
warranty. In no event will the author be held liable for any damages |
8 |
arising from the use of this software. |
9 |
|
10 |
Permission is granted to anyone to use this software for any purpose, |
11 |
including commercial applications, and to alter it and redistribute it |
12 |
freely, subject to the following restrictions: |
13 |
|
14 |
1. The origin of this software must not be misrepresented; you must not |
15 |
claim that you wrote the original software. If you use this software |
16 |
in a product, an acknowledgment in the product documentation would be |
17 |
appreciated but is not required. |
18 |
2. Altered source versions must be plainly marked as such, and must not be |
19 |
misrepresented as being the original software. |
20 |
3. This notice may not be removed or altered from any source distribution. |
21 |
|
22 |
Mark Adler madler@alumni.caltech.edu |
23 |
*/ |
24 |
|
25 |
/* |
26 |
* Change history: |
27 |
* |
28 |
* 1.0 11 Dec 2004 - First version |
29 |
* 1.1 12 Jun 2005 - Changed ssize_t to long for portability |
30 |
* 1.2 14 Aug 2012 - Clean up for z_const usage |
31 |
*/ |
32 |
|
33 |
/* |
34 |
gzjoin takes one or more gzip files on the command line and writes out a |
35 |
single gzip file that will uncompress to the concatenation of the |
36 |
uncompressed data from the individual gzip files. gzjoin does this without |
37 |
having to recompress any of the data and without having to calculate a new |
38 |
crc32 for the concatenated uncompressed data. gzjoin does however have to |
39 |
decompress all of the input data in order to find the bits in the compressed |
40 |
data that need to be modified to concatenate the streams. |
41 |
|
42 |
gzjoin does not do an integrity check on the input gzip files other than |
43 |
checking the gzip header and decompressing the compressed data. They are |
44 |
otherwise assumed to be complete and correct. |
45 |
|
46 |
Each joint between gzip files removes at least 18 bytes of previous trailer |
47 |
and subsequent header, and inserts an average of about three bytes to the |
48 |
compressed data in order to connect the streams. The output gzip file |
49 |
has a minimal ten-byte gzip header with no file name or modification time. |
50 |
|
51 |
This program was written to illustrate the use of the Z_BLOCK option of |
52 |
inflate() and the crc32_combine() function. gzjoin will not compile with |
53 |
versions of zlib earlier than 1.2.3. |
54 |
*/ |
55 |
|
56 |
#include <stdio.h> /* fputs(), fprintf(), fwrite(), putc() */ |
57 |
#include <stdlib.h> /* exit(), malloc(), free() */ |
58 |
#include <fcntl.h> /* open() */ |
59 |
#include <unistd.h> /* close(), read(), lseek() */ |
60 |
#include "zlib.h" |
61 |
/* crc32(), crc32_combine(), inflateInit2(), inflate(), inflateEnd() */ |
62 |
|
63 |
#define local static |
64 |
|
65 |
/* exit with an error (return a value to allow use in an expression) */ |
66 |
local int bail(char *why1, char *why2) |
67 |
{ |
68 |
fprintf(stderr, "gzjoin error: %s%s, output incomplete\n", why1, why2); |
69 |
exit(1); |
70 |
return 0; |
71 |
} |
72 |
|
73 |
/* -- simple buffered file input with access to the buffer -- */ |
74 |
|
75 |
#define CHUNK 32768 /* must be a power of two and fit in unsigned */ |
76 |
|
77 |
/* bin buffered input file type */ |
78 |
typedef struct { |
79 |
char *name; /* name of file for error messages */ |
80 |
int fd; /* file descriptor */ |
81 |
unsigned left; /* bytes remaining at next */ |
82 |
unsigned char *next; /* next byte to read */ |
83 |
unsigned char *buf; /* allocated buffer of length CHUNK */ |
84 |
} bin; |
85 |
|
86 |
/* close a buffered file and free allocated memory */ |
87 |
local void bclose(bin *in) |
88 |
{ |
89 |
if (in != NULL) { |
90 |
if (in->fd != -1) |
91 |
close(in->fd); |
92 |
if (in->buf != NULL) |
93 |
free(in->buf); |
94 |
free(in); |
95 |
} |
96 |
} |
97 |
|
98 |
/* open a buffered file for input, return a pointer to type bin, or NULL on |
99 |
failure */ |
100 |
local bin *bopen(char *name) |
101 |
{ |
102 |
bin *in; |
103 |
|
104 |
in = malloc(sizeof(bin)); |
105 |
if (in == NULL) |
106 |
return NULL; |
107 |
in->buf = malloc(CHUNK); |
108 |
in->fd = open(name, O_RDONLY, 0); |
109 |
if (in->buf == NULL || in->fd == -1) { |
110 |
bclose(in); |
111 |
return NULL; |
112 |
} |
113 |
in->left = 0; |
114 |
in->next = in->buf; |
115 |
in->name = name; |
116 |
return in; |
117 |
} |
118 |
|
119 |
/* load buffer from file, return -1 on read error, 0 or 1 on success, with |
120 |
1 indicating that end-of-file was reached */ |
121 |
local int bload(bin *in) |
122 |
{ |
123 |
long len; |
124 |
|
125 |
if (in == NULL) |
126 |
return -1; |
127 |
if (in->left != 0) |
128 |
return 0; |
129 |
in->next = in->buf; |
130 |
do { |
131 |
len = (long)read(in->fd, in->buf + in->left, CHUNK - in->left); |
132 |
if (len < 0) |
133 |
return -1; |
134 |
in->left += (unsigned)len; |
135 |
} while (len != 0 && in->left < CHUNK); |
136 |
return len == 0 ? 1 : 0; |
137 |
} |
138 |
|
139 |
/* get a byte from the file, bail if end of file */ |
140 |
#define bget(in) (in->left ? 0 : bload(in), \ |
141 |
in->left ? (in->left--, *(in->next)++) : \ |
142 |
bail("unexpected end of file on ", in->name)) |
143 |
|
144 |
/* get a four-byte little-endian unsigned integer from file */ |
145 |
local unsigned long bget4(bin *in) |
146 |
{ |
147 |
unsigned long val; |
148 |
|
149 |
val = bget(in); |
150 |
val += (unsigned long)(bget(in)) << 8; |
151 |
val += (unsigned long)(bget(in)) << 16; |
152 |
val += (unsigned long)(bget(in)) << 24; |
153 |
return val; |
154 |
} |
155 |
|
156 |
/* skip bytes in file */ |
157 |
local void bskip(bin *in, unsigned skip) |
158 |
{ |
159 |
/* check pointer */ |
160 |
if (in == NULL) |
161 |
return; |
162 |
|
163 |
/* easy case -- skip bytes in buffer */ |
164 |
if (skip <= in->left) { |
165 |
in->left -= skip; |
166 |
in->next += skip; |
167 |
return; |
168 |
} |
169 |
|
170 |
/* skip what's in buffer, discard buffer contents */ |
171 |
skip -= in->left; |
172 |
in->left = 0; |
173 |
|
174 |
/* seek past multiples of CHUNK bytes */ |
175 |
if (skip > CHUNK) { |
176 |
unsigned left; |
177 |
|
178 |
left = skip & (CHUNK - 1); |
179 |
if (left == 0) { |
180 |
/* exact number of chunks: seek all the way minus one byte to check |
181 |
for end-of-file with a read */ |
182 |
lseek(in->fd, skip - 1, SEEK_CUR); |
183 |
if (read(in->fd, in->buf, 1) != 1) |
184 |
bail("unexpected end of file on ", in->name); |
185 |
return; |
186 |
} |
187 |
|
188 |
/* skip the integral chunks, update skip with remainder */ |
189 |
lseek(in->fd, skip - left, SEEK_CUR); |
190 |
skip = left; |
191 |
} |
192 |
|
193 |
/* read more input and skip remainder */ |
194 |
bload(in); |
195 |
if (skip > in->left) |
196 |
bail("unexpected end of file on ", in->name); |
197 |
in->left -= skip; |
198 |
in->next += skip; |
199 |
} |
200 |
|
201 |
/* -- end of buffered input functions -- */ |
202 |
|
203 |
/* skip the gzip header from file in */ |
204 |
local void gzhead(bin *in) |
205 |
{ |
206 |
int flags; |
207 |
|
208 |
/* verify gzip magic header and compression method */ |
209 |
if (bget(in) != 0x1f || bget(in) != 0x8b || bget(in) != 8) |
210 |
bail(in->name, " is not a valid gzip file"); |
211 |
|
212 |
/* get and verify flags */ |
213 |
flags = bget(in); |
214 |
if ((flags & 0xe0) != 0) |
215 |
bail("unknown reserved bits set in ", in->name); |
216 |
|
217 |
/* skip modification time, extra flags, and os */ |
218 |
bskip(in, 6); |
219 |
|
220 |
/* skip extra field if present */ |
221 |
if (flags & 4) { |
222 |
unsigned len; |
223 |
|
224 |
len = bget(in); |
225 |
len += (unsigned)(bget(in)) << 8; |
226 |
bskip(in, len); |
227 |
} |
228 |
|
229 |
/* skip file name if present */ |
230 |
if (flags & 8) |
231 |
while (bget(in) != 0) |
232 |
; |
233 |
|
234 |
/* skip comment if present */ |
235 |
if (flags & 16) |
236 |
while (bget(in) != 0) |
237 |
; |
238 |
|
239 |
/* skip header crc if present */ |
240 |
if (flags & 2) |
241 |
bskip(in, 2); |
242 |
} |
243 |
|
244 |
/* write a four-byte little-endian unsigned integer to out */ |
245 |
local void put4(unsigned long val, FILE *out) |
246 |
{ |
247 |
putc(val & 0xff, out); |
248 |
putc((val >> 8) & 0xff, out); |
249 |
putc((val >> 16) & 0xff, out); |
250 |
putc((val >> 24) & 0xff, out); |
251 |
} |
252 |
|
253 |
/* Load up zlib stream from buffered input, bail if end of file */ |
254 |
local void zpull(z_streamp strm, bin *in) |
255 |
{ |
256 |
if (in->left == 0) |
257 |
bload(in); |
258 |
if (in->left == 0) |
259 |
bail("unexpected end of file on ", in->name); |
260 |
strm->avail_in = in->left; |
261 |
strm->next_in = in->next; |
262 |
} |
263 |
|
264 |
/* Write header for gzip file to out and initialize trailer. */ |
265 |
local void gzinit(unsigned long *crc, unsigned long *tot, FILE *out) |
266 |
{ |
267 |
fwrite("\x1f\x8b\x08\0\0\0\0\0\0\xff", 1, 10, out); |
268 |
*crc = crc32(0L, Z_NULL, 0); |
269 |
*tot = 0; |
270 |
} |
271 |
|
272 |
/* Copy the compressed data from name, zeroing the last block bit of the last |
273 |
block if clr is true, and adding empty blocks as needed to get to a byte |
274 |
boundary. If clr is false, then the last block becomes the last block of |
275 |
the output, and the gzip trailer is written. crc and tot maintains the |
276 |
crc and length (modulo 2^32) of the output for the trailer. The resulting |
277 |
gzip file is written to out. gzinit() must be called before the first call |
278 |
of gzcopy() to write the gzip header and to initialize crc and tot. */ |
279 |
local void gzcopy(char *name, int clr, unsigned long *crc, unsigned long *tot, |
280 |
FILE *out) |
281 |
{ |
282 |
int ret; /* return value from zlib functions */ |
283 |
int pos; /* where the "last block" bit is in byte */ |
284 |
int last; /* true if processing the last block */ |
285 |
bin *in; /* buffered input file */ |
286 |
unsigned char *start; /* start of compressed data in buffer */ |
287 |
unsigned char *junk; /* buffer for uncompressed data -- discarded */ |
288 |
z_off_t len; /* length of uncompressed data (support > 4 GB) */ |
289 |
z_stream strm; /* zlib inflate stream */ |
290 |
|
291 |
/* open gzip file and skip header */ |
292 |
in = bopen(name); |
293 |
if (in == NULL) |
294 |
bail("could not open ", name); |
295 |
gzhead(in); |
296 |
|
297 |
/* allocate buffer for uncompressed data and initialize raw inflate |
298 |
stream */ |
299 |
junk = malloc(CHUNK); |
300 |
strm.zalloc = Z_NULL; |
301 |
strm.zfree = Z_NULL; |
302 |
strm.opaque = Z_NULL; |
303 |
strm.avail_in = 0; |
304 |
strm.next_in = Z_NULL; |
305 |
ret = inflateInit2(&strm, -15); |
306 |
if (junk == NULL || ret != Z_OK) |
307 |
bail("out of memory", ""); |
308 |
|
309 |
/* inflate and copy compressed data, clear last-block bit if requested */ |
310 |
len = 0; |
311 |
zpull(&strm, in); |
312 |
start = in->next; |
313 |
last = start[0] & 1; |
314 |
if (last && clr) |
315 |
start[0] &= ~1; |
316 |
strm.avail_out = 0; |
317 |
for (;;) { |
318 |
/* if input used and output done, write used input and get more */ |
319 |
if (strm.avail_in == 0 && strm.avail_out != 0) { |
320 |
fwrite(start, 1, strm.next_in - start, out); |
321 |
start = in->buf; |
322 |
in->left = 0; |
323 |
zpull(&strm, in); |
324 |
} |
325 |
|
326 |
/* decompress -- return early when end-of-block reached */ |
327 |
strm.avail_out = CHUNK; |
328 |
strm.next_out = junk; |
329 |
ret = inflate(&strm, Z_BLOCK); |
330 |
switch (ret) { |
331 |
case Z_MEM_ERROR: |
332 |
bail("out of memory", ""); |
333 |
case Z_DATA_ERROR: |
334 |
bail("invalid compressed data in ", in->name); |
335 |
} |
336 |
|
337 |
/* update length of uncompressed data */ |
338 |
len += CHUNK - strm.avail_out; |
339 |
|
340 |
/* check for block boundary (only get this when block copied out) */ |
341 |
if (strm.data_type & 128) { |
342 |
/* if that was the last block, then done */ |
343 |
if (last) |
344 |
break; |
345 |
|
346 |
/* number of unused bits in last byte */ |
347 |
pos = strm.data_type & 7; |
348 |
|
349 |
/* find the next last-block bit */ |
350 |
if (pos != 0) { |
351 |
/* next last-block bit is in last used byte */ |
352 |
pos = 0x100 >> pos; |
353 |
last = strm.next_in[-1] & pos; |
354 |
if (last && clr) |
355 |
in->buf[strm.next_in - in->buf - 1] &= ~pos; |
356 |
} |
357 |
else { |
358 |
/* next last-block bit is in next unused byte */ |
359 |
if (strm.avail_in == 0) { |
360 |
/* don't have that byte yet -- get it */ |
361 |
fwrite(start, 1, strm.next_in - start, out); |
362 |
start = in->buf; |
363 |
in->left = 0; |
364 |
zpull(&strm, in); |
365 |
} |
366 |
last = strm.next_in[0] & 1; |
367 |
if (last && clr) |
368 |
in->buf[strm.next_in - in->buf] &= ~1; |
369 |
} |
370 |
} |
371 |
} |
372 |
|
373 |
/* update buffer with unused input */ |
374 |
in->left = strm.avail_in; |
375 |
in->next = in->buf + (strm.next_in - in->buf); |
376 |
|
377 |
/* copy used input, write empty blocks to get to byte boundary */ |
378 |
pos = strm.data_type & 7; |
379 |
fwrite(start, 1, in->next - start - 1, out); |
380 |
last = in->next[-1]; |
381 |
if (pos == 0 || !clr) |
382 |
/* already at byte boundary, or last file: write last byte */ |
383 |
putc(last, out); |
384 |
else { |
385 |
/* append empty blocks to last byte */ |
386 |
last &= ((0x100 >> pos) - 1); /* assure unused bits are zero */ |
387 |
if (pos & 1) { |
388 |
/* odd -- append an empty stored block */ |
389 |
putc(last, out); |
390 |
if (pos == 1) |
391 |
putc(0, out); /* two more bits in block header */ |
392 |
fwrite("\0\0\xff\xff", 1, 4, out); |
393 |
} |
394 |
else { |
395 |
/* even -- append 1, 2, or 3 empty fixed blocks */ |
396 |
switch (pos) { |
397 |
case 6: |
398 |
putc(last | 8, out); |
399 |
last = 0; |
400 |
case 4: |
401 |
putc(last | 0x20, out); |
402 |
last = 0; |
403 |
case 2: |
404 |
putc(last | 0x80, out); |
405 |
putc(0, out); |
406 |
} |
407 |
} |
408 |
} |
409 |
|
410 |
/* update crc and tot */ |
411 |
*crc = crc32_combine(*crc, bget4(in), len); |
412 |
*tot += (unsigned long)len; |
413 |
|
414 |
/* clean up */ |
415 |
inflateEnd(&strm); |
416 |
free(junk); |
417 |
bclose(in); |
418 |
|
419 |
/* write trailer if this is the last gzip file */ |
420 |
if (!clr) { |
421 |
put4(*crc, out); |
422 |
put4(*tot, out); |
423 |
} |
424 |
} |
425 |
|
426 |
/* join the gzip files on the command line, write result to stdout */ |
427 |
int main(int argc, char **argv) |
428 |
{ |
429 |
unsigned long crc, tot; /* running crc and total uncompressed length */ |
430 |
|
431 |
/* skip command name */ |
432 |
argc--; |
433 |
argv++; |
434 |
|
435 |
/* show usage if no arguments */ |
436 |
if (argc == 0) { |
437 |
fputs("gzjoin usage: gzjoin f1.gz [f2.gz [f3.gz ...]] > fjoin.gz\n", |
438 |
stderr); |
439 |
return 0; |
440 |
} |
441 |
|
442 |
/* join gzip files on command line and write to stdout */ |
443 |
gzinit(&crc, &tot, stdout); |
444 |
while (argc--) |
445 |
gzcopy(*argv++, argc, &crc, &tot, stdout); |
446 |
|
447 |
/* done */ |
448 |
return 0; |
449 |
} |