source: zzuf/trunk/test/zzcat.c @ 4121

Last change on this file since 4121 was 4121, checked in by Sam Hocevar, 11 years ago

Rewrite zzcat to make it a lot more configurable.

  • Property svn:keywords set to Id
File size: 13.4 KB
Line 
1/*
2 *  zzcat - various cat reimplementations for testing purposes
3 *  Copyright (c) 2006-2009 Sam Hocevar <sam@hocevar.net>
4 *                All Rights Reserved
5 *
6 *  $Id: zzcat.c 4121 2009-12-14 02:55:31Z sam $
7 *
8 *  This program is free software. It comes without any warranty, to
9 *  the extent permitted by applicable law. You can redistribute it
10 *  and/or modify it under the terms of the Do What The Fuck You Want
11 *  To Public License, Version 2, as published by Sam Hocevar. See
12 *  http://sam.zoy.org/wtfpl/COPYING for more details.
13 */
14
15/*
16 * TODO: fsetpos64, fgetln
17 */
18
19#include "config.h"
20
21/* Needed for lseek64() */
22#define _LARGEFILE64_SOURCE
23/* Needed for O_RDONLY on HP-UX */
24#define _INCLUDE_POSIX_SOURCE
25/* Needed for fgets_unlocked() */
26#define _GNU_SOURCE
27
28#if defined HAVE_STDINT_H
29#   include <stdint.h>
30#elif defined HAVE_INTTYPES_H
31#   include <inttypes.h>
32#endif
33#include <sys/types.h>
34#include <sys/stat.h>
35#include <fcntl.h>
36#if defined HAVE_UNISTD_H
37#   include <unistd.h>
38#endif
39#if defined HAVE_SYS_MMAN_H
40#   include <sys/mman.h>
41#endif
42#include <stdlib.h>
43#include <stdio.h>
44#include <string.h>
45
46static inline unsigned int myrand(void)
47{
48    static int seed = 1;
49    int x, y;
50    x = (seed + 0x12345678) << 11;
51    y = (seed + 0xfedcba98) >> 21;
52    seed = x * 1010101 + y * 343434;
53    return seed;
54}
55
56#define FOPEN(cmd) \
57    do { \
58        cmd; \
59        if (!f) \
60        { \
61            fprintf(stderr, "E: zzcat: cannot open `%s'\n", file); \
62            return EXIT_FAILURE; \
63        } \
64        retoff = 0; \
65        p = strchr(p, ')') + 1; \
66    } while(0)
67
68#define FCLOSE(cmd) \
69    do { \
70        cmd; \
71        f = NULL; \
72        p = strchr(p, ')') + 1; \
73    } while(0)
74
75#define MERGE(address, cnt, off) \
76    do { \
77        size_t _cnt = cnt, _off = off; \
78        if (_cnt && retoff + _cnt > retlen) \
79        { \
80            retlen = retoff + _cnt; \
81            retbuf = realloc(retbuf, retlen); \
82        } \
83        if (_cnt > 0) \
84            memcpy(retbuf + retoff, address, _cnt); \
85        retoff += _off; \
86    } while(0)
87
88#define FREAD(cmd, buf, cnt) FCALL(cmd, buf, cnt, cnt)
89#define FSEEK(cmd, off) FCALL(cmd, /* unused */ "", 0, off)
90
91#define FCALL(cmd, buf, cnt, off) \
92    do { \
93        if (!f) \
94        { \
95            f = fopen(file, "r"); \
96            if (!f) \
97            { \
98                fprintf(stderr, "E: zzcat: cannot open `%s'\n", file); \
99                return EXIT_FAILURE; \
100            } \
101        } \
102        cmd; \
103        MERGE(buf, cnt, off); \
104        p = strchr(p, ')') + 1; \
105    } while(0)
106
107/*
108 * Command parser. We rewrite fmt by replacing the last character with
109 * '%c' and check that the sscanf() call returns the expected number of
110 * matches plus one (for the last character). We use this macro trick to
111 * avoid using vsscanf() which does not exist on all platforms.
112 */
113
114struct parser
115{
116    char tmpfmt[1024], ch, lastch;
117};
118
119static int make_fmt(struct parser *p, char const *fmt)
120{
121    char const *tmp;
122    size_t len;
123    int ret = 0;
124
125    len = strlen(fmt);
126    p->lastch = fmt[len - 1];
127
128    memcpy(p->tmpfmt, fmt, len - 1);
129    p->tmpfmt[len - 1] = '%';
130    p->tmpfmt[len] = 'c';
131    p->tmpfmt[len + 1] = '\0';
132
133    for (tmp = p->tmpfmt; *tmp; tmp++)
134        if (*tmp == '%')
135            tmp++, ret++;
136
137    return ret;
138}
139
140#define PARSECMD(fmt, arg...) \
141    make_fmt(&parser, fmt) == sscanf(p, parser.tmpfmt, ##arg, &parser.ch) \
142        && parser.ch == parser.lastch
143
144/*
145 * File reader. We parse a command line and perform all the operations it
146 * contains on the specified file.
147 */
148
149static int cat_file(char const *p, char const *file)
150{
151    struct { char const *p; int count; } loops[128];
152    char *retbuf = NULL, *tmp;
153    FILE *f = NULL;
154    size_t retlen = 0, retoff = 0;
155    int nloops = 0, fd = -1;
156
157    /* Allocate 32MB for our temporary buffer. Any larger value will crash. */
158    tmp = malloc(32 * 1024 * 1024);
159
160    while (*p)
161    {
162        struct parser parser;
163        long int l1, l2;
164        char *s, *lineptr = NULL;
165        size_t k;
166        ssize_t l;
167        int n;
168        char ch;
169
170        /* Ignore punctuation */
171        if (strchr(" \t,;\r\n", *p))
172            p++;
173
174        /* Loop handling */
175        else if (PARSECMD("repeat ( %li ,", &l1))
176        {
177            p = strchr(p, ',') + 1;
178            loops[nloops].p = p;
179            loops[nloops].count = l1;
180            nloops++;
181        }
182        else if (PARSECMD(")"))
183        {
184            if (nloops == 0)
185            {
186                fprintf(stderr, "E: zzcat: ')' outside a loop\n");
187                return EXIT_FAILURE;
188            }
189            loops[nloops - 1].count--;
190            if (loops[nloops - 1].count <= 0)
191            {
192                nloops--;
193                p = strchr(p, ')') + 1;
194            }
195            else
196            {
197                p = loops[nloops - 1].p;
198            }
199        }
200
201        /* FILE * opening functions */
202        else if (PARSECMD("fopen ( )"))
203            FOPEN(f = fopen(file, "r"));
204#if defined HAVE_FOPEN64
205        else if (PARSECMD("fopen64 ( )"))
206            FOPEN(f = fopen64(file, "r"));
207#endif
208#if defined HAVE___FOPEN64
209        else if (PARSECMD("__fopen64 ( )"))
210            FOPEN(f = __fopen64(file, "r"));
211#endif
212        else if (PARSECMD("freopen ( )"))
213            FOPEN(f = freopen(file, "r", f));
214#if defined HAVE_FREOPEN64
215        else if (PARSECMD("freopen64 ( )"))
216            FOPEN(f = freopen64(file, "r", f));
217#endif
218#if defined HAVE___FREOPEN64
219        else if (PARSECMD("__freopen64 ( )"))
220            FOPEN(f = __freopen64(file, "r", f));
221#endif
222
223        /* FILE * closing functions */
224        else if (PARSECMD("fclose ( )"))
225            FCLOSE(fclose(f));
226
227        /* FILE * reading functions */
228        else if (PARSECMD("fread ( %li , %li )", &l1, &l2))
229            FREAD(l = fread(tmp, l1, l2, f), tmp, l > 0 ? l * l1 : 0);
230        else if (PARSECMD("getc ( )"))
231            FREAD(ch = (n = getc(f)), &ch, (n != EOF));
232        else if (PARSECMD("fgetc ( )"))
233            FREAD(ch = (n = fgetc(f)), &ch, (n != EOF));
234        else if (PARSECMD("fgets ( %li )", &l1))
235            FREAD(s = fgets(tmp, l1, f), tmp, s ? strlen(tmp) : 0);
236#if defined HAVE__IO_GETC
237        else if (PARSECMD("_IO_getc ( )"))
238            FREAD(ch = (n = _IO_getc(f)), &ch, (n != EOF));
239#endif
240#if defined HAVE_FREAD_UNLOCKED
241        else if (PARSECMD("fread_unlocked ( %li , %li )", &l1, &l2))
242            FREAD(l = fread_unlocked(tmp, l1, l2, f), tmp, l > 0 ? l * l1 : 0);
243#endif
244#if defined HAVE_FGETS_UNLOCKED
245        else if (PARSECMD("fgets_unlocked ( %li )", &l1))
246            FREAD(s = fgets_unlocked(tmp, l1, f), tmp, s ? strlen(tmp) : 0);
247#endif
248#if defined HAVE_GETC_UNLOCKED
249        else if (PARSECMD("getc_unlocked ( )"))
250            FREAD(ch = (n = getc_unlocked(f)), &ch, (n != EOF));
251#endif
252#if defined HAVE_FGETC_UNLOCKED
253        else if (PARSECMD("fgetc_unlocked ( )"))
254            FREAD(ch = (n = fgetc_unlocked(f)), &ch, (n != EOF));
255#endif
256
257        /* FILE * getdelim functions */
258#if defined HAVE_GETLINE
259        else if (PARSECMD("getline ( )"))
260            FREAD(l = getline(&lineptr, &k, f), lineptr, l >= 0 ? l : 0);
261#endif
262#if defined HAVE_GETDELIM
263        else if (PARSECMD("getdelim ( '%c' )", &ch))
264            FREAD(l = getdelim(&lineptr, &k, ch, f), lineptr, l >= 0 ? l : 0);
265        else if (PARSECMD("getdelim ( %i )", &n))
266            FREAD(l = getdelim(&lineptr, &k, n, f), lineptr, l >= 0 ? l : 0);
267#endif
268#if defined HAVE___GETDELIM
269        else if (PARSECMD("__getdelim ( '%c' )", &ch))
270            FREAD(l = __getdelim(&lineptr, &k, ch, f), lineptr, l >= 0 ? l : 0);
271        else if (PARSECMD("__getdelim ( %i )", &n))
272            FREAD(l = __getdelim(&lineptr, &k, n, f), lineptr, l >= 0 ? l : 0);
273#endif
274
275        /* FILE * seeking functions */
276        else if (PARSECMD("fseek ( %li , SEEK_CUR )", &l1))
277            FSEEK(l = fseek(f, l1, SEEK_CUR),
278                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
279        else if (PARSECMD("fseek ( %li , SEEK_SET )", &l1))
280            FSEEK(l = fseek(f, l1, SEEK_SET),
281                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
282        else if (PARSECMD("fseek ( %li , SEEK_END )", &l1))
283            FSEEK(l = fseek(f, l1, SEEK_END),
284                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
285#if defined HAVE_FSEEKO
286        else if (PARSECMD("fseeko ( %li , SEEK_CUR )", &l1))
287            FSEEK(l = fseeko(f, l1, SEEK_CUR),
288                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
289        else if (PARSECMD("fseeko ( %li , SEEK_SET )", &l1))
290            FSEEK(l = fseeko(f, l1, SEEK_SET),
291                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
292        else if (PARSECMD("fseeko ( %li , SEEK_END )", &l1))
293            FSEEK(l = fseeko(f, l1, SEEK_END),
294                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
295#endif
296#if defined HAVE_FSEEKO64
297        else if (PARSECMD("fseeko64 ( %li , SEEK_CUR )", &l1))
298            FSEEK(l = fseeko64(f, l1, SEEK_CUR),
299                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
300        else if (PARSECMD("fseeko64 ( %li , SEEK_SET )", &l1))
301            FSEEK(l = fseeko64(f, l1, SEEK_SET),
302                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
303        else if (PARSECMD("fseeko64 ( %li , SEEK_END )", &l1))
304            FSEEK(l = fseeko64(f, l1, SEEK_END),
305                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
306#endif
307#if defined HAVE___FSEEKO64
308        else if (PARSECMD("__fseeko64 ( %li , SEEK_CUR )", &l1))
309            FSEEK(l = __fseeko64(f, l1, SEEK_CUR),
310                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
311        else if (PARSECMD("__fseeko64 ( %li , SEEK_SET )", &l1))
312            FSEEK(l = __fseeko64(f, l1, SEEK_SET),
313                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
314        else if (PARSECMD("__fseeko64 ( %li , SEEK_END )", &l1))
315            FSEEK(l = __fseeko64(f, l1, SEEK_END),
316                  ftell(f) >= 0 ? ftell(f) - retoff : 0);
317#endif
318        else if (PARSECMD("rewind ( )"))
319            FSEEK(rewind(f), -retlen);
320        else if (PARSECMD("ungetc ( )"))
321            FSEEK(if(retoff) ungetc((unsigned char)retbuf[retoff - 1], f),
322                  retoff ? -1 : 0);
323
324        /* Unrecognised sequence */
325        else
326        {
327            char buf[16];
328            snprintf(buf, 16, strlen(p) < 16 ? "%s" : "%.12s...", p);
329            fprintf(stderr, "E: zzcat: syntax error near `%s'\n", buf);
330            return EXIT_FAILURE;
331        }
332
333        /* Clean up our mess */
334        if (lineptr)
335            free(lineptr);
336    }
337
338    if (f)
339        fclose(f);
340
341    if (fd >= 0)
342        close(fd);
343
344    fwrite(retbuf, retlen, 1, stdout);
345
346    free(retbuf);
347    free(tmp);
348
349    return EXIT_SUCCESS;
350}
351
352/*
353 * Main program.
354 */
355
356int main(int argc, char *argv[])
357{
358    int i;
359
360    if (argc < 2)
361    {
362        fprintf(stderr, "E: zzcat: too few arguments\n");
363        return EXIT_FAILURE;
364    }
365
366    if (argc == 2)
367        return cat_file("fread(1,33554432)", argv[1]);
368
369    for (i = 2; i < argc; i++)
370    {
371        int ret = cat_file(argv[1], argv[i]);
372        if (ret)
373            return ret;
374    }
375
376    return EXIT_SUCCESS;
377}
378
379#if 0
380/* Only read() calls */
381static int zzcat_read(char const *name, unsigned char *data, int64_t len,
382                      int64_t chunk)
383{
384    int i, fd = open(name, O_RDONLY);
385    if(fd < 0)
386        return EXIT_FAILURE;
387    for(i = 0; i < len; i += chunk)
388        read(fd, data + i, chunk);
389    close(fd);
390    return EXIT_SUCCESS;
391}
392
393/* Socket seeks and reads */
394static int zzcat_random_socket(char const *name, unsigned char *data,
395                               int64_t len)
396{
397    int i, j, fd = open(name, O_RDONLY);
398    if(fd < 0)
399        return EXIT_FAILURE;
400    for(i = 0; i < 128; i++)
401    {
402        lseek(fd, myrand() % len, SEEK_SET);
403        for(j = 0; j < 4; j++)
404            read(fd, data + lseek(fd, 0, SEEK_CUR), myrand() % 4096);
405#ifdef HAVE_LSEEK64
406        lseek64(fd, myrand() % len, SEEK_SET);
407        for(j = 0; j < 4; j++)
408            read(fd, data + lseek(fd, 0, SEEK_CUR), myrand() % 4096);
409#endif
410    }
411    close(fd);
412    return EXIT_SUCCESS;
413}
414
415/* Standard stream seeks and reads */
416static int zzcat_random_stream(char const *name, unsigned char *data,
417                               int64_t len)
418{
419    FILE *stream = fopen(name, "r");
420    int i, j;
421    if(!stream)
422        return EXIT_FAILURE;
423    for(i = 0; i < 128; i++)
424    {
425        long int now;
426        fseek(stream, myrand() % len, SEEK_SET);
427        for(j = 0; j < 4; j++)
428            fread(data + ftell(stream),
429                  myrand() % (len - ftell(stream)), 1, stream);
430        fseek(stream, myrand() % len, SEEK_SET);
431        now = ftell(stream);
432        for(j = 0; j < 16; j++)
433            data[now + j] = getc(stream);
434        now = ftell(stream);
435        for(j = 0; j < 16; j++)
436            data[now + j] = fgetc(stream);
437    }
438    fclose(stream);
439    return EXIT_SUCCESS;
440}
441
442#ifdef HAVE_MMAP
443/* mmap() followed by random memory reads */
444static int zzcat_random_mmap(char const *name, unsigned char *data,
445                               int64_t len)
446{
447    int i, j, fd = open(name, O_RDONLY);
448    if(fd < 0)
449        return EXIT_FAILURE;
450    for(i = 0; i < 128; i++)
451    {
452        char *map;
453        int moff, mlen, pgsz = len + 1;
454#ifdef HAVE_GETPAGESIZE
455        pgsz = getpagesize();
456#endif
457        moff = len < pgsz ? 0 : (myrand() % (len / pgsz)) * pgsz;
458        mlen = 1 + (myrand() % (len - moff));
459        map = mmap(NULL, mlen, PROT_READ, MAP_PRIVATE, fd, moff);
460        if(map == MAP_FAILED)
461            return EXIT_FAILURE;
462        for(j = 0; j < 128; j++)
463        {
464            int x = myrand() % mlen;
465            data[moff + x] = map[x];
466        }
467        munmap(map, mlen);
468    }
469    close(fd);
470    return EXIT_SUCCESS;
471}
472#endif
473#endif
474
Note: See TracBrowser for help on using the repository browser.