fsort.c revision 1.25 1 /* $NetBSD: fsort.c,v 1.25 2003/08/07 11:15:54 agc Exp $ */
2
3 /*-
4 * Copyright (c) 1993
5 * The Regents of the University of California. All rights reserved.
6 *
7 * This code is derived from software contributed to Berkeley by
8 * Peter McIlroy.
9 *
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
12 * are met:
13 * 1. Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 * notice, this list of conditions and the following disclaimer in the
17 * documentation and/or other materials provided with the distribution.
18 * 3. Neither the name of the University nor the names of its contributors
19 * may be used to endorse or promote products derived from this software
20 * without specific prior written permission.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * SUCH DAMAGE.
33 */
34
35 /*
36 * Read in the next bin. If it fits in one segment sort it;
37 * otherwise refine it by segment deeper by one character,
38 * and try again on smaller bins. Sort the final bin at this level
39 * of recursion to keep the head of fstack at 0.
40 * After PANIC passes, abort to merge sort.
41 */
42 #include "sort.h"
43 #include "fsort.h"
44
45 #ifndef lint
46 __RCSID("$NetBSD: fsort.c,v 1.25 2003/08/07 11:15:54 agc Exp $");
47 __SCCSID("@(#)fsort.c 8.1 (Berkeley) 6/6/93");
48 #endif /* not lint */
49
50 #include <stdlib.h>
51 #include <string.h>
52
53 static const u_char **keylist = 0;
54 u_char *buffer = 0, *linebuf = 0;
55 size_t bufsize = DEFBUFSIZE;
56 size_t linebuf_size;
57 #define FSORTMAX 4
58 int PANIC = FSORTMAX;
59
60 struct tempfile fstack[MAXFCT];
61 #define MSTART (MAXFCT - MERGE_FNUM)
62 #define CHECKFSTACK(n) \
63 if (n >= MAXFCT) \
64 errx(2, "fstack: too many temporary files; use -H or sort in pieces")
65
66 #define SALIGN(n) ((n+sizeof(length_t)-1) & ~(sizeof(length_t)-1))
67
68 void
69 fsort(binno, depth, top, filelist, nfiles, outfp, ftbl)
70 int binno, depth, top;
71 struct filelist *filelist;
72 int nfiles;
73 FILE *outfp;
74 struct field *ftbl;
75 {
76 const u_char **keypos;
77 u_char *bufend, *tmpbuf;
78 u_char *weights;
79 int ntfiles, mfct = 0, total, i, maxb, lastb, panic = 0;
80 int c, nelem, base;
81 long sizes [NBINS+1];
82 get_func_t get;
83 struct recheader *crec;
84 struct field tfield[2];
85 FILE *prevfp, *tailfp[FSORTMAX+1];
86
87 memset(tailfp, 0, sizeof(tailfp));
88 prevfp = outfp;
89 memset(tfield, 0, sizeof(tfield));
90 if (ftbl[0].flags & R)
91 tfield[0].weights = Rascii;
92 else
93 tfield[0].weights = ascii;
94 tfield[0].icol.num = 1;
95 weights = ftbl[0].weights;
96 if (!buffer) {
97 buffer = malloc(bufsize);
98 keylist = malloc(MAXNUM * sizeof(u_char *));
99 memset(keylist, 0, MAXNUM * sizeof(u_char *));
100 if (!SINGL_FLD) {
101 linebuf_size = DEFLLEN;
102 if ((linebuf = malloc(linebuf_size)) == NULL)
103 errx(2, "cannot allocate memory");
104 }
105 }
106 bufend = buffer + bufsize;
107 if (binno >= 0) {
108 base = top + nfiles;
109 get = getnext;
110 } else {
111 base = 0;
112 if (SINGL_FLD)
113 get = makeline;
114 else
115 get = makekey;
116 }
117 for (;;) {
118 memset(sizes, 0, sizeof(sizes));
119 c = ntfiles = 0;
120 if (binno == weights[REC_D] &&
121 !(SINGL_FLD && ftbl[0].flags & F)) { /* pop */
122 rd_append(weights[REC_D], top,
123 nfiles, prevfp, buffer, bufend);
124 break;
125 } else if (binno == weights[REC_D]) {
126 depth = 0; /* start over on flat weights */
127 ftbl = tfield;
128 weights = ftbl[0].weights;
129 }
130 while (c != EOF) {
131 keypos = keylist;
132 nelem = 0;
133 crec = (RECHEADER *) buffer;
134
135 do_read:
136 while((c = get(binno, top, filelist, nfiles, crec,
137 bufend, ftbl)) == 0) {
138 *keypos++ = crec->data + depth;
139 if (++nelem == MAXNUM) {
140 c = BUFFEND;
141 break;
142 }
143 crec =(RECHEADER *) ((char *) crec +
144 SALIGN(crec->length) + sizeof(TRECHEADER));
145 }
146
147 if (c == BUFFEND && nelem < MAXNUM
148 && bufsize < MAXBUFSIZE) {
149 const u_char **keyp;
150 u_char *oldb = buffer;
151
152 /* buffer was too small for data, allocate
153 * bigger buffer */
154 bufsize *= 2;
155 buffer = realloc(buffer, bufsize);
156 if (!buffer) {
157 err(2, "failed to realloc buffer to %ld bytes",
158 (unsigned long) bufsize);
159 }
160 bufend = buffer + bufsize;
161
162 /* patch up keylist[] */
163 for(keyp = &keypos[-1]; keyp >= keylist; keyp--)
164 *keyp = buffer + (*keyp - oldb);
165
166 crec = (RECHEADER *) (buffer + ((u_char *)crec - oldb));
167 goto do_read;
168 }
169
170 if (c != BUFFEND && !ntfiles && !mfct) {
171 /* do not push */
172 continue;
173 }
174
175 /* push */
176 if (panic >= PANIC) {
177 fstack[MSTART + mfct].fp = ftmp();
178 if ((stable_sort)
179 ? sradixsort(keylist, nelem,
180 weights, REC_D)
181 : radixsort(keylist, nelem,
182 weights, REC_D) )
183 err(2, NULL);
184 append(keylist, nelem, depth,
185 fstack[MSTART + mfct].fp, putrec,
186 ftbl);
187 mfct++;
188 /* reduce number of open files */
189 if (mfct == MERGE_FNUM ||(c == EOF && ntfiles)) {
190 /*
191 * Only copy extra incomplete crec
192 * data if there are any.
193 */
194 int nodata = (bufend >= (u_char *)crec
195 && bufend <= crec->data);
196
197 if (!nodata) {
198 tmpbuf = malloc(bufend -
199 crec->data);
200 memmove(tmpbuf, crec->data,
201 bufend - crec->data);
202 }
203
204 CHECKFSTACK(base + ntfiles);
205 fstack[base + ntfiles].fp = ftmp();
206 fmerge(0, MSTART, filelist,
207 mfct, geteasy,
208 fstack[base + ntfiles].fp,
209 putrec, ftbl);
210 ntfiles++;
211 mfct = 0;
212
213 if (!nodata) {
214 memmove(crec->data, tmpbuf,
215 bufend - crec->data);
216 free(tmpbuf);
217 }
218 }
219 } else {
220 CHECKFSTACK(base + ntfiles);
221 fstack[base + ntfiles].fp = ftmp();
222 onepass(keylist, depth, nelem, sizes,
223 weights, fstack[base + ntfiles].fp);
224 ntfiles++;
225 }
226 }
227 if (!ntfiles && !mfct) { /* everything in memory--pop */
228 if (nelem > 1
229 && ((stable_sort)
230 ? sradixsort(keylist, nelem, weights, REC_D)
231 : radixsort(keylist, nelem, weights, REC_D) ))
232 err(2, NULL);
233 if (nelem > 0)
234 append(keylist, nelem, depth, outfp, putline, ftbl);
235 break; /* pop */
236 }
237 if (panic >= PANIC) {
238 if (!ntfiles)
239 fmerge(0, MSTART, filelist, mfct, geteasy,
240 outfp, putline, ftbl);
241 else
242 fmerge(0, base, filelist, ntfiles, geteasy,
243 outfp, putline, ftbl);
244 break;
245
246 }
247 total = maxb = lastb = 0; /* find if one bin dominates */
248 for (i = 0; i < NBINS; i++)
249 if (sizes[i]) {
250 if (sizes[i] > sizes[maxb])
251 maxb = i;
252 lastb = i;
253 total += sizes[i];
254 }
255 if (sizes[maxb] < max((total / 2) , BUFSIZE))
256 maxb = lastb; /* otherwise pop after last bin */
257 fstack[base].lastb = lastb;
258 fstack[base].maxb = maxb;
259
260 /* start refining next level. */
261 getnext(-1, base, NULL, ntfiles, crec, bufend, 0); /* rewind */
262 for (i = 0; i < maxb; i++) {
263 if (!sizes[i]) /* bin empty; step ahead file offset */
264 getnext(i, base, NULL,ntfiles, crec, bufend, 0);
265 else
266 fsort(i, depth+1, base, filelist, ntfiles,
267 outfp, ftbl);
268 }
269
270 get = getnext;
271
272 if (lastb != maxb) {
273 if (prevfp != outfp)
274 tailfp[panic] = prevfp;
275 prevfp = ftmp();
276 for (i = maxb+1; i <= lastb; i++)
277 if (!sizes[i])
278 getnext(i, base, NULL, ntfiles, crec,
279 bufend,0);
280 else
281 fsort(i, depth+1, base, filelist,
282 ntfiles, prevfp, ftbl);
283 }
284
285 /* sort biggest (or last) bin at this level */
286 depth++;
287 panic++;
288 binno = maxb;
289 top = base;
290 nfiles = ntfiles; /* so overwrite them */
291 }
292 if (prevfp != outfp) {
293 concat(outfp, prevfp);
294 fclose(prevfp);
295 }
296 for (i = panic; i >= 0; --i)
297 if (tailfp[i]) {
298 concat(outfp, tailfp[i]);
299 fclose(tailfp[i]);
300 }
301
302 /* If on top level, free our structures */
303 if (depth == 0) {
304 free(keylist), keylist = NULL;
305 free(buffer), buffer = NULL;
306 }
307 }
308
309 /*
310 * This is one pass of radix exchange, dumping the bins to disk.
311 */
312 #define swap(a, b, t) t = a, a = b, b = t
313 void
314 onepass(a, depth, n, sizes, tr, fp)
315 const u_char **a;
316 int depth;
317 long n, sizes[];
318 u_char *tr;
319 FILE *fp;
320 {
321 size_t tsizes[NBINS+1];
322 const u_char **bin[257], ***bp, ***bpmax, **top[256], ***tp;
323 static int histo[256];
324 int *hp;
325 int c;
326 const u_char **an, *t, **aj;
327 const u_char **ak, *r;
328
329 memset(tsizes, 0, sizeof(tsizes));
330 depth += sizeof(TRECHEADER);
331 an = &a[n];
332 for (ak = a; ak < an; ak++) {
333 histo[c = tr[**ak]]++;
334 tsizes[c] += ((const RECHEADER *) (*ak -= depth))->length;
335 }
336
337 bin[0] = a;
338 bpmax = bin + 256;
339 tp = top, hp = histo;
340 for (bp = bin; bp < bpmax; bp++) {
341 *tp++ = *(bp+1) = *bp + (c = *hp);
342 *hp++ = 0;
343 if (c <= 1)
344 continue;
345 }
346 for (aj = a; aj < an; *aj = r, aj = bin[c+1])
347 for (r = *aj; aj < (ak = --top[c = tr[r[depth]]]) ;)
348 swap(*ak, r, t);
349
350 for (ak = a, c = 0; c < 256; c++) {
351 an = bin[c+1];
352 n = an - ak;
353 tsizes[c] += n * sizeof(TRECHEADER);
354 /* tell getnext how many elements in this bin, this segment. */
355 EWRITE(&tsizes[c], sizeof(size_t), 1, fp);
356 sizes[c] += tsizes[c];
357 for (; ak < an; ++ak)
358 putrec((const RECHEADER *) *ak, fp);
359 }
360 }
361