1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
|
/*-
* Copyright (c) 1988, 1989, 1990, 1993
* The Regents of the University of California. All rights reserved.
* Copyright (c) 1988, 1989 by Adam de Boor
* Copyright (c) 1989 by Berkeley Softworks
* All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Adam de Boor.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)dir.c 8.2 (Berkeley) 1/2/94
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/*-
* dir.c --
* Directory searching using wildcards and/or normal names...
* Used both for source wildcarding in the Makefile and for finding
* implicit sources.
*
* The interface for this module is:
* Dir_Init Initialize the module.
*
* Dir_HasWildcards Returns TRUE if the name given it needs to
* be wildcard-expanded.
*
* Dir_Expand Given a pattern and a path, return a Lst of names
* which match the pattern on the search path.
*
* Dir_FindFile Searches for a file on a given search path.
* If it exists, the entire path is returned.
* Otherwise NULL is returned.
*
* Dir_MTime Return the modification time of a node. The file
* is searched for along the default search path.
* The path and mtime fields of the node are filled
* in.
*
* Dir_AddDir Add a directory to a search path.
*
* Dir_MakeFlags Given a search path and a command flag, create
* a string with each of the directories in the path
* preceded by the command flag and all of them
* separated by a space.
*
* Dir_Destroy Destroy an element of a search path. Frees up all
* things that can be freed for the element as long
* as the element is no longer referenced by any other
* search path.
* Dir_ClearPath Resets a search path to the empty list.
*
* For debugging:
* Dir_PrintDirectories Print stats about the directory cache.
*/
#include <sys/types.h>
#include <sys/stat.h>
#include <dirent.h>
#include <err.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include "arch.h"
#include "dir.h"
#include "globals.h"
#include "GNode.h"
#include "hash.h"
#include "lst.h"
#include "make.h"
#include "str.h"
#include "targ.h"
#include "util.h"
/*
* A search path consists of a Lst of Path structures. A Path structure
* has in it the name of the directory and a hash table of all the files
* in the directory. This is used to cut down on the number of system
* calls necessary to find implicit dependents and their like. Since
* these searches are made before any actions are taken, we need not
* worry about the directory changing due to creation commands. If this
* hampers the style of some makefiles, they must be changed.
*
* A list of all previously-read directories is kept in the
* openDirectories Lst. This list is checked first before a directory
* is opened.
*
* The need for the caching of whole directories is brought about by
* the multi-level transformation code in suff.c, which tends to search
* for far more files than regular make does. In the initial
* implementation, the amount of time spent performing "stat" calls was
* truly astronomical. The problem with hashing at the start is,
* of course, that pmake doesn't then detect changes to these directories
* during the course of the make. Three possibilities suggest themselves:
*
* 1) just use stat to test for a file's existence. As mentioned
* above, this is very inefficient due to the number of checks
* engendered by the multi-level transformation code.
* 2) use readdir() and company to search the directories, keeping
* them open between checks. I have tried this and while it
* didn't slow down the process too much, it could severely
* affect the amount of parallelism available as each directory
* open would take another file descriptor out of play for
* handling I/O for another job. Given that it is only recently
* that UNIX OS's have taken to allowing more than 20 or 32
* file descriptors for a process, this doesn't seem acceptable
* to me.
* 3) record the mtime of the directory in the Path structure and
* verify the directory hasn't changed since the contents were
* hashed. This will catch the creation or deletion of files,
* but not the updating of files. However, since it is the
* creation and deletion that is the problem, this could be
* a good thing to do. Unfortunately, if the directory (say ".")
* were fairly large and changed fairly frequently, the constant
* rehashing could seriously degrade performance. It might be
* good in such cases to keep track of the number of rehashes
* and if the number goes over a (small) limit, resort to using
* stat in its place.
*
* An additional thing to consider is that pmake is used primarily
* to create C programs and until recently pcc-based compilers refused
* to allow you to specify where the resulting object file should be
* placed. This forced all objects to be created in the current
* directory. This isn't meant as a full excuse, just an explanation of
* some of the reasons for the caching used here.
*
* One more note: the location of a target's file is only performed
* on the downward traversal of the graph and then only for terminal
* nodes in the graph. This could be construed as wrong in some cases,
* but prevents inadvertent modification of files when the "installed"
* directory for a file is provided in the search path.
*
* Another data structure maintained by this module is an mtime
* cache used when the searching of cached directories fails to find
* a file. In the past, Dir_FindFile would simply perform an access()
* call in such a case to determine if the file could be found using
* just the name given. When this hit, however, all that was gained
* was the knowledge that the file existed. Given that an access() is
* essentially a stat() without the copyout() call, and that the same
* filesystem overhead would have to be incurred in Dir_MTime, it made
* sense to replace the access() with a stat() and record the mtime
* in a cache for when Dir_MTime was actually called.
*/
/* main search path */
Lst dirSearchPath = Lst_Initializer(dirSearchPath);
/* the list of all open directories */
static Lst openDirectories = Lst_Initializer(openDirectories);
/*
* Variables for gathering statistics on the efficiency of the hashing
* mechanism.
*/
static int hits; /* Found in directory cache */
static int misses; /* Sad, but not evil misses */
static int nearmisses; /* Found under search path */
static int bigmisses; /* Sought by itself */
static Path *dot; /* contents of current directory */
/* Results of doing a last-resort stat in Dir_FindFile --
* if we have to go to the system to find the file, we might as well
* have its mtime on record.
* XXX: If this is done way early, there's a chance other rules will
* have already updated the file, in which case we'll update it again.
* Generally, there won't be two rules to update a single file, so this
* should be ok, but...
*/
static Hash_Table mtimes;
static int DirPrintWord(void *, void *);
static int DirPrintDir(void *, void *);
/*-
*-----------------------------------------------------------------------
* Dir_Init --
* initialize things for this module
*
* Results:
* none
*
* Side Effects:
* none
*-----------------------------------------------------------------------
*/
void
Dir_Init(void)
{
Hash_InitTable(&mtimes, 0);
}
/*-
*-----------------------------------------------------------------------
* Dir_InitDot --
* initialize the "." directory
*
* Results:
* none
*
* Side Effects:
* some directories may be opened.
*-----------------------------------------------------------------------
*/
void
Dir_InitDot(void)
{
LstNode *ln;
Dir_AddDir(&openDirectories, ".");
if ((ln = Lst_Last(&openDirectories)) == NULL)
err(1, "cannot open current directory");
dot = Lst_Datum(ln);
/*
* We always need to have dot around, so we increment its
* reference count to make sure it's not destroyed.
*/
dot->refCount += 1;
}
/*-
*-----------------------------------------------------------------------
* DirFindName --
* See if the Path structure describes the same directory as the
* given one by comparing their names. Called from Dir_AddDir via
* Lst_Find when searching the list of open directories.
*
* Results:
* 0 if it is the same. Non-zero otherwise
*
* Side Effects:
* None
*-----------------------------------------------------------------------
*/
static int
DirFindName(const void *p, const void *dname)
{
return (strcmp(((const Path *)p)->name, dname));
}
/*-
*-----------------------------------------------------------------------
* Dir_HasWildcards --
* See if the given name has any wildcard characters in it.
*
* Results:
* returns TRUE if the word should be expanded, FALSE otherwise
*
* Side Effects:
* none
*-----------------------------------------------------------------------
*/
Boolean
Dir_HasWildcards(const char *name)
{
const char *cp;
int wild = 0, brace = 0, bracket = 0;
for (cp = name; *cp; cp++) {
switch (*cp) {
case '{':
brace++;
wild = 1;
break;
case '}':
brace--;
break;
case '[':
bracket++;
wild = 1;
break;
case ']':
bracket--;
break;
case '?':
case '*':
wild = 1;
break;
default:
break;
}
}
return (wild && bracket == 0 && brace == 0);
}
/*-
*-----------------------------------------------------------------------
* DirMatchFiles --
* Given a pattern and a Path structure, see if any files
* match the pattern and add their names to the 'expansions' list if
* any do. This is incomplete -- it doesn't take care of patterns like
* src / *src / *.c properly (just *.c on any of the directories), but it
* will do for now.
*
* Results:
* Always returns 0
*
* Side Effects:
* File names are added to the expansions lst. The directory will be
* fully hashed when this is done.
*-----------------------------------------------------------------------
*/
static int
DirMatchFiles(const char *pattern, const Path *p, Lst *expansions)
{
Hash_Search search; /* Index into the directory's table */
Hash_Entry *entry; /* Current entry in the table */
Boolean isDot; /* TRUE if the directory being searched is . */
isDot = (*p->name == '.' && p->name[1] == '\0');
for (entry = Hash_EnumFirst(&p->files, &search);
entry != NULL;
entry = Hash_EnumNext(&search)) {
/*
* See if the file matches the given pattern. Note we follow
* the UNIX convention that dot files will only be found if
* the pattern begins with a dot (note also that as a side
* effect of the hashing scheme, .* won't match . or ..
* since they aren't hashed).
*/
if (Str_Match(entry->name, pattern) &&
((entry->name[0] != '.') ||
(pattern[0] == '.'))) {
Lst_AtEnd(expansions, (isDot ? estrdup(entry->name) :
str_concat(p->name, entry->name, STR_ADDSLASH)));
}
}
return (0);
}
/*-
*-----------------------------------------------------------------------
* DirExpandCurly --
* Expand curly braces like the C shell. Does this recursively.
* Note the special case: if after the piece of the curly brace is
* done there are no wildcard characters in the result, the result is
* placed on the list WITHOUT CHECKING FOR ITS EXISTENCE. The
* given arguments are the entire word to expand, the first curly
* brace in the word, the search path, and the list to store the
* expansions in.
*
* Results:
* None.
*
* Side Effects:
* The given list is filled with the expansions...
*
*-----------------------------------------------------------------------
*/
static void
DirExpandCurly(const char *word, const char *brace, Lst *path, Lst *expansions)
{
const char *end; /* Character after the closing brace */
const char *cp; /* Current position in brace clause */
const char *start; /* Start of current piece of brace clause */
int bracelevel; /* Number of braces we've seen. If we see a right brace
* when this is 0, we've hit the end of the clause. */
char *file; /* Current expansion */
int otherLen; /* The length of the other pieces of the expansion
* (chars before and after the clause in 'word') */
char *cp2; /* Pointer for checking for wildcards in
* expansion before calling Dir_Expand */
start = brace + 1;
/*
* Find the end of the brace clause first, being wary of nested brace
* clauses.
*/
for (end = start, bracelevel = 0; *end != '\0'; end++) {
if (*end == '{')
bracelevel++;
else if ((*end == '}') && (bracelevel-- == 0))
break;
}
if (*end == '\0') {
Error("Unterminated {} clause \"%s\"", start);
return;
} else
end++;
otherLen = brace - word + strlen(end);
for (cp = start; cp < end; cp++) {
/*
* Find the end of this piece of the clause.
*/
bracelevel = 0;
while (*cp != ',') {
if (*cp == '{')
bracelevel++;
else if ((*cp == '}') && (bracelevel-- <= 0))
break;
cp++;
}
/*
* Allocate room for the combination and install the
* three pieces.
*/
file = emalloc(otherLen + cp - start + 1);
if (brace != word)
strncpy(file, word, brace - word);
if (cp != start)
strncpy(&file[brace - word], start, cp - start);
strcpy(&file[(brace - word) + (cp - start)], end);
/*
* See if the result has any wildcards in it. If we find one,
* call Dir_Expand right away, telling it to place the result
* on our list of expansions.
*/
for (cp2 = file; *cp2 != '\0'; cp2++) {
switch (*cp2) {
case '*':
case '?':
case '{':
case '[':
Dir_Expand(file, path, expansions);
goto next;
default:
break;
}
}
if (*cp2 == '\0') {
/*
* Hit the end w/o finding any wildcards, so stick
* the expansion on the end of the list.
*/
Lst_AtEnd(expansions, file);
} else {
next:
free(file);
}
start = cp + 1;
}
}
/*-
*-----------------------------------------------------------------------
* DirExpandInt --
* Internal expand routine. Passes through the directories in the
* path one by one, calling DirMatchFiles for each. NOTE: This still
* doesn't handle patterns in directories... Works given a word to
* expand, a path to look in, and a list to store expansions in.
*
* Results:
* None.
*
* Side Effects:
* Things are added to the expansions list.
*
*-----------------------------------------------------------------------
*/
static void
DirExpandInt(const char *word, Lst *path, Lst *expansions)
{
LstNode *ln; /* Current node */
for (ln = Lst_First(path); ln != NULL; ln = Lst_Succ(ln))
DirMatchFiles(word, (Path *)Lst_Datum(ln), expansions);
}
/*-
*-----------------------------------------------------------------------
* DirPrintWord --
* Print a word in the list of expansions. Callback for Dir_Expand
* when DEBUG(DIR), via Lst_ForEach.
*
* Results:
* === 0
*
* Side Effects:
* The passed word is printed, followed by a space.
*
*-----------------------------------------------------------------------
*/
static int
DirPrintWord(void *word, void *dummy __unused)
{
DEBUGF(DIR, ("%s ", (char *)word));
return (0);
}
/*-
*-----------------------------------------------------------------------
* Dir_Expand --
* Expand the given word into a list of words by globbing it looking
* in the directories on the given search path.
*
* Results:
* A list of words consisting of the files which exist along the search
* path matching the given pattern is placed in expansions.
*
* Side Effects:
* Directories may be opened. Who knows?
*-----------------------------------------------------------------------
*/
void
Dir_Expand(char *word, Lst *path, Lst *expansions)
{
char *cp;
DEBUGF(DIR, ("expanding \"%s\"...", word));
cp = strchr(word, '{');
if (cp != NULL)
DirExpandCurly(word, cp, path, expansions);
else {
cp = strchr(word, '/');
if (cp != NULL) {
/*
* The thing has a directory component -- find the
* first wildcard in the string.
*/
for (cp = word; *cp != '\0'; cp++) {
if (*cp == '?' || *cp == '[' ||
*cp == '*' || *cp == '{') {
break;
}
}
if (*cp == '{') {
/*
* This one will be fun.
*/
DirExpandCurly(word, cp, path, expansions);
return;
} else if (*cp != '\0') {
/*
* Back up to the start of the component
*/
char *dirpath;
while (cp > word && *cp != '/')
cp--;
if (cp != word) {
char sc;
/*
* If the glob isn't in the first
* component, try and find all the
* components up to the one with a
* wildcard.
*/
sc = cp[1];
cp[1] = '\0';
dirpath = Dir_FindFile(word, path);
cp[1] = sc;
/*
* dirpath is null if can't find the
* leading component
* XXX: Dir_FindFile won't find internal
* components. i.e. if the path contains
* ../Etc/Object and we're looking for
* Etc, * it won't be found. Ah well.
* Probably not important.
*/
if (dirpath != NULL) {
char *dp =
&dirpath[strlen(dirpath)
- 1];
Lst tp = Lst_Initializer(tp);
if (*dp == '/')
*dp = '\0';
Dir_AddDir(&tp, dirpath);
DirExpandInt(cp + 1, &tp,
expansions);
Lst_Destroy(&tp, NOFREE);
}
} else {
/*
* Start the search from the local
* directory
*/
DirExpandInt(word, path, expansions);
}
} else {
/*
* Return the file -- this should never happen.
*/
DirExpandInt(word, path, expansions);
}
} else {
/*
* First the files in dot
*/
DirMatchFiles(word, dot, expansions);
/*
* Then the files in every other directory on the path.
*/
DirExpandInt(word, path, expansions);
}
}
if (DEBUG(DIR)) {
Lst_ForEach(expansions, DirPrintWord, (void *)NULL);
DEBUGF(DIR, ("\n"));
}
}
/*-
*-----------------------------------------------------------------------
* Dir_FindFile --
* Find the file with the given name along the given search path.
*
* Results:
* The path to the file or NULL. This path is guaranteed to be in a
* different part of memory than name and so may be safely free'd.
*
* Side Effects:
* If the file is found in a directory which is not on the path
* already (either 'name' is absolute or it is a relative path
* [ dir1/.../dirn/file ] which exists below one of the directories
* already on the search path), its directory is added to the end
* of the path on the assumption that there will be more files in
* that directory later on. Sometimes this is true. Sometimes not.
*-----------------------------------------------------------------------
*/
char *
Dir_FindFile(char *name, Lst *path)
{
char *p1; /* pointer into p->name */
char *p2; /* pointer into name */
LstNode *ln; /* a list element */
char *file; /* the current filename to check */
Path *p; /* current path member */
char *cp; /* final component of the name */
Boolean hasSlash; /* true if 'name' contains a / */
struct stat stb; /* Buffer for stat, if necessary */
Hash_Entry *entry; /* Entry for mtimes table */
/*
* Find the final component of the name and note whether it has a
* slash in it (the name, I mean)
*/
cp = strrchr(name, '/');
if (cp != NULL) {
hasSlash = TRUE;
cp += 1;
} else {
hasSlash = FALSE;
cp = name;
}
DEBUGF(DIR, ("Searching for %s...", name));
/*
* No matter what, we always look for the file in the current directory
* before anywhere else and we *do not* add the ./ to it if it exists.
* This is so there are no conflicts between what the user specifies
* (fish.c) and what pmake finds (./fish.c).
*/
if ((!hasSlash || (cp - name == 2 && *name == '.')) &&
(Hash_FindEntry(&dot->files, cp) != NULL)) {
DEBUGF(DIR, ("in '.'\n"));
hits += 1;
dot->hits += 1;
return (estrdup(name));
}
/*
* We look through all the directories on the path seeking one which
* contains the final component of the given name and whose final
* component(s) match the name's initial component(s). If such a beast
* is found, we concatenate the directory name and the final component
* and return the resulting string. If we don't find any such thing,
* we go on to phase two...
*/
for (ln = Lst_First(path); ln != NULL; ln = Lst_Succ(ln)) {
p = Lst_Datum(ln);
DEBUGF(DIR, ("%s...", p->name));
if (Hash_FindEntry(&p->files, cp) != NULL) {
DEBUGF(DIR, ("here..."));
if (hasSlash) {
/*
* If the name had a slash, its initial
* components and p's final components must
* match. This is false if a mismatch is
* encountered before all of the initial
* components have been checked (p2 > name at
* the end of the loop), or we matched only
* part of one of the components of p
* along with all the rest of them (*p1 != '/').
*/
p1 = p->name + strlen(p->name) - 1;
p2 = cp - 2;
while (p2 >= name && p1 >= p->name &&
*p1 == *p2) {
p1 -= 1; p2 -= 1;
}
if (p2 >= name || (p1 >= p->name &&
*p1 != '/')) {
DEBUGF(DIR, ("component mismatch -- "
"continuing..."));
continue;
}
}
file = str_concat(p->name, cp, STR_ADDSLASH);
DEBUGF(DIR, ("returning %s\n", file));
p->hits += 1;
hits += 1;
return (file);
} else if (hasSlash) {
/*
* If the file has a leading path component and that
* component exactly matches the entire name of the
* current search directory, we assume the file
* doesn't exist and return NULL.
*/
for (p1 = p->name, p2 = name; *p1 && *p1 == *p2;
p1++, p2++)
continue;
if (*p1 == '\0' && p2 == cp - 1) {
if (*cp == '\0' || ISDOT(cp) || ISDOTDOT(cp)) {
DEBUGF(DIR, ("returning %s\n", name));
return (estrdup(name));
} else {
DEBUGF(DIR, ("must be here but isn't --"
" returning NULL\n"));
return (NULL);
}
}
}
}
/*
* We didn't find the file on any existing members of the directory.
* If the name doesn't contain a slash, that means it doesn't exist.
* If it *does* contain a slash, however, there is still hope: it
* could be in a subdirectory of one of the members of the search
* path. (eg. /usr/include and sys/types.h. The above search would
* fail to turn up types.h in /usr/include, but it *is* in
* /usr/include/sys/types.h) If we find such a beast, we assume there
* will be more (what else can we assume?) and add all but the last
* component of the resulting name onto the search path (at the
* end). This phase is only performed if the file is *not* absolute.
*/
if (!hasSlash) {
DEBUGF(DIR, ("failed.\n"));
misses += 1;
return (NULL);
}
if (*name != '/') {
Boolean checkedDot = FALSE;
DEBUGF(DIR, ("failed. Trying subdirectories..."));
for (ln = Lst_First(path); ln != NULL; ln = Lst_Succ(ln)) {
p = Lst_Datum(ln);
if (p != dot) {
file = str_concat(p->name, name, STR_ADDSLASH);
} else {
/*
* Checking in dot -- DON'T put a leading ./
* on the thing.
*/
file = estrdup(name);
checkedDot = TRUE;
}
DEBUGF(DIR, ("checking %s...", file));
if (stat(file, &stb) == 0) {
DEBUGF(DIR, ("got it.\n"));
/*
* We've found another directory to search. We
* know there's a slash in 'file' because we put
* one there. We nuke it after finding it and
* call Dir_AddDir to add this new directory
* onto the existing search path. Once that's
* done, we restore the slash and triumphantly
* return the file name, knowing that should a
* file in this directory every be referenced
* again in such a manner, we will find it
* without having to do numerous numbers of
* access calls. Hurrah!
*/
cp = strrchr(file, '/');
*cp = '\0';
Dir_AddDir(path, file);
*cp = '/';
/*
* Save the modification time so if
* it's needed, we don't have to fetch it again.
*/
DEBUGF(DIR, ("Caching %s for %s\n",
Targ_FmtTime(stb.st_mtime), file));
entry = Hash_CreateEntry(&mtimes, file,
(Boolean *)NULL);
Hash_SetValue(entry,
(void *)(long)stb.st_mtime);
nearmisses += 1;
return (file);
} else {
free(file);
}
}
DEBUGF(DIR, ("failed. "));
if (checkedDot) {
/*
* Already checked by the given name, since . was in
* the path, so no point in proceeding...
*/
DEBUGF(DIR, ("Checked . already, returning NULL\n"));
return (NULL);
}
}
/*
* Didn't find it that way, either. Sigh. Phase 3. Add its directory
* onto the search path in any case, just in case, then look for the
* thing in the hash table. If we find it, grand. We return a new
* copy of the name. Otherwise we sadly return a NULL pointer. Sigh.
* Note that if the directory holding the file doesn't exist, this will
* do an extra search of the final directory on the path. Unless
* something weird happens, this search won't succeed and life will
* be groovy.
*
* Sigh. We cannot add the directory onto the search path because
* of this amusing case:
* $(INSTALLDIR)/$(FILE): $(FILE)
*
* $(FILE) exists in $(INSTALLDIR) but not in the current one.
* When searching for $(FILE), we will find it in $(INSTALLDIR)
* b/c we added it here. This is not good...
*/
#ifdef notdef
cp[-1] = '\0';
Dir_AddDir(path, name);
cp[-1] = '/';
bigmisses += 1;
ln = Lst_Last(path);
if (ln == NULL)
return (NULL);
p = Lst_Datum(ln);
if (Hash_FindEntry(&p->files, cp) != NULL) {
return (estrdup(name));
return (NULL);
#else /* !notdef */
DEBUGF(DIR, ("Looking for \"%s\"...", name));
bigmisses += 1;
entry = Hash_FindEntry(&mtimes, name);
if (entry != NULL) {
DEBUGF(DIR, ("got it (in mtime cache)\n"));
return (estrdup(name));
} else if (stat (name, &stb) == 0) {
entry = Hash_CreateEntry(&mtimes, name, (Boolean *)NULL);
DEBUGF(DIR, ("Caching %s for %s\n",
Targ_FmtTime(stb.st_mtime), name));
Hash_SetValue(entry, (void *)(long)stb.st_mtime);
return (estrdup(name));
} else {
DEBUGF(DIR, ("failed. Returning NULL\n"));
return (NULL);
}
#endif /* notdef */
}
/*-
*-----------------------------------------------------------------------
* Dir_MTime --
* Find the modification time of the file described by gn along the
* search path dirSearchPath.
*
* Results:
* The modification time or 0 if it doesn't exist
*
* Side Effects:
* The modification time is placed in the node's mtime slot.
* If the node didn't have a path entry before, and Dir_FindFile
* found one for it, the full name is placed in the path slot.
*-----------------------------------------------------------------------
*/
int
Dir_MTime(GNode *gn)
{
char *fullName; /* the full pathname of name */
struct stat stb; /* buffer for finding the mod time */
Hash_Entry *entry;
if (gn->type & OP_ARCHV)
return (Arch_MTime(gn));
else if (gn->path == NULL)
fullName = Dir_FindFile(gn->name, &dirSearchPath);
else
fullName = gn->path;
if (fullName == NULL)
fullName = estrdup(gn->name);
entry = Hash_FindEntry(&mtimes, fullName);
if (entry != NULL) {
/*
* Only do this once -- the second time folks are checking to
* see if the file was actually updated, so we need to
* actually go to the filesystem.
*/
DEBUGF(DIR, ("Using cached time %s for %s\n",
Targ_FmtTime((time_t)(long)Hash_GetValue(entry)),
fullName));
stb.st_mtime = (time_t)(long)Hash_GetValue(entry);
Hash_DeleteEntry(&mtimes, entry);
} else if (stat(fullName, &stb) < 0) {
if (gn->type & OP_MEMBER) {
if (fullName != gn->path)
free(fullName);
return (Arch_MemMTime(gn));
} else {
stb.st_mtime = 0;
}
}
if (fullName && gn->path == (char *)NULL)
gn->path = fullName;
gn->mtime = stb.st_mtime;
return (gn->mtime);
}
/*-
*-----------------------------------------------------------------------
* Dir_AddDir --
* Add the given name to the end of the given path. The order of
* the arguments is backwards so ParseDoDependency can do a
* Lst_ForEach of its list of paths...
*
* Results:
* none
*
* Side Effects:
* A structure is added to the list and the directory is
* read and hashed.
*-----------------------------------------------------------------------
*/
void
Dir_AddDir(Lst *path, const char *name)
{
LstNode *ln; /* node in case Path structure is found */
Path *p; /* pointer to new Path structure */
DIR *d; /* for reading directory */
struct dirent *dp; /* entry in directory */
ln = Lst_Find(&openDirectories, name, DirFindName);
if (ln != NULL) {
p = Lst_Datum(ln);
if (Lst_Member(path, p) == NULL) {
p->refCount += 1;
Lst_AtEnd(path, p);
}
} else {
DEBUGF(DIR, ("Caching %s...", name));
if ((d = opendir(name)) != NULL) {
p = emalloc(sizeof(Path));
p->name = estrdup(name);
p->hits = 0;
p->refCount = 1;
Hash_InitTable(&p->files, -1);
while ((dp = readdir(d)) != NULL) {
#if defined(sun) && defined(d_ino) /* d_ino is a sunos4 #define for d_fileno */
/*
* The sun directory library doesn't check for
* a 0 inode (0-inode slots just take up space),
* so we have to do it ourselves.
*/
if (dp->d_fileno == 0)
continue;
#endif /* sun && d_ino */
/* Skip the '.' and '..' entries by checking
* for them specifically instead of assuming
* readdir() reuturns them in that order when
* first going through a directory. This is
* needed for XFS over NFS filesystems since
* SGI does not guarantee that these are the
* first two entries returned from readdir().
*/
if (ISDOT(dp->d_name) || ISDOTDOT(dp->d_name))
continue;
Hash_CreateEntry(&p->files, dp->d_name,
(Boolean *)NULL);
}
closedir(d);
Lst_AtEnd(&openDirectories, p);
if (path != &openDirectories)
Lst_AtEnd(path, p);
}
DEBUGF(DIR, ("done\n"));
}
}
/*-
*-----------------------------------------------------------------------
* Dir_CopyDir --
* Callback function for duplicating a search path via Lst_Duplicate.
* Ups the reference count for the directory.
*
* Results:
* Returns the Path it was given.
*
* Side Effects:
* The refCount of the path is incremented.
*
*-----------------------------------------------------------------------
*/
void *
Dir_CopyDir(void *p)
{
((Path *)p)->refCount += 1;
return (p);
}
/*-
*-----------------------------------------------------------------------
* Dir_MakeFlags --
* Make a string by taking all the directories in the given search
* path and preceding them by the given flag. Used by the suffix
* module to create variables for compilers based on suffix search
* paths.
*
* Results:
* The string mentioned above. Note that there is no space between
* the given flag and each directory. The empty string is returned if
* Things don't go well.
*
* Side Effects:
* None
*-----------------------------------------------------------------------
*/
char *
Dir_MakeFlags(const char *flag, Lst *path)
{
char *str; /* the string which will be returned */
char *tstr; /* the current directory preceded by 'flag' */
char *nstr;
LstNode *ln; /* the node of the current directory */
Path *p; /* the structure describing the current directory */
str = estrdup("");
for (ln = Lst_First(path); ln != NULL; ln = Lst_Succ(ln)) {
p = Lst_Datum(ln);
tstr = str_concat(flag, p->name, 0);
nstr = str_concat(str, tstr, STR_ADDSPACE);
free(str);
free(tstr);
str = nstr;
}
return (str);
}
/*-
*-----------------------------------------------------------------------
* Dir_Destroy --
* Nuke a directory descriptor, if possible. Callback procedure
* for the suffixes module when destroying a search path.
*
* Results:
* None.
*
* Side Effects:
* If no other path references this directory (refCount == 0),
* the Path and all its data are freed.
*
*-----------------------------------------------------------------------
*/
void
Dir_Destroy(void *pp)
{
Path *p = pp;
p->refCount -= 1;
if (p->refCount == 0) {
LstNode *ln;
if ((ln = Lst_Member(&openDirectories, p)) != NULL)
Lst_Remove(&openDirectories, ln);
Hash_DeleteTable(&p->files);
free(p->name);
free(p);
}
}
/*-
*-----------------------------------------------------------------------
* Dir_ClearPath --
* Clear out all elements of the given search path. This is different
* from destroying the list, notice.
*
* Results:
* None.
*
* Side Effects:
* The path is set to the empty list.
*
*-----------------------------------------------------------------------
*/
void
Dir_ClearPath(Lst *path)
{
Path *p;
while (!Lst_IsEmpty(path)) {
p = Lst_DeQueue(path);
Dir_Destroy(p);
}
}
/*-
*-----------------------------------------------------------------------
* Dir_Concat --
* Concatenate two paths, adding the second to the end of the first.
* Makes sure to avoid duplicates.
*
* Results:
* None
*
* Side Effects:
* Reference counts for added dirs are upped.
*
*-----------------------------------------------------------------------
*/
void
Dir_Concat(Lst *path1, Lst *path2)
{
LstNode *ln;
Path *p;
for (ln = Lst_First(path2); ln != NULL; ln = Lst_Succ(ln)) {
p = Lst_Datum(ln);
if (Lst_Member(path1, p) == NULL) {
p->refCount += 1;
Lst_AtEnd(path1, p);
}
}
}
/********** DEBUG INFO **********/
void
Dir_PrintDirectories(void)
{
LstNode *ln;
Path *p;
printf("#*** Directory Cache:\n");
printf("# Stats: %d hits %d misses %d near misses %d losers (%d%%)\n",
hits, misses, nearmisses, bigmisses,
(hits + bigmisses + nearmisses ?
hits * 100 / (hits + bigmisses + nearmisses) : 0));
printf("# %-20s referenced\thits\n", "directory");
for (ln = Lst_First(&openDirectories); ln != NULL; ln = Lst_Succ(ln)) {
p = Lst_Datum(ln);
printf("# %-20s %10d\t%4d\n", p->name, p->refCount, p->hits);
}
}
static int
DirPrintDir(void *p, void *dummy __unused)
{
printf("%s ", ((Path *)p)->name);
return (0);
}
void
Dir_PrintPath(Lst *path)
{
Lst_ForEach(path, DirPrintDir, (void *)NULL);
}
|