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
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
|
# -*-Makefile-*-
# This Makefile fragment tries to be general-purpose enough to be
# used by many projects via the gnulib maintainer-makefile module.
## Copyright (C) 2001-2014 Free Software Foundation, Inc.
##
## This program is free software: you can redistribute it and/or modify
## it under the terms of the GNU General Public License as published by
## the Free Software Foundation, either version 3 of the License, or
## (at your option) any later version.
##
## This program is distributed in the hope that it will be useful,
## but WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
## GNU General Public License for more details.
##
## You should have received a copy of the GNU General Public License
## along with this program. If not, see <http://www.gnu.org/licenses/>.
# This is reported not to work with make-3.79.1
# ME := $(word $(words $(MAKEFILE_LIST)),$(MAKEFILE_LIST))
ME := maint.mk
# Helper variables.
_empty =
_sp = $(_empty) $(_empty)
# _equal,S1,S2
# ------------
# If S1 == S2, return S1, otherwise the empty string.
_equal = $(and $(findstring $(1),$(2)),$(findstring $(2),$(1)))
# member-check,VARIABLE,VALID-VALUES
# ----------------------------------
# Check that $(VARIABLE) is in the space-separated list of VALID-VALUES, and
# return it. Die otherwise.
member-check = \
$(strip \
$(if $($(1)), \
$(if $(findstring $(_sp),$($(1))), \
$(error invalid $(1): '$($(1))', expected $(2)), \
$(or $(findstring $(_sp)$($(1))$(_sp),$(_sp)$(2)$(_sp)), \
$(error invalid $(1): '$($(1))', expected $(2)))), \
$(error $(1) undefined)))
# Do not save the original name or timestamp in the .tar.gz file.
# Use --rsyncable if available.
gzip_rsyncable := \
$(shell gzip --help 2>/dev/null|grep rsyncable >/dev/null \
&& printf %s --rsyncable)
GZIP_ENV = '--no-name --best $(gzip_rsyncable)'
GIT = git
VC = $(GIT)
VC_LIST = $(srcdir)/$(_build-aux)/vc-list-files -C $(srcdir)
# You can override this variable in cfg.mk to set your own regexp
# matching files to ignore.
VC_LIST_ALWAYS_EXCLUDE_REGEX ?= ^$$
# This is to preprocess robustly the output of $(VC_LIST), so that even
# when $(srcdir) is a pathological name like "....", the leading sed command
# removes only the intended prefix.
_dot_escaped_srcdir = $(subst .,\.,$(srcdir))
# Post-process $(VC_LIST) output, prepending $(srcdir)/, but only
# when $(srcdir) is not ".".
ifeq ($(srcdir),.)
_prepend_srcdir_prefix =
else
_prepend_srcdir_prefix = | $(SED) 's|^|$(srcdir)/|'
endif
# In order to be able to consistently filter "."-relative names,
# (i.e., with no $(srcdir) prefix), this definition is careful to
# remove any $(srcdir) prefix, and to restore what it removes.
_sc_excl = \
$(or $(exclude_file_name_regexp--$@),^$$)
VC_LIST_EXCEPT = \
$(VC_LIST) | $(SED) 's|^$(_dot_escaped_srcdir)/||' \
| if test -f $(srcdir)/.x-$@; then grep -vEf $(srcdir)/.x-$@; \
else grep -Ev -e "$${VC_LIST_EXCEPT_DEFAULT-ChangeLog}"; fi \
| grep -Ev -e '($(VC_LIST_ALWAYS_EXCLUDE_REGEX)|$(_sc_excl))' \
$(_prepend_srcdir_prefix)
ifeq ($(origin prev_version_file), undefined)
prev_version_file = $(srcdir)/.prev-version
endif
PREV_VERSION := $(shell cat $(prev_version_file) 2>/dev/null)
VERSION_REGEXP = $(subst .,\.,$(VERSION))
PREV_VERSION_REGEXP = $(subst .,\.,$(PREV_VERSION))
ifeq ($(VC),$(GIT))
this-vc-tag = v$(VERSION)
this-vc-tag-regexp = v$(VERSION_REGEXP)
else
tag-package = $(shell echo "$(PACKAGE)" | tr '[:lower:]' '[:upper:]')
tag-this-version = $(subst .,_,$(VERSION))
this-vc-tag = $(tag-package)-$(tag-this-version)
this-vc-tag-regexp = $(this-vc-tag)
endif
my_distdir = $(PACKAGE)-$(VERSION)
# Prevent programs like 'sort' from considering distinct strings to be equal.
# Doing it here saves us from having to set LC_ALL elsewhere in this file.
# NOTE: commented out for https://github.com/ostreedev/ostree/issues/1101
# export LC_ALL = C
## --------------- ##
## Sanity checks. ##
## --------------- ##
ifneq ($(_gl-Makefile),)
_cfg_mk := $(wildcard $(srcdir)/cfg.mk)
# Collect the names of rules starting with 'sc_'.
syntax-check-rules := $(sort $(shell $(SED) -n \
's/^\(sc_[a-zA-Z0-9_-]*\):.*/\1/p' $(srcdir)/$(ME) $(_cfg_mk)))
.PHONY: $(syntax-check-rules)
ifeq ($(shell $(VC_LIST) >/dev/null 2>&1; echo $$?),0)
local-checks-available += $(syntax-check-rules)
else
local-checks-available += no-vc-detected
no-vc-detected:
@echo "No version control files detected; skipping syntax check"
endif
.PHONY: $(local-checks-available)
# Arrange to print the name of each syntax-checking rule just before running it.
$(syntax-check-rules): %: %.m
sc_m_rules_ = $(patsubst %, %.m, $(syntax-check-rules))
.PHONY: $(sc_m_rules_)
$(sc_m_rules_):
@echo $(patsubst sc_%.m, %, $@)
@date +%s.%N > .sc-start-$(basename $@)
# Compute and print the elapsed time for each syntax-check rule.
sc_z_rules_ = $(patsubst %, %.z, $(syntax-check-rules))
.PHONY: $(sc_z_rules_)
$(sc_z_rules_): %.z: %
@end=$$(date +%s.%N); \
start=$$(cat .sc-start-$*); \
rm -f .sc-start-$*; \
awk -v s=$$start -v e=$$end \
'END {printf "%.2f $(patsubst sc_%,%,$*)\n", e - s}' < /dev/null
# The patsubst here is to replace each sc_% rule with its sc_%.z wrapper
# that computes and prints elapsed time.
local-check := \
$(patsubst sc_%, sc_%.z, \
$(filter-out $(local-checks-to-skip), $(local-checks-available)))
syntax-check: $(local-check)
endif
# _sc_search_regexp
#
# This macro searches for a given construct in the selected files and
# then takes some action.
#
# Parameters (shell variables):
#
# prohibit | require
#
# Regular expression (ERE) denoting either a forbidden construct
# or a required construct. Those arguments are exclusive.
#
# exclude
#
# Regular expression (ERE) denoting lines to ignore that matched
# a prohibit construct. For example, this can be used to exclude
# comments that mention why the nearby code uses an alternative
# construct instead of the simpler prohibited construct.
#
# in_vc_files | in_files
#
# grep-E-style regexp selecting the files to check. For in_vc_files,
# the regexp is used to select matching files from the list of all
# version-controlled files; for in_files, it's from the names printed
# by "find $(srcdir)". When neither is specified, use all files that
# are under version control.
#
# containing | non_containing
#
# Select the files (non) containing strings matching this regexp.
# If both arguments are specified then CONTAINING takes
# precedence.
#
# with_grep_options
#
# Extra options for grep.
#
# ignore_case
#
# Ignore case.
#
# halt
#
# Message to display before to halting execution.
#
# Finally, you may exempt files based on an ERE matching file names.
# For example, to exempt from the sc_space_tab check all files with the
# .diff suffix, set this Make variable:
#
# exclude_file_name_regexp--sc_space_tab = \.diff$
#
# Note that while this functionality is mostly inherited via VC_LIST_EXCEPT,
# when filtering by name via in_files, we explicitly filter out matching
# names here as well.
# Initialize each, so that envvar settings cannot interfere.
export require =
export prohibit =
export exclude =
export in_vc_files =
export in_files =
export containing =
export non_containing =
export halt =
export with_grep_options =
# By default, _sc_search_regexp does not ignore case.
export ignore_case =
_ignore_case = $$(test -n "$$ignore_case" && printf %s -i || :)
define _sc_say_and_exit
dummy=; : so we do not need a semicolon before each use; \
{ printf '%s\n' "$(ME): $$msg" 1>&2; exit 1; };
endef
define _sc_search_regexp
dummy=; : so we do not need a semicolon before each use; \
\
: Check arguments; \
test -n "$$prohibit" && test -n "$$require" \
&& { msg='Cannot specify both prohibit and require' \
$(_sc_say_and_exit) } || :; \
test -z "$$prohibit" && test -z "$$require" \
&& { msg='Should specify either prohibit or require' \
$(_sc_say_and_exit) } || :; \
test -z "$$prohibit" && test -n "$$exclude" \
&& { msg='Use of exclude requires a prohibit pattern' \
$(_sc_say_and_exit) } || :; \
test -n "$$in_vc_files" && test -n "$$in_files" \
&& { msg='Cannot specify both in_vc_files and in_files' \
$(_sc_say_and_exit) } || :; \
test "x$$halt" != x \
|| { msg='halt not defined' $(_sc_say_and_exit) }; \
\
: Filter by file name; \
if test -n "$$in_files"; then \
files=$$(find $(srcdir) | grep -E "$$in_files" \
| grep -Ev '$(_sc_excl)'); \
else \
files=$$($(VC_LIST_EXCEPT)); \
if test -n "$$in_vc_files"; then \
files=$$(echo "$$files" | grep -E "$$in_vc_files"); \
fi; \
fi; \
\
: Filter by content; \
test -n "$$files" && test -n "$$containing" \
&& { files=$$(grep -l "$$containing" $$files); } || :; \
test -n "$$files" && test -n "$$non_containing" \
&& { files=$$(grep -vl "$$non_containing" $$files); } || :; \
\
: Check for the construct; \
if test -n "$$files"; then \
if test -n "$$prohibit"; then \
grep $$with_grep_options $(_ignore_case) -nE "$$prohibit" $$files \
| grep -vE "$${exclude:-^$$}" \
&& { msg="$$halt" $(_sc_say_and_exit) } || :; \
else \
grep $$with_grep_options $(_ignore_case) -LE "$$require" $$files \
| grep . \
&& { msg="$$halt" $(_sc_say_and_exit) } || :; \
fi \
else :; \
fi || :;
endef
sc_avoid_if_before_free:
@$(srcdir)/$(_build-aux)/useless-if-before-free \
$(useless_free_options) \
$$($(VC_LIST_EXCEPT) | grep -v useless-if-before-free) && \
{ echo '$(ME): found useless "if" before "free" above' 1>&2; \
exit 1; } || :
sc_cast_of_argument_to_free:
@prohibit='\<free *\( *\(' halt="don't cast free argument" \
$(_sc_search_regexp)
sc_cast_of_x_alloc_return_value:
@prohibit='\*\) *x(m|c|re)alloc\>' \
halt="don't cast x*alloc return value" \
$(_sc_search_regexp)
sc_cast_of_alloca_return_value:
@prohibit='\*\) *alloca\>' \
halt="don't cast alloca return value" \
$(_sc_search_regexp)
sc_space_tab:
@prohibit='[ ] ' \
halt='found SPACE-TAB sequence; remove the SPACE' \
$(_sc_search_regexp)
# Don't use *scanf or the old ato* functions in "real" code.
# They provide no error checking mechanism.
# Instead, use strto* functions.
sc_prohibit_atoi_atof:
@prohibit='\<([fs]?scanf|ato([filq]|ll)) *\(' \
halt='do not use *scan''f, ato''f, ato''i, ato''l, ato''ll or ato''q' \
$(_sc_search_regexp)
# Use STREQ rather than comparing strcmp == 0, or != 0.
sp_ = strcmp *\(.+\)
sc_prohibit_strcmp:
@prohibit='! *strcmp *\(|\<$(sp_) *[!=]=|[!=]= *$(sp_)' \
exclude='# *define STRN?EQ\(' \
halt='replace strcmp calls above with STREQ/STRNEQ' \
$(_sc_search_regexp)
# Really. You don't want to use this function.
# It may fail to NUL-terminate the destination,
# and always NUL-pads out to the specified length.
sc_prohibit_strncpy:
@prohibit='\<strncpy *\(' \
halt='do not use strncpy, period' \
$(_sc_search_regexp)
# Pass EXIT_*, not number, to usage, exit, and error (when exiting)
# Convert all uses automatically, via these two commands:
# git grep -l '\<exit *(1)' \
# | grep -vEf .x-sc_prohibit_magic_number_exit \
# | xargs --no-run-if-empty \
# perl -pi -e 's/(^|[^.])\b(exit ?)\(1\)/$1$2(EXIT_FAILURE)/'
# git grep -l '\<exit *(0)' \
# | grep -vEf .x-sc_prohibit_magic_number_exit \
# | xargs --no-run-if-empty \
# perl -pi -e 's/(^|[^.])\b(exit ?)\(0\)/$1$2(EXIT_SUCCESS)/'
sc_prohibit_magic_number_exit:
@prohibit='(^|[^.])\<(usage|exit|error) ?\(-?[0-9]+[,)]' \
exclude='exit \(77\)|error ?\(((0|77),|[^,]*)' \
halt='use EXIT_* values rather than magic number' \
$(_sc_search_regexp)
# Using EXIT_SUCCESS as the first argument to error is misleading,
# since when that parameter is 0, error does not exit. Use '0' instead.
sc_error_exit_success:
@prohibit='error *\(EXIT_SUCCESS,' \
in_vc_files='\.[chly]$$' \
halt='found error (EXIT_SUCCESS' \
$(_sc_search_regexp)
# "FATAL:" should be fully upper-cased in error messages
# "WARNING:" should be fully upper-cased, or fully lower-cased
sc_error_message_warn_fatal:
@grep -nEA2 '[^rp]error *\(' $$($(VC_LIST_EXCEPT)) \
| grep -E '"Warning|"Fatal|"fatal' && \
{ echo '$(ME): use FATAL, WARNING or warning' 1>&2; \
exit 1; } || :
# Error messages should not start with a capital letter
sc_error_message_uppercase:
@grep -nEA2 '[^rp]error *\(' $$($(VC_LIST_EXCEPT)) \
| grep -E '"[A-Z]' \
| grep -vE '"FATAL|"WARNING|"Java|"C#|PRIuMAX' && \
{ echo '$(ME): found capitalized error message' 1>&2; \
exit 1; } || :
# Error messages should not end with a period
sc_error_message_period:
@grep -nEA2 '[^rp]error *\(' $$($(VC_LIST_EXCEPT)) \
| grep -E '[^."]\."' && \
{ echo '$(ME): found error message ending in period' 1>&2; \
exit 1; } || :
sc_file_system:
@prohibit=file''system \
ignore_case=1 \
halt='found use of "file''system"; spell it "file system"' \
$(_sc_search_regexp)
# Don't use cpp tests of this symbol. All code assumes config.h is included.
sc_prohibit_have_config_h:
@prohibit='^# *if.*HAVE''_CONFIG_H' \
halt='found use of HAVE''_CONFIG_H; remove' \
$(_sc_search_regexp)
# Nearly all .c files must include <config.h>. However, we also permit this
# via inclusion of a package-specific header, if cfg.mk specified one.
# config_h_header must be suitable for grep -E.
config_h_header ?= <config\.h>
sc_require_config_h:
@require='^# *include $(config_h_header)' \
in_vc_files='\.c$$' \
halt='the above files do not include <config.h>' \
$(_sc_search_regexp)
# You must include <config.h> before including any other header file.
# This can possibly be via a package-specific header, if given by cfg.mk.
sc_require_config_h_first:
@if $(VC_LIST_EXCEPT) | grep '\.c$$' > /dev/null; then \
fail=0; \
for i in $$($(VC_LIST_EXCEPT) | grep '\.c$$'); do \
grep '^# *include\>' $$i | $(SED) 1q \
| grep -E '^# *include $(config_h_header)' > /dev/null \
|| { echo $$i; fail=1; }; \
done; \
test $$fail = 1 && \
{ echo '$(ME): the above files include some other header' \
'before <config.h>' 1>&2; exit 1; } || :; \
else :; \
fi
sc_prohibit_HAVE_MBRTOWC:
@prohibit='\bHAVE_MBRTOWC\b' \
halt="do not use $$prohibit; it is always defined" \
$(_sc_search_regexp)
# To use this "command" macro, you must first define two shell variables:
# h: the header name, with no enclosing <> or ""
# re: a regular expression that matches IFF something provided by $h is used.
define _sc_header_without_use
dummy=; : so we do not need a semicolon before each use; \
h_esc=`echo '[<"]'"$$h"'[">]'|$(SED) 's/\./\\\\./g'`; \
if $(VC_LIST_EXCEPT) | grep '\.c$$' > /dev/null; then \
files=$$(grep -l '^# *include '"$$h_esc" \
$$($(VC_LIST_EXCEPT) | grep '\.c$$')) && \
grep -LE "$$re" $$files | grep . && \
{ echo "$(ME): the above files include $$h but don't use it" \
1>&2; exit 1; } || :; \
else :; \
fi
endef
# Prohibit the inclusion of assert.h without an actual use of assert.
sc_prohibit_assert_without_use:
@h='assert.h' re='\<assert *\(' $(_sc_header_without_use)
# Prohibit the inclusion of close-stream.h without an actual use.
sc_prohibit_close_stream_without_use:
@h='close-stream.h' re='\<close_stream *\(' $(_sc_header_without_use)
# Prohibit the inclusion of getopt.h without an actual use.
sc_prohibit_getopt_without_use:
@h='getopt.h' re='\<getopt(_long)? *\(' $(_sc_header_without_use)
# Don't include quotearg.h unless you use one of its functions.
sc_prohibit_quotearg_without_use:
@h='quotearg.h' re='\<quotearg(_[^ ]+)? *\(' $(_sc_header_without_use)
# Don't include quote.h unless you use one of its functions.
sc_prohibit_quote_without_use:
@h='quote.h' re='\<quote((_n)? *\(|_quoting_options\>)' \
$(_sc_header_without_use)
# Don't include this header unless you use one of its functions.
sc_prohibit_long_options_without_use:
@h='long-options.h' re='\<parse_long_options *\(' \
$(_sc_header_without_use)
# Don't include this header unless you use one of its functions.
sc_prohibit_inttostr_without_use:
@h='inttostr.h' re='\<(off|[iu]max|uint)tostr *\(' \
$(_sc_header_without_use)
# Don't include this header unless you use one of its functions.
sc_prohibit_ignore_value_without_use:
@h='ignore-value.h' re='\<ignore_(value|ptr) *\(' \
$(_sc_header_without_use)
# Don't include this header unless you use one of its functions.
sc_prohibit_error_without_use:
@h='error.h' \
re='\<error(_at_line|_print_progname|_one_per_line|_message_count)? *\('\
$(_sc_header_without_use)
# Don't include xalloc.h unless you use one of its functions.
# Consider these symbols:
# perl -lne '/^# *define (\w+)\(/ and print $1' lib/xalloc.h|grep -v '^__';
# perl -lne '/^(?:extern )?(?:void|char) \*?(\w+) *\(/ and print $1' lib/xalloc.h
# Divide into two sets on case, and filter each through this:
# | sort | perl -MRegexp::Assemble -le \
# 'print Regexp::Assemble->new(file => "/dev/stdin")->as_string'|sed 's/\?://g'
# Note this was produced by the above:
# _xa1 = \
#x(((2n?)?re|c(har)?|n(re|m)|z)alloc|alloc_(oversized|die)|m(alloc|emdup)|strdup)
# But we can do better, in at least two ways:
# 1) take advantage of two "dup"-suffixed strings:
# x(((2n?)?re|c(har)?|n(re|m)|[mz])alloc|alloc_(oversized|die)|(mem|str)dup)
# 2) notice that "c(har)?|[mz]" is equivalent to the shorter and more readable
# "char|[cmz]"
# x(((2n?)?re|char|n(re|m)|[cmz])alloc|alloc_(oversized|die)|(mem|str)dup)
_xa1 = x(((2n?)?re|char|n(re|m)|[cmz])alloc|alloc_(oversized|die)|(mem|str)dup)
_xa2 = X([CZ]|N?M)ALLOC
sc_prohibit_xalloc_without_use:
@h='xalloc.h' \
re='\<($(_xa1)|$(_xa2)) *\('\
$(_sc_header_without_use)
# Extract function names:
# perl -lne '/^(?:extern )?(?:void|char) \*?(\w+) *\(/ and print $1' lib/hash.h
_hash_re = \
clear|delete|free|get_(first|next)|insert|lookup|print_statistics|reset_tuning
_hash_fn = \<($(_hash_re)) *\(
_hash_struct = (struct )?\<[Hh]ash_(table|tuning)\>
sc_prohibit_hash_without_use:
@h='hash.h' \
re='$(_hash_fn)|$(_hash_struct)'\
$(_sc_header_without_use)
sc_prohibit_cloexec_without_use:
@h='cloexec.h' re='\<(set_cloexec_flag|dup_cloexec) *\(' \
$(_sc_header_without_use)
sc_prohibit_posixver_without_use:
@h='posixver.h' re='\<posix2_version *\(' $(_sc_header_without_use)
sc_prohibit_same_without_use:
@h='same.h' re='\<same_name *\(' $(_sc_header_without_use)
sc_prohibit_hash_pjw_without_use:
@h='hash-pjw.h' \
re='\<hash_pjw\>' \
$(_sc_header_without_use)
sc_prohibit_safe_read_without_use:
@h='safe-read.h' re='(\<SAFE_READ_ERROR\>|\<safe_read *\()' \
$(_sc_header_without_use)
sc_prohibit_argmatch_without_use:
@h='argmatch.h' \
re='(\<(ARRAY_CARDINALITY|X?ARGMATCH(|_TO_ARGUMENT|_VERIFY))\>|\<(invalid_arg|argmatch(_exit_fn|_(in)?valid)?) *\()' \
$(_sc_header_without_use)
sc_prohibit_canonicalize_without_use:
@h='canonicalize.h' \
re='CAN_(EXISTING|ALL_BUT_LAST|MISSING)|canonicalize_(mode_t|filename_mode|file_name)' \
$(_sc_header_without_use)
sc_prohibit_root_dev_ino_without_use:
@h='root-dev-ino.h' \
re='(\<ROOT_DEV_INO_(CHECK|WARN)\>|\<get_root_dev_ino *\()' \
$(_sc_header_without_use)
sc_prohibit_openat_without_use:
@h='openat.h' \
re='\<(openat_(permissive|needs_fchdir|(save|restore)_fail)|l?(stat|ch(own|mod))at|(euid)?accessat|(FCHMOD|FCHOWN|STAT)AT_INLINE)\>' \
$(_sc_header_without_use)
# Prohibit the inclusion of c-ctype.h without an actual use.
ctype_re = isalnum|isalpha|isascii|isblank|iscntrl|isdigit|isgraph|islower\
|isprint|ispunct|isspace|isupper|isxdigit|tolower|toupper
sc_prohibit_c_ctype_without_use:
@h='c-ctype.h' re='\<c_($(ctype_re)) *\(' \
$(_sc_header_without_use)
# The following list was generated by running:
# man signal.h|col -b|perl -ne '/bsd_signal.*;/.../sigwaitinfo.*;/ and print' \
# | perl -lne '/^\s+(?:int|void).*?(\w+).*/ and print $1' | fmt
_sig_functions = \
bsd_signal kill killpg pthread_kill pthread_sigmask raise sigaction \
sigaddset sigaltstack sigdelset sigemptyset sigfillset sighold sigignore \
siginterrupt sigismember signal sigpause sigpending sigprocmask sigqueue \
sigrelse sigset sigsuspend sigtimedwait sigwait sigwaitinfo
_sig_function_re = $(subst $(_sp),|,$(strip $(_sig_functions)))
# The following were extracted from "man signal.h" manually.
_sig_types_and_consts = \
MINSIGSTKSZ SA_NOCLDSTOP SA_NOCLDWAIT SA_NODEFER SA_ONSTACK \
SA_RESETHAND SA_RESTART SA_SIGINFO SIGEV_NONE SIGEV_SIGNAL \
SIGEV_THREAD SIGSTKSZ SIG_BLOCK SIG_SETMASK SIG_UNBLOCK SS_DISABLE \
SS_ONSTACK mcontext_t pid_t sig_atomic_t sigevent siginfo_t sigset_t \
sigstack sigval stack_t ucontext_t
# generated via this:
# perl -lne '/^#ifdef (SIG\w+)/ and print $1' lib/sig2str.c|sort -u|fmt -70
_sig_names = \
SIGABRT SIGALRM SIGALRM1 SIGBUS SIGCANCEL SIGCHLD SIGCLD SIGCONT \
SIGDANGER SIGDIL SIGEMT SIGFPE SIGFREEZE SIGGRANT SIGHUP SIGILL \
SIGINFO SIGINT SIGIO SIGIOT SIGKAP SIGKILL SIGKILLTHR SIGLOST SIGLWP \
SIGMIGRATE SIGMSG SIGPHONE SIGPIPE SIGPOLL SIGPRE SIGPROF SIGPWR \
SIGQUIT SIGRETRACT SIGSAK SIGSEGV SIGSOUND SIGSTKFLT SIGSTOP SIGSYS \
SIGTERM SIGTHAW SIGTRAP SIGTSTP SIGTTIN SIGTTOU SIGURG SIGUSR1 \
SIGUSR2 SIGVIRT SIGVTALRM SIGWAITING SIGWINCH SIGWIND SIGWINDOW \
SIGXCPU SIGXFSZ
_sig_syms_re = $(subst $(_sp),|,$(strip $(_sig_names) $(_sig_types_and_consts)))
# Prohibit the inclusion of signal.h without an actual use.
sc_prohibit_signal_without_use:
@h='signal.h' \
re='\<($(_sig_function_re)) *\(|\<($(_sig_syms_re))\>' \
$(_sc_header_without_use)
# Don't include stdio--.h unless you use one of its functions.
sc_prohibit_stdio--_without_use:
@h='stdio--.h' re='\<((f(re)?|p)open|tmpfile) *\(' \
$(_sc_header_without_use)
# Don't include stdio-safer.h unless you use one of its functions.
sc_prohibit_stdio-safer_without_use:
@h='stdio-safer.h' re='\<((f(re)?|p)open|tmpfile)_safer *\(' \
$(_sc_header_without_use)
# Prohibit the inclusion of strings.h without a sensible use.
# Using the likes of bcmp, bcopy, bzero, index or rindex is not sensible.
sc_prohibit_strings_without_use:
@h='strings.h' \
re='\<(strn?casecmp|ffs(ll)?)\>' \
$(_sc_header_without_use)
# Get the list of symbol names with this:
# perl -lne '/^# *define ([A-Z]\w+)\(/ and print $1' lib/intprops.h|fmt
_intprops_names = \
TYPE_IS_INTEGER TYPE_TWOS_COMPLEMENT TYPE_ONES_COMPLEMENT \
TYPE_SIGNED_MAGNITUDE TYPE_SIGNED TYPE_MINIMUM TYPE_MAXIMUM \
INT_BITS_STRLEN_BOUND INT_STRLEN_BOUND INT_BUFSIZE_BOUND \
INT_ADD_RANGE_OVERFLOW INT_SUBTRACT_RANGE_OVERFLOW \
INT_NEGATE_RANGE_OVERFLOW INT_MULTIPLY_RANGE_OVERFLOW \
INT_DIVIDE_RANGE_OVERFLOW INT_REMAINDER_RANGE_OVERFLOW \
INT_LEFT_SHIFT_RANGE_OVERFLOW INT_ADD_OVERFLOW INT_SUBTRACT_OVERFLOW \
INT_NEGATE_OVERFLOW INT_MULTIPLY_OVERFLOW INT_DIVIDE_OVERFLOW \
INT_REMAINDER_OVERFLOW INT_LEFT_SHIFT_OVERFLOW
_intprops_syms_re = $(subst $(_sp),|,$(strip $(_intprops_names)))
# Prohibit the inclusion of intprops.h without an actual use.
sc_prohibit_intprops_without_use:
@h='intprops.h' \
re='\<($(_intprops_syms_re)) *\(' \
$(_sc_header_without_use)
_stddef_syms_re = NULL|offsetof|ptrdiff_t|size_t|wchar_t
# Prohibit the inclusion of stddef.h without an actual use.
sc_prohibit_stddef_without_use:
@h='stddef.h' \
re='\<($(_stddef_syms_re))\>' \
$(_sc_header_without_use)
_de1 = dirfd|(close|(fd)?open|read|rewind|seek|tell)dir(64)?(_r)?
_de2 = (versionsort|struct dirent|getdirentries|alphasort|scandir(at)?)(64)?
_de3 = MAXNAMLEN|DIR|ino_t|d_ino|d_fileno|d_namlen
_dirent_syms_re = $(_de1)|$(_de2)|$(_de3)
# Prohibit the inclusion of dirent.h without an actual use.
sc_prohibit_dirent_without_use:
@h='dirent.h' \
re='\<($(_dirent_syms_re))\>' \
$(_sc_header_without_use)
# Prohibit the inclusion of verify.h without an actual use.
sc_prohibit_verify_without_use:
@h='verify.h' \
re='\<(verify(true|expr)?|static_assert) *\(' \
$(_sc_header_without_use)
# Don't include xfreopen.h unless you use one of its functions.
sc_prohibit_xfreopen_without_use:
@h='xfreopen.h' re='\<xfreopen *\(' $(_sc_header_without_use)
sc_obsolete_symbols:
@prohibit='\<(HAVE''_FCNTL_H|O''_NDELAY)\>' \
halt='do not use HAVE''_FCNTL_H or O'_NDELAY \
$(_sc_search_regexp)
# FIXME: warn about definitions of EXIT_FAILURE, EXIT_SUCCESS, STREQ
# Each nonempty ChangeLog line must start with a year number, or a TAB.
sc_changelog:
@prohibit='^[^12 ]' \
in_vc_files='^ChangeLog$$' \
halt='found unexpected prefix in a ChangeLog' \
$(_sc_search_regexp)
# Ensure that each .c file containing a "main" function also
# calls set_program_name.
sc_program_name:
@require='set_program_name *\(.*\);' \
in_vc_files='\.c$$' \
containing='\<main *(' \
halt='the above files do not call set_program_name' \
$(_sc_search_regexp)
# Ensure that each .c file containing a "main" function also
# calls bindtextdomain.
sc_bindtextdomain:
@require='bindtextdomain *\(' \
in_vc_files='\.c$$' \
containing='\<main *(' \
halt='the above files do not call bindtextdomain' \
$(_sc_search_regexp)
# Require that the final line of each test-lib.sh-using test be this one:
# Exit $fail
# Note: this test requires GNU grep's --label= option.
Exit_witness_file ?= tests/test-lib.sh
Exit_base := $(notdir $(Exit_witness_file))
sc_require_test_exit_idiom:
@if test -f $(srcdir)/$(Exit_witness_file); then \
die=0; \
for i in $$(grep -l -F 'srcdir/$(Exit_base)' \
$$($(VC_LIST) tests)); do \
tail -n1 $$i | grep '^Exit .' > /dev/null \
&& : || { die=1; echo $$i; } \
done; \
test $$die = 1 && \
{ echo 1>&2 '$(ME): the final line in each of the above is not:'; \
echo 1>&2 'Exit something'; \
exit 1; } || :; \
fi
sc_trailing_blank:
@prohibit='[ ]$$' \
halt='found trailing blank(s)' \
exclude='^Binary file .* matches$$' \
$(_sc_search_regexp)
# Match lines like the following, but where there is only one space
# between the options and the description:
# -D, --all-repeated[=delimit-method] print all duplicate lines\n
longopt_re = --[a-z][0-9A-Za-z-]*(\[?=[0-9A-Za-z-]*\]?)?
sc_two_space_separator_in_usage:
@prohibit='^ *(-[A-Za-z],)? $(longopt_re) [^ ].*\\$$' \
halt='help2man requires at least two spaces between an option and its description'\
$(_sc_search_regexp)
# A regexp matching function names like "error" that may be used
# to emit translatable messages.
_gl_translatable_diag_func_re ?= error
# Look for diagnostics that aren't marked for translation.
# This won't find any for which error's format string is on a separate line.
sc_unmarked_diagnostics:
@prohibit='\<$(_gl_translatable_diag_func_re) *\([^"]*"[^"]*[a-z]{3}' \
exclude='(_|ngettext ?)\(' \
halt='found unmarked diagnostic(s)' \
$(_sc_search_regexp)
# Avoid useless parentheses like those in this example:
# #if defined (SYMBOL) || defined (SYM2)
sc_useless_cpp_parens:
@prohibit='^# *if .*defined *\(' \
halt='found useless parentheses in cpp directive' \
$(_sc_search_regexp)
sc_prohibit_defined_have_decl_tests:
@prohibit='(#[ ]*ifn?def|\<defined)\>[ (]+HAVE_DECL_' \
halt='HAVE_DECL macros are always defined' \
$(_sc_search_regexp)
# ==================================================================
gl_other_headers_ ?= \
intprops.h \
openat.h \
stat-macros.h
# Perl -lne code to extract "significant" cpp-defined symbols from a
# gnulib header file, eliminating a few common false-positives.
# The exempted names below are defined only conditionally in gnulib,
# and hence sometimes must/may be defined in application code.
gl_extract_significant_defines_ = \
/^\# *define ([^_ (][^ (]*)(\s*\(|\s+\w+)/\
&& $$2 !~ /(?:rpl_|_used_without_)/\
&& $$1 !~ /^(?:NSIG|ENODATA)$$/\
&& $$1 !~ /^(?:SA_RESETHAND|SA_RESTART)$$/\
and print $$1
# Create a list of regular expressions matching the names
# of macros that are guaranteed to be defined by parts of gnulib.
define def_sym_regex
gen_h=$(gl_generated_headers_); \
(cd $(gnulib_dir)/lib; \
for f in *.in.h $(gl_other_headers_); do \
test -f $$f \
&& perl -lne '$(gl_extract_significant_defines_)' $$f; \
done; \
) | sort -u \
| $(SED) 's/^/^ *# *(define|undef) */;s/$$/\\>/'
endef
# Don't define macros that we already get from gnulib header files.
sc_prohibit_always-defined_macros:
@if test -d $(gnulib_dir); then \
case $$(echo all: | grep -l -f - Makefile) in Makefile);; *) \
echo '$(ME): skipping $@: you lack GNU grep' 1>&2; exit 0;; \
esac; \
$(def_sym_regex) | grep -E -f - $$($(VC_LIST_EXCEPT)) \
&& { echo '$(ME): define the above via some gnulib .h file' \
1>&2; exit 1; } || :; \
fi
# ==================================================================
# Prohibit checked in backup files.
sc_prohibit_backup_files:
@$(VC_LIST) | grep '~$$' && \
{ echo '$(ME): found version controlled backup file' 1>&2; \
exit 1; } || :
# Require the GPL.
sc_GPL_version:
@prohibit='either ''version [^23]' \
halt='GPL vN, N!=3' \
$(_sc_search_regexp)
# Require the latest GFDL. Two regexp, since some .texi files end up
# line wrapping between 'Free Documentation License,' and 'Version'.
_GFDL_regexp = (Free ''Documentation.*Version 1\.[^3]|Version 1\.[^3] or any)
sc_GFDL_version:
@prohibit='$(_GFDL_regexp)' \
halt='GFDL vN, N!=3' \
$(_sc_search_regexp)
# Don't use Texinfo's @acronym{}.
# http://lists.gnu.org/archive/html/bug-gnulib/2010-03/msg00321.html
texinfo_suffix_re_ ?= \.(txi|texi(nfo)?)$$
sc_texinfo_acronym:
@prohibit='@acronym\{' \
in_vc_files='$(texinfo_suffix_re_)' \
halt='found use of Texinfo @acronym{}' \
$(_sc_search_regexp)
cvs_keywords = \
Author|Date|Header|Id|Name|Locker|Log|RCSfile|Revision|Source|State
sc_prohibit_cvs_keyword:
@prohibit='\$$($(cvs_keywords))\$$' \
halt='do not use CVS keyword expansion' \
$(_sc_search_regexp)
# This Perl code is slightly obfuscated. Not only is each "$" doubled
# because it's in a Makefile, but the $$c's are comments; we cannot
# use "#" due to the way the script ends up concatenated onto one line.
# It would be much more concise, and would produce better output (including
# counts) if written as:
# perl -ln -0777 -e '/\n(\n+)$/ and print "$ARGV: ".length $1' ...
# but that would be far less efficient, reading the entire contents
# of each file, rather than just the last two bytes of each.
# In addition, while the code below detects both blank lines and a missing
# newline at EOF, the above detects only the former.
#
# This is a perl script that is expected to be the single-quoted argument
# to a command-line "-le". The remaining arguments are file names.
# Print the name of each file that does not end in exactly one newline byte.
# I.e., warn if there are blank lines (2 or more newlines), or if the
# last byte is not a newline. However, currently we don't complain
# about any file that contains exactly one byte.
# Exit nonzero if at least one such file is found, otherwise, exit 0.
# Warn about, but otherwise ignore open failure. Ignore seek/read failure.
#
# Use this if you want to remove trailing empty lines from selected files:
# perl -pi -0777 -e 's/\n\n+$/\n/' files...
#
require_exactly_one_NL_at_EOF_ = \
foreach my $$f (@ARGV) \
{ \
open F, "<", $$f or (warn "failed to open $$f: $$!\n"), next; \
my $$p = sysseek (F, -2, 2); \
my $$c = "seek failure probably means file has < 2 bytes; ignore"; \
my $$last_two_bytes; \
defined $$p and $$p = sysread F, $$last_two_bytes, 2; \
close F; \
$$c = "ignore read failure"; \
$$p && ($$last_two_bytes eq "\n\n" \
|| substr ($$last_two_bytes,1) ne "\n") \
and (print $$f), $$fail=1; \
} \
END { exit defined $$fail }
sc_prohibit_empty_lines_at_EOF:
@perl -le '$(require_exactly_one_NL_at_EOF_)' $$($(VC_LIST_EXCEPT)) \
|| { echo '$(ME): empty line(s) or no newline at EOF' \
1>&2; exit 1; } || :
# Make sure we don't use st_blocks. Use ST_NBLOCKS instead.
# This is a bit of a kludge, since it prevents use of the string
# even in comments, but for now it does the job with no false positives.
sc_prohibit_stat_st_blocks:
@prohibit='[.>]st_blocks' \
halt='do not use st_blocks; use ST_NBLOCKS' \
$(_sc_search_regexp)
# Make sure we don't define any S_IS* macros in src/*.c files.
# They're already defined via gnulib's sys/stat.h replacement.
sc_prohibit_S_IS_definition:
@prohibit='^ *# *define *S_IS' \
halt='do not define S_IS* macros; include <sys/stat.h>' \
$(_sc_search_regexp)
# Perl block to convert a match to FILE_NAME:LINENO:TEST,
# that is shared by two definitions below.
perl_filename_lineno_text_ = \
-e ' {' \
-e ' $$n = ($$` =~ tr/\n/\n/ + 1);' \
-e ' ($$v = $$&) =~ s/\n/\\n/g;' \
-e ' print "$$ARGV:$$n:$$v\n";' \
-e ' }'
prohibit_doubled_word_RE_ ?= \
/\b(then?|[iao]n|i[fst]|but|f?or|at|and|[dt]o)\s+\1\b/gims
prohibit_doubled_word_ = \
-e 'while ($(prohibit_doubled_word_RE_))' \
$(perl_filename_lineno_text_)
# Define this to a regular expression that matches
# any filename:dd:match lines you want to ignore.
# The default is to ignore no matches.
ignore_doubled_word_match_RE_ ?= ^$$
sc_prohibit_doubled_word:
@perl -n -0777 $(prohibit_doubled_word_) $$($(VC_LIST_EXCEPT)) \
| grep -vE '$(ignore_doubled_word_match_RE_)' \
| grep . && { echo '$(ME): doubled words' 1>&2; exit 1; } || :
# A regular expression matching undesirable combinations of words like
# "can not"; this matches them even when the two words appear on different
# lines, but not when there is an intervening delimiter like "#" or "*".
# Similarly undesirable, "See @xref{...}", since an @xref should start
# a sentence. Explicitly prohibit any prefix of "see" or "also".
# Also prohibit a prefix matching "\w+ +".
# @pxref gets the same see/also treatment and should be parenthesized;
# presume it must *not* start a sentence.
bad_xref_re_ ?= (?:[\w,:;] +|(?:see|also)\s+)\@xref\{
bad_pxref_re_ ?= (?:[.!?]|(?:see|also))\s+\@pxref\{
prohibit_undesirable_word_seq_RE_ ?= \
/(?:\bcan\s+not\b|$(bad_xref_re_)|$(bad_pxref_re_))/gims
prohibit_undesirable_word_seq_ = \
-e 'while ($(prohibit_undesirable_word_seq_RE_))' \
$(perl_filename_lineno_text_)
# Define this to a regular expression that matches
# any filename:dd:match lines you want to ignore.
# The default is to ignore no matches.
ignore_undesirable_word_sequence_RE_ ?= ^$$
sc_prohibit_undesirable_word_seq:
@perl -n -0777 $(prohibit_undesirable_word_seq_) \
$$($(VC_LIST_EXCEPT)) \
| grep -vE '$(ignore_undesirable_word_sequence_RE_)' | grep . \
&& { echo '$(ME): undesirable word sequence' >&2; exit 1; } || :
# Except for shell files and for loops, double semicolon is probably a mistake
sc_prohibit_double_semicolon:
@prohibit='; *;[ {} \]*(/[/*]|$$)' \
in_vc_files='\.[chly]$$' \
exclude='\bfor *\(.*\)' \
halt="Double semicolon detected" \
$(_sc_search_regexp)
_ptm1 = use "test C1 && test C2", not "test C1 -''a C2"
_ptm2 = use "test C1 || test C2", not "test C1 -''o C2"
# Using test's -a and -o operators is not portable.
# We prefer test over [, since the latter is spelled [[ in configure.ac.
sc_prohibit_test_minus_ao:
@prohibit='(\<test| \[+) .+ -[ao] ' \
halt='$(_ptm1); $(_ptm2)' \
$(_sc_search_regexp)
# Avoid a test bashism.
sc_prohibit_test_double_equal:
@prohibit='(\<test| \[+) .+ == ' \
containing='#! */bin/[a-z]*sh' \
halt='use "test x = x", not "test x =''= x"' \
$(_sc_search_regexp)
# Each program that uses proper_name_utf8 must link with one of the
# ICONV libraries. Otherwise, some ICONV library must appear in LDADD.
# The perl -0777 invocation below extracts the possibly-multi-line
# definition of LDADD from the appropriate Makefile.am and exits 0
# when it contains "ICONV".
sc_proper_name_utf8_requires_ICONV:
@progs=$$(grep -l 'proper_name_utf8 ''("' $$($(VC_LIST_EXCEPT)));\
if test "x$$progs" != x; then \
fail=0; \
for p in $$progs; do \
dir=$$(dirname "$$p"); \
perl -0777 \
-ne 'exit !(/^LDADD =(.+?[^\\]\n)/ms && $$1 =~ /ICONV/)' \
$$dir/Makefile.am && continue; \
base=$$(basename "$$p" .c); \
grep "$${base}_LDADD.*ICONV)" $$dir/Makefile.am > /dev/null \
|| { fail=1; echo 1>&2 "$(ME): $$p uses proper_name_utf8"; }; \
done; \
test $$fail = 1 && \
{ echo 1>&2 '$(ME): the above do not link with any ICONV library'; \
exit 1; } || :; \
fi
# Warn about "c0nst struct Foo const foo[]",
# but not about "char const *const foo" or "#define const const".
sc_redundant_const:
@prohibit='\bconst\b[[:space:][:alnum:]]{2,}\bconst\b' \
halt='redundant "const" in declarations' \
$(_sc_search_regexp)
sc_const_long_option:
@prohibit='^ *static.*struct option ' \
exclude='const struct option|struct option const' \
halt='add "const" to the above declarations' \
$(_sc_search_regexp)
sc_prohibit_gs_strfreev:
@prohibit='\bgs_strfreev\b' \
halt="do not use gs_strfreev; use g_autoptr instead" \
$(_sc_search_regexp)
sc_prohibit_gs_unref:
@prohibit='\bgs_unref_(keyfile|variant*|ptrarray|hashtable|bytes|object)\b' \
halt="do not use gs_unref_*; use g_autoptr instead" \
$(_sc_search_regexp)
NEWS_hash = \
$$($(SED) -n '/^\*.* $(PREV_VERSION_REGEXP) ([0-9-]*)/,$$p' \
$(srcdir)/NEWS \
| perl -0777 -pe \
's/^Copyright.+?Free\sSoftware\sFoundation,\sInc\.\n//ms' \
| md5sum - \
| $(SED) 's/ .*//')
# Ensure that we use only the standard $(VAR) notation,
# not @...@ in Makefile.am, now that we can rely on automake
# to emit a definition for each substituted variable.
# However, there is still one case in which @VAR@ use is not just
# legitimate, but actually required: when augmenting an automake-defined
# variable with a prefix. For example, gettext uses this:
# MAKEINFO = env LANG= LC_MESSAGES= LC_ALL= LANGUAGE= @MAKEINFO@
# otherwise, makeinfo would put German or French (current locale)
# navigation hints in the otherwise-English documentation.
#
# Allow the package to add exceptions via a hook in cfg.mk;
# for example, @PRAGMA_SYSTEM_HEADER@ can be permitted by
# setting this to ' && !/PRAGMA_SYSTEM_HEADER/'.
_makefile_at_at_check_exceptions ?=
sc_makefile_at_at_check:
@perl -ne '/\@\w+\@/' \
-e ' && !/(\w+)\s+=.*\@\1\@$$/' \
-e ''$(_makefile_at_at_check_exceptions) \
-e 'and (print "$$ARGV:$$.: $$_"), $$m=1; END {exit !$$m}' \
$$($(VC_LIST_EXCEPT) | grep -E '(^|/)(Makefile\.am|[^/]+\.mk)$$') \
&& { echo '$(ME): use $$(...), not @...@' 1>&2; exit 1; } || :
news-check: NEWS
$(AM_V_GEN)if $(SED) -n $(news-check-lines-spec)p $< \
| grep -E $(news-check-regexp) >/dev/null; then \
:; \
else \
echo 'NEWS: $$(news-check-regexp) failed to match' 1>&2; \
exit 1; \
fi
sc_makefile_TAB_only_indentation:
@prohibit='^ [ ]{8}' \
in_vc_files='akefile|\.mk$$' \
halt='found TAB-8-space indentation' \
$(_sc_search_regexp)
sc_m4_quote_check:
@prohibit='(AC_DEFINE(_UNQUOTED)?|AC_DEFUN)\([^[]' \
in_vc_files='(^configure\.ac|\.m4)$$' \
halt='quote the first arg to AC_DEF*' \
$(_sc_search_regexp)
fix_po_file_diag = \
'you have changed the set of files with translatable diagnostics;\n\
apply the above patch\n'
# Verify that all source files using _() (more specifically, files that
# match $(_gl_translatable_string_re)) are listed in po/POTFILES.in.
po_file ?= $(srcdir)/po/POTFILES.in
generated_files ?= $(srcdir)/lib/*.[ch]
_gl_translatable_string_re ?= \b(N?_|gettext *)\([^)"]*("|$$)
sc_po_check:
@if test -f $(po_file); then \
grep -E -v '^(#|$$)' $(po_file) \
| grep -v '^src/false\.c$$' | sort > $@-1; \
files=; \
for file in $$($(VC_LIST_EXCEPT)) $(generated_files); do \
test -r $$file || continue; \
case $$file in \
*.m4|*.mk) continue ;; \
*.?|*.??) ;; \
*) continue;; \
esac; \
case $$file in \
*.[ch]) \
base=`expr " $$file" : ' \(.*\)\..'`; \
{ test -f $$base.l || test -f $$base.y; } && continue;; \
esac; \
files="$$files $$file"; \
done; \
grep -E -l '$(_gl_translatable_string_re)' $$files \
| $(SED) 's|^$(_dot_escaped_srcdir)/||' | sort -u > $@-2; \
diff -u -L $(po_file) -L $(po_file) $@-1 $@-2 \
|| { printf '$(ME): '$(fix_po_file_diag) 1>&2; exit 1; }; \
rm -f $@-1 $@-2; \
fi
# Sometimes it is useful to change the PATH environment variable
# in Makefiles. When doing so, it's better not to use the Unix-centric
# path separator of ':', but rather the automake-provided '$(PATH_SEPARATOR)'.
msg = 'Do not use ":" above; use $$(PATH_SEPARATOR) instead'
sc_makefile_path_separator_check:
@prohibit='PATH[=].*:' \
in_vc_files='akefile|\.mk$$' \
halt=$(msg) \
$(_sc_search_regexp)
v_etc_file = $(gnulib_dir)/lib/version-etc.c
sample-test = tests/sample-test
texi = doc/$(PACKAGE).texi
# Make sure that the copyright date in $(v_etc_file) is up to date.
# Do the same for the $(sample-test) and the main doc/.texi file.
sc_copyright_check:
@require='enum { COPYRIGHT_YEAR = '$$(date +%Y)' };' \
in_files=$(v_etc_file) \
halt='out of date copyright in $(v_etc_file); update it' \
$(_sc_search_regexp)
@require='# Copyright \(C\) '$$(date +%Y)' Free' \
in_vc_files=$(sample-test) \
halt='out of date copyright in $(sample-test); update it' \
$(_sc_search_regexp)
@require='Copyright @copyright\{\} .*'$$(date +%Y) \
in_vc_files=$(texi) \
halt='out of date copyright in $(texi); update it' \
$(_sc_search_regexp)
sc_missing_cmd_prefix:
@prohibit='^ostree ' \
halt='found missing ${CMD_PREFIX}' \
in_vc_files='$tests/[^/]*\.sh$$' \
$(_sc_search_regexp)
# If tests/help-version exists and seems to be new enough, assume that its
# use of init.sh and path_prepend_ is correct, and ensure that every other
# use of init.sh is identical.
# This is useful because help-version cross-checks prog --version
# with $(VERSION), which verifies that its path_prepend_ invocation
# sets PATH correctly. This is an inexpensive way to ensure that
# the other init.sh-using tests also get it right.
_hv_file ?= $(srcdir)/tests/help-version
_hv_regex_weak ?= ^ *\. .*/init\.sh"
# Fix syntax-highlighters "
_hv_regex_strong ?= ^ *\. "\$${srcdir=\.}/init\.sh"
sc_cross_check_PATH_usage_in_tests:
@if test -f $(_hv_file); then \
grep -l 'VERSION mismatch' $(_hv_file) >/dev/null \
|| { echo "$@: skipped: no such file: $(_hv_file)" 1>&2; \
exit 0; }; \
grep -lE '$(_hv_regex_strong)' $(_hv_file) >/dev/null \
|| { echo "$@: $(_hv_file) lacks conforming use of init.sh" 1>&2; \
exit 1; }; \
good=$$(grep -E '$(_hv_regex_strong)' $(_hv_file)); \
grep -LFx "$$good" \
$$(grep -lE '$(_hv_regex_weak)' $$($(VC_LIST_EXCEPT))) \
| grep . && \
{ echo "$(ME): the above files use path_prepend_ inconsistently" \
1>&2; exit 1; } || :; \
fi
# BRE regex of file contents to identify a test script.
_test_script_regex ?= \<init\.sh\>
# In tests, use "compare expected actual", not the reverse.
sc_prohibit_reversed_compare_failure:
@prohibit='\<compare [^ ]+ ([^ ]*exp|/dev/null)' \
containing='$(_test_script_regex)' \
halt='reversed compare arguments' \
$(_sc_search_regexp)
# #if HAVE_... will evaluate to false for any non numeric string.
# That would be flagged by using -Wundef, however gnulib currently
# tests many undefined macros, and so we can't enable that option.
# So at least preclude common boolean strings as macro values.
sc_Wundef_boolean:
@prohibit='^#define.*(yes|no|true|false)$$' \
in_files='$(CONFIG_INCLUDE)' \
halt='Use 0 or 1 for macro values' \
$(_sc_search_regexp)
# Even if you use pathmax.h to guarantee that PATH_MAX is defined, it might
# not be constant, or might overflow a stack. In general, use PATH_MAX as
# a limit, not an array or alloca size.
sc_prohibit_path_max_allocation:
@prohibit='(\balloca *\([^)]*|\[[^]]*)\bPATH_MAX' \
halt='Avoid stack allocations of size PATH_MAX' \
$(_sc_search_regexp)
sc_vulnerable_makefile_CVE-2009-4029:
@prohibit='perm -777 -exec chmod a\+rwx|chmod 777 \$$\(distdir\)' \
in_files='(^|/)Makefile\.in$$' \
halt=$$(printf '%s\n' \
'the above files are vulnerable; beware of running' \
' "make dist*" rules, and upgrade to fixed automake' \
' see http://bugzilla.redhat.com/542609 for details') \
$(_sc_search_regexp)
sc_vulnerable_makefile_CVE-2012-3386:
@prohibit='chmod a\+w \$$\(distdir\)' \
in_files='(^|/)Makefile\.in$$' \
halt=$$(printf '%s\n' \
'the above files are vulnerable; beware of running' \
' "make distcheck", and upgrade to fixed automake' \
' see http://bugzilla.redhat.com/CVE-2012-3386 for details') \
$(_sc_search_regexp)
vc-diff-check:
$(AM_V_GEN)(unset CDPATH; cd $(srcdir) && $(VC) diff) > vc-diffs || :
$(AM_V_at)if test -s vc-diffs; then \
cat vc-diffs; \
echo "Some files are locally modified:" 1>&2; \
exit 1; \
else \
rm vc-diffs; \
fi
.PHONY: no-submodule-changes
no-submodule-changes:
$(AM_V_GEN)if test -d $(srcdir)/.git \
&& git --version >/dev/null 2>&1; then \
diff=$$(cd $(srcdir) && git submodule -q foreach \
git diff-index --name-only HEAD) \
|| exit 1; \
case $$diff in '') ;; \
*) echo '$(ME): submodule files are locally modified:'; \
echo "$$diff"; exit 1;; esac; \
else \
: ; \
fi
submodule-checks ?= no-submodule-changes public-submodule-commit
# Ensure that each sub-module commit we're using is public.
# Without this, it is too easy to tag and release code that
# cannot be built from a fresh clone.
.PHONY: public-submodule-commit
public-submodule-commit:
$(AM_V_GEN)if test -d $(srcdir)/.git \
&& git --version >/dev/null 2>&1; then \
cd $(srcdir) && \
git submodule --quiet foreach \
'test "$$(git rev-parse "$$sha1")" \
= "$$(git merge-base origin "$$sha1")"' \
|| { echo '$(ME): found non-public submodule commit' >&2; \
exit 1; }; \
else \
: ; \
fi
# Code Coverage
init-coverage:
$(MAKE) $(AM_MAKEFLAGS) clean
lcov --directory . --zerocounters
COVERAGE_CCOPTS ?= "-g --coverage"
COVERAGE_OUT ?= doc/coverage
build-coverage:
$(MAKE) $(AM_MAKEFLAGS) CFLAGS=$(COVERAGE_CCOPTS) CXXFLAGS=$(COVERAGE_CCOPTS)
$(MAKE) $(AM_MAKEFLAGS) CFLAGS=$(COVERAGE_CCOPTS) CXXFLAGS=$(COVERAGE_CCOPTS) check
mkdir -p $(COVERAGE_OUT)
lcov --directory . --output-file $(COVERAGE_OUT)/$(PACKAGE).info \
--capture
gen-coverage:
genhtml --output-directory $(COVERAGE_OUT) \
$(COVERAGE_OUT)/$(PACKAGE).info \
--highlight --frames --legend \
--title "$(PACKAGE_NAME)"
coverage: init-coverage build-coverage gen-coverage
# Some projects carry local adjustments for gnulib modules via patches in
# a gnulib patch directory whose default name is gl/ (defined in bootstrap
# via local_gl_dir=gl). Those patches become stale as the originals evolve
# in gnulib. Use this rule to refresh any stale patches. It applies each
# patch to the original in $(gnulib_dir) and uses the temporary result to
# generate a fuzz-free .diff file. If you customize the name of your local
# gnulib patch directory via bootstrap.conf, this rule detects that name.
# Run this from a non-VPATH (i.e., srcdir) build directory.
.PHONY: refresh-gnulib-patches
refresh-gnulib-patches:
gl=gl; \
if test -f bootstrap.conf; then \
t=$$(perl -lne '/^\s*local_gl_dir=(\S+)/ and $$d=$$1;' \
-e 'END{defined $$d and print $$d}' bootstrap.conf); \
test -n "$$t" && gl=$$t; \
fi; \
for diff in $$(cd $$gl; git ls-files | grep '\.diff$$'); do \
b=$$(printf %s "$$diff"|$(SED) 's/\.diff$$//'); \
VERSION_CONTROL=none \
patch "$(gnulib_dir)/$$b" "$$gl/$$diff" || exit 1; \
( cd $(gnulib_dir) || exit 1; \
git diff "$$b" > "../$$gl/$$diff"; \
git checkout $$b ) || exit 1; \
done
# Update gettext files.
PACKAGE ?= $(shell basename $(PWD))
PO_DOMAIN ?= $(PACKAGE)
POURL = http://translationproject.org/latest/$(PO_DOMAIN)/
PODIR ?= po
refresh-po:
rm -f $(PODIR)/*.po && \
echo "$(ME): getting translations into po (please ignore the robots.txt ERROR 404)..." && \
wget --no-verbose --directory-prefix $(PODIR) --no-directories --recursive --level 1 --accept .po --accept .po.1 $(POURL) && \
echo 'en@boldquot' > $(PODIR)/LINGUAS && \
echo 'en@quot' >> $(PODIR)/LINGUAS && \
ls $(PODIR)/*.po | $(SED) 's/\.po//;s,$(PODIR)/,,' | \
sort >> $(PODIR)/LINGUAS
# Running indent once is not idempotent, but running it twice is.
INDENT_SOURCES ?= $(C_SOURCES)
.PHONY: indent
indent:
indent $(INDENT_SOURCES)
indent $(INDENT_SOURCES)
# If you want to set UPDATE_COPYRIGHT_* environment variables,
# put the assignments in this variable.
update-copyright-env ?=
# Run this rule once per year (usually early in January)
# to update all FSF copyright year lists in your project.
# If you have an additional project-specific rule,
# add it in cfg.mk along with a line 'update-copyright: prereq'.
# By default, exclude all variants of COPYING; you can also
# add exemptions (such as ChangeLog..* for rotated change logs)
# in the file .x-update-copyright.
.PHONY: update-copyright
update-copyright:
$(AM_V_GEN)grep -l -w Copyright \
$$(export VC_LIST_EXCEPT_DEFAULT=COPYING && $(VC_LIST_EXCEPT)) \
| $(update-copyright-env) xargs $(srcdir)/$(_build-aux)/$@
# This tight_scope test is skipped with a warning if $(_gl_TS_headers) is not
# overridden and $(_gl_TS_dir)/Makefile.am does not mention noinst_HEADERS.
# NOTE: to override any _gl_TS_* default value, you must
# define the variable(s) using "export" in cfg.mk.
_gl_TS_dir ?= src
ALL_RECURSIVE_TARGETS += sc_tight_scope
sc_tight_scope: tight-scope.mk
@fail=0; \
if ! grep '^ *export _gl_TS_headers *=' $(srcdir)/cfg.mk \
> /dev/null \
&& ! grep -w noinst_HEADERS $(srcdir)/$(_gl_TS_dir)/Makefile.am \
> /dev/null 2>&1; then \
echo '$(ME): skipping $@'; \
else \
$(MAKE) -s -C $(_gl_TS_dir) \
-f Makefile \
-f $(abs_top_srcdir)/cfg.mk \
-f $(abs_top_builddir)/$< \
_gl_tight_scope \
|| fail=1; \
fi; \
rm -f $<; \
exit $$fail
tight-scope.mk: $(ME)
@rm -f $@ $@-t
@perl -ne '/^# TS-start/.../^# TS-end/ and print' $(srcdir)/$(ME) > $@-t
@chmod a=r $@-t && mv $@-t $@
ifeq (a,b)
# TS-start
# Most functions should have static scope.
# Any that don't must be marked with 'extern', but 'main'
# and 'usage' are exceptions: they're always extern, but
# do not need to be marked. Symbols matching '__.*' are
# reserved by the compiler, so are automatically excluded below.
_gl_TS_unmarked_extern_functions ?= main usage
_gl_TS_function_match ?= /^(?:$(_gl_TS_extern)) +.*?(\S+) *\(/
# If your project uses a macro like "XTERN", then put
# the following in cfg.mk to override this default:
# export _gl_TS_extern = extern|XTERN
_gl_TS_extern ?= extern
# The second nm|grep checks for file-scope variables with 'extern' scope.
# Without gnulib's progname module, you might put program_name here.
# Symbols matching '__.*' are reserved by the compiler,
# so are automatically excluded below.
_gl_TS_unmarked_extern_vars ?=
# NOTE: the _match variables are perl expressions -- not mere regular
# expressions -- so that you can extend them to match other patterns
# and easily extract matched variable names.
# For example, if your project declares some global variables via
# a macro like this: GLOBAL(type, var_name, initializer), then you
# can override this definition to automatically extract those names:
# export _gl_TS_var_match = \
# /^(?:$(_gl_TS_extern)) .*?\**(\w+)(\[.*?\])?;/ || /\bGLOBAL\(.*?,\s*(.*?),/
_gl_TS_var_match ?= /^(?:$(_gl_TS_extern)) .*?(\w+)(\[.*?\])?;/
# The names of object files in (or relative to) $(_gl_TS_dir).
_gl_TS_obj_files ?= *.$(OBJEXT)
# Files in which to search for the one-line style extern declarations.
# $(_gl_TS_dir)-relative.
_gl_TS_headers ?= $(noinst_HEADERS)
_gl_TS_other_headers ?= *.h
.PHONY: _gl_tight_scope
_gl_tight_scope: $(bin_PROGRAMS)
t=exceptions-$$$$; \
trap 's=$$?; rm -f $$t; exit $$s' 0; \
for sig in 1 2 3 13 15; do \
eval "trap 'v=`expr $$sig + 128`; (exit $$v); exit $$v' $$sig"; \
done; \
src=`for f in $(SOURCES); do \
test -f $$f && d= || d=$(srcdir)/; echo $$d$$f; done`; \
hdr=`for f in $(_gl_TS_headers); do \
test -f $$f && d= || d=$(srcdir)/; echo $$d$$f; done`; \
( printf '^%s$$\n' '__.*' $(_gl_TS_unmarked_extern_functions); \
grep -h -A1 '^extern .*[^;]$$' $$src \
| grep -vE '^(extern |--)' | $(SED) 's/ .*//'; \
perl -lne \
'$(_gl_TS_function_match) and print "^$$1\$$"' $$hdr; \
) | sort -u > $$t; \
nm -e $(_gl_TS_obj_files)|$(SED) -n 's/.* T //p'|grep -Ev -f $$t \
&& { echo the above functions should have static scope >&2; \
exit 1; } || : ; \
( printf '^%s$$\n' '__.*' $(_gl_TS_unmarked_extern_vars); \
perl -lne '$(_gl_TS_var_match) and print "^$$1\$$"' \
$$hdr $(_gl_TS_other_headers) \
) | sort -u > $$t; \
nm -e $(_gl_TS_obj_files) | $(SED) -n 's/.* [BCDGRS] //p' \
| sort -u | grep -Ev -f $$t \
&& { echo the above variables should have static scope >&2; \
exit 1; } || :
# TS-end
endif
|