forked from emscripten-core/emscripten
-
Notifications
You must be signed in to change notification settings - Fork 0
/
emcc.py
executable file
·3317 lines (2814 loc) · 137 KB
/
emcc.py
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
#!/usr/bin/env python3
# Copyright 2011 The Emscripten Authors. All rights reserved.
# Emscripten is available under two separate licenses, the MIT license and the
# University of Illinois/NCSA Open Source License. Both these licenses can be
# found in the LICENSE file.
"""emcc - compiler helper script
=============================
emcc is a drop-in replacement for a compiler like gcc or clang.
See emcc --help for details.
emcc can be influenced by a few environment variables:
EMCC_DEBUG - "1" will log out useful information during compilation, as well as
save each compiler step as an emcc-* file in the temp dir
(by default /tmp/emscripten_temp). "2" will save additional emcc-*
steps, that would normally not be separately produced (so this
slows down compilation).
EMMAKEN_NO_SDK - Will tell emcc *not* to use the emscripten headers. Instead
your system headers will be used.
"""
import json
import logging
import os
import re
import shlex
import stat
import sys
import time
import base64
from enum import Enum
from subprocess import PIPE
import emscripten
from tools import shared, system_libs
from tools import colored_logger, diagnostics, building
from tools.shared import unsuffixed, unsuffixed_basename, WINDOWS, safe_move, safe_copy
from tools.shared import run_process, asbytes, read_and_preprocess, exit_with_error, DEBUG
from tools.shared import do_replace
from tools.response_file import substitute_response_files
from tools.minimal_runtime_shell import generate_minimal_runtime_html
import tools.line_endings
from tools.toolchain_profiler import ToolchainProfiler
from tools import js_manipulation
from tools import wasm2c
from tools import webassembly
from tools import config
if __name__ == '__main__':
ToolchainProfiler.record_process_start()
try:
from urllib.parse import quote
except ImportError:
# Python 2 compatibility
from urllib import quote
logger = logging.getLogger('emcc')
# endings = dot + a suffix, safe to test by filename.endswith(endings)
C_ENDINGS = ('.c', '.i')
CXX_ENDINGS = ('.cpp', '.cxx', '.cc', '.c++', '.CPP', '.CXX', '.C', '.CC', '.C++', '.ii')
OBJC_ENDINGS = ('.m', '.mi')
OBJCXX_ENDINGS = ('.mm', '.mii')
ASSEMBLY_CPP_ENDINGS = ('.S',)
SPECIAL_ENDINGLESS_FILENAMES = (os.devnull,)
SOURCE_ENDINGS = C_ENDINGS + CXX_ENDINGS + OBJC_ENDINGS + OBJCXX_ENDINGS + SPECIAL_ENDINGLESS_FILENAMES + ASSEMBLY_CPP_ENDINGS
C_ENDINGS = C_ENDINGS + SPECIAL_ENDINGLESS_FILENAMES # consider the special endingless filenames like /dev/null to be C
EXECUTABLE_ENDINGS = ('.wasm', '.html', '.js', '.mjs', '.out', '')
DYNAMICLIB_ENDINGS = ('.dylib', '.so') # Windows .dll suffix is not included in this list, since those are never linked to directly on the command line.
STATICLIB_ENDINGS = ('.a',)
ASSEMBLY_ENDINGS = ('.ll', '.s')
HEADER_ENDINGS = ('.h', '.hxx', '.hpp', '.hh', '.H', '.HXX', '.HPP', '.HH')
# Supported LLD flags which we will pass through to the linker.
SUPPORTED_LINKER_FLAGS = (
'--start-group', '--end-group',
'-(', '-)',
'--whole-archive', '--no-whole-archive',
'-whole-archive', '-no-whole-archive'
)
# Unsupported LLD flags which we will ignore.
# Maps to true if the flag takes an argument.
UNSUPPORTED_LLD_FLAGS = {
# macOS-specific linker flag that libtool (ltmain.sh) will if macOS is detected.
'-bind_at_load': False,
'-M': False,
# wasm-ld doesn't support soname or other dynamic linking flags (yet). Ignore them
# in order to aid build systems that want to pass these flags.
'-soname': True,
'-allow-shlib-undefined': False,
'-rpath': True,
'-rpath-link': True,
'-version-script': True,
}
LIB_PREFIXES = ('', 'lib')
DEFAULT_ASYNCIFY_IMPORTS = [
'emscripten_sleep', 'emscripten_wget', 'emscripten_wget_data', 'emscripten_idb_load',
'emscripten_idb_store', 'emscripten_idb_delete', 'emscripten_idb_exists',
'emscripten_idb_load_blob', 'emscripten_idb_store_blob', 'SDL_Delay',
'emscripten_scan_registers', 'emscripten_lazy_load_code',
'emscripten_fiber_swap',
'wasi_snapshot_preview1.fd_sync', '__wasi_fd_sync', '_emval_await']
# Mapping of emcc opt levels to llvm opt levels. We use llvm opt level 3 in emcc
# opt levels 2 and 3 (emcc 3 is unsafe opts, so unsuitable for the only level to
# get llvm opt level 3, and speed-wise emcc level 2 is already the slowest/most
# optimizing level)
LLVM_OPT_LEVEL = {
0: ['-O0'],
1: ['-O1'],
2: ['-O3'],
3: ['-O3'],
}
# Target options
final_js = None
UBSAN_SANITIZERS = {
'alignment',
'bool',
'builtin',
'bounds',
'enum',
'float-cast-overflow',
'float-divide-by-zero',
'function',
'implicit-unsigned-integer-truncation',
'implicit-signed-integer-truncation',
'implicit-integer-sign-change',
'integer-divide-by-zero',
'nonnull-attribute',
'null',
'nullability-arg',
'nullability-assign',
'nullability-return',
'object-size',
'pointer-overflow',
'return',
'returns-nonnull-attribute',
'shift',
'signed-integer-overflow',
'unreachable',
'unsigned-integer-overflow',
'vla-bound',
'vptr',
'undefined',
'undefined-trap',
'implicit-integer-truncation',
'implicit-integer-arithmetic-value-change',
'implicit-conversion',
'integer',
'nullability',
}
VALID_ENVIRONMENTS = ('web', 'webview', 'worker', 'node', 'shell')
# this function uses the global 'final' variable, which contains the current
# final output file. if a method alters final, and calls this method, then it
# must modify final globally (i.e. it can't receive final as a param and
# return it)
# TODO: refactor all this, a singleton that abstracts over the final output
# and saving of intermediates
def save_intermediate(name, suffix='js'):
if not DEBUG:
return
if not final_js:
logger.debug('(not saving intermediate %s because not generating JS)' % name)
return
building.save_intermediate(final_js, name + '.' + suffix)
def save_intermediate_with_wasm(name, wasm_binary):
if not DEBUG:
return
save_intermediate(name) # save the js
building.save_intermediate(wasm_binary, name + '.wasm')
class TimeLogger(object):
last = time.time()
@staticmethod
def update():
TimeLogger.last = time.time()
def log_time(name):
"""Log out times for emcc stages"""
if DEBUG:
now = time.time()
logger.debug('emcc step "%s" took %.2f seconds', name, now - TimeLogger.last)
TimeLogger.update()
def base64_encode(b):
b64 = base64.b64encode(b)
if type(b64) == bytes:
return b64.decode('ascii')
else:
return b64
class OFormat(Enum):
WASM = 1
JS = 2
MJS = 3
HTML = 4
BARE = 5
class EmccOptions(object):
def __init__(self):
self.output_file = None
self.post_link = False
self.executable = False
self.compiler_wrapper = None
self.oformat = None
self.requested_debug = ''
self.profiling = False
self.profiling_funcs = False
self.tracing = False
self.emit_symbol_map = False
self.llvm_opts = None
self.use_closure_compiler = None
self.closure_args = []
self.js_transform = None
self.pre_js = '' # before all js
self.post_js = '' # after all js
self.extern_pre_js = '' # before all js, external to optimized code
self.extern_post_js = '' # after all js, external to optimized code
self.preload_files = []
self.embed_files = []
self.exclude_files = []
self.ignore_dynamic_linking = False
self.shell_path = shared.path_from_root('src', 'shell.html')
self.source_map_base = ''
self.emrun = False
self.cpu_profiler = False
self.thread_profiler = False
self.memory_profiler = False
self.memory_init_file = None
self.use_preload_cache = False
self.use_preload_plugins = False
self.proxy_to_worker = False
self.default_object_extension = '.o'
self.valid_abspaths = []
self.cfi = False
# Specifies the line ending format to use for all generated text files.
# Defaults to using the native EOL on each platform (\r\n on Windows, \n on
# Linux & MacOS)
self.output_eol = os.linesep
self.no_entry = False
self.shared = False
self.relocatable = False
def will_metadce():
# The metadce JS parsing code does not currently support the JS that gets generated
# when assertions are enabled.
if shared.Settings.ASSERTIONS:
return False
return shared.Settings.OPT_LEVEL >= 3 or shared.Settings.SHRINK_LEVEL >= 1
def setup_environment_settings():
# Environment setting based on user input
environments = shared.Settings.ENVIRONMENT.split(',')
if any([x for x in environments if x not in VALID_ENVIRONMENTS]):
exit_with_error('Invalid environment specified in "ENVIRONMENT": ' + shared.Settings.ENVIRONMENT + '. Should be one of: ' + ','.join(VALID_ENVIRONMENTS))
shared.Settings.ENVIRONMENT_MAY_BE_WEB = not shared.Settings.ENVIRONMENT or 'web' in environments
shared.Settings.ENVIRONMENT_MAY_BE_WEBVIEW = not shared.Settings.ENVIRONMENT or 'webview' in environments
shared.Settings.ENVIRONMENT_MAY_BE_NODE = not shared.Settings.ENVIRONMENT or 'node' in environments
shared.Settings.ENVIRONMENT_MAY_BE_SHELL = not shared.Settings.ENVIRONMENT or 'shell' in environments
# The worker case also includes Node.js workers when pthreads are
# enabled and Node.js is one of the supported environments for the build to
# run on. Node.js workers are detected as a combination of
# ENVIRONMENT_IS_WORKER and ENVIRONMENT_IS_NODE.
shared.Settings.ENVIRONMENT_MAY_BE_WORKER = \
not shared.Settings.ENVIRONMENT or \
'worker' in environments or \
(shared.Settings.ENVIRONMENT_MAY_BE_NODE and shared.Settings.USE_PTHREADS)
if not shared.Settings.ENVIRONMENT_MAY_BE_WORKER and shared.Settings.PROXY_TO_WORKER:
exit_with_error('If you specify --proxy-to-worker and specify a "-s ENVIRONMENT=" directive, it must include "worker" as a target! (Try e.g. -s ENVIRONMENT=web,worker)')
if not shared.Settings.ENVIRONMENT_MAY_BE_WORKER and shared.Settings.USE_PTHREADS:
exit_with_error('When building with multithreading enabled and a "-s ENVIRONMENT=" directive is specified, it must include "worker" as a target! (Try e.g. -s ENVIRONMENT=web,worker)')
def minify_whitespace():
return shared.Settings.OPT_LEVEL >= 2 and shared.Settings.DEBUG_LEVEL == 0
def embed_memfile():
return (shared.Settings.SINGLE_FILE or
(shared.Settings.MEM_INIT_METHOD == 0 and
(not shared.Settings.MAIN_MODULE and
not shared.Settings.SIDE_MODULE and
not shared.Settings.GENERATE_SOURCE_MAP)))
def expand_byte_size_suffixes(value):
"""Given a string with KB/MB size suffixes, such as "32MB", computes how
many bytes that is and returns it as an integer.
"""
value = value.strip()
match = re.match(r'^(\d+)\s*([kmgt]?b)?$', value, re.I)
if not match:
exit_with_error("invalid byte size `%s`. Valid suffixes are: kb, mb, gb, tb" % value)
value, suffix = match.groups()
value = int(value)
if suffix:
size_suffixes = {suffix: 1024 ** i for i, suffix in enumerate(['b', 'kb', 'mb', 'gb', 'tb'])}
value *= size_suffixes[suffix.lower()]
return value
def apply_settings(changes):
"""Take a list of settings in form `NAME=VALUE` and apply them to the global
Settings object.
"""
def standardize_setting_change(key, value):
# boolean NO_X settings are aliases for X
# (note that *non*-boolean setting values have special meanings,
# and we can't just flip them, so leave them as-is to be
# handled in a special way later)
if key.startswith('NO_') and value in ('0', '1'):
key = key[3:]
value = str(1 - int(value))
return key, value
for change in changes:
key, value = change.split('=', 1)
key, value = standardize_setting_change(key, value)
if key in shared.Settings.internal_settings:
exit_with_error('%s is an internal setting and cannot be set from command line', key)
# map legacy settings which have aliases to the new names
# but keep the original key so errors are correctly reported via the `setattr` below
user_key = key
if key in shared.Settings.legacy_settings and key in shared.Settings.alt_names:
key = shared.Settings.alt_names[key]
# In those settings fields that represent amount of memory, translate suffixes to multiples of 1024.
if key in ('TOTAL_STACK', 'INITIAL_MEMORY', 'MEMORY_GROWTH_LINEAR_STEP', 'MEMORY_GROWTH_GEOMETRIC_STEP',
'GL_MAX_TEMP_BUFFER_SIZE', 'MAXIMUM_MEMORY', 'DEFAULT_PTHREAD_STACK_SIZE'):
value = str(expand_byte_size_suffixes(value))
if value and value[0] == '@':
filename = value[1:]
if not os.path.exists(filename):
exit_with_error('%s: file not found parsing argument: %s' % (filename, change))
value = open(filename).read()
else:
value = value.replace('\\', '\\\\')
try:
value = parse_value(value)
except Exception as e:
exit_with_error('a problem occurred in evaluating the content after a "-s", specifically "%s": %s', change, str(e))
# Do some basic type checking by comparing to the existing settings.
# Sadly we can't do this generically in the SettingsManager since there are settings
# that so change types internally over time.
existing = getattr(shared.Settings, user_key, None)
if existing is not None:
# We only currently worry about lists vs non-lists.
if (type(existing) == list) != (type(value) == list):
exit_with_error('setting `%s` expects `%s` but got `%s`' % (user_key, type(existing), type(value)))
setattr(shared.Settings, user_key, value)
if key == 'EXPORTED_FUNCTIONS':
# used for warnings in emscripten.py
shared.Settings.USER_EXPORTED_FUNCTIONS = shared.Settings.EXPORTED_FUNCTIONS[:]
# TODO(sbc): Remove this legacy way.
if key == 'WASM_OBJECT_FILES':
shared.Settings.LTO = 0 if value else 'full'
def is_ar_file_with_missing_index(archive_file):
# We parse the archive header outselves because llvm-nm --print-armap is slower and less
# reliable.
# See: https://github.com/emscripten-core/emscripten/issues/10195
archive_header = b'!<arch>\n'
file_header_size = 60
with open(archive_file, 'rb') as f:
header = f.read(len(archive_header))
if header != archive_header:
# This is not even an ar file
return False
file_header = f.read(file_header_size)
if len(file_header) != file_header_size:
# We don't have any file entires at all so we don't consider the index missing
return False
name = file_header[:16].strip()
# If '/' is the name of the first file we have an index
return name != b'/'
def ensure_archive_index(archive_file):
# Fastcomp linking works without archive indexes.
if not shared.Settings.AUTO_ARCHIVE_INDEXES:
return
if is_ar_file_with_missing_index(archive_file):
diagnostics.warning('emcc', '%s: archive is missing an index; Use emar when creating libraries to ensure an index is created', archive_file)
diagnostics.warning('emcc', '%s: adding index', archive_file)
run_process([shared.LLVM_RANLIB, archive_file])
def get_all_js_syms():
# Runs the js compiler to generate a list of all symbols available in the JS
# libraries. This must be done separately for each linker invokation since the
# list of symbols depends on what settings are used.
# TODO(sbc): Find a way to optimize this. Potentially we could add a super-set
# mode of the js compiler that would generate a list of all possible symbols
# that could be checked in.
old_full = shared.Settings.INCLUDE_FULL_LIBRARY
try:
# Temporarily define INCLUDE_FULL_LIBRARY since we want a full list
# of all available JS library functions.
shared.Settings.INCLUDE_FULL_LIBRARY = True
shared.Settings.ONLY_CALC_JS_SYMBOLS = True
emscripten.generate_struct_info()
glue, forwarded_data = emscripten.compile_settings()
forwarded_json = json.loads(forwarded_data)
library_fns = forwarded_json['Functions']['libraryFunctions']
library_fns_list = []
for name in library_fns:
if shared.is_c_symbol(name):
name = shared.demangle_c_symbol_name(name)
library_fns_list.append(name)
finally:
shared.Settings.ONLY_CALC_JS_SYMBOLS = False
shared.Settings.INCLUDE_FULL_LIBRARY = old_full
return library_fns_list
def filter_link_flags(flags, using_lld):
def is_supported(f):
if using_lld:
for flag, takes_arg in UNSUPPORTED_LLD_FLAGS.items():
# lld allows various flags to have either a single -foo or double --foo
if f.startswith(flag) or f.startswith('-' + flag):
diagnostics.warning('linkflags', 'ignoring unsupported linker flag: `%s`', f)
return False, takes_arg
return True, False
else:
if f in SUPPORTED_LINKER_FLAGS:
return True, False
# Silently ignore -l/-L flags when not using lld. If using lld allow
# them to pass through the linker
if f.startswith('-l') or f.startswith('-L'):
return False, False
diagnostics.warning('linkflags', 'ignoring unsupported linker flag: `%s`', f)
return False, False
results = []
skip_next = False
for f in flags:
if skip_next:
skip_next = False
continue
keep, skip_next = is_supported(f[1])
if keep:
results.append(f)
return results
def fix_windows_newlines(text):
# Avoid duplicating \r\n to \r\r\n when writing out text.
if WINDOWS:
text = text.replace('\r\n', '\n')
return text
def cxx_to_c_compiler(cxx):
# Convert C++ compiler name into C compiler name
dirname, basename = os.path.split(cxx)
basename = basename.replace('clang++', 'clang').replace('g++', 'gcc').replace('em++', 'emcc')
return os.path.join(dirname, basename)
def get_binaryen_passes():
# run the binaryen optimizer in -O2+. in -O0 we don't need it obviously, while
# in -O1 we don't run it as the LLVM optimizer has been run, and it does the
# great majority of the work; not running the binaryen optimizer in that case
# keeps -O1 mostly-optimized while compiling quickly and without rewriting
# DWARF etc.
run_binaryen_optimizer = shared.Settings.OPT_LEVEL >= 2
passes = []
# safe heap must run before post-emscripten, so post-emscripten can apply the sbrk ptr
if shared.Settings.SAFE_HEAP:
passes += ['--safe-heap']
if shared.Settings.MEMORY64 == 2:
passes += ['--memory64-lowering']
if run_binaryen_optimizer:
passes += ['--post-emscripten']
if not shared.Settings.EXIT_RUNTIME:
passes += ['--no-exit-runtime']
if run_binaryen_optimizer:
passes += [building.opt_level_to_str(shared.Settings.OPT_LEVEL, shared.Settings.SHRINK_LEVEL)]
elif shared.Settings.STANDALONE_WASM:
# even if not optimizing, make an effort to remove all unused imports and
# exports, to make the wasm as standalone as possible
passes += ['--remove-unused-module-elements']
# when optimizing, use the fact that low memory is never used (1024 is a
# hardcoded value in the binaryen pass)
if run_binaryen_optimizer and shared.Settings.GLOBAL_BASE >= 1024:
passes += ['--low-memory-unused']
if shared.Settings.AUTODEBUG:
# adding '--flatten' here may make these even more effective
passes += ['--instrument-locals']
passes += ['--log-execution']
passes += ['--instrument-memory']
if shared.Settings.LEGALIZE_JS_FFI:
# legalize it again now, as the instrumentation may need it
passes += ['--legalize-js-interface']
if shared.Settings.EMULATE_FUNCTION_POINTER_CASTS:
# note that this pass must run before asyncify, as if it runs afterwards we only
# generate the byn$fpcast_emu functions after asyncify runs, and so we wouldn't
# be able to further process them.
passes += ['--fpcast-emu']
if shared.Settings.ASYNCIFY:
passes += ['--asyncify']
if shared.Settings.ASSERTIONS:
passes += ['--pass-arg=asyncify-asserts']
if shared.Settings.ASYNCIFY_ADVISE:
passes += ['--pass-arg=asyncify-verbose']
if shared.Settings.ASYNCIFY_IGNORE_INDIRECT:
passes += ['--pass-arg=asyncify-ignore-indirect']
passes += ['--pass-arg=asyncify-imports@%s' % ','.join(shared.Settings.ASYNCIFY_IMPORTS)]
# shell escaping can be confusing; try to emit useful warnings
def check_human_readable_list(items):
for item in items:
if item.count('(') != item.count(')'):
logger.warning('''emcc: ASYNCIFY list contains an item without balanced parentheses ("(", ")"):''')
logger.warning(''' ''' + item)
logger.warning('''This may indicate improper escaping that led to splitting inside your names.''')
logger.warning('''Try to quote the entire argument, like this: -s 'ASYNCIFY_ONLY=["foo(int, char)", "bar"]' ''')
break
if shared.Settings.ASYNCIFY_REMOVE:
check_human_readable_list(shared.Settings.ASYNCIFY_REMOVE)
passes += ['--pass-arg=asyncify-removelist@%s' % ','.join(shared.Settings.ASYNCIFY_REMOVE)]
if shared.Settings.ASYNCIFY_ADD:
check_human_readable_list(shared.Settings.ASYNCIFY_ADD)
passes += ['--pass-arg=asyncify-addlist@%s' % ','.join(shared.Settings.ASYNCIFY_ADD)]
if shared.Settings.ASYNCIFY_ONLY:
check_human_readable_list(shared.Settings.ASYNCIFY_ONLY)
passes += ['--pass-arg=asyncify-onlylist@%s' % ','.join(shared.Settings.ASYNCIFY_ONLY)]
if shared.Settings.BINARYEN_IGNORE_IMPLICIT_TRAPS:
passes += ['--ignore-implicit-traps']
# normally we can assume the memory, if imported, has not been modified
# beforehand (in fact, in most cases the memory is not even imported anyhow,
# but it is still safe to pass the flag), and is therefore filled with zeros.
# the one exception is dynamic linking of a side module: the main module is ok
# as it is loaded first, but the side module may be assigned memory that was
# previously used.
if run_binaryen_optimizer and not shared.Settings.SIDE_MODULE:
passes += ['--zero-filled-memory']
if shared.Settings.BINARYEN_EXTRA_PASSES:
# BINARYEN_EXTRA_PASSES is comma-separated, and we support both '-'-prefixed and
# unprefixed pass names
extras = shared.Settings.BINARYEN_EXTRA_PASSES.split(',')
passes += [('--' + p) if p[0] != '-' else p for p in extras if p]
return passes
def make_js_executable(script):
src = open(script).read()
cmd = shared.shlex_join(config.JS_ENGINE)
if not os.path.isabs(config.JS_ENGINE[0]):
# TODO: use whereis etc. And how about non-*NIX?
cmd = '/usr/bin/env -S ' + cmd
logger.debug('adding `#!` to JavaScript file: %s' % cmd)
# add shebang
with open(script, 'w') as f:
f.write('#!%s\n' % cmd)
f.write(src)
try:
os.chmod(script, stat.S_IMODE(os.stat(script).st_mode) | stat.S_IXUSR) # make executable
except OSError:
pass # can fail if e.g. writing the executable to /dev/null
def do_split_module(wasm_file):
os.rename(wasm_file, wasm_file + '.orig')
args = ['--instrument']
building.run_binaryen_command('wasm-split', wasm_file + '.orig', outfile=wasm_file, args=args)
def is_dash_s_for_emcc(args, i):
# -s OPT=VALUE or -s OPT or -sOPT are all interpreted as emscripten flags.
# -s by itself is a linker option (alias for --strip-all)
if args[i] == '-s':
if len(args) <= i + 1:
return False
arg = args[i + 1]
else:
arg = args[i][2:]
arg = arg.split('=')[0]
return arg.isidentifier() and arg.isupper()
def parse_s_args(args):
settings_changes = []
for i in range(len(args)):
if args[i].startswith('-s'):
if is_dash_s_for_emcc(args, i):
if args[i] == '-s':
key = args[i + 1]
args[i + 1] = ''
else:
key = args[i][2:]
args[i] = ''
# If not = is specified default to 1
if '=' not in key:
key += '=1'
# Special handling of browser version targets. A version -1 means that the specific version
# is not supported at all. Replace those with INT32_MAX to make it possible to compare e.g.
# #if MIN_FIREFOX_VERSION < 68
if re.match(r'MIN_.*_VERSION(=.*)?', key):
try:
if int(key.split('=')[1]) < 0:
key = key.split('=')[0] + '=0x7FFFFFFF'
except Exception:
pass
settings_changes.append(key)
newargs = [a for a in args if a]
return (settings_changes, newargs)
def calc_cflags(options):
# Flags we pass to the compiler when building C/C++ code
# We add these to the user's flags (newargs), but not when building .s or .S assembly files
cflags = []
if options.tracing:
cflags.append('-D__EMSCRIPTEN_TRACING__=1')
if shared.Settings.USE_PTHREADS:
cflags.append('-D__EMSCRIPTEN_PTHREADS__=1')
if not shared.Settings.STRICT:
# The preprocessor define EMSCRIPTEN is deprecated. Don't pass it to code
# in strict mode. Code should use the define __EMSCRIPTEN__ instead.
cflags.append('-DEMSCRIPTEN')
# if exception catching is disabled, we can prevent that code from being
# generated in the frontend
if shared.Settings.DISABLE_EXCEPTION_CATCHING == 1 and not shared.Settings.EXCEPTION_HANDLING:
cflags.append('-fignore-exceptions')
if shared.Settings.INLINING_LIMIT:
cflags.append('-fno-inline-functions')
if shared.Settings.RELOCATABLE:
cflags.append('-fPIC')
cflags.append('-fvisibility=default')
if shared.Settings.LTO:
cflags.append('-flto=' + shared.Settings.LTO)
else:
# With LTO mode these args get passed instead
# at link time when the backend runs.
for a in building.llvm_backend_args():
cflags += ['-mllvm', a]
return cflags
def get_file_suffix(filename):
"""Parses the essential suffix of a filename, discarding Unix-style version
numbers in the name. For example for 'libz.so.1.2.8' returns '.so'"""
if filename in SPECIAL_ENDINGLESS_FILENAMES:
return filename
while filename:
filename, suffix = os.path.splitext(filename)
if not suffix[1:].isdigit():
return suffix
return ''
def in_temp(name):
temp_dir = shared.get_emscripten_temp_dir()
return os.path.join(temp_dir, os.path.basename(name))
run_via_emxx = False
#
# Main run() function
#
def run(args):
target = None
# Additional compiler flags that we treat as if they were passed to us on the
# commandline
EMCC_CFLAGS = os.environ.get('EMCC_CFLAGS')
if DEBUG:
cmd = shared.shlex_join(args)
if EMCC_CFLAGS:
cmd += ' + ' + EMCC_CFLAGS
logger.warning('invocation: ' + cmd + ' (in ' + os.getcwd() + ')')
if EMCC_CFLAGS:
args.extend(shlex.split(EMCC_CFLAGS))
# Strip args[0] (program name)
args = args[1:]
misc_temp_files = shared.configuration.get_temp_files()
# Handle some global flags
# read response files very early on
try:
args = substitute_response_files(args)
except IOError as e:
exit_with_error(e)
if '--help' in args:
# Documentation for emcc and its options must be updated in:
# site/source/docs/tools_reference/emcc.rst
# This then gets built (via: `make -C site text`) to:
# site/build/text/docs/tools_reference/emcc.txt
# This then needs to be copied to its final home in docs/emcc.txt from where
# we read it here. We have CI rules that ensure its always up-to-date.
with open(shared.path_from_root('docs', 'emcc.txt'), 'r') as f:
print(f.read())
print('''
------------------------------------------------------------------
emcc: supported targets: llvm bitcode, WebAssembly, NOT elf
(autoconf likes to see elf above to enable shared object support)
''')
return 0
if '--version' in args:
# if the emscripten folder is not a git repo, don't run git show - that can
# look up and find the revision in a parent directory that is a git repo
revision = ''
if os.path.exists(shared.path_from_root('.git')):
revision = run_process(['git', 'rev-parse', 'HEAD'], stdout=PIPE, stderr=PIPE, cwd=shared.path_from_root()).stdout.strip()
elif os.path.exists(shared.path_from_root('emscripten-revision.txt')):
revision = open(shared.path_from_root('emscripten-revision.txt')).read().strip()
if revision:
revision = ' (%s)' % revision
print('''%s%s
Copyright (C) 2014 the Emscripten authors (see AUTHORS.txt)
This is free and open source software under the MIT license.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
''' % (version_string(), revision))
return 0
if run_via_emxx:
clang = shared.CLANG_CXX
else:
clang = shared.CLANG_CC
if len(args) == 1 and args[0] == '-v': # -v with no inputs
# autoconf likes to see 'GNU' in the output to enable shared object support
print(version_string(), file=sys.stderr)
return shared.check_call([clang, '-v'] + shared.get_clang_flags(), check=False).returncode
if '-dumpmachine' in args:
print(shared.get_llvm_target())
return 0
if '-dumpversion' in args: # gcc's doc states "Print the compiler version [...] and don't do anything else."
print(shared.EMSCRIPTEN_VERSION)
return 0
if '--cflags' in args:
# fake running the command, to see the full args we pass to clang
args = [x for x in args if x != '--cflags']
with misc_temp_files.get_file(suffix='.o') as temp_target:
input_file = 'hello_world.c'
cmd = [shared.PYTHON, sys.argv[0], shared.path_from_root('tests', input_file), '-v', '-c', '-o', temp_target] + args
proc = run_process(cmd, stderr=PIPE, check=False)
if proc.returncode != 0:
print(proc.stderr)
exit_with_error('error getting cflags')
lines = [x for x in proc.stderr.splitlines() if clang in x and input_file in x]
parts = shlex.split(lines[0].replace('\\', '\\\\'))
parts = [x for x in parts if x not in ['-c', '-o', '-v', '-emit-llvm'] and input_file not in x and temp_target not in x]
print(shared.shlex_join(parts[1:]))
return 0
shared.check_sanity()
def get_language_mode(args):
return_next = False
for item in args:
if return_next:
return item
if item == '-x':
return_next = True
continue
if item.startswith('-x'):
return item[2:]
return ''
language_mode = get_language_mode(args)
EMMAKEN_CFLAGS = os.environ.get('EMMAKEN_CFLAGS')
if EMMAKEN_CFLAGS:
args += shlex.split(EMMAKEN_CFLAGS)
# ---------------- Utilities ---------------
seen_names = {}
def uniquename(name):
if name not in seen_names:
seen_names[name] = str(len(seen_names))
return unsuffixed(name) + '_' + seen_names[name] + shared.suffix(name)
# ---------------- End configs -------------
def optimizing(opts):
return '-O0' not in opts
with ToolchainProfiler.profile_block('parse arguments and setup'):
## Parse args
newargs = list(args)
# Scan and strip emscripten specific cmdline warning flags.
# This needs to run before other cmdline flags have been parsed, so that
# warnings are properly printed during arg parse.
newargs = diagnostics.capture_warnings(newargs)
if not config.config_file:
diagnostics.warning('deprecated', 'Specifying EM_CONFIG as a python literal is deprecated. Please use a file instead.')
for i in range(len(newargs)):
if newargs[i] in ('-l', '-L', '-I'):
# Scan for individual -l/-L/-I arguments and concatenate the next arg on
# if there is no suffix
newargs[i] += newargs[i + 1]
newargs[i + 1] = ''
options, settings_changes, user_js_defines, newargs = parse_args(newargs)
if options.post_link or options.oformat == OFormat.BARE:
diagnostics.warning('experimental', '--oformat=base/--post-link are experimental and subject to change.')
if '-print-search-dirs' in newargs:
return run_process([clang, '-print-search-dirs'], check=False).returncode
if options.emrun:
options.pre_js += open(shared.path_from_root('src', 'emrun_prejs.js')).read() + '\n'
options.post_js += open(shared.path_from_root('src', 'emrun_postjs.js')).read() + '\n'
# emrun mode waits on program exit
shared.Settings.EXIT_RUNTIME = 1
if options.cpu_profiler:
options.post_js += open(shared.path_from_root('src', 'cpuprofiler.js')).read() + '\n'
if options.memory_profiler:
shared.Settings.MEMORYPROFILER = 1
if options.thread_profiler:
options.post_js += open(shared.path_from_root('src', 'threadprofiler.js')).read() + '\n'
if options.llvm_opts is None:
options.llvm_opts = LLVM_OPT_LEVEL[shared.Settings.OPT_LEVEL]
elif type(options.llvm_opts) == int:
options.llvm_opts = ['-O%d' % options.llvm_opts]
if options.memory_init_file is None:
options.memory_init_file = shared.Settings.OPT_LEVEL >= 2
# TODO: support source maps with js_transform
if options.js_transform and shared.Settings.DEBUG_LEVEL >= 4:
logger.warning('disabling source maps because a js transform is being done')
shared.Settings.DEBUG_LEVEL = 3
explicit_settings_changes, newargs = parse_s_args(newargs)
settings_changes += explicit_settings_changes
settings_key_changes = {}
for s in settings_changes:
key, value = s.split('=', 1)
settings_key_changes[key] = value
# Find input files
# These three arrays are used to store arguments of different types for
# type-specific processing. In order to shuffle the arguments back together
# after processing, all of these arrays hold tuples (original_index, value).
# Note that the index part of the tuple can have a fractional part for input
# arguments that expand into multiple processed arguments, as in -Wl,-f1,-f2.
input_files = []
libs = []
link_flags = []
has_header_inputs = False
lib_dirs = []
has_dash_c = '-c' in newargs
has_dash_S = '-S' in newargs
has_dash_E = '-E' in newargs
compile_only = has_dash_c or has_dash_S or has_dash_E
def add_link_flag(i, f):
if f.startswith('-l'):
libs.append((i, f[2:]))
if f.startswith('-L'):
lib_dirs.append(f[2:])
link_flags.append((i, f))
# find input files with a simple heuristic. we should really analyze
# based on a full understanding of gcc params, right now we just assume that
# what is left contains no more |-x OPT| things
skip = False
for i in range(len(newargs)):
if skip:
skip = False
continue
arg = newargs[i]
if arg in ('-MT', '-MF', '-MJ', '-MQ', '-D', '-U', '-o', '-x',
'-Xpreprocessor', '-include', '-imacros', '-idirafter',
'-iprefix', '-iwithprefix', '-iwithprefixbefore',
'-isysroot', '-imultilib', '-A', '-isystem', '-iquote',
'-install_name', '-compatibility_version',
'-current_version', '-I', '-L', '-include-pch',
'-Xlinker', '-Xclang'):
skip = True
if not arg.startswith('-'):
# os.devnul should always be reported as existing but there is bug in windows
# python before 3.8:
# https://bugs.python.org/issue1311
if not os.path.exists(arg) and arg != os.devnull:
exit_with_error('%s: No such file or directory ("%s" was expected to be an input file, based on the commandline arguments provided)', arg, arg)
file_suffix = get_file_suffix(arg)
if file_suffix in SOURCE_ENDINGS + DYNAMICLIB_ENDINGS + ASSEMBLY_ENDINGS + HEADER_ENDINGS or building.is_ar(arg):
# we already removed -o <target>, so all these should be inputs
newargs[i] = ''
if file_suffix in SOURCE_ENDINGS or (has_dash_c and file_suffix == '.bc'):
input_files.append((i, arg))
elif file_suffix in HEADER_ENDINGS:
input_files.append((i, arg))
has_header_inputs = True
elif file_suffix in ASSEMBLY_ENDINGS or building.is_bitcode(arg) or building.is_ar(arg):
input_files.append((i, arg))
elif building.is_wasm(arg):
input_files.append((i, arg))
elif file_suffix in (STATICLIB_ENDINGS + DYNAMICLIB_ENDINGS):
# if it's not, and it's a library, just add it to libs to find later
libname = unsuffixed_basename(arg)
for prefix in LIB_PREFIXES:
if not prefix:
continue
if libname.startswith(prefix):
libname = libname[len(prefix):]
break
libs.append((i, libname))
elif file_suffix in STATICLIB_ENDINGS:
if not building.is_ar(arg):
if building.is_bitcode(arg):
message = arg + ': File has a suffix of a static library ' + str(STATICLIB_ENDINGS) + ', but instead is an LLVM bitcode file! When linking LLVM bitcode files use .bc or .o.'
else:
message = arg + ': Unknown format, not a static library!'
exit_with_error(message)