9aedf65 | Alon Zakai | 15 March 2017, 23:24:31 UTC | 1.37.6 | 15 March 2017, 23:24:31 UTC |
abf5add | jgravelle-google | 15 March 2017, 20:03:31 UTC | Wast single space fix (#5035) * Fix parsing metadata from a wast with single spaces * More robust fix | 15 March 2017, 20:03:31 UTC |
9163134 | Alon Zakai | 15 March 2017, 17:03:15 UTC | if verifying the linked module, use -disable-debug-info-type-map so we have the same behavior in opt as in llvm-link (#5021) | 15 March 2017, 17:03:15 UTC |
0b13449 | Alon Zakai | 15 March 2017, 17:01:26 UTC | don't minify in emterpretify code if in binaryen, it must minify only at the very end (#5017) | 15 March 2017, 17:01:26 UTC |
6fc1485 | Alon Zakai (kripken) | 15 March 2017, 00:59:15 UTC | fix whitespace | 15 March 2017, 00:59:59 UTC |
f6bd50f | Alon Zakai (kripken) | 15 March 2017, 00:59:00 UTC | Merge branch 'patch-1' of https://github.com/Deamon87/emscripten into incoming | 15 March 2017, 00:59:00 UTC |
734685a | Evan Wallace | 14 March 2017, 23:24:26 UTC | force absolute paths in Building.llvm_nm to maximize cache usage (#5025) I noticed that emscripten always loads all object files twice, once with the relative path and once with the absolute path. Always using the absolute path results in a 4.5s speedup on our codebase. | 14 March 2017, 23:24:26 UTC |
6104001 | jgravelle-google | 14 March 2017, 21:42:58 UTC | Split strtod and strtold tests, support wasm_backend's f128 output (#4904) | 14 March 2017, 21:42:58 UTC |
8c9a3fd | jgravelle-google | 14 March 2017, 21:42:43 UTC | Add leading underscore to implemented functions in wasm_backend (#4854) | 14 March 2017, 21:42:43 UTC |
edf90d0 | Deamon87 | 14 March 2017, 21:15:01 UTC | fix | 14 March 2017, 21:15:01 UTC |
2903980 | Deamon87 | 14 March 2017, 21:14:21 UTC | Merge branch 'incoming' into patch-1 | 14 March 2017, 21:14:21 UTC |
02b5191 | Deamon87 | 14 March 2017, 21:13:13 UTC | Update AUTHORS | 14 March 2017, 21:13:13 UTC |
c471557 | juj | 14 March 2017, 17:30:10 UTC | Merge pull request #5027 from gouletr/fix/fetch-store-propagate-error Propagate failed cache store errors in returned fetch status | 14 March 2017, 17:30:10 UTC |
7b0f92a | juj | 14 March 2017, 17:26:28 UTC | Merge pull request #4995 from gouletr/fix/fetch-delete-cached-data Add missing emscripten_start_fetch dependency | 14 March 2017, 17:26:28 UTC |
2d45637 | Robert Goulet | 14 March 2017, 14:36:47 UTC | Fix incorrect stringToUTF8 parameter order to set statusText | 14 March 2017, 14:36:47 UTC |
21bada7 | Robert Goulet | 14 March 2017, 14:34:54 UTC | Run xhr callbacks on cache storage events to propagate result up | 14 March 2017, 14:34:54 UTC |
01abc80 | Aleksander Guryanov | 13 March 2017, 21:23:30 UTC | Remove unnecessary Array.fill (#5020) | 13 March 2017, 21:23:30 UTC |
31cf2bb | jgravelle-google | 13 March 2017, 16:11:40 UTC | Update musl to 1.1.15 (#4813) * Revert all emscripten musl changes * Update musl to 1.1.5 * Reapply emscripten patches * Fix up libs to partially compile * Revert emscripten musl include folder * Update musl include folder to 1.1.5 * Reapply emscripten libc include changes * Get compiling again! * Restore alltypes.h to have emscripten-specific changes alltypes.h is autogenerated, and we've since overridden the types Longer-term, we should figure out a way to either autogenerate the types to match what emscripten expects, or change emscripten's expectation to support musl's assumptions. Probably worth including with merging the asm&wasm ABIs. * Update expected langinfo test output for fixes in upstream musl * Canonicalize nans in hyperbolic test Nans are now printed out with their sign if negative. Asm and wasm treat these differently, so canonicalize how they print. * Tweak how musl forwards syscalls to emscripten Before: pass all syscall args as varargs, args are int32 After: cast all syscall args to int32, pass as varargs This fixes and prevents aliasing issues, and is more consistent with other architectures * Change __syscall to syscall in isatty to set errno POSIX specifies that isatty sets errno in the case that it fails. __syscall doesn't set errno, and syscall does. * Update struct_info again * Always allocate pthread_self, even in the stub * sys_open should call __syscall_ret * Fix up pthread js for optimized code (remove __asm: true) * Canonicalize more nans in testsuite * Re-set expected langinfo output * Update align_moar expected addresses * Disable assert in wasm_backend llseek syscall * Update now-passing wasm_backend tests * Update other.test_locale to non-stub setlocale * Fix compiling with -s USE_PTHREADS=1 * Initialize __pthread_self() when using pthread_stub * Canonicalize nans for test_zerodiv * Update count of globals in test_dlfcn_self * Rework how we initialize pthread_self for non-threaded applications * Prefer 'EMSCRIPTEN_START_FUNCS' in test_memorygrowth, because 'var TOTAL_MEMORY' can move around * Now that CURRENT_LOCALE isn't reading off NULL, test_langinfo should expect ASCII * Re-disable test_demangle_stacks - was spuriously passing earlier * Explicitly pass 0 as the high bits for SYS__llseek * Enable v8 for test_readdir now that it supports printErr * Only use __syscall_emscripten if __EMSCRIPTEN__ is defined * Move O_LARGEFILE as default back to libc * libc #if to #ifdef * Rename pthread import funcs to not rely on alias * Consistently reference pthread_setcancelstate by alias in libc * Stub out fcntl64 for NO_FILESYSTEM case * Update emscripten-specific pthread struct fields * Consistently reference pthread_testcancel by alias in libc * Stub out SYS_rt_sigqueueinfo syscall * Stub out SYS_sendmmsg syscall * Stub out SYS_recvmmsg syscall * Revert "Consistently reference pthread_testcancel by alias in libc" This reverts commit 43563a5b141a47d4a7aa7e48c059ed0e98597fbb. * Revert "Consistently reference pthread_setcancelstate by alias in libc" This reverts commit 2ca2eded8a0d50bb3847cda0a99c900532f6937b. * Correctly fix unresolved symbol: __pthread_setcancelstate * Update library_pthread.c pthread_testcancel to match musl's expectations * test_strftime.c -> test_strftime.cpp * library_pthread_stub pthread_exit should call exit, not nothing * TEMP: init pthread_self->locale with pthreads enabled * Initialize pthread global data from libc in JS * Move _b_lock and _b_waiters back to not emscripten-only * Remove untrue __asm annotation from threaded pthread_self, make stub pthread_self non-asm * Remove old unneeded pthread init exported functions * Move pthread_self in stub case to libc C code * Move emscripten_pthread initialization to its own file, out from locale_map.c * Statically allocate the pthread struct and not just a pointer in the stub case * Remove commented-out test args * Remove misleading __asm annotation for _register_pthread_ptr * Do module-side pthread initialization in-module * Don't cast away volatile when calling emscripten_futex_wake in __wake * Expose libc struct to emscripten, use that for non-main pthread initialization * Re-add pthread_impl.h import to library_pthread.c * Update align_moar alignment post-rebase * Update emscripten header include in libc.c * Remove obsolete locale preallocating code in newlocale We previously statically allocated a single locale in the event that we needed to dynamically allocate one during startup. musl changed their implementation to statically allocate C_LOCALE and UTF8_LOCALE, so we already get that behavior. In addition this was causing invalid locales to get allocated instead of returning NULL, because the preceding logic changed. That was caught by other.test_locale_wrong * Update test_llvm_lto * Bump version to 1.37.5 | 13 March 2017, 16:11:40 UTC |
bda1131 | Jacob Gravelle | 13 March 2017, 15:54:36 UTC | 1.37.4 | 13 March 2017, 15:54:36 UTC |
de8b0f2 | Dmitry Tolmachov | 13 March 2017, 15:18:00 UTC | Implement test for readdir inside worker_fs_read | 13 March 2017, 15:18:00 UTC |
d508994 | jgravelle-google | 10 March 2017, 23:04:02 UTC | Test runner main factoring (#5003) * Break some overlong lines * Scope main test runner logic into a function * Replace exec with equivalent setattr Non-global exec assignment isn't allowed due to variable scoping shenanigans, this avoids that and is a bit more direct with intent * Move get modules closer to first use * Start splitting main runner into sections * Continue factoring into functions * Refactor random test selection * Split long import to multiple lines * Split loading suites and running tests * Add missing modules argument to skip_requested_tests * Actually skip tests when requesting skipping Note: this brings skipme in line with the state of test skipping on incoming, namely that it doesn't work, but it does correctly replace the test case with the skipme method. | 10 March 2017, 23:04:02 UTC |
b66454d | Deamon87 | 10 March 2017, 09:49:13 UTC | Implement readdir for WORKERFS | 10 March 2017, 09:49:13 UTC |
51350b0 | Alon Zakai (kripken) | 09 March 2017, 23:38:10 UTC | update binaryen to version_30 | 09 March 2017, 23:38:10 UTC |
df993f9 | Alon Zakai | 09 March 2017, 23:37:22 UTC | Fix outlining + growth + wasm (#5015) Minify JS at the end when building to wasm, as in intermediate steps we don't want minified code (can interfere with asm2wasm parsing) | 09 March 2017, 23:37:22 UTC |
5da32ac | Alon Zakai | 09 March 2017, 21:31:26 UTC | Support Binaryen's 3 trap modes (#5006) * support binaryen's 3 trap modes: js compatibility, clamp, and allow trapping | 09 March 2017, 21:31:26 UTC |
e6e11c0 | Alon Zakai | 09 March 2017, 00:49:39 UTC | increase a test timeout | 09 March 2017, 00:49:39 UTC |
67fabbc | Alon Zakai | 24 February 2017, 20:25:53 UTC | fix regression from c4c9d4b68ce8ada83c8caafe2a39ba352409ed98 in ONLY_MY_CODE, in that mode we should not export anything that is not the user's code | 09 March 2017, 00:49:24 UTC |
6c728c9 | Alon Zakai | 08 March 2017, 23:18:57 UTC | clean up temp files in other.test_binaryen_debug | 08 March 2017, 23:18:57 UTC |
93a082b | Jukka Jylanki | 08 March 2017, 05:43:06 UTC | Demote llvm-nm failures from exceptions to logged debug messages, because it can be expected that llvm-nm might fail, and processing that failure is done proper by examining the returncode of the run later, so eagerly stopping the build here is not needed. Fixes #5007. | 08 March 2017, 05:43:06 UTC |
28d43cc | Jukka Jylanki | 08 March 2017, 03:39:49 UTC | When feeding and updating queued audio buffers in OpenAL, the buffers might have been created with different OpenAL audio contexts, so audio update code should not rely on a single active AL context, but update each source with the audio context that was used to create that audio source. Also fixes the issue where code un-makecurrents OpenAL audio context in a thread when there are audio sources playing. | 08 March 2017, 03:39:49 UTC |
800c110 | juj | 07 March 2017, 22:18:55 UTC | Merge pull request #5004 from juj/fix_llvm_nm_and_ar_race_conditions fix_llvm_nm_and_ar_race_condition | 07 March 2017, 22:18:55 UTC |
972fb4e | Jukka Jylanki | 07 March 2017, 12:35:02 UTC | Propagate errors from llvm-nm and llvm-ar calls back up to main process from subprocesses. Clean up temporary directory created for llvm-ar at process exit. Fix race condition in llvm-ar temporary directory creation. Fixes #4960. | 07 March 2017, 13:16:33 UTC |
eac8e0b | Alon Zakai | 07 March 2017, 04:07:52 UTC | Minify JS in wasm mode too (#4999) * minify the JS in wasm mode when possible. even if the compiled code is in a side wasm file, in small projects the JS can be of significant size | 07 March 2017, 04:07:52 UTC |
ebf0980 | Alon Zakai (kripken) | 06 March 2017, 22:25:20 UTC | make other.test_binaryen_warn_mem synchronous, as it is not compatible with async compilation yet, see #4971 | 06 March 2017, 22:25:20 UTC |
c8305b2 | AlexPerrot | 06 March 2017, 18:25:46 UTC | Add support for policies when calling val::as. (#4992) * Add support for policies when calling val::as. This allows to use allow_raw_pointers() with val::as to get a pointer. * Add test for support of val::as<Instance*>(allow_raw_pointers()). | 06 March 2017, 18:25:46 UTC |
e4ade2d | AlexPerrot | 05 March 2017, 02:09:05 UTC | Allow embind to work with async wasm compilation (#5001) * Use Module["asm"] instead of asm in embind to enable use of embind in wasm. * change test_core.test_embind_2 to work asynchronously. This allows the test to work with wasm async compilation. * reactivate embind when using BINARYEN_ASYNC_COMPILATION | 05 March 2017, 02:09:05 UTC |
cb1e780 | Alon Zakai (kripken) | 03 March 2017, 19:48:38 UTC | csmith tweaks | 03 March 2017, 19:48:38 UTC |
59b8f23 | Alon Zakai (kripken) | 03 March 2017, 05:50:06 UTC | update binaryen to version_29 | 03 March 2017, 05:50:06 UTC |
58f9587 | Robert Goulet | 02 March 2017, 22:21:26 UTC | Add missing __emscripten_fetch_delete_cached_data to emscripten_start_fetch dependencies. Fixes #4994 | 02 March 2017, 22:21:26 UTC |
5933064 | Alon Zakai (kripken) | 02 March 2017, 21:16:13 UTC | fix --output_eol docs #4972 | 02 March 2017, 21:16:13 UTC |
04255b4 | Alon Zakai | 28 February 2017, 23:30:27 UTC | ignore the 'no running' warning in test_files | 28 February 2017, 23:30:46 UTC |
47c2c82 | jgravelle-google | 28 February 2017, 21:18:54 UTC | Generify message for test_python, add message for test_cases (#4987) | 28 February 2017, 21:18:54 UTC |
88dba7b | jgravelle-google | 28 February 2017, 21:18:37 UTC | Enable test_math_lgamma for wasm_backend (fixed in clang) (#4988) | 28 February 2017, 21:18:37 UTC |
cee49ec | jgravelle-google | 28 February 2017, 21:18:16 UTC | Add expected wasm_backend alignment to align_moar test (#4989) | 28 February 2017, 21:18:16 UTC |
94ac79b | Alon Zakai | 28 February 2017, 19:01:34 UTC | setTempRet0's return value is received in some cases, even though it is ignored. JS engines no longer accept an undefined as a valid value now, so we must return an int properly (#4976) | 28 February 2017, 19:01:34 UTC |
142adaf | Tasuku SUENAGA a.k.a. gunyarakun | 28 February 2017, 00:08:05 UTC | Fix handling node types on JSDCE() in js-optimizer.js (#4695) * Add a test case of js-optimizer for special object keys * Check variable names and object keys strictly on JSDCE * Move a variable definition and add comments to clarify what's done * Add Tasuku SUENAGA to AUTHORS | 28 February 2017, 00:08:05 UTC |
f920c92 | Alon Zakai (kripken) | 25 February 2017, 23:45:18 UTC | fix asm2.test_webidl, work around a closure compiler issue where a singleton object's methods were called without the instance | 25 February 2017, 23:45:18 UTC |
47d4b6f | Alon Zakai (kripken) | 25 February 2017, 03:05:43 UTC | fix memory growth in binaryen interpreter, broken in #4918 | 25 February 2017, 03:05:43 UTC |
59d0ec8 | ImplOfAnImpl | 25 February 2017, 00:45:28 UTC | Fix for issue 4926 (#4965) * Fix for issue #4926: ELIMINATE_DUPLICATE_FUNCTIONS=1 produces invalid asm.js In function eliminator literals 0.0 are replaced with ZERO$DOT$ZERO before running Uglify * Added myself to authors | 25 February 2017, 00:45:28 UTC |
b39bb50 | Derek Schuff | 24 February 2017, 19:05:00 UTC | Always use async compile unless overridden (#4971) We want to avoid blocking the main thread whenever possible, even when debugging or not-optimizing. Having different behavior here between optimization levels is really weird and surprising. To enforce this preference to avoid blocking the main thread, Chrome will limit synchronous compiles to a pretty small module size; small enough that even emscripten's unoptimized hello world won't compile. | 24 February 2017, 19:05:00 UTC |
428b790 | juj | 24 February 2017, 13:51:51 UTC | Merge pull request #4907 from Jura-Z/incoming WebGL 2 bugfixes for uniform blocks | 24 February 2017, 13:51:51 UTC |
288f8d1 | Jukka Jylänki | 24 February 2017, 12:47:23 UTC | Fix typo in testing whether garbage free WebGL 2 entry points exist. | 24 February 2017, 12:47:23 UTC |
e7e3216 | Jukka Jylänki | 24 February 2017, 11:51:47 UTC | Fix browser.test_sdl_alloctext to allocate a valid heap size. | 24 February 2017, 12:33:34 UTC |
86cd045 | Jukka Jylänki | 24 February 2017, 11:50:30 UTC | Update test_mmap after changing page size from 4k to 16k. | 24 February 2017, 12:33:34 UTC |
52ee9cd | Jukka Jylänki | 24 February 2017, 11:45:34 UTC | Fix sysconf(_SC_PHYS_PAGES) test after changing page size from 4096 bytes to 16384 bytes. Fix sysconf(_SC_PHYS_PAGES) to return correct size when ALLOW_MEMORY_GROWTH=1 is specified or if BINARYEN_MEM_MAX=xxx is specified. | 24 February 2017, 12:33:34 UTC |
db346f8 | juj | 24 February 2017, 11:48:58 UTC | Merge pull request #4969 from juj/opt_redundant_webaudio_pitch_and_gain opt_redundant_webaudio_pitch_and_gain | 24 February 2017, 11:48:58 UTC |
c4c9d4b | Alon Zakai | 23 February 2017, 22:38:35 UTC | Fix Runtime.stackSave etc. in wasm async compilation mode (#4966) * fix wasm async compilation error, we were exporting stackSave etc. in a weird way, that didn't get the benefits of the wrapper that handles async compilation. standardize those exports, add a test, and move binaryen2 to test async compilation, which matches how the feature is actually used * fix makeStructuralReturn * fix setThrew * use a decorator * embind is not yet compatible with async compilation | 23 February 2017, 22:38:35 UTC |
4f56761 | jgravelle-google | 23 February 2017, 22:17:45 UTC | Fix EM_ASM test regression (other.test_split_memory_faking) (#4970) | 23 February 2017, 22:17:45 UTC |
84c9c65 | juj | 23 February 2017, 19:49:56 UTC | Multiprocessing pool cwd fix (#4959) The issue with reusing the Multiprocessing pool in the test harness was that the pool was created to a CWD that was deleted in between tests, which would cause all later subprocess spawns from the pool workers to fail because the worker processes had CWD pointing to a nonexisting directory. To fix, root the pool subprocesses to a known safe directory (EMSCRIPTEN_TEMP) that is very unlikely to die throughout the main process run. Reverts PR #4957 because that addressed a workaround only and had a bug that close_multiprocessing_pool() would not bind to closing the right pool each time if multiple ones had been created. Also reverts PR #4958 because I don't think TMP=. is a feature we should support. Though tested that TMP=. emcc tests/hello_world.c does build fine without errors after this PR. | 23 February 2017, 19:49:56 UTC |
44c9b72 | Jukka Jylänki | 23 February 2017, 16:32:38 UTC | Some applications unconditionally perform pitch and volume changes in OpenAL, which shows up in profilers as a few % overhead, so avoid calling out to Web Audio API if it's not necessary. | 23 February 2017, 16:32:38 UTC |
816b90b | jgravelle-google | 23 February 2017, 15:27:24 UTC | Fix EM_ASM for wasm backend (#4949) * Change EM_ASM implementation to not use varargs for wasm_backend * Move test_em_asm_parameter_pack out to a file * Add test for several combinations of args and return types for EM_ASM * Enable all-but-one of the no_wasm_backend tests that needed EM_ASM * Add explanation for EM_ASM declarations. Reorder C++ asm/wasm cases for readability * asmjs always expects vararg em_asms, always leave them as vararg | 23 February 2017, 15:27:24 UTC |
b70a251 | Bailey Hayes | 22 February 2017, 23:16:03 UTC | add WarningManager (#4775) Use it to allow suppression of default warnings for absolute include/link paths (-Wno-absolute-paths/-Wno-warn-absolute-paths), separating asm (-Wno-separate-asm), and missing optimizations due to allowing memory growth (-Wno-almost-asm). From #4473 | 22 February 2017, 23:16:03 UTC |
86e321c | Alon Zakai (kripken) | 21 February 2017, 21:02:20 UTC | add some preamble docs #4964 | 21 February 2017, 21:02:20 UTC |
91e56cb | Alon Zakai (kripken) | 21 February 2017, 05:14:33 UTC | fix comment | 21 February 2017, 05:14:33 UTC |
4268613 | Alon Zakai (kripken) | 21 February 2017, 05:13:25 UTC | Merge branch 'gorry_fix_20170215' of https://github.com/gorry/emscripten into incoming | 21 February 2017, 05:13:25 UTC |
ee866d1 | Charlie Birks | 20 February 2017, 21:10:07 UTC | Add a test for SDL2 custom cursors (#4955) * Add a test for SDL2 custom cursors * Update SDL2 port to version 13 | 20 February 2017, 21:10:07 UTC |
0539f3f | juj | 20 February 2017, 10:25:02 UTC | Merge pull request #4807 from juj/cleanup_asmjs cleanup_asmjs | 20 February 2017, 10:25:02 UTC |
6acd3a6 | Jukka Jylänki | 14 February 2017, 11:54:24 UTC | use temp_files.note() to clean up the temporary .asm.js file. | 20 February 2017, 10:21:25 UTC |
55e0a58 | Jukka Jylänki | 13 February 2017, 13:37:11 UTC | When outputting wasm only, don't generate .asm.js. Also, don't overwrite a pre-existing .asm.js file in that directory if one wants to deploy both asm.js and wasm. | 20 February 2017, 10:21:24 UTC |
97d9566 | juj | 20 February 2017, 10:20:27 UTC | Merge pull request #4939 from juj/optimize_albufferdata optimize_albufferdata | 20 February 2017, 10:20:27 UTC |
94cf240 | Jukka Jylänki | 12 February 2017, 18:51:57 UTC | Optimize alBufferData() by tightening inner loops. Gives ~5x faster performance. Also fix an issue if page is on the background and a next loop of an audio source is due to start, it is possible that the audio buffer has not yet been set to play. In that case, alSourcef() does not need to adjust existing buffers when changing pitch. | 20 February 2017, 10:07:38 UTC |
713507d | juj | 20 February 2017, 10:05:38 UTC | Merge pull request #4918 from juj/64k_pages 64k_pages | 20 February 2017, 10:05:38 UTC |
30867e8 | Jukka Jylänki | 19 February 2017, 13:23:52 UTC | Fix tests to allocate a valid amount of total memory. | 19 February 2017, 13:30:29 UTC |
d771580 | Jukka Jylänki | 19 February 2017, 12:40:26 UTC | Add check for BINARYEN_MEM_MAX being a multiple of 64KB at compile time. | 19 February 2017, 13:24:50 UTC |
11093c6 | Jukka Jylänki | 17 February 2017, 10:53:18 UTC | Import alignUp to fetch worker. | 19 February 2017, 13:24:49 UTC |
74acd7c | Jukka Jylänki | 17 February 2017, 10:43:07 UTC | Add helper expand_byte_size_suffixes() to expand arithmetic and KB/MB suffixes to integer. | 19 February 2017, 13:24:49 UTC |
8ee6142 | Jukka Jylänki | 16 February 2017, 13:39:12 UTC | Remove redundant memory size tests in preamble.js. Add a test. | 19 February 2017, 13:24:49 UTC |
4e9bca9 | Jukka Jylänki | 14 February 2017, 11:47:00 UTC | Make page size 16KB. | 19 February 2017, 13:24:49 UTC |
65ab9db | Jukka Jylänki | 14 February 2017, 11:35:54 UTC | Stop automatically resizing TOTAL_MEMORY at startup if it is wrong, but require at compile time that the size is specified appropriately. Fixes #4850. | 19 February 2017, 13:24:49 UTC |
0a1f5ce | Jukka Jylänki | 14 February 2017, 10:00:52 UTC | Assert that BINARYEN_MEM_MAX is divisible by page size. | 19 February 2017, 13:24:49 UTC |
f9f29ad | Jukka Jylänki | 14 February 2017, 09:57:57 UTC | Apply code formatting. | 19 February 2017, 13:24:48 UTC |
9092db4 | Jukka Jylänki | 05 February 2017, 01:20:14 UTC | WebAssembly.Memory.grow() API call takes in the delta of the amount of memory to grow, instead of an absolute value for the new buffer, which the previous code misassumed. | 19 February 2017, 13:24:48 UTC |
ed7d2bd | Jukka Jylänki | 04 February 2017, 05:32:32 UTC | Use WebAssembly.Memory.grow() from JS side instead of the Wasm instruction, since it is a bit cleaner because Module['reallocBuffer']() is already JS code anyways. | 19 February 2017, 13:24:48 UTC |
5a581ff | Jukka Jylänki | 04 February 2017, 05:28:26 UTC | Growing memory in wasm can actually give a larger heap than what was asked, so prepare for that as well. | 19 February 2017, 13:24:48 UTC |
51c509d | Jukka Jylänki | 04 February 2017, 03:17:34 UTC | Switch to 64KB page sizes from 4KB, since Wasm pages are 64KB. Revise the heap grow strategy to specifically cap to 2GB-64KB/2GB-16MB for wasm/asm.js. Only report memory growth being slow if running asm.js. Pass maximum WebAssembly Memory size when BINARYEN_MEM_MAX is specified. Fix memory growth to work in binaryen wasm+asm.js builds. (exports['__growWasmMemory'] does not exist in asm.js in that case) | 19 February 2017, 13:24:48 UTC |
481a4ea | Jukka Jylänki | 19 February 2017, 12:33:43 UTC | Guard HTML5 API .staticInit() for the case when window object is not present. | 19 February 2017, 12:34:35 UTC |
da2d623 | Sam Clegg | 19 February 2017, 00:29:51 UTC | Fix bug in extract_archive_contents when TMP=. (#4958) Remove redundant call to os.path.join(temp_dir). This code is running with CWD=temp_dir already. In most cases this is harmless since temp_dir is normally absolute and therefore the resulting path is absolute. However, setting TMP=. will break this as python's gettempdir() will then return '.'. To reproduce the problem simply run "TMP=. emcc hello.cc". The effect is that all the libraries apprar to be empty and you will get undefined symbols. | 19 February 2017, 00:29:51 UTC |
1a1166f | Alon Zakai | 18 February 2017, 19:37:17 UTC | add --ignore-implicit-traps support for binaryen (#4954) | 18 February 2017, 19:37:17 UTC |
6beabbc | Derek Schuff | 18 February 2017, 00:03:04 UTC | Expose Building method to get a new multiprocessing pool (#4957) Use it for Building.read_link_inputs to work around #4941 | 18 February 2017, 00:03:04 UTC |
345394a | Jukka Jylänki | 21 December 2016, 12:41:20 UTC | SharedArrayBuffer shouldn't be present in transfer list anymore. | 17 February 2017, 11:08:35 UTC |
651eacd | juj | 17 February 2017, 10:57:54 UTC | Merge pull request #4921 from juj/no_unconnected_gamepad_garbage no_unconnected_gamepad_garbage | 17 February 2017, 10:57:54 UTC |
1822370 | Alon Zakai (kripken) | 16 February 2017, 23:36:12 UTC | 1.37.3 | 16 February 2017, 23:36:12 UTC |
4a88854 | Alon Zakai | 16 February 2017, 23:34:22 UTC | update binaryen to version_28: 0x01, new type system (#4953) | 16 February 2017, 23:34:22 UTC |
222b375 | juj | 16 February 2017, 15:29:30 UTC | Merge pull request #4911 from juj/avoid_unconditional_asmjs_download avoid_unconditional_asmjs_download | 16 February 2017, 15:29:30 UTC |
42ffd7a | juj | 16 February 2017, 13:17:51 UTC | Merge pull request #4945 from juj/chrome_webgl2_workaround chrome_webgl2_workaround | 16 February 2017, 13:17:51 UTC |
a69a0d5 | Hiroaki GOTO as GORRY | 15 February 2017, 07:16:05 UTC | add my name to AUTHORS | 15 February 2017, 07:16:05 UTC |
92910bf | Hiroaki GOTO as GORRY | 15 February 2017, 07:12:37 UTC | fix compare operator to "===" | 15 February 2017, 07:12:37 UTC |
65c7a28 | Hiroaki GOTO as GORRY | 15 February 2017, 07:10:32 UTC | add comment "why set posz = -.5" | 15 February 2017, 07:10:32 UTC |
084c7b9 | Hiroaki GOTO as GORRY | 14 February 2017, 18:08:44 UTC | initialize panning to center for Chrome | 14 February 2017, 18:08:44 UTC |
9a91242 | Hiroaki GOTO as GORRY | 14 February 2017, 18:04:26 UTC | fix crashing on joystick detached | 14 February 2017, 18:04:26 UTC |