| 1690a58 | Sam Clegg | 30 January 2021, 02:29:24 UTC | 2.0.13 (#13391) | 30 January 2021, 02:29:24 UTC |
| 681cdf3 | Sam Clegg | 29 January 2021, 20:37:31 UTC | Use cache directory when building system libraries (#13370) This avoids the use of the temp directory when building system libraries. Instead we use a directory inside the cache. At the point when we are building the system libray we are already holding the cache lock. This is part of a sequence of changes to improve FROZEN_CACHE mode. | 29 January 2021, 20:37:31 UTC |
| df83597 | Sam Clegg | 29 January 2021, 19:39:07 UTC | Update llvm dependency to 13.0 (#13383) Followup on #13358. Remove exceptions now that llvm 13 has rolled into emscripten-releases. | 29 January 2021, 19:39:07 UTC |
| ee11dcc | Sam Clegg | 29 January 2021, 18:28:33 UTC | Use <<< >>> as late-replacement pattern for MINIMAL_RUNTIME. NFC (#13381) See #13265 | 29 January 2021, 18:28:33 UTC |
| a58a9b9 | Allan Regush | 29 January 2021, 17:42:41 UTC | [Docs] Fix broken documentation link (#13346) Fixes #13345 | 29 January 2021, 17:42:41 UTC |
| e74e0af | Sam Clegg | 29 January 2021, 16:06:22 UTC | Avoid writing even sanity.txt if FROZEN_CACHE is enabled (#13316) These improvements to the `FROZE_CACHE` setting allow for a pre-populated, read-only cache. In this scenario the cache is cannot ever be locked, and that we don't even write the sanity file. Users can still force a sanity check with `emcc --check` but no sanity file will ever be written. Fixes: #13369 | 29 January 2021, 16:06:22 UTC |
| a8d41c0 | Sam Clegg | 29 January 2021, 16:04:28 UTC | Use a specific lock file to avoid temp directory contention (#13371) Previously in debug mode we could try to avoid temp file contension by taking and holding the cache lock for the duration of the process. This change uses a specific lock file within the temp directory which avoid needing a writable cache directory. Its also more logical. Partial fix for #13369 | 29 January 2021, 16:04:28 UTC |
| 4928f85 | waterlike86 | 29 January 2021, 03:17:19 UTC | Allow loadWebAssemblyModule to load a compiled wasm (#13252) Enables loadWebAssemblyModule to use both compiled wasm and binary. expose mergelibSymbols so it can be used postInstantiate. For performance reasons we fetch and compile a side module on a separate thread and post it over to the worker thread with the main module. Fixes: #13262 | 29 January 2021, 03:17:19 UTC |
| 4c19197 | Thomas Lively | 29 January 2021, 00:59:07 UTC | [UBSan] Fix infinite recursion in type_info::operator== (#13367) Under UBSan, comparing type_infos triggers a lookup in UBSan's typeinfo cache. When this cache is not warm, this causes a cache miss, and the code path that handles that ends up calling __dynamic_cast. __dynamic_cast itself ends up comparing type_infos and this recursion ends up exhausting the stack. ``` RangeError: Maximum call stack size exceeded ... at std::type_info::operator==(std::type_info const&) const at is_equal(std::type_info const*, std::type_info const*, bool) at __dynamic_cast (<anonymous>:wasm-function[40]:0x8fb) at __ubsan::checkDynamicType(void*, void*, unsigned long) at HandleDynamicTypeCacheMiss(__ubsan::DynamicTypeCacheMissData*, unsigned long, unsigned long, __ubsan::ReportOptions) at __ubsan_handle_dynamic_type_cache_miss at std::type_info::operator==(std::type_info const&) const ... ``` This is not a problem on native platforms because inlining and follow-on optimizations remove the call to type_info::operator== under __dynamic_cast, so UBSan never re-enters its type checking routine. In Emscripten, however, we build libc++abi with -Oz, so these optimizations do not take place. The fix is to add __attribute__((always_inline)) to type_info::operator==, which is enough to get the recursion optimized out in -Oz. Fixes #13330. Fixes #13324. | 29 January 2021, 00:59:07 UTC |
| 06825ec | Sam Clegg | 28 January 2021, 23:57:25 UTC | Don't export all symbols in RELOCATABLE mode. (#13366) Since we don't have many users who use RELOCATABLE directly it went unnoticed that setting this option on its own was causing all symbols to be included and exported. Fixes #13029 | 28 January 2021, 23:57:25 UTC |
| 96ffe11 | Sam Clegg | 28 January 2021, 21:49:07 UTC | Remove complexity in caching of llvm-nm results. NFC (#13360) This extra argument (include_internal) and the extra complexity it entailed had only a single call site in the test suite. Calling llvm-nm directly in the test suite is much simpler. | 28 January 2021, 21:49:07 UTC |
| 88a3b47 | jprendes | 28 January 2021, 21:46:07 UTC | Deterministic dynamic libs order (#13308) When using pthread with more than one dynamic library the order in which the libraries are linked when the module is loaded in the browser's main thread is non deterministic. This makes the function pointers between the main thread and the workers not compatible. With this change the main thread links the dynamic libraries in the same deterministic order as in the workers. This fix only applies to libraries linked when the module is loaded and adds a test to lock down this behavior. Libraries linked only through dlopen remains unsupported. Fix #13303 | 28 January 2021, 21:46:07 UTC |
| deab778 | Sam Clegg | 28 January 2021, 19:46:45 UTC | Update expected llvm version to 13.0 (#13358) This change temporarily disables the version check to allow the llvm change to roll in. Copied from the previous time around: #11637. | 28 January 2021, 19:46:45 UTC |
| ae86492 | Sam Clegg | 28 January 2021, 01:42:06 UTC | Remove undef Browser.safeCallback and Browser.safeSetInterval (#13333) These were not used to tested anywhere in emscripten. Its possible that external developers could be accessing them so I mentioned the removal in the changelog. | 28 January 2021, 01:42:06 UTC |
| 7c0e898 | Sam Clegg | 28 January 2021, 01:40:01 UTC | tests: remove get_async_args function. NFC (#13350) Its cleaner just to inline these settings like all the others. | 28 January 2021, 01:40:01 UTC |
| d82c215 | Sam Clegg | 27 January 2021, 21:26:55 UTC | Remove inline fake emscripten API implemention from emscripten.h (#13344) If folks find this fake implementation useful we could consider shipping it as fake_emscripten.c perhaps (and add some tests), but having it inline in the header seems messy and makes the headers harder to read. | 27 January 2021, 21:26:55 UTC |
| bae6cf7 | Sam Clegg | 27 January 2021, 21:23:41 UTC | Cleanup test_emscripten_async_wget2 (#13348) - Name the next more meaningfully - Merge into a single file - Remove tabs - Fix fomatting - Put trivial methods in class | 27 January 2021, 21:23:41 UTC |
| 052693a | Sam Clegg | 27 January 2021, 21:19:38 UTC | Remove extra arguments from emscripten_set_main_loop JS function. NFC (#13342) This function is only used by native code and it takes only 3 arguments. | 27 January 2021, 21:19:38 UTC |
| c5fc4f2 | Sam Clegg | 27 January 2021, 21:18:50 UTC | Build cached items directly to expected locations. NFC. (#13329) When populating the cache with the creator functions that are passed to `Cache.get` and `Cache.get_lib`, pass the name of the expected output file to the creator function and assert that it creates it. This avoid unnessarily using a temporary file in most cases and saves some code size/complexity. Also add mpg123 to embuilder where it was missing. | 27 January 2021, 21:18:50 UTC |
| 4f9c218 | Sam Clegg | 27 January 2021, 16:31:39 UTC | Use JS naming convention for __set_network_callbac. NFC. (#13336) | 27 January 2021, 16:31:39 UTC |
| 7ff719c | juj | 27 January 2021, 16:12:59 UTC | Restore the dynCall() and dynCall_sig() API into the build (#13296) * Restore the dynCall() and dynCall_sig() API into the build when -s USE_LEGACY_DYNCALLS=1 is passed. Add support for dynCall() in MINIMAL_RUNTIME builds (that was not implemented before). * Add ChangeLog entry. * Rename USE_LEGACY_DYNCALLS to DYNCALLS. * Inline getDynCaller. * Address review. * Merge new dynCall test with old. * Reformat parseTools.js | 27 January 2021, 16:12:59 UTC |
| 110a856 | Sam Clegg | 27 January 2021, 01:15:36 UTC | Override get_num_cores when in EMCC_DEBUG mode (#13331) This means that its not just the mutliprocessing pool that we limit to single core but also any places we do things like "make -j" in ports to testing. | 27 January 2021, 01:15:36 UTC |
| c92f413 | Sam Clegg | 27 January 2021, 00:33:50 UTC | Use arrow functions in internal JS code. NFC. (#13319) | 27 January 2021, 00:33:50 UTC |
| 6dcfcb4 | Sam Clegg | 26 January 2021, 23:15:37 UTC | Revert accidental change to log messages in #13299. NFC. (#13328) | 26 January 2021, 23:15:37 UTC |
| 01ef6f5 | mehdisadeghi | 26 January 2021, 22:07:11 UTC | Add HarfBuzz flag to SDL2_ttf rev 703ebc7c66fd. (#12868) * Add HarfBuzz flag to SDL2_ttf rev 703ebc7c66fd. `TTF_USE_HARFBUZ` flag adds complex script support to SDL_ttf, hence the update to the latest revesion. Fixes #12845. * Add a new test for Arabic shaping with HarfBuzz. This commit adds a test which uses SDL2_ttf for Arabic Shaping with HarfBuzz. In order to do so, a font with Arabic Glyphs was necessary. Therefore NotoNaskhArabic font was added. | 26 January 2021, 22:07:11 UTC |
| ff5e9f6 | Alon Zakai | 26 January 2021, 21:36:16 UTC | Document memory profiling options (#13307) | 26 January 2021, 21:36:16 UTC |
| 67866c3 | Sam Clegg | 26 January 2021, 21:00:12 UTC | Fix embuilder --force for ports (#13325) This was broken at some point (possibly by #9353) but we didn't have a test for it. | 26 January 2021, 21:00:12 UTC |
| c8c46c3 | Sam Clegg | 26 January 2021, 20:58:39 UTC | Add eslint for linting our build-time JS code (#13305) Initially I chose a single file from the JS compiler and made it conform to the style. There is currently no plan to extend this to libray code or JS code that runs on the client. Partly for practical reasons, since that code all run through a pre-processor, but also because we have very different JS version requirements there. | 26 January 2021, 20:58:39 UTC |
| f58f42b | Sam Clegg | 26 January 2021, 18:50:42 UTC | Refactor Ports.fetch_port to avoid locking the cache where possible (#13321) This is in service of improving support for FROZEN_CACHE and is needed as part of #13316. The logic here mirrors the lock in cache.get() where we first attempt to verify the contents of the cache without holding the lock and early out in the common case the what we want is already here. | 26 January 2021, 18:50:42 UTC |
| 830e1cc | Sam Clegg | 26 January 2021, 18:44:05 UTC | Remove some python3 compat code from system_libs.py. NFC. (#13322) | 26 January 2021, 18:44:05 UTC |
| ac0a4b3 | Adam Bujalski | 26 January 2021, 18:38:47 UTC | Disabling flaky test_pthread_cond_signal_1_1 (#13259) test_pthread_cond_signal_1_1 has very tight timing requirements, and expects that side threads will be woken and do some action in very short time. This causes occasional spurious fail of this test not because side thread(s) hasn't been woken by pthread_cond_signal, but because this threads hasn't been scheduled and hasn't get CPU on time. See #13283 | 26 January 2021, 18:38:47 UTC |
| 536a14e | Sam Clegg | 26 January 2021, 00:41:57 UTC | Avoid using EM_IGNORE_SANITY in testcode. (#13317) This isn't needed provided: a. the fake tools are good enough to fool our checks b. we clear the sanity file explicitly between each test iteration. | 26 January 2021, 00:41:57 UTC |
| 5d0d5cb | Sam Clegg | 25 January 2021, 23:37:04 UTC | Remove symbols for last 3 libraries that have them: gl, al, html5 (#11340) Rather than including these libraries on the link line based on running `llvm-nm` on the inputs we now include them unconditionally. This behavior can be disabled using AUTO_NATIVE_LIBRARIES=0. This change should not any effect on normal build since including these libraries should be a nps. For MAIN_MODULE=1 build it could be a slightly increased code size, but MAIN_MODULE=1 builds are already enormous and MAIN_MODULE=2 is the path to make them smaller. Since these are the last libraries to use these symbols files we can also remove all the code surrounding the processing of them! | 25 January 2021, 23:37:04 UTC |
| d38ebe3 | Kagami Sascha Rosylight | 25 January 2021, 21:21:45 UTC | Use Cache.get_lib for giflib (#13311) #13209 added get_lib but it was missed in #13139. | 25 January 2021, 21:21:45 UTC |
| 38485ce | Sam Clegg | 22 January 2021, 18:22:36 UTC | Remove unused async callback system from library_browser.py (#13301) This looks like it was only used by EMTERPRETIFY_ASYNC which was removed in #11020. | 22 January 2021, 18:22:36 UTC |
| 1f4a748 | Sam Clegg | 22 January 2021, 14:32:13 UTC | tests: Remove force_c paramater from btest method. NFC (#13300) Since we started passing the filename all the way through to the compiler (See #13112) this argument has been completely ignored. | 22 January 2021, 14:32:13 UTC |
| f236e5d | Sam Clegg | 22 January 2021, 01:15:47 UTC | More cleanup of heap access code in parseToolsj.js. NFC (#13292) Followup on #13290. We only deal with a single heap these days so we can remove some of the complexity here. | 22 January 2021, 01:15:47 UTC |
| 904a66e | Sam Clegg | 22 January 2021, 01:15:19 UTC | Acquire the cache lock before checking for sanity file (#13299) Without this check there is a race to create that sanity file when emscripten if first used. I noticed this strange behavior when I would run the test suite just after clearig the cache. On my machine the test suite would launch many emcc processes and each of them would notice the sanity file as missing and run the checks. Some of them would see an empty sanify file that one of the other processes was in the middle of creating (before they flushed it), and when this happens that process would try to clear the cache! This resulted in the cache being clearing by one process while it was being populated by another. Not good. The fix is the hold the cache lock while we check for the sanity file and while we run the sanity checks. This problem has always existed but was recently exacerbated by the fact that we populate the sysroot headers as soon as we start emcc. Also remove superfluous ensure() calls. Since we call this during the constructor these days the cache can be assumed to always exist. | 22 January 2021, 01:15:19 UTC |
| 6928ece | iamahuman | 21 January 2021, 23:49:14 UTC | emcc: support merging static library into a single object (#7683) Remove the optimization (again) for the special case of single object input, which skips the linker and directly copies the input object file to output. This behavior interferes with linking (merging) all objects in a library archive into a single object, done with the following command: cc -o lib.o -r -Wl,--whole-archive lib.a The optimization was once useful when no clear distinction was made between compile mode and link-only mode. It was also required for compatibility (see #9510 and #9571). However, #9510 was addressed by #9511, and #9600 introduced the '-r' flag that allows to explicitly indicate linking to an object, making the optimization less useful. | 21 January 2021, 23:49:14 UTC |
| 6111548 | Sam Clegg | 21 January 2021, 22:29:56 UTC | parseTools: remove unused variable (#13293) Looks like the actual usage of printType was removed back in 66bd874d9bcc3a0ef2b9892253ee8359525ce48d. | 21 January 2021, 22:29:56 UTC |
| 7c1d639 | Sam Clegg | 21 January 2021, 22:29:35 UTC | parseTools: mark args as unused in makeGetValue/makeSetValue (#13294) | 21 January 2021, 22:29:35 UTC |
| 7c2417d | Sam Clegg | 21 January 2021, 01:32:38 UTC | Cleanup makeGetSlabs in src/parseTools.js. NFC. (#13290) - Make fall throughs more obvious. - Remove break after return. - Remove unused final return (it was unreachable). | 21 January 2021, 01:32:38 UTC |
| c62e44a | Sam Clegg | 20 January 2021, 23:13:36 UTC | Add musl update script. NFC. (#13269) I'm in the process of updateing musl using a git mirror that holds both upstream musl and our local emscripten changes. This scripte can be used to copy new versions into place. | 20 January 2021, 23:13:36 UTC |
| b0da998 | Patric Stout | 20 January 2021, 21:39:52 UTC | Zero sin_zero when writing a sockaddr_in (#13003) Although not strictly defined by POSIX, it is done by most (all?) other implementations, most often in patches over the years as places got forgotten. This is mostly useful to allow quick comparing between sockaddr_in to see if they are the same address. Fixes #12998 | 20 January 2021, 21:39:52 UTC |
| 094e639 | Sam Clegg | 20 January 2021, 19:18:35 UTC | Rebaseline other.test_gen_struct_info. NFC (#13291) Of course right when I landed this new test if crossed paths with a change the json output. Doh! | 20 January 2021, 19:18:35 UTC |
| 585c5e6 | Sam Clegg | 20 January 2021, 16:52:37 UTC | Refactor gen_struct_info to allow separate compilation. NFC. (#13270) As part of the pending musl update we can no longer (or at least it becomes problematic) to compile internal musl headers alongside the public headers. To avoid this, this change updates gen_struct_info.py to compile internal headers separately. To facilitate this, this change adds support for compiling groups of headers and merging the results into a single json output. As part of debugging this I added support for compiling every header file in isolation. While this is slow, it uncovered a couple of issues where structs and defines that were listed under the wrong header and also a header that could not be compiled on its own (SDL_pixels.h). | 20 January 2021, 16:52:37 UTC |
| 462ddb4 | Sam Clegg | 20 January 2021, 15:55:13 UTC | Implement fstatat flag: AT_EMPTY_PATH (#13282) Also, add new test for fstatat. I notived this issue while updating to the new musl version that uses this internally. | 20 January 2021, 15:55:13 UTC |
| 0c32a99 | Sam Clegg | 20 January 2021, 14:27:58 UTC | Rebaseline code size tests after llvm change. NFC (#13286) | 20 January 2021, 14:27:58 UTC |
| 4b78241 | Sam Clegg | 19 January 2021, 20:34:55 UTC | Remove redundant zero'ing of pthread fields. NFC. (#13273) The spawnThread function is only even called with struct that have already been zero-initialized in pthread_create. | 19 January 2021, 20:34:55 UTC |
| a0e9b4e | Sam Clegg | 19 January 2021, 20:33:40 UTC | Remove redundant USE_THREADS macro in system_libs.py. NFC (#13274) We already automatically define `__EMSCRIPTEN_THREADS__` in emcc.py whenever `-s USE_PTHREADS` is enabled. | 19 January 2021, 20:33:40 UTC |
| d7c7aed | Sam Clegg | 19 January 2021, 14:50:16 UTC | Use individual socket syscalls. (#13272) Remove the SYSCALL_USE_SOCKETCALL definition which was causing musl to use a single syscall for all socket calls (a mostly legacy linux convention apparently). This is much better for emscripten because it allows DCE and more precise dependencies. | 19 January 2021, 14:50:16 UTC |
| 39c848c | Sam Clegg | 19 January 2021, 14:11:38 UTC | Add generic musl arch headers. NFC. (#13280) This allows us to then remove some of our emscripten-specific arch headers that are not longer needed (since they do not differ from the generic version). This is part of making musl updates easier. See #7279 | 19 January 2021, 14:11:38 UTC |
| ee483fc | Sam Clegg | 18 January 2021, 17:48:00 UTC | test_other: use self.do_runf where it makes sense. NFC. (#13275) | 18 January 2021, 17:48:00 UTC |
| 540d3ba | Sam Clegg | 18 January 2021, 16:40:41 UTC | Fix definition of mtx_t and cond_t (#13271) The convention is musl is to duplicate these types rather than typedef them. Fixes: #5343 | 18 January 2021, 16:40:41 UTC |
| 4df8093 | Sam Clegg | 18 January 2021, 16:20:59 UTC | Move emscripten-specific source files out of musl directory (#13268) This allows for us to more easily manage/maintain/upgrade the musl directory as a fork of upstream. | 18 January 2021, 16:20:59 UTC |
| 0e92f93 | Sam Clegg | 18 January 2021, 16:18:22 UTC | Improve error message when check for multiple of page size (#13279) Also: - remove duplication of wasm page size specification - remove old reference to asm.js Fixes: #13277 | 18 January 2021, 16:18:22 UTC |
| 279b7f8 | Sam Clegg | 18 January 2021, 15:51:53 UTC | test: remove two unneeded include paths. NFC. (#13276) | 18 January 2021, 15:51:53 UTC |
| ba2b891 | Sam Clegg | 18 January 2021, 15:25:55 UTC | Remove out copyright from some upstream sources. NFC (#13278) | 18 January 2021, 15:25:55 UTC |
| 72d6ae3 | Sam Clegg | 15 January 2021, 17:02:57 UTC | Remove need to `getQuoted` by using seperate template pattern (#13265) Use <<< >>> for late-substitution replacements as opposed to {{{ }}} which is used to by the JS compiler. This avoids the somewhat confusing getQuoted function which works as an escape hatch. | 15 January 2021, 17:02:57 UTC |
| ae09c9d | Sam Clegg | 15 January 2021, 15:18:53 UTC | Don't include webgl2.c in regular gl library (#13261) This fixes big where linking MAIN_MODULE and libgl was bring in undefined references to GL2 symbols. Normally including this extra `webgl.c` file in the library would be harmless since it will be ignored if no GL2 symbols are used. However, with MAIN_MODULE mode all objects from all libraries are included which means that webgl2.o is included which itself references all webgl2 symbols. The problem specifically is that `webgl2.c` includes `emscripten_webgl2_get_proc_address` which in turn references all the `emscripten_XXX` versions of the GL2 API such as `emscripten_glReadBuffer` which are defined in `library_webgl2.js` Fixes: #13101 | 15 January 2021, 15:18:53 UTC |
| 3307679 | Sam Clegg | 15 January 2021, 14:06:07 UTC | Simplify handling of forwarded JS compiler metadata. NFC. (#13264) This class/code used to do a lot more in the past so it made sense to have a separate class. Now that the logic is simpler we can just directly die the metadata for a further simplification. | 15 January 2021, 14:06:07 UTC |
| 36a3fa5 | Dexter Chua | 14 January 2021, 21:44:43 UTC | [NFC] Be clearer about not generating DataRequest if lz4 is used (#13218) | 14 January 2021, 21:44:43 UTC |
| 7684d4a | juj | 14 January 2021, 21:24:45 UTC | Remove old leftover MEMFS code (#13255) * In PR #7918, the option MEMFS_APPEND_TO_TYPED_ARRAYS=0 was removed, dropping the support for using regular JS Arrays to back MEMFS file storage. In this PR, complete the removal of old code that related to JS Array backing, since all files are now typed array backed. * Migrate tests to use the MEMFS typed array API. * Fix tests. | 14 January 2021, 21:24:45 UTC |
| 5ee8fe3 | Sam Clegg | 14 January 2021, 20:32:29 UTC | Initial support for pthreads + dynamic linking (#13245) This support is still experimental and has some major caveats: - Only supports load-time dynamic linking - No support for sharing TLS data between dylibs. The major change is that workers now also call the module `run` function although they exit early, once dynamic libraries have all been loaded. See: #3494 | 14 January 2021, 20:32:29 UTC |
| 5fb1d65 | Sam Clegg | 14 January 2021, 19:56:12 UTC | Use addAtExit in __postset. NFC. (#13254) When registering atexit handlers in __postset functions the convention elsewhere seems to be to prefer this helper function (which does nothing when EXIT_RUNTIME is not set). | 14 January 2021, 19:56:12 UTC |
| a577171 | Sam Clegg | 14 January 2021, 19:34:32 UTC | Avoid special replacements for STACK_BASE/STACK_MAX/HEAP_BASE (#13253) Rather then using `getQuoted` and then special-case replacements for pre/post we can just use internal settings. This has the advantage that these settings can be included in library code as we as the pre/post js. After this change getQuoted now only has a couple of remainging uses for ATINITS, ATEXITS, ATMAINS and WASM_BINARY_DATA | 14 January 2021, 19:34:32 UTC |
| 259e2ee | Raffarti | 14 January 2021, 19:33:57 UTC | Fix `alcCaptureSamples` (#13178) - Properly read from the circular buffer (fixes #13173) - Fix requested frame count check (it was failing when requested frequency was higher than source frequency) | 14 January 2021, 19:33:57 UTC |
| 9feb504 | Alon Zakai | 14 January 2021, 01:15:29 UTC | Warn on failure to load from a file URI. Fixes #13207 (#13215) | 14 January 2021, 01:15:29 UTC |
| 428f33a | Alon Zakai | 14 January 2021, 01:14:59 UTC | Do not process main thread queued calls while returning a worker to the pool (#13233) Fixes #12970 | 14 January 2021, 01:14:59 UTC |
| 97f13a1 | Kagami Sascha Rosylight | 13 January 2021, 22:56:33 UTC | Add libgif port (#13139) This is useful in building the JPEG XL reference software, for example. | 13 January 2021, 22:56:33 UTC |
| c5cde8a | Sam Clegg | 13 January 2021, 22:26:11 UTC | Cleanup test_size.cpp. NFC. (#13247) - Fix memory leak. - Re-enable in safe heap mode (seems to pass just fine). - Fix indentation. | 13 January 2021, 22:26:11 UTC |
| ea208ff | Sam Clegg | 13 January 2021, 22:09:18 UTC | Re-enable some disabled browser tests. NFC (#13244) | 13 January 2021, 22:09:18 UTC |
| 04e7685 | Sam Clegg | 13 January 2021, 14:06:49 UTC | Extract stackCheckInit() function from run(). NFC (#13246) Split out from #13245. | 13 January 2021, 14:06:49 UTC |
| f581350 | Alon Zakai | 12 January 2021, 22:18:57 UTC | Disable a flaky recently-added test, asan.test_pthread_cxx_threads (#13242) | 12 January 2021, 22:18:57 UTC |
| 63422ac | Adam Leskis | 12 January 2021, 20:31:41 UTC | [docs] Update http to https (#13225) | 12 January 2021, 20:31:41 UTC |
| 6df2aae | Sam Clegg | 12 January 2021, 18:46:39 UTC | Use JS naming convention for fetch JS functions (#13238) The strange mixed naming convention used here exposed a bug in jsifier. See: #13230 | 12 January 2021, 18:46:39 UTC |
| 3c7a866 | Sam Clegg | 12 January 2021, 18:45:57 UTC | Fix MODULARIZE + PTHREADS + USE_OFFSET_CONVERTER (asan) (#13236) Fixes: #13138 | 12 January 2021, 18:45:57 UTC |
| 1050ed3 | Sam Clegg | 12 January 2021, 13:34:10 UTC | Add basic/empty test of C++ threads. NFC. (#13237) | 12 January 2021, 13:34:10 UTC |
| c168284 | Alon Zakai | 11 January 2021, 23:43:33 UTC | [docs] Some fixes for packaging docs (#13234) | 11 January 2021, 23:43:33 UTC |
| d1d1147 | Kleis Auke Wolthuizen | 11 January 2021, 22:48:06 UTC | Cleanup pthread tests. NFC. (#13211) | 11 January 2021, 22:48:06 UTC |
| 8e885f7 | Amin Yahyaabadi | 11 January 2021, 22:45:57 UTC | [docs] Add Windows package to the packaging section (#13149) Fixes emscripten-core/emsdk#272 I added the packages to the bottom of the downloads page, instead of a hidden mention in a very large note. | 11 January 2021, 22:45:57 UTC |
| c978e1e | Alon Zakai | 11 January 2021, 22:44:39 UTC | [docs] Mention SINGLE_FILE alongside the local webserver docs. see #13207 (#13232) | 11 January 2021, 22:44:39 UTC |
| e099eca | Sam Clegg | 11 January 2021, 22:44:25 UTC | Avoid mangling + demangling in jsifier.js (#13230) We had a bug that only showed up with symbols that start with `$_`. These happen to exist in `library_fetch.js` e.g: `$__emscripten_fetch_xhr`. In this case, when we mangle, we remove the leading `$` giving `__emscripten_fetch_xhr`, and then if we demangle we do that wrong thing and end up removing the `_` instead of re-adding the `$`. This change avoids the round trip by always storing both the mangled and demangled version of a given symbol. See #12268 | 11 January 2021, 22:44:25 UTC |
| ad43082 | Sam Clegg | 11 January 2021, 22:44:07 UTC | Fix strange/broken syntax in __resumeException__deps (#13231) Its hard for me to tell exactly what this was originally trying to express, but right now it looks broken. Prior to bcf60cea I think it might have made more sense when the function was just one element of the list, but then with bcf60cea it became a function with the list inside it as expressions that do nothing. | 11 January 2021, 22:44:07 UTC |
| 75095c1 | Sam Clegg | 11 January 2021, 19:20:13 UTC | Move simd compatability headers to `compat` (#13170) We want these files to take presence over the clang builtin include so we put them in the compat directory which is explictly added to include path and therefore comes before clang's builtin include directory. | 11 January 2021, 19:20:13 UTC |
| dee59ba | Sam Clegg | 11 January 2021, 17:41:01 UTC | Restore orginal file layout for musl and libcxx. NFC. (#13154) Now that we build the sysroot, including the header tree programatically we can restore the original layout of the musl and libcxx directories. This will make future updates and comparisons with upstream easier. | 11 January 2021, 17:41:01 UTC |
| bbec228 | Sam Clegg | 11 January 2021, 16:39:44 UTC | Allow overriding STRICT mode setting (#13229) STRICT mode enables a bunch of other settings but it should still be possible to individually re-enable them. | 11 January 2021, 16:39:44 UTC |
| 9a1d041 | Sam Clegg | 11 January 2021, 14:46:47 UTC | tests: switch more tests to btest_exit. NFC. (#13018) Thus avoiding the need for REPORT_RESULT. | 11 January 2021, 14:46:47 UTC |
| 44add0c | Sam Clegg | 11 January 2021, 12:51:40 UTC | tests: fix check_dlfcn check (#13228) I broke this change when I landed #13201. | 11 January 2021, 12:51:40 UTC |
| ba88a7c | Sam Clegg | 11 January 2021, 12:16:52 UTC | Bump binaryen version (#13222) | 11 January 2021, 12:16:52 UTC |
| 298b2af | Sam Clegg | 10 January 2021, 12:09:02 UTC | Build a complete sysroot in the cache directory (#13090) Rather than adding various include paths, copy any needed headers into the sysroot along with any libraries. This means that emscripten can work a lot more like the traditional cross compiler (e.g. clang -target=xxx --sysroot=yyy), and we can start to think of the emscripten driver as a seperate thing to the sysroot. Fixes: #9353 | 10 January 2021, 12:09:02 UTC |
| ae6f40d | Kleis Auke Wolthuizen | 09 January 2021, 19:22:05 UTC | Update posixtestsuite and disable newly failing tests. NFC. (#13059) Previously the testsuite had been modified to disable them. See: https://github.com/emscripten-core/posixtestsuite/pull/4. | 09 January 2021, 19:22:05 UTC |
| 61721ba | Sam Clegg | 09 January 2021, 16:23:50 UTC | Feedback from #13213 (#13216) | 09 January 2021, 16:23:50 UTC |
| 38f85ee | Sam Clegg | 09 January 2021, 16:21:25 UTC | 2.0.12 (#13221) | 09 January 2021, 16:21:25 UTC |
| d0e647b | Sam Clegg | 08 January 2021, 23:25:07 UTC | Use single cache rather then reconfiguring/moving it (#13209) This was split out from the change to use a single sysroot (#13090). I think it cleaner this way: There is single cache, with single lock file and the root doesn't change, but the libraries live in sub-directories within the cache. | 08 January 2021, 23:25:07 UTC |
| 91a177b | Sam Clegg | 08 January 2021, 22:45:48 UTC | Remove remnands of WebVR support (#13210) See initial removal or the library in #10460. | 08 January 2021, 22:45:48 UTC |
| 96bd26c | Sam Clegg | 08 January 2021, 22:45:30 UTC | tests: use `delete_contents` helper function. NFC. (#13208) | 08 January 2021, 22:45:30 UTC |
| dd827d6 | Dexter Chua | 08 January 2021, 22:43:35 UTC | Fix LZ4 file packaging support when MODULARIZE=1 (#13161) This is based on work by @mdboom in #7107 | 08 January 2021, 22:43:35 UTC |
| 5dc63ff | Sam Clegg | 08 January 2021, 19:41:12 UTC | Remove use of `initializer` metadata produced by finalize (#13213) With the wasm backend there is only ever at most one initializer function do we don't need complex handling, we can do it all in the pre/postable. Also ignore two other unused metadata fields: `simd` and `implementedFunctions`. Both of these are not longer output by binaryen. | 08 January 2021, 19:41:12 UTC |
| 2458fe3 | Sam Clegg | 08 January 2021, 18:47:07 UTC | Combine check_call and print_compiler_stack. NFC. (#13203) This means that `emcc -v` will consistently print all sub-commands that run via `check_call`. | 08 January 2021, 18:47:07 UTC |
