summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Merge pull request #91 from WebAssembly/sanitizersAlon Zakai2016-01-1119-90/+249
|\ | | | | Add Travis builds with sanitizers
| * Check for negative overflow.JF Bastien2016-01-111-30/+50
| |
| * Safe integer: assert before converting double to integer.JF Bastien2016-01-101-0/+5
| |
| * Asm2WasmBuilder: allow building u/s int32.JF Bastien2016-01-101-2/+3
| |
| * Add Travis builds with sanitizersJF Bastien2016-01-1019-77/+210
|/ | | | | | | | | | | | | | | | This triggers 5 independent build / test runs: - clang, no sanitizer; - clang, UB sanitizer; - clang, address sanitizer (disabled for now); - clang, thread sanitizer (disabled for now); - GCC. Enabling UBSan led to these changes: - Fix a bunch of undefined behavior throughout the code base. - Fix some tests that relied on that undefined behavior. - Make some of the tests easier to debug by printing their command line. - Add ubsan blacklist to work around libstdc++ bug. - Example testcase also needs sanitizer because libsupport.a uses it.
* Merge pull request #90 from WebAssembly/more-travisAlon Zakai2016-01-102-8/+5
|\ | | | | A few Travis change
| * A few Travis changeJF Bastien2016-01-102-8/+5
|/ | | | | | | Initialize the git submodules to the index (I'm not sure what they're initialized to right now). Move exports to before_script, and print out compiler version. Invoke check directly (the shebang and x permissions should take care of it). Remove the sleep on warnings, it slows down Travis.
* Merge pull request #89 from WebAssembly/more-travisAlon Zakai2016-01-092-3/+4
|\ | | | | Start to add testing in travis
| * start to add testing in travisAlon Zakai2016-01-091-2/+3
| |
| * use CXX for example testcases, if definedAlon Zakai2016-01-091-1/+1
|/
* Merge pull request #88 from WebAssembly/TravisJF Bastien2016-01-093-3/+30
|\ | | | | Add basic Travis CI build
| * Add basic Travis CI buildJF Bastien2016-01-093-3/+30
|/ | | | No setup or tests yet, I first want to get the basics working properly.
* Update test/waterfallJF Bastien2016-01-091-0/+0
| | | | There's a fix that should make some of the testing faster when there are failures.
* use simplify-locals by defaultAlon Zakai2016-01-099-12/+4
|
* start work on SimplifyLocalsAlon Zakai2016-01-095-1/+113
|
* add test for -OAlon Zakai2016-01-082-0/+46057
|
* add -O option to shell that runs default optimization passesAlon Zakai2016-01-081-1/+6
|
* Merge pull request #87 from WebAssembly/not-just-lkgrJF Bastien2016-01-082-20/+35
|\ | | | | Allow force-updating to the latest waterfall hash.
| * Allow force-updating to the latest waterfall hash.JF Bastien2016-01-082-20/+35
|/ | | | Allow overriding hash.
* RemoveUnneededBrs: remove all code in a block after a breakAlon Zakai2016-01-082-58/+60
|
* improve RemoveUnusedBrs: unify an if-else's brs into one br outsideAlon Zakai2016-01-083-672/+669
|
* add a testcaseAlon Zakai2016-01-082-0/+115
|
* update waterfallAlon Zakai2016-01-081056-7264/+18121
|
* Use waterfall instead of experimental, now the repo has moved.JF Bastien2016-01-085-5/+5
|
* s2wasm torture test: delete output dir when done.JF Bastien2016-01-082-1/+2
|
* pr24716 is failing sexpr-wasm, not s2wasmJF Bastien2016-01-081-4/+0
|
* Merge pull request #86 from WebAssembly/jfbastien-patch-1Alon Zakai2016-01-081-1/+1
|\ | | | | s2wasm: only create output file when successful
| * s2wasm: only create output file when successfulJF Bastien2016-01-081-1/+1
|/ | | Previously the file was created but was empty when s2wasm failed. This caused sexpr-wasm to be sad because it tried to use an empty file.
* Undo the TODO, waterfall seems un-borked nowJF Bastien2016-01-081-8/+6
|
* simplify print output in passes testsAlon Zakai2016-01-087-287/+2
|
* optimize if_else in RemoveUnusedBrsAlon Zakai2016-01-083-8/+186
|
* Temporarily mark tests as passing, even if they shouldn't beJF Bastien2016-01-081-6/+12
|
* updat experimental submoduleAlon Zakai2016-01-071-0/+0
|
* optimize RemoveUnusedBrs some more - handle values #84Alon Zakai2016-01-073-6/+168
|
* Merge pull request #83 from WebAssembly/skipAlon Zakai2016-01-071-1/+1
|\ | | | | Handle .skip
| * Handle .skipJF Bastien2016-01-071-1/+1
|/ | | | LLVM will generate this now.
* handle .zero x, yAlon Zakai2016-01-072-2/+6
|
* Merge branch 'private-label-names'Alon Zakai2016-01-071097-17104/+17104
|\ | | | | | | | | | | Conflicts: test/llvm_autogenerated/cfg-stackify.wast test/llvm_autogenerated/switch.wast
| * LLVM is now emitting private labels with a ".L" prefix. Update binaryen.Dan Gohman2016-01-071097-17104/+17104
| | | | | | | | | | The .L prefix convention is admittedly not the prettiest convention possible, but it is widely used in the ELF ecosystem.
* | fix default being a br in switch #70Alon Zakai2016-01-074-10/+5
|/
* Merge pull request #78 from WebAssembly/update-if-neededAlon Zakai2016-01-061-18/+40
|\ | | | | update.py: update only if needed
| * update.py: update only if neededJF Bastien2016-01-061-18/+40
|/ | | | @kripken asked if test/lkgr could be updated only if the downloaded content is different. This patch does just that.
* fix switch parsing and semantics, for br in the tableAlon Zakai2016-01-062-9/+5
|
* update switch outputs #70Alon Zakai2016-01-062-3/+3
|
* update auto_update_tests.pyAlon Zakai2016-01-061-12/+11
|
* fix switch br/case printing #70Alon Zakai2016-01-061-1/+5
|
* Merge pull request #74 from WebAssembly/llvm_autogeneratedDan Gohman2016-01-0653-1809/+1552
|\ | | | | Replace the experimental/prototype-wasmate/test tests.
| * Remove .wast files for tests that are currently blacklisted or removed.Dan Gohman2016-01-063-176/+0
| |
| * Print nan literals as `nan:0x...` rather than `nan:...` to match the spec.Dan Gohman2016-01-062-6/+6
| |
| * Remove .wast files that now live in llvm_autogenerated.Dan Gohman2016-01-0648-9656/+13
| |
generated by cgit v1.2.3 (git 2.39.1) at 2025-08-17 03:09:36 +0000