| Age | Commit message (Collapse) | Author |
|---|
|
optimizer
|
|
|
|
|
|
|
|
Conflicts:
tools/js-optimizer.js
|
|
|
|
|
|
|
|
|
|
Conflicts:
emcc
tools/js-optimizer.js
tools/js_optimizer.py
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
other.test_chunking on Windows.
On Windows, prints to stdout/stderr do not appear coherently synchronized per-line if multiple threads print data simultaneously to console. This caused other.test_chunking tests to fail when they receive output like
emcc: step took 5.09 seconds
emcc: running pre-closure post-opts
emcc: running post-closure post-opts
emcc: applying js optimization passes: ['eliminate', 'simplifyExpressionsPre', 'simplifyExpressionsPost', 'last']
splitting up js o.
ptimization into 2 chunks of size 1048576, using 2 cores (total: 1.42 MB)
emcc: step took 2.37 seconds
emcc: total time: 11.75 seconds
where the test looks for the string 'splitting up js optimization into' from the output. Note that the chunk processor thread has printed a single '.' in between the critical line, creating a race condition that randomly makes the test pass or fail, depending on where in the output the '.' lands in.
|
|
|
|
|
|
|
|
|
|
|
|
test_embind
|
|
|
|
|
|
EMCC_DEBUG=1 and 2
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
the generated functions and send only them, or we don't know them and send all the code, in either case js-optimizer.js does not need a list of generated functions
|
|
|
|
|
|
|
|
|
|
|
|
to at least add params to node/v8 in ~/.emscripten by defining a list there and not just a string
|
|
|
