Feature/34/support wasm

[azriel91]

(this is based over #33)

Closes #34.

Attempt to support WASM. This builds and runs in a WASM application, but requires the following PRs:

• Adds checked_add and checked_sub to wasm::Instant. sebcrozet/instant#17 -- provide Instant::checked_sub. Needs instant release > 0.1.3 (Release 0.1.4 sebcrozet/instant#18)
• Uses instant::Instant in place of std::time::Instant. Amanieu/parking_lot#231 -- use instant::Instant. Needs parking_lot release > 0.10.2.

[antifuchs]

Holy heck! That's amazing!!

Thanks so much for these contributions - I'm somewhat time-constrained this week, but I'll try to get them merged this weekend!

[azriel91]

Sweet ✌️, both the PRs are merged, simply waiting on their release, so we can refer to the released versions in here 🚀

[antifuchs]

So, I love this! Two things that I'd like to see in this PR still:

• some way to test this in CI, would hate to accidentally undo this work (if you're not sure how to do that, feel free to send me a snippet that I can try with - I have zero experience with wasm-targeted rust code, but have an idea how to integrate it in the ci system)

• a Changelog entry crediting you! (:

[azriel91]

Heya, quality checkboxes, my favourite ✌️. I'm halfway through writing CI tests, but it needs a change in wasm-pack (rustwasm/wasm-pack#851) so that wasm-pack doesn't try to compile native crates on the wasm32-unknown-unknown target.

[azriel91]

Heya, I got tests running through wasm-pack:

$ wasm-pack test --headless --firefox -- --no-default-features --features "std wasm-bindgen"

Running target/wasm32-unknown-unknown/debug/deps/direct-8375365da32962c5.wasm
Set timeout to 20 seconds...
Running headless tests in Firefox on `http://127.0.0.1:39841/`
Try find `webdriver.json` for configure browser's capabilities:
Not found
running 7 tests                                   

test direct::correct_wait_time ... ok
test direct::all_capacity_check_rejects_excess ... ok
test direct::rejects_too_many_all ... ok
test direct::never_allows_more_than_capacity_all ... ok
test direct::all_1_identical_to_1 ... ok
test direct::rejects_too_many ... ok
test direct::accepts_first_cell ... ok

test result: ok. 7 passed; 0 failed; 0 ignored
# .. more output

Something that happens in the sinks and streams tests is the elapsed time is not strictly greater than the boundary value, so the > assertion sometimes fails. I've set it to be >= in a6c15a5, but I think for true correctness it should be strictly greater than?

Also, the error message from the browser web driver only includes the panic message and no true line number, which is why I copied the assertion operation, so we get:

    error output:
-        panicked at 'elapsed: 200ms', tests/sinks.rs:1:1
+        panicked at 'expected i.elapsed() > Duration::from_millis(200), elapsed: 200ms', tests/sinks.rs:1:1

I haven't updated .circleci config yet, but may call it a day today.

[azriel91]

I've updated circleci config, though I wouldn't know if it works until it's run. Also, there is a way to run tests without wasm-pack, which I've tried and works, so if the wasm-pack PR doesn't have any activity for a while then I may switch the implementation.

[azriel91]

Heya, long break, but the two dependencies above are published. I've taken them in and rebased over master.
The one item still pending is CI, the PR to update wasm-pack is still pending -- I've just refreshed that as well, so hopefully it's accepted.

[antifuchs]

Oooh, that's super exciting - I'll check your changes out this weekend, hopefully we can get wasm support shipped soon!

[antifuchs]

bors try

[bors]

try

Build failed:

• continuous_integration

[antifuchs]

It looks like the wasm tests you added are failing because it doesn't have jq available... this is weird.

Since the circleci config is auto-generated by cargo template-ci (and so will get overwritten unless we perform trickery), maybe it's easier to install a github action to run the tests in a wasm environment? Maybe https://github.com/marketplace/actions/wasm-pack-action could work there?

[azriel91]

Gotcha, shall try that in the next two days ✌️

[azriel91]

Heya, been spending time on getting wasm-pack to pass through features correctly, need to get that working for the download action to pick it up. Shall come back to this when that one is resolved.

[codecov]

Codecov Report

Merging #35 (d4984e5) into master (dbc4ce5) will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #35      +/-   ##
==========================================
+ Coverage   99.35%   99.36%   +0.01%     
==========================================
  Files          30       31       +1     
  Lines        2011     2043      +32     
==========================================
+ Hits         1998     2030      +32     
  Misses         13       13              

Impacted Files	Coverage Δ
src/clock/with_std.rs	100.00% <ø> (ø)
src/gcra.rs	100.00% <ø> (ø)
src/jitter.rs	100.00% <ø> (ø)
src/lib.rs	100.00% <ø> (ø)
tests/direct.rs	100.00% <ø> (ø)
tests/keyed_hashmap.rs	100.00% <ø> (ø)
tests/future.rs	100.00% <100.00%> (ø)
tests/keyed_dashmap.rs	100.00% <100.00%> (ø)
tests/macros.rs	100.00% <100.00%> (ø)
tests/memory_leaks.rs	100.00% <100.00%> (ø)
... and 4 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update dbc4ce5...d4984e5. Read the comment docs.