CI: run tests on i686 and wasm32-wasi

[pitdicker]

Fixes #352.

[codecov]

Codecov Report

Merging #1237 (34ca0b4) into 0.4.x (243d26d) will decrease coverage by 0.28%.
Report is 1 commits behind head on 0.4.x.
The diff coverage is 66.66%.

@@            Coverage Diff             @@
##            0.4.x    #1237      +/-   ##
==========================================
- Coverage   86.10%   85.82%   -0.28%     
==========================================
  Files          37       37              
  Lines       13522    13517       -5     
==========================================
- Hits        11643    11601      -42     
- Misses       1879     1916      +37     

Files Changed	Coverage Δ
src/datetime/tests.rs	96.53% <ø> (ø)
src/offset/local/tz_info/timezone.rs	85.69% <66.66%> (-0.14%)	⬇️

... and 3 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[djc]

Nit: I'd like the comments to go before the attributes, keeping the attributes right next to the code.

[djc]

Does this also fix #1236?

[pitdicker]

No. I'll reply to that issue.

[pitdicker]

In the end I opted for using cross instead of installing the cross-compiling toolchains in Ubuntu on the CI.

This hit another problem: the test_time_zone_from_posix_tz test expects there to always be a timezone database available on Unix, but that is not true in the docker containers used by cross.

@djc Do you want to give it a second look?

[pitdicker]

I have no clue why codecov thinks our coverage in local/unix/timezone.rs has massively changed from 85.69% to 66.67%, while at the same time saying the change in that file is -0.14%.

For now codecov seems mostly trouble, but maybe we get some use out of it eventually...

[djc]

LGTM, thanks!