Has anyone got the tests to work?

[TimJentzsch]

When using cargo run --package gentest to run the tests, I get the following error:

   Compiling gentest v0.1.0 (/home/tim/Documents/stretch/scripts/gentest)
    Finished dev [unoptimized + debuginfo] target(s) in 2.16s
     Running `target/debug/gentest`
Starting ChromeDriver 101.0.4951.41 (93c720db8323b3ec10d056025ab95c23a31997c9-refs/branch-heads/4951@{#904}) on port 4444
Only local connections are allowed.
Please see https://chromedriver.chromium.org/security-considerations for suggestions on keeping ChromeDriver safe.
ChromeDriver was started successfully.
thread 'tokio-runtime-worker-4' panicked at 'unexpected webdriver error; webdriver returned error: unknown error: DevToolsActivePort file doesn't exist', /home/tim/.cargo/registry/src/github.com-1ecc6299db9ec823/fantoccini-0.11.9/src/session.rs:328:17
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
thread 'tokio-runtime-worker-3' panicked at 'called `Option::unwrap()` on a `None` value', /home/tim/.cargo/registry/src/github.com-1ecc6299db9ec823/fantoccini-0.11.9/src/session.rs:282:68

Did anyone get this to work and might be able to help me out?

[TimJentzsch]

Docs for reference: https://github.com/DioxusLabs/stretch#testing

[alice-i-cecile]

For future readers, this is discussed further in #45.

[TimJentzsch]

This is actually a different issue, #45 is about installing chromedriver, which is already completed in this one (see the Starting ChromeDriver... line in the log).

I wasn't able to resolve this particular issue yet.

[CobaltCause]

I've hit a different error while trying to use this tool:

$ cargo run -p gentest
    Finished dev [unoptimized + debuginfo] target(s) in 0.09s
     Running `target/debug/gentest`
Starting ChromeDriver 101.0.4951.41 (93c720db8323b3ec10d056025ab95c23a31997c9-refs/branch-heads/4951@{#904}) on port 4444
Only local connections are allowed.
Please see https://chromedriver.chromium.org/security-considerations for suggestions on keeping ChromeDriver safe.
ChromeDriver was started successfully.
░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░ 1/238
thread 'tokio-runtime-worker-1' panicked at 'called `Option::unwrap()` on a `None` value', scripts/gentest/src/main.rs:173:61

Extra things I needed to install were chromedriver, chromium, pkg-config, and openssl; and then I was able to build the package but running it produced the above error.

[NickAcPT]

This isn't a direct answer to this - although I've tried generating the tests but couldn't also get it to work.

While working on my Kotlin port of this same library, I decided to automatically generate the tests myself, based on those same Yoga test files, but while doing so, I came across some inconsistencies with the results given by Sprawl and the ones given by a Chrome browser.

For full transparency, here's the Chrome version I am using:

Chrome is up to date
Version 102.0.5005.63 (Official Build) (64-bit)

A test that shows me these kinds of inconsistencies is the min_height test.

Comparison of results in min_height

Just to make my point (and suffering) clear, here's my comparison of the assertions vs what Chrome tells me.

Root element

assert_eq!(sprawl.layout(node).unwrap().size.width, 100f32);
assert_eq!(sprawl.layout(node).unwrap().size.height, 100f32);
assert_eq!(sprawl.layout(node).unwrap().location.x, 0f32);
assert_eq!(sprawl.layout(node).unwrap().location.y, 0f32);

Clearly the root is 100x100 and at 0x0, this checks out.

Root's first child (node 0)

assert_eq!(sprawl.layout(node0).unwrap().size.width, 100f32);
assert_eq!(sprawl.layout(node0).unwrap().size.height, 80f32);
assert_eq!(sprawl.layout(node0).unwrap().location.x, 0f32);
assert_eq!(sprawl.layout(node0).unwrap().location.y, 0f32);

This is.. odd? Sprawl is giving the wrong value here, since the test passes, when compared to what Chrome returns.

Conclusion

I think it's that time of the year to fix the test generation and regenerate those tests and make sure we're still up-to-spec.

Please someone help, I've spent 3 hours yesterday trying to debug this issue.

[alice-i-cecile]

I think it's that time of the year to fix the test generation and regenerate those tests and make sure we're still up-to-spec.

Agreed. I don't love using Chrome as a source of truth, but it's not a terrible plan.

[NickAcPT]

I think it's that time of the year to fix the test generation and regenerate those tests and make sure we're still up-to-spec.

Agreed. I don't love using Chrome as a source of truth, but it's not a terrible plan.

Well, apparently that's already being done from what I can see. It just appears to be out-of-date.

[alice-i-cecile]

Yes; it's just that this is an inherited library (from the dead stretch crate). There are quite a few design choices that I feel... differently about.

[NickAcPT]

Yes; it's just that this is an inherited library (from the dead stretch crate). There are quite a few design choices that I feel... differently about.

Well, fair. Now it's just a matter of whether or not it's worth the effort to change some of those design choices and refactor stuff.

[alice-i-cecile]

Yep. Next course of action will be to add unit tests, for the refactored layout algorithm in #47, and then sit down and decide on our overall test strategy.

[Weibye]

Update on this:

@mockersf has helpfully fixed the gentest now in #145, and I have improved the error message in #147

As a policy: Whenever making changes / refactors that affects these, please run the gentest at the end then commit them separately so that the PR can be reviewed by commit without all the noise.