#19576 introduced a regression where the CLI reporters displayed some
times with way too many decimals. e.g. 7.123456789ms.
Prior to #19576, there were two monotonicTime implementations; that PR
updated the reporters to use the common definition that had existed in
utils.ts. However, that introduced a regression in the base.ts reporters
which used the ms duration humanizing package which did not account for
the more precise decimals used by the shared monotonicTime function.
This fix removes the dependency on the third-party ms package and now
consistently uses Pavel's humanize function which the HTML reporter had
been using and proved to have better defaults for decimals.
Additionally, we add more test coverage to limit future regressions
since this was caught in passing.
Closes#19556.
Relates #19576.
Note: this keeps existing behavior of `undefined` in `test.use()`
reverting to the config value and not to the original default value.
References #19615.
- When disposing recursively, only the root dispatcher received
`_dispose()` call, while some dispatchers need `_onDispose()` to clean
things up.
- When reusing the context, pages should be notified with `_onClose()`
so that all client-side waiting promises could reject.
Fixes#19216.
<img width="1676" alt="Screenshot 2022-12-14 at 2 05 29 PM"
src="https://user-images.githubusercontent.com/11915034/207726274-167d9ee7-c01b-4379-aff1-37dad979d23b.png">
Flakiness dashboard reveals on Window's variants of Chromium, the
decimal for actual vs. expected was slightly off (e.g. 1705568002.677066
vs. 1705568002.674173).
This test fix assumes it's acceptable to suppress the roundtrip floating
point difference (which appears to be constrained to Windows).
Signed-off-by: Ross Wollman <rwoll@users.noreply.github.com>
- `<loopback>` for local interfaces;
- `123.123.123.123` for IPv4;
- `[1:2:3:4:5:6:7:8]` for IPv6;
- `*.example.com` and `.example.com` for subdomains;
- `example.com` for domains;
- `anything:3000` for port matching.
This patch adds a reverse proxy in front of novnc and playwright
server inside the container.
As a result:
- container exposes a single HTTP port to the host
- all Playwright clients can connect using this exposed port, e.g.
`http://127.0.0.1:5400`
- navigating to the exposed port in the browser lands on a nice HTML
page
with a link to view container screen
- users can also manually navigate to `http://127.0.0.1:5400/screen` to
view screen
This patch adds a grid border around the image so that the SSIM
resolution doesn't drop for the border pixels.
We also add a test with WebKit rendering artifacts to make sure
new approach helps to fight this.
While migrating Playwright tests from pixelmatch comparator to the
ssim-cie94 comparator, we noticed a rendering artifact that happens
on playwright screenshot tests on WebKit Linux.
This PR adds a set of cropped-down examples and tunes the ssim-cie94
comparator to accept these rendering artifacts.
This patch implements a new image comparison function, codenamed
"ssim-cie94". The goal of the new comparison function is to cancel out
browser non-determenistic rendering.
To use the new comparison function:
```ts
await expect(page).toHaveScreenshot({
comparator: 'ssim-cie94',
});
```
As of Nov 30, 2022, we identified the following sources of
non-determenistic rendering for Chromium:
- Anti-aliasing for certain shapes might be different due to the
way skia rasterizes certain shapes.
- Color blending might be different on `x86` and `aarch64`
architectures.
The new function employs a few heuristics to fight these
differences.
Consider two non-equal image pixels `(r1, g1, b1)` and `(r2, g2, b2)`:
1. If the [CIE94] metric is less then 1.0, then we consider these pixels
**EQUAL**. (The value `1.0` is the [just-noticeable difference] for
[CIE94].). Otherwise, proceed to next step.
1. If all the 8 neighbors of the first pixel match its color, or
if the 8 neighbors of the second pixel match its color, then these
pixels are **DIFFERENT**. (In case of anti-aliasing, some of the
direct neighbors have to be blended up or down.) Otherwise, proceed
to next step.
1. If SSIM in some locality around the different pixels is more than
0.99, then consider this pixels to be **EQUAL**. Otherwise, mark them
as **DIFFERENT**. (Local SSIM for anti-aliased pixels turns out to be
very close to 1.0).
[CIE94]: https://en.wikipedia.org/wiki/Color_difference#CIE94
[just-noticeable difference]:
https://en.wikipedia.org/wiki/Just-noticeable_difference
This prevents shared workers from stalling upon restart.
We receive `Inspector.targetCrashed` and
`Inspector.targetReloadedAfterCrash` events that assume
`Runtime.runIfWaitingForDebugger` from any attached client. It is easier
and more stable to just detach from shared workers, because we do not
inspect them.
For service workers, we should actually issue
`Runtime.runIfWaitingForDebugger` in such cases, because we attach to
them.
Fixes#18932.
The behavior regarding filters (both in config, command line and .only)
is the following:
- if some of tests match and none of setup match then we'll run all
setup files and all matching tests
- otherwise the filters apply to setup files the same way as to regular
tests
- Properly convert coordinates for iframes with non-zero borders.
- IFrames that have `transform` anywhere in the ancestors skip
`hitPoint`-based check because we cannot reliably translate the viewport
point into frame document's coordinates.
Fixes#18245.
Running `npx playwright test file:123` will have the following behavior
- if only test files match then only matching subset of tests will run
but all setup files will run as well
- if only setup files match the filter then only those setup tests will
run
- if both setup and test files match an error will be thrown
Following the `getByText()` and other methods:
- By default, matching is substring and case-insensitive. Before, it was
only case-insensitive, but not substring.
- With new option `exact: true`, matching is full string and
case-sensitive.
- Matching always normalizes whitespace.
- Codegen generates `exact: false` by default.
- `internal:role` treats `[name="foo"i]` as non-exact match.
Various fixes:
- Updated `getByRole` docs to match the reality.
- Locator generator edge cases.
This patch:
- updates documentation to lead users from `TestConfig.snapshotDir` and
`testInfo.snapshotSuffix` to `TestConfig.snapshotPathTemplate` as a
better and more flexible alternative.
- drops `{snapshotSuffix}` from documentation
- stops using `snapshotSuffix = ''` in our own tests and switches us
to the `snapshotPathTemplate`.
- adds `{testName}` token.
- Use `snapshotPathTemplate` for docker screenshots in html-reporter
- Mark the snapshot path template test as slow since it re-spawns
worker for each project.
- Fix docker smoke tests
- prefer `role=checkbox` over `input[type=checkbox]`
- prefer `#id` over `input[type=checkbox]` and `role=checkbox`
- prefer `text=foo` over `internal:has-text=foo`
- ignore `none` and `presentation` roles
- remove non-strict support
This configuration option allows to set a string with template
values for precise control over snapshot path location.
An example of `snapshotPathTemplate` usage:
```ts
// playwright.config.ts
// Notice the `testDir` configuration!
export default {
testDir: './tests',
snapshotPathTemplate: './__screenshots__/{platform}/{projectName}/{testFilePath}/{arg}{ext}',
}
```
Currently supported "magic tokens" inside the `snapshotPathTemplate`
are:
- `{testDir}` - project's `testDir`
- `{snapshotDir}` - project's `snapshotDir`
- `{platform}` - `process.platform`
- `{projectName}` - Project's sanitized name
- `{testFileDir}` - Directories in relative path from `testDir` to test
file path (e.g. `page/` in the example below)
- `{testFileName}` - Test file name (with extension) (e.g.
`page-click.spec.ts` in the example below)
- `{testFilePath}` - Relative path from `testDir` to test file path
(e.g. `page/page-click.spec.ts` in the example below)
- `{ext}` - snapshot extension (with dots)
- `{arg}` - joined snapshot name parts, without extension (e.g.
`foo/bar/baz` in the example below)
- `{snapshotSuffix}` - `testInfo.snapshotSuffix` value.
Consider the following file structure:
```
playwright.config.ts
tests/
└── page/
└── page-click.spec.ts
```
The following `page-click.spec.ts`:
```ts
// page-click.spec.ts
import { test, expect } from '@playwright/test';
test('should work', async ({ page }) => {
await expect(page).toHaveScreenshot(['foo', 'bar', 'baz.png']);
});
```
Fixes#7792
- Separate worker for some Android tests.
- Use png comparator for some screenshots tests instead of buffer
equality.
- Skip drag&drop tests on Android.
- Various timeout fixes.
Per discussion in #12950, adding sanitized name to the output filename
prefix. This can make debugging easier, and the filename structure more
human friendly.
This patch implements a new mode of network tethering for Playwright
server & its clients.
With this patch:
- playwright server could be launched with the
`--browser-proxy-mode=tether` flag to engage in the new mode
- a new type of client, "Network Tethering Client" can connect to the
server to provide network traffic to the browsers
- all clients that connect to the server with the `x-playwright-proxy:
*` header will get traffic from the "Network Tethering Client"
This patch also adds an environment variable
`PW_OWNED_BY_TETHER_CLIENT`. With this env, playwright server will
auto-close when the network tethering client disconnects. It will also
auto-close if the network client does not connect to the server in the
first 10 seconds of the server existence. This way we can ensure that
`npx playwright docker start` blocks terminal & controls the lifetime of
the started container.
Previously, we considered root when selector has `:scope` modifier, but
did not actually match it with other css specifiers, like in
`:scope.selected`.
Fixes#17824.
We have to reliably know whether test timed out or not, and soft expect
error could have marked it with `status=failed` but it would still time
out. Now we have a separate `_didTimeout` flag for this.
Fixes#18023.
