Commit Graph

3247 Commits

Author SHA1 Message Date
Andrey Lushnikov
08024c82e8
feat(firefox): roll to r1188 (#4091)
This brings in a new Firefox 82 beta.

Fixes #3995
2020-10-07 17:46:10 -07:00
Andrey Lushnikov
6a7d24469a
devops: fix firefox build (#4088)
New firefox build requires a pre-created python virtual environment.
We should detect it and create if necessary.

References #3995
2020-10-07 14:34:58 -07:00
Andrey Lushnikov
1ccce09a02
browser(firefox): roll Firefox to beta Oct 7, 2020 (#4087)
References #3995
2020-10-07 14:12:19 -07:00
Yury Semikhatsky
09906949c2
docs(page): clarify page.close({runBeforeUnload: true}) behavior (#4086) 2020-10-07 13:27:23 -07:00
Andrey Lushnikov
6bb212ceed
feat(firefox): roll Firefox to r1186 (#4085)
This roll includes two important fixes:
- await search service startup when launching browser
- report `"load"` event differently to support `window.stop()` in future

References #3995
2020-10-07 13:24:58 -07:00
Max Schmitt
fd769ec911
chore: remove test dependency on pw itself (#4078) 2020-10-07 11:08:30 -07:00
Andrey Lushnikov
6beef551d9
browser(firefox): wait for search and addon manager initialization (#4081)
It looks like terminating browser when search service or addon manager is
not fully initialized results in a broken shutdown sequence. As of
today, this results in multiple errors in the browser STDERR. In future,
this might also result in browser stalling instead of terminating.

This starts awaiting search and addon manager termination.

References #3995
2020-10-07 10:24:16 -07:00
Andrey Lushnikov
3b4232864d
browser(firefox): a different way to emit 'load' event (#4080)
Using WebProgressListener events works in all cases. Currently
used `pageshow` event will stop being emitted in future when loading
was stopped with `window.stop()` api.

References #3995
2020-10-07 09:22:56 -07:00
Arjun Attam
aafe5dac0d
docs: add videos to verification doc (#4071)
* docs: add videos to verification doc

* review changes
2020-10-07 09:04:53 -07:00
Pavel Feldman
ad58e49201
Revert "feat(firefox): migrate to the pipe channel (#4068)" (#4073)
Mac sporadically hangs on browser close.
2020-10-06 21:16:50 -07:00
Pavel Feldman
1fe3c783b4
test: roll test runner 0.9.22 (#4072) 2020-10-06 15:51:18 -07:00
Pavel Feldman
ce7aa7a608
feat(firefox): migrate to the pipe channel (#4068) 2020-10-06 14:40:09 -07:00
Andrey Lushnikov
e6869edf98
browser(firefox): fix typo in dispatcher teardown (#4069) 2020-10-06 12:06:50 -07:00
Andrey Lushnikov
5d1291528f
feat(firefox): roll to r1182 (#4067)
This revision enables `documentchannel` feature by default.

References #3995
2020-10-06 03:03:33 -07:00
Andrey Lushnikov
4ab66a4fe5
browser(firefox): follow-up with assorted simplifications (#4066)
This patch:
- moves `SimpleChannel` to synchronously dispatch buffered commands
  instead of a `await Promise.resolve()` hack
- moves dialog & screencast handling from `PageHandler` to
  `TargetManager`. This leaves `PageHandler` to be concerned solely about
  protocol.
- removes `attach` and `detach` methods for worker channels: since
  channels are buffering messages until the namespace registers, there's
  no chance to loose any events.
- slightly simplifies `PageNetwork` class: it's lifetime is now
  identical to the lifetime of the associated `PageTarget`, so a lot can
  be simplified later on.

References #3995
2020-10-06 01:53:25 -07:00
Andrey Lushnikov
c8a64b88e1
browser(firefox): enable document channel (#4065)
In the current tip-of-tree Firefox, document channel is enabled by
default, so we have to enable it in order to roll further.

This patch:
1. Removes content disposition sniffing from content process since it
   crashes renderer with document channel.
2. Merges all page-related handlers in a single `PageHandler` and
   serializes network events wrt the `Page.frameAttached` event.

The serialization mentioned in (2) is necessary: frame attachment is
reported from the content process, and network events are reported from
the browsers process. This is an inherent race, that becomes exposed by
the document channel.

On a side note, (2) makes it possible to synchronously report all
buffered events in `SimpleChannel` (cc offline discussion with @dgozman
that highlighted an unsighty approach that we currently employ there: reporting
events in a subsequent microtask.)

References #3995
2020-10-06 00:15:24 -07:00
Dmitry Gozman
e403fd3912
test: update fixtures to new syntax (#4063)
- Some simplifications around defineParameter.
- overrideTestFixtures must be chained as well.
2020-10-05 21:40:56 -07:00
Pavel Feldman
0db09f8ed4
test: roll test runner to 0.9.20 (#4062) 2020-10-05 17:03:24 -07:00
Thomas Chaplin
2df6425254
chore(typo): resolve typo in src/progress.ts (#4041) 2020-10-05 15:21:35 -07:00
Pavel Feldman
857abcfc10
browser(firefox): make pipe work on Windows (#4058) 2020-10-05 10:58:56 -07:00
Dmitry Gozman
a7beaf657c
browser(chromium): roll to 813607 (#4052) 2020-10-05 04:23:35 -07:00
Dmitry Gozman
d31cbc21e5
fix(video): wait for videos when closing persistent context (#4040)
To achieve this, we close all the pages one by one, then wait
for the videos to finish processing, and then close the browser.
2020-10-04 18:18:05 -07:00
Yury Semikhatsky
fbe0fb2977
fix(api.json): do not copy documentation from base class members (#4048) 2020-10-02 19:19:19 -07:00
Dmitry Gozman
e214f795e0
feat(video): support videos in remote browser (#4042) 2020-10-02 17:27:56 -07:00
Dmitry Gozman
133de10a47
browser(firefox): start screencast in existing pages upon setScreencastOptions (#4045)
This makes it work in persistent context.

To achieve this, we have to move screencast logic into PageTarget and
make PageHandler listen to an event.
2020-10-02 17:16:49 -07:00
Pavel Feldman
1f16ce2632
test: remove failing expectations from a passing test (#4036) 2020-10-02 11:01:04 -07:00
Andrey Lushnikov
ec36eef3a2
feat(firefox): roll Firefox to r1179 (#4044)
References #3995
2020-10-02 09:27:24 -07:00
Andrey Lushnikov
2c11b10598
browser(firefox): remove multisession logic (#4039)
This patch:
1. Changes `SimpleChannel` to buffer messages to the namespace that
   hasn't been registered yet. This allows us to create `SimpleChannel`
   per target on the browser side right away.
2. Removes multisession support. Now there's only one `PageAgent` in the
   content process, which talks to a single `PageHandler` on the browser
   side. Both ends can be created as-soon-as-needed; thanks to
   `SimpleChannel` bufferring, no messages will be lost and all messages
   will be delivered in proper order. (This is currently the reason why
   build 1178 flakes on windows).
3. Straightens up the target reporting. Targets are reported as soon
   as they appear on the browser side.

**NOTE:** this doesn't yet remove sessions from protocol.

References #3995
2020-10-02 04:13:42 -07:00
Dmitry Gozman
5e42029fce
api: allow exposeBinding to pass handles (#4030)
This adds an option `{ handle: true }` to pass a single handle instead of arbitrary json values.
2020-10-01 22:47:31 -07:00
Tom Jenkinson
c2171218fa
test: add a test for request interception with redirects (#3994) 2020-10-01 19:55:39 -07:00
Pavel Feldman
81c1daed73
test: roll test runner 0.9.17 (#4035) 2020-10-01 13:32:52 -07:00
Max Schmitt
318ab28171
chore: include api.json in NPM package (#4034) 2020-10-01 12:47:16 -07:00
Dmitry Gozman
4a77363af7
api: update videos api, hide tracing (#4015) 2020-10-01 11:06:19 -07:00
Dmitry Gozman
920cc7c886
test: run crash tests with chromium wire (#4026) 2020-10-01 05:30:27 -07:00
Andrey Lushnikov
b74a6b78ef
browser(firefox): do not double-attach session to the same target (#4027)
We currently might double-attach to the target in `BrowserHandler` since we iterate over all targets, and then subscribe to the additional event when target is getting initialized.

This patch fixes this race condition and should unblock the roll to r1177.

References #3995
2020-09-30 23:50:02 -07:00
Dmitry Gozman
f885d07cb9
fix(close): fix a race during context.close and page.close (#4018)
There is a race between "close" event coming from the server and
"close" command issued from the client.

This is similar to calling close after disconnect, so added tests.
2020-09-30 21:17:30 -07:00
Andrey Lushnikov
b9dcfb9909
test: minor fixes (#4024)
- missing await
- missing `defaultBrowserOptions` that's important to respect `FFPATH` (and other custom browsers)
2020-09-30 17:23:13 -07:00
Dmitry Gozman
13d48da84a
test: record trace/videos on retries (#4009) 2020-09-30 16:52:21 -07:00
Andrey Lushnikov
974358442d
browser(firefox): move user agent emulation to browser side (#4016)
References #3995
2020-09-30 08:02:22 -07:00
Dmitry Gozman
ab2714ed29
test: unflake click-timeout-4 (#4012)
We used to animate for 5 seconds, and also timeout after 5 seconds.
Presumably, this made the click work sometimes due to different timeout
schedulers in node vs browser.
2020-09-30 04:42:08 -07:00
Dmitry Gozman
ccc827cd87
test: add a test for screen.avail{Width,Height} emulation (#4011)
This currently does not work in WebKit.
2020-09-30 04:14:29 -07:00
Andrey Lushnikov
e28083974d
browser(firefox): simplify PageTarget lifecycle (#4014)
As of today, we create `PageTarget` instances whenever we get a
sync IPC from the content process. This, however, breaks an invariant
that `browserContext.pages` always has all pages (and *browsing contexts* - not to be confused with *browser contexts*), associated with browser context. This invariant will be especially important when we move
user agent emulation to browser-side.

This patch makes `PageTarget` lifecycle symmetrical:
- `PageTarget` instance is created when tab is opened
- `PageTarget` is destroyed when tab is crashed or closed

This should also fix a bunch of race conditions with persistent mode, since sometimes we arrive to the window after its
initialization.

Drive-by: straighten viewport management and put a nice descriptive comment.
2020-09-30 02:10:34 -07:00
Andrey Lushnikov
24bc0e39b8
browser(firefox): remove the hack around setting viewport size (#4010)
Juggler code had a bug where we subscribed to window and tab
events, but did not iterate collections of current windows and tabs.

As a result, we were sometimes failing to set viewport size for the
initial window, and implemented an artificial promise to workaround
the problem.

This patch:
- starts calling `onOpenWindow` and `onOpenTabListener` callbacks
  for *all* windows and tabs - current and future, eliminating the
  race condition.

This worked too well and we started overriding window sizes that
were set by users with `window.open(url, 'width=300;height=400')` (we
have a test for this). To fix this, we now plumb `CHROME_WITH_SIZE`
flag from appWindow and override viewport iff this flag is not set.

After this patch, we will use the `onTabOpened` event to move user
agent emulation to the browser-side.

References #3995
2020-09-30 00:36:46 -07:00
Andrey Lushnikov
a20c0e0adb
roll(firefox): roll Firefox to r1174 (#4005)
The r1174 starts using cross-process frameIds. This
simplifies Juggler a lot, but regresses a rare usecase:
network requests from workers from subframes are attributed
to main frame.

This adds a test to annotate this regression. Note that this
doesn't also work in Chromium.

References #3995
2020-09-29 21:07:25 -07:00
Pavel Feldman
d658b687ca
chore: refactor screencast tests (#4007) 2020-09-29 18:52:30 -07:00
Pavel Feldman
20b83ee0b4
fix(electron): do not use --require to throttle startup (#4006) 2020-09-29 18:01:14 -07:00
Pavel Feldman
6dccd2735a
fix(wire): fix the wire mode (#4008) 2020-09-29 18:00:56 -07:00
Pavel Feldman
7ccdc5176d
chore: include api.json into the Playwright package (#4003) 2020-09-29 13:48:24 -07:00
Pavel Feldman
c30b894f27
chore: add missing image test expectation (#4004) 2020-09-29 13:47:57 -07:00
Pavel Feldman
de1e63dff1
test: roll test runner 0.9.16 (#3998) 2020-09-29 11:51:00 -07:00