Cypress: Cypress headless hangs in second spec after previous goes back in history

[dwelle]

Current behavior:

In headless electron (headed electron works fine), the process hangs forever when doing the below steps.

Basically, you do something like this in a.spec.js:

cy.visit(`https://google.com`)
cy.window().then(win => {
  win.history.pushState({}, null, `/`)
});
cy.go(-1);

and in b.spec.js you retrieve an elem:

cy.get('.item');

I know that the 2nd spec file isn't really a valid use case (since we haven't yet loaded a page), but I've wanted to create the smallest reproducible test case.

I'm not really sure what's going on, but there are several significant things that must be done:

1. load some page
2. push a new state to history
3. go back window.location.go(-1)

4. and in the next spec (must be a different file so that Cypress runner closes & opens new window), retrieve a DOM elem.

   This last step is weird because before I managed to narrow it down to this smallest reproduction, I've been doing much more in my real-world spec, and it failed only on one particular DOM retrieval. Don't have a clue, but it also reproduces when I simply query the DOM, without loading a page.

5. hangs, forever (?). DEBUG=cypress:* doesn't show anything interesting, I think --- no new messages are logged after it hangs.

Desired behavior:

Should not hang.

Steps to reproduce: (app code and test code)

https://github.com/dwelle/cypress-test-tiny/tree/dwelle_hang

and run:

npx cypress run --spec cypress/integration/a.spec.js,cypress/integration/b.spec.js

The url links to the branch directly. Refer to README, if need be.

Versions

• cypress 3.5.0 (3.4.1 works)
• Win 7

[jennifer-shehane]

I can confirm, this hangs forever during cypress run (does not happen when --browser chrome. This does not hang in 3.4.1.

I had to manually close the Cypress application to get it to stop running.

DEBUG logs at hang

 cypress:server:project received runnables { id: 'r1', title: '', root: true, type: 'suite', tests: [ { id: 'r2', title: 'history get item', body: "function () {\n  cy.get('.item');\n}", type: 'test' } ], suites: [] } +813ms
  cypress:server:reporter trying to load reporter: spec +2s
  cypress:server:reporter spec is Mocha reporter +0ms
  cypress:server:project onMocha start +5ms
  cypress:server:reporter got mocha event 'start' with args: [ { start: '2019-10-28T16:27:19.404Z' } ] +5ms

  cypress:server:project onMocha suite +1ms
  cypress:server:reporter got mocha event 'suite' with args: [ { id: 'r1', title: '', root: true, type: 'suite' } ] +1ms

  cypress:server:project onMocha test +24ms
  cypress:server:reporter got mocha event 'test' with args: [ { id: 'r2', title: 'history get item', body: "function () {\n  cy.get('.item');\n}", type: 'test' } ] +24ms
  cypress:server:socket automation:request get:cookies { domain: 'localhost' } +63ms
  cypress:server:cookies getting:cookies { domain: 'localhost' } +5s
  cypress:server:cookies received get:cookies [] +1ms
cypress:browserify finished bundling: /Users/jennifer/Library/Application Support/Cypress/cy/production/projects/cypress-transform-test-74d5161f4a1005f86b0d85a657ddcf07/bundles/cypress/integration/spec2.js +16ms
  cypress:server:controllers:spec sending spec { filePath: '/Users/jennifer/Library/Application Support/Cypress/cy/production/projects/cypress-transform-test-74d5161f4a1005f86b0d85a657ddcf07/bundles/cypress/integration/spec2.js' } +10ms
  cypress:server:project received runnables { id: 'r1', title: '', root: true, type: 'suite', tests: [ { id: 'r2', title: 'history get item', body: "function () {\n  cy.get('.item');\n}", type: 'test' } ], suites: [] } +813ms
  cypress:server:reporter trying to load reporter: spec +2s
  cypress:server:reporter spec is Mocha reporter +0ms
  cypress:server:project onMocha start +5ms
  cypress:server:reporter got mocha event 'start' with args: [ { start: '2019-10-28T16:27:19.404Z' } ] +5ms

  cypress:server:project onMocha suite +1ms
  cypress:server:reporter got mocha

specfile1.js

it('history go', () => {
  cy.visit(`https://google.com`)
  cy.window().then(win => {
    win.history.pushState({}, null, `/`)
  });
  cy.go('back');
})

specfile2.js

it('history get item', () => {
  cy.get('.item')
})

Workaround

Run tests in cypress run with the --browser flag specifying an installed version of Chrome.

[psoares-resilient]

Hi, just updated to the 3.5.0 as well and the same behaviour happens.

 Cypress:    3.5.0                                                                              │
  │ Browser:    Electron 73                                                                          │
  │ Specs:      6 found (data/investigate.spec.ts, data/missing-numbers.spec.ts, data/pagination.s │
  │             pec.ts, data/search.spec.ts, data/sort.spec.ts, data/text-filter.spec.ts)          │
  │ Searched:   cypress/integration/data/* 

In our case, it is always hanging after the first it in the second matched spec (whatever order it may be). Switching to chrome ( 72) also let the tests continue.

[dwelle]

Can we get an update on this? It's a blocking regression that prevents from upgrading from 3.4.1 (I don't know of any workaround).

[cormack00]

Is there any chance this will be fixed in 3.6.1? Causing issues here too.

[jinder]

I think this is a broader issue than the comment suggests - HTML5 navigation in general seems to cause Cypress to hang.

[jennifer-shehane]

Maybe related: https://github.com/cypress-io/cypress/issues/5645

[jennifer-shehane]

Also likely related: https://github.com/cypress-io/cypress/issues/5596

[brian-mann]

I spent a few hours debugging this - I believe it is a bug in electron core. I have tried dozens of different possible workarounds involving clearing session, cache, partitions, etc and nothing is working as of yet.

[brian-mann]

Confirmed this is a bug in electron. It is still failing in 6.1.5 but it is working in 7.1.2.

Looks like we can fix this by upgrading electron.

[cypress-bot[bot]]

The code for this is done in cypress-io/cypress#5849, but has yet to be released.
We'll update this issue and reference the changelog when it's released.

[jennifer-shehane]

Released in 3.8.0.

[raghugitrepo]

@jennifer-shehane sorry for commenting on closed ticket. I am kind of facing similar issue running tests in docker. After the first test spec cypress hangs on forever.
I am using "cypress/browsers:node13.6.0-chrome80-ff72" image. Also pasting the logs of cypress in debug mode.

[34;1mcypress:server:browsers:chrome connecting to Chrome remote interface at random port 44619 +15ms
cypress:server:protocol Getting WS connection to CRI on port 44619 +2s
cypress:launcher chrome stderr: [2265:2280:0423/035613.707142:ERROR:bus.cc(393)] Failed to connect to the bus: Failed to connect to socket /var/run/dbus/system_bus_socket: No such file or directory +148ms
cypress:network:connect received error on connect, retrying { iteration: 0, delay: 100, err: Error: connect ECONNREFUSED 127.0.0.1:44619  at TCPConnectWrap.afterConnect [as oncomplete] (net.js:1128:14) { errno: 'ECONNREFUSED', code: 'ECONNREFUSED', syscall: 'connect', address: '127.0.0.1', port: 44619 } } +260ms
cypress:launcher chrome stderr: [2265:2265:0423/035613.708675:ERROR:edid_parser.cc(102)] Too short EDID data: manufacturer id +1ms
cypress:launcher chrome stderr: [2265:2289:0423/035613.731195:ERROR:bus.cc(393)] Failed to connect to the bus: Address does not contain a colon +23ms
cypress:launcher chrome stderr: DevTools listening on ws://127.0.0.1:44619/devtools/browser/bf00290b-61c9-4229-808a-d016583652d7 +7ms
cypress:launcher chrome stderr: [2265:2289:0423/035613.741804:ERROR:bus.cc(393)] Failed to connect to the bus: Address does not contain a colon
[2265:2289:0423/035613.741826:ERROR:bus.cc(393)] Failed to connect to the bus: Address does not contain a colon +4ms
cypress:server:server Got CONNECT request from www.gstatic.com:443 +234ms

[ycwdaaaa]

@jennifer-shehane sorry for commenting on closed ticket. I am kind of facing similar issue running tests in docker. After the first test spec cypress hangs on forever.
I am using "cypress/browsers:node13.6.0-chrome80-ff72" image. Also pasting the logs of cypress in debug mode.

[34;1mcypress:server:browsers:chrome �[0mconnecting to Chrome remote interface at random port 44619 �[34m+15ms�[0m
�[33;1mcypress:server:protocol �[0mGetting WS connection to CRI on port 44619 �[33m+2s�[0m
�[31;1mcypress:launcher �[0mchrome stderr: [2265:2280:0423/035613.707142:ERROR:bus.cc(393)] Failed to connect to the bus: Failed to connect to socket /var/run/dbus/system_bus_socket: No such file or directory �[31m+148ms�[0m
�[32;1mcypress:network:connect �[0mreceived error on connect, retrying { iteration: �[33m0�[39m, delay: �[33m100�[39m, err: Error: connect ECONNREFUSED 127.0.0.1:44619 �[90m at TCPConnectWrap.afterConnect [as oncomplete] (net.js:1128:14)�[39m { errno: �[32m'ECONNREFUSED'�[39m, code: �[32m'ECONNREFUSED'�[39m, syscall: �[32m'connect'�[39m, address: �[32m'127.0.0.1'�[39m, port: �[33m44619�[39m } } �[32m+260ms�[0m
�[31;1mcypress:launcher �[0mchrome stderr: [2265:2265:0423/035613.708675:ERROR:edid_parser.cc(102)] Too short EDID data: manufacturer id �[31m+1ms�[0m
�[31;1mcypress:launcher �[0mchrome stderr: [2265:2289:0423/035613.731195:ERROR:bus.cc(393)] Failed to connect to the bus: Address does not contain a colon �[31m+23ms�[0m
�[31;1mcypress:launcher �[0mchrome stderr: DevTools listening on ws://127.0.0.1:44619/devtools/browser/bf00290b-61c9-4229-808a-d016583652d7 �[31m+7ms�[0m
�[31;1mcypress:launcher �[0mchrome stderr: [2265:2289:0423/035613.741804:ERROR:bus.cc(393)] Failed to connect to the bus: Address does not contain a colon
[2265:2289:0423/035613.741826:ERROR:bus.cc(393)] Failed to connect to the bus: Address does not contain a colon �[31m+4ms�[0m
�[31;1mcypress:server:server �[0mGot CONNECT request from www.gstatic.com:443 �[31m+234ms�[0m

I meet the same error, is there any update?

[jennifer-shehane]

This issue will be closed to further comment as the exact issue here was resolved and tested.

If you're experiencing a bug similar to this in Cypress, please open a new issue with a fully reproducible example that we can run. There may be a specific edge case with the issue that we need more detail to fix.