Don't set readyState on connection error #4562
Conversation
|
Please also check this into 0.61 |
I planned to do a full backport after the devmain submission happened and stabilized (checkpoint released), to avoid doing continuous fixes in three branches. Is there a motive to do this sooner? |
|
If we're checking into both 0.60 and master at the same time, there's risk we forget to get this into 0.61 and see failures when upgrading. I'd personally be a bit less anxious if we kept the contents of the branches in sync. |
While I'd argue I'll NOT forget about such a task, the request is fair. |
JunielKatarn
added
the
AutoMerge
|
Hello @JunielKatarn! Because this pull request has the Do note that I've been instructed to only help merge pull requests of this repository that have been opened for at least 60 minutes, a condition that will be fulfilled in about 41 minutes. No worries though, I will be back when the time is right! 😉 p.s. you can customize the way I help with merging this pull request, such as holding this pull request until a specific person approves. Simply @mention me (
|
|
The alternative to checking into all branches at the same time would be to just make these in master and backport after they're checked in. That might be the easiest, and is the general practice. I would honestly reccomend that method to avoid extra work, forked review comments, etc. I specifically just worry about having it before and after a version. Branch management has already been a bit messy, so any standard process that eliminates the possibility of error can help. |
Agreed. Ideally we should submit to master, wait for approval and then backport. In the rare case where we need to push to a stable branch at the same time (and this has to be a very rare case that is time critical), we should push to all supported branches at the same time. |
ghost
merged commit 
* Don't set readyState on connection error * Change files
* Publish fix * Creating new Microsoft.ReactNative.sln (#4535) * Created new Microsoft.ReactNative.sln * Update PR to build vnext/Microsoft.ReactNative.sln * Add params to disable ReactUWP nuget pack during PR * applying package updates ***NO_CI*** * Move playground app to use Microsoft.ReactNative (#4534) * Move playground app to use Microsoft.ReactNative * formatting * Bunch of code review feedback * Change files * formatting * Delete unneeded gitignore * Minor removal of diff * Minor changes to try to avoid having to update cert all the time * Fix bundle command * No need to build rntester seperately, the release build of playground will do it from msbuild * Remove test property * Bump beachball from 1.28.0 to 1.28.1 (#4539) * applying package updates ***NO_CI*** * applying package updates ***NO_CI*** * Enabling managed unit tests in CI loop (#4532) * Disable .net native for managed unit tests * Change files * Enable managed unit tests in ci loop * re-merge unit test tasks, can't run arm anyway * Remove prefer 32bit from 64bit flavors * Update mstest dependencies * Managed tests are separate again * Add comment * Add test adapter path * x86 debug again * Update test adapter folder * Try again * remove frameworkuap10 * Switch to appxrecipie * Try to switch to installed version of vstest * Remove installer task * applying package updates ***NO_CI*** * Remove 'Delete Android SDK' from E2ETest (#4554) * Enable WebSocketModule unit tests (#4538) * Prototype mockable MessageQueueThread and JSExecutor * Declare MockJSExecutor::CallFunctionFunctor * Add missing using std::unique_ptr * Use WebSocketModule::MethodId * clang format * Make WebSocketModule methods override instead of virtual. * Allow WebSocketModule to use custom resource factory * Rename *Impl to *Mock * Move Instance mocks into separate source files * Move Instance mocks into separate source files * Change files * clang format * Define WebSocketModule::SetResourceFactory * applying package updates ***NO_CI*** * Improve inner loop and error reporting (local CLI) (#4536) -Surface msbuild errors (shows as undefined right now) -downloading nuget with Invoke-WebRequest will usually take a long time because of the PS progress bar (which sadly gets updated for every byte), so I'm disabling the progress bar (it only takes a second to download when the progress bar is turned off) retry once if nuget failed since if you cancel at just the right time, you could end up with a truncated nuget.exe and are wedged unless you know which file to delete. Learned that the hard way. -factor out calling onto powershell functions -we were not exposing whether the function failed or not * applying package updates ***NO_CI*** * Don't set readyState on connection error (#4562) * Don't set readyState on connection error * Change files * applying package updates ***NO_CI*** * Bump @babel/core from 7.8.4 to 7.9.0 (#4565) * Simplified C++ macros and improved their comments (#4568) * Simplified C++ macros and improved their comments * Change files * applying package updates ***NO_CI*** * Handle syntax errors in RedBox * Change files * format * clean up Co-authored-by: Andrew Coates (REDMOND) <acoates@microsoft.com> Co-authored-by: Jon Thysell <jthysell@microsoft.com> Co-authored-by: React-Native-Windows Bot <53619745+rnbot@users.noreply.github.com> Co-authored-by: Andrew Coates <30809111+acoates-ms@users.noreply.github.com> Co-authored-by: dependabot-preview[bot] <27856297+dependabot-preview[bot]@users.noreply.github.com> Co-authored-by: Julio César Rocha <JunielKatarn@users.noreply.github.com> Co-authored-by: Vladimir Morozov <vmoroz@users.noreply.github.com>
Setting this may prevent
PerformClosefrom getting called, which leads to a hang.Microsoft Reviewers: Open in CodeFlow