Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

x/tools/gopls: changes cause canceled diagnostics for files in unrelated packages #36144

Closed
myitcv opened this issue Dec 14, 2019 · 8 comments
Closed

Comments

@myitcv
Copy link
Member

@myitcv myitcv commented Dec 14, 2019

What version of Go are you using (go version)?

$ go version
go version devel +da4d58587e Sat Dec 7 15:57:30 2019 +0000 linux/amd64
$ go list -m golang.org/x/tools
golang.org/x/tools v0.0.0-20191213221258-04c2e8eff935 => github.com/myitcvforks/tools v0.0.0-20191214162001-758e20ac67bf
$ go list -m golang.org/x/tools/gopls
golang.org/x/tools/gopls v0.1.8-0.20191213221258-04c2e8eff935 => github.com/myitcvforks/tools/gopls v0.0.0-20191214162001-758e20ac67bf

Note this version is a temporary fork with fixes as described in #35638 (comment)

Does this issue reproduce with the latest release?

Yes

What operating system and processor architecture are you using (go env)?

go env Output
$ go env
GO111MODULE="on"
GOARCH="amd64"
GOBIN=""
GOCACHE="/home/myitcv/.cache/go-build"
GOENV="/home/myitcv/.config/go/env"
GOEXE=""
GOFLAGS=""
GOHOSTARCH="amd64"
GOHOSTOS="linux"
GOINSECURE=""
GONOPROXY=""
GONOSUMDB=""
GOOS="linux"
GOPATH="/home/myitcv/gostuff"
GOPRIVATE=""
GOPROXY="https://proxy.golang.org,direct"
GOROOT="/home/myitcv/gos"
GOSUMDB="sum.golang.org"
GOTMPDIR=""
GOTOOLDIR="/home/myitcv/gos/pkg/tool/linux_amd64"
GCCGO="gccgo"
AR="ar"
CC="gcc"
CXX="g++"
CGO_ENABLED="1"
GOMOD="/home/myitcv/gostuff/src/github.com/myitcv/govim/go.mod"
CGO_CFLAGS="-g -O2"
CGO_CPPFLAGS=""
CGO_CXXFLAGS="-g -O2"
CGO_FFLAGS="-g -O2"
CGO_LDFLAGS="-g -O2"
PKG_CONFIG="pkg-config"
GOGCCFLAGS="-fPIC -m64 -pthread -fmessage-length=0 -fdebug-prefix-map=/tmp/go-build705453102=/tmp/go-build -gno-record-gcc-switches"

What did you do?

We have a govim test that starts by opening a file and then calling Formatting.

But before calling Formatting we first wait for the initial diagnostics to be sent for the file.

However very occasionally we do not see the initial diagnostics for the file and the test times out (and fails) waiting.

Here is a failing and passing log of this test:

On the occasions that this test fails, changes to go.mod appear to be related. govim currently supports file watching by simulating the file that has changed being open in the editor. For example, with go.mod changes, govim detects the file has changed, signals DidOpen to gopls on the first change, then DidChange thereafter.

As you can see from the failing log above, the calculation of diagnostics for file.go gets almost to the point of publishing, but then its context gets cancelled. That in itself isn't a problem except that the diagnostics are not subsequently sent.

What did you expect to see?

Initial diagnostics to be sent for a file.

What did you see instead?

As above.


cc @stamblerre @heschik @ridersofrohan

@heschik

This comment has been minimized.

Copy link
Contributor

@heschik heschik commented Dec 16, 2019

This seems like a fairly deep-rooted problem. As I understand it, if we're diagnosing package A, and I edit a file in unrelated package B, we're going to cancel the first diagnostics run and then not re-diagnose A. The fact that it was a go.mod file this time may be something of a red herring.

@myitcv

This comment has been minimized.

Copy link
Member Author

@myitcv myitcv commented Dec 28, 2019

FWIW, this is now the main cause of test failures in the govim test suite.

cc @findleyr

myitcv added a commit to myitcvforks/govim that referenced this issue Jan 1, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these
myitcv added a commit to myitcvforks/govim that referenced this issue Jan 1, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these
myitcv added a commit to govim/govim that referenced this issue Jan 4, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these
myitcv added a commit to govim/govim that referenced this issue Jan 5, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these
myitcv added a commit to govim/govim that referenced this issue Jan 5, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these
myitcv added a commit to govim/govim that referenced this issue Jan 5, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these
myitcv added a commit to govim/govim that referenced this issue Jan 5, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these
myitcv added a commit to govim/govim that referenced this issue Jan 5, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 6, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 6, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 6, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 6, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 6, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 6, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 7, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 7, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 7, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 7, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 8, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 8, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 8, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 8, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
myitcv added a commit to govim/govim that referenced this issue Jan 8, 2020
This uber PR contains a number of changes to our testscript tests:

* Drop all errlogmatch's looking for assertions where we really should
  not need them. Historically we needed these because the threading
  model of gopls was a bit broken. i.e. despite us notifying of a change
  to a file, unless we waited for diagnostics to be available then we
  couldn't guarantee that, for example, completions would be correct
* Fix up all go.mod files to include go directives so that we don't fall
  foul of golang/go#36144
* Where we do require errlogmatch's, make sure we follow the existing
  semantics per golang/go#36243 and
  golang/go#36340
* Fix up some scripts so that we don't have initial errors in the file
  and hence don't need to soak up initial diagnostics
* Use sleep $DEFAULT_ERRLOGMATCH_WAIT where we expect there to be no
  diagnostic errors
* Only use errlogmatch on diagnostic publications where we need to
  know that a file has been opened, e.g. file watching, or where we
  need to soak up initial diagnostic notifications (see
  golang/go#36243)
* Move all commented-out errlogmatch commands matching on the number of
  errors to the end of each script. We still can't enable these

Add some initial tests that help to verify our expectations around
diagnostics being published by gopls.

Set GOPATH and GOCACHE consistently for the install testscript scripts
to speed them up.

Call t.Parallel in a couple of places to speed up the entire test run.

Add a test (that is skipped) which capture the essence of
golang/go#36144.
@stamblerre stamblerre changed the title x/tools/gopls: initial diagnostics not sent for file x/tools/gopls: changes cause canceled diagnostics for files in unrelated packages Jan 15, 2020
@myitcv

This comment has been minimized.

Copy link
Member Author

@myitcv myitcv commented Jan 16, 2020

@stamblerre I'm minded to close this given the massive number of updates recently, unless you/@heschik would like to keep it open?

myitcv added a commit to govim/govim that referenced this issue Jan 16, 2020
@myitcv

This comment has been minimized.

Copy link
Member Author

@myitcv myitcv commented Jan 16, 2020

FWIW I'm re-enabling a test against current master to stress this out a bit: govim/govim#684

@stamblerre

This comment has been minimized.

Copy link
Contributor

@stamblerre stamblerre commented Jan 16, 2020

Yes, I think it's reasonable to close this now that we diagnose entire workspace on every file change. Even if one diagnostics run gets canceled, another one will immediately be started and will complete once the user stops typing.

@myitcv

This comment has been minimized.

Copy link
Member Author

@myitcv myitcv commented Jan 17, 2020

Given the test case originally described in this issue, I think #36463 is actually more relevant:

The initial workspace load must not be canceled

Once we get a stable master (following a conclusion of #36601) we can easily reenable our test that verifies whether things are working/not here.

myitcv added a commit to govim/govim that referenced this issue Jan 17, 2020
@stamblerre

This comment has been minimized.

Copy link
Contributor

@stamblerre stamblerre commented Jan 17, 2020

Sounds good. Please follow up here once you get a chance to test this.

myitcv added a commit to govim/govim that referenced this issue Jan 18, 2020
myitcv added a commit to govim/govim that referenced this issue Jan 20, 2020
myitcv added a commit to govim/govim that referenced this issue Jan 20, 2020
@myitcv

This comment has been minimized.

Copy link
Member Author

@myitcv myitcv commented Jan 20, 2020

This does indeed appear to be fixed now; we at least have a passing CI! If we see any further issues we can re-open/create a new issue. Thanks again for all the work on fixes @stamblerre!

@myitcv myitcv closed this Jan 20, 2020
myitcv added a commit to govim/govim that referenced this issue Jan 21, 2020
myitcv added a commit to govim/govim that referenced this issue Jan 22, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
4 participants
You can’t perform that action at this time.