Go: all: golang.org/x repos need to stay green

[rsc]

We need to get the subrepos green consistently for 1.5 and moving forward.

[gopherbot]

CL https://golang.org/cl/12788 mentions this issue.

[gopherbot]

CL https://golang.org/cl/12830 mentions this issue.

[gopherbot]

CL https://golang.org/cl/12832 mentions this issue.

[gopherbot]

CL https://golang.org/cl/12765 mentions this issue.

[gopherbot]

CL https://golang.org/cl/12838 mentions this issue.

[gopherbot]

CL https://golang.org/cl/13008 mentions this issue.

[gopherbot]

CL https://golang.org/cl/13032 mentions this issue.

[gopherbot]

CL https://golang.org/cl/13009 mentions this issue.

[gopherbot]

CL https://golang.org/cl/12916 mentions this issue.

[gopherbot]

CL https://golang.org/cl/13252 mentions this issue.

[gopherbot]

CL https://golang.org/cl/13268 mentions this issue.

[gopherbot]

CL https://golang.org/cl/13280 mentions this issue.

[gopherbot]

CL https://golang.org/cl/13300 mentions this issue.

[gopherbot]

CL https://golang.org/cl/13695 mentions this issue.

[quentinmit]

Looking at the current status, I see several failing subrepos.

• "sync" is failing because it has no tests to run
• "arch" is failing on all the 386 platforms with #12840:

../../gopath/src/golang.org/x/arch/vendor/rsc.io/pdf/page.go:250: internal compiler error: out of fixed registers

• "exp" is failing due to missing build tags to exclude unsupported packages (linux i2c on windows, OpenGL on freebsd, etc.)
• "mobile" appears to have broken tests:

    seq_test.go:98: gomobile: toolchain partially installed, run `gomobile init`
    seq_test.go:99: failed to run gomobile bind: exit status 1

• "oauth2" is failing to find its dependencies:

cannot find package "google.golang.org/appengine"

• "term" has nothing to build

I'm not sure how useful it is to have an umbrella bug for all the different subrepos. All of these except the compiler error seem like simple fixes, however, so I'm going to tag this NeedsFix.

[gopherbot]

CL https://golang.org/cl/23318 mentions this issue.

[quentinmit]

"text" is now failing due to a2a4db7375204c22898559d3b96a748c6df96294 (upgrading tip to Unicode 9)

It looks like there's already golang.org/cl/24540 to fix it. @mpvl is there anything blocking submission of that CL?

[gopherbot]

CL https://golang.org/cl/24942 mentions this issue.

[gopherbot]

CL https://golang.org/cl/24945 mentions this issue.

[gopherbot]

CL https://golang.org/cl/24946 mentions this issue.

[gopherbot]

CL https://golang.org/cl/24944 mentions this issue.

[gopherbot]

CL https://golang.org/cl/24970 mentions this issue.

[gopherbot]

CL https://golang.org/cl/24972 mentions this issue.

[bradfitz]

Sent a bunch of CLs. Kicking down the road to Go 1.8.

[gopherbot]

CL https://golang.org/cl/30600 mentions this issue.

[gopherbot]

CL https://golang.org/cl/34274 mentions this issue.

[gopherbot]

CL https://golang.org/cl/34275 mentions this issue.

[gopherbot]

CL https://golang.org/cl/34288 mentions this issue.

[bradfitz]

Pretty green for Go 1.8.

[broady]

Looking green for 1.9.

[bradfitz]

All green for 1.10 except #23166 which is already marked for 1.10. (/cc @alandonovan) So pushing this to bug for reuse in Go 1.11.

[gopherbot]

Change https://golang.org/cl/125816 mentions this issue: godoc: fix tests on Go 1.9

[dmitshur]

The x/tools subrepo tests are currently passing on tip but failing on 1.10 and 1.9 according to https://build.golang.org/. The failures are related to the new x/tools/go/packages package have started after x/tools/cmd/callgraph began using it as of CL 128695. See https://build.golang.org/?repo=golang.org%2fx%2ftools.

/cc @alandonovan FYI.

[gopherbot]

Change https://golang.org/cl/129175 mentions this issue: cmd/callgraph: temporarily disable failing tests on go1.10

[dmitshur]

Subrepos are green on first-class ports, other than #29660 (x/tools failure on s390x) and #17538 (x/text failure on s390x).

Edit: #29660 (x/tools failure on s390x) has been resolved.

That's green enough for Go 1.12 RC 1, but we should try to fix them for final release (or reconsider what qualifies as first-class port on build.golang.org).

[andybons]

LGTM for 1.12

[gopherbot]

Change https://golang.org/cl/165797 mentions this issue: misc/android: enable many more tests on GOOS=android

[gopherbot]

Change https://golang.org/cl/171033 mentions this issue: shiny/driver/gldriver: define GL_SILENCE_DEPRECATION

[gopherbot]

Change https://golang.org/cl/171034 mentions this issue: shiny/driver/gldriver: fix 2 warnings on macOS 10.14

[bcmills]

The subrepos seem to be generally ok except on solaris-amd64-oraclerel and aix-ppc64.

(I guess this thread is probably relevant for Solaris?)

[bcmills]

I did a sweep of the dashboard for 1.13.
The bad news is, there are still a lot of failures.

The good news is, many of the failures are specific to one platform or arch (openbsd-arm, illumos, and js-wasm are the most common), and a few of them are sporadic rather than consistent.

New failures include:

• x/build
  
  
  
  • #32838
  
  
• x/crypto
  
  
  
  • #32840
  
  
• x/debug
  
  
  
  • #32839
  
  
• x/exp
  
  
  
  • #32837
  
  
• x/image
  
  
  
  • #32835
  
  
• x/net
  
  
  
  • #32842
  
  
  • #32841
  
  
• x/review
  
  
  
  • #32836
  
  
• x/text
  
  
  
  • #32844
  
  
  • #32843
  
  
• x/tools
  
  
  
  • #32834
  
  
  • #32833
  
  
  • #32832
  
  
  • #32831
  
  

[gopherbot]

Change https://golang.org/cl/38050 mentions this issue: term: add IsTerminal func

[bcmills]

x/blog: #33940

[gopherbot]

Change https://golang.org/cl/192327 mentions this issue: dashboard: disable failing repos on misconfigured builders

[bcmills]

x/website: #33945

[bcmills]

x/website again: #33946

[bcmills]

x/tools:

• #33950
• #33951

[bcmills]

x/perf: #33952

[bcmills]

x/tools: #33956

[bcmills]

x/sys/unix, but only on go1.12 (so not a release blocker for go1.13): #33982

[bcmills]

Current red cells for release-branch.go1.13:

• #33951 (x/tools consistently failing on windows-amd64-race)
• #34034 (x/tools/internal/imports timing out on dragonfly-amd64)
• Various failures on the misconfigured openbsd-arm builder (see https://github.com/golang/go/issues/25498#issuecomment-526695541).

Since the openbsd-arm builder seems to be misconfigured, I think we can safely ignore that column.

The windows-amd64-race issue seems to be an out-of-memory condition, and the other -race builders are passing so it's not likely to be masking any other failures. We should address it soon, but I don't think it needs to block the release.

The dragonfly-amd64 issue may just be a slow builder, and dragonfly/amd64 is not a first class port anyway, so it probably does not need to block the release either.

[gopherbot]

Change https://golang.org/cl/208579 mentions this issue: dashboard: don't build x/exp and x/mobile on the linux-arm-arm5spacemonkey builder

[gopherbot]

Change https://golang.org/cl/209338 mentions this issue: dashboard: don't build x/exp and x/mobile on the plan9 builders

[gopherbot]

Change https://golang.org/cl/209477 mentions this issue: dashboard: don't build x/exp and x/mobile on freebsd 10.3, 10.4 and 11.1 builders

[gopherbot]

Change https://golang.org/cl/213037 mentions this issue: godoc/static: update copyright year to 2020

[gopherbot]

Change https://golang.org/cl/213119 mentions this issue: godoc/static: update copyright year to 2013

[gopherbot]

Change https://golang.org/cl/213121 mentions this issue: content/static: generate constant 2019 copyright year

[gopherbot]

Change https://golang.org/cl/215299 mentions this issue: kubernetes/gke: skip 4 tests on builders

[cagedmantis]

@dmitshur and I have verrified that the subrepos are green for the release branches of golang.org/x repos that are used in the cmd and std modules:

github.com/google/pprof v0.0.0-20191105193234-27840fff0d09
github.com/ianlancetaylor/demangle v0.0.0-20180524225900-fc6590592b44
golang.org/x/arch v0.0.0-20190815191158-8a70ba74b3a1
golang.org/x/crypto v0.0.0-20200128174031-69ecbb4d6d5d
golang.org/x/mod v0.2.0
golang.org/x/net v0.0.0-20190620200207-3b0461eec859
golang.org/x/sync v0.0.0-20190423024810-112230192c58
golang.org/x/sys v0.0.0-20200131233351-9e5cf931a04b
golang.org/x/text v0.3.0
golang.org/x/tools v0.0.0-20200131233409-575de47986ce
golang.org/x/xerrors v0.0.0-20191011141410-1b5146add898
rsc.io/pdf v0.1.1

golang.org/x/crypto v0.0.0-20200128174031-69ecbb4d6d5d
golang.org/x/net v0.0.0-20191126235420-ef20fe5d7933
golang.org/x/sys v0.0.0-20200201011859-915c9c3d4ccf
golang.org/x/text v0.3.3-0.20191031172631-4b67af870c6f
golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e

We are considering the sub repos to be green for the purposes of cutting the go1.14rc1.

[dmitshur]

This task has been completed for the Go 1.14 cycle at the time of cutting the go1.14rc1 release, see above comment for details.

Moving this issue to the Go 1.15 milestone.

[gopherbot]

Change https://golang.org/cl/227379 mentions this issue: dashboard, app/appengine: set a known issue for staticlockranking builder

[dmitshur]

The golang.org/x repos are green for first-class ports at this time.

The only failures are in x/build (#36163), and x/mobile and x/exp (#39156), none of which contribute packages to the main Go distribution.

The golang.org/x repos need to continue to stay green for the rest of 1.15 release cycle, but this isn't blocking the Go 1.15 Beta 1 release.

[gopherbot]

Change https://golang.org/cl/240512 mentions this issue: dashboard: stop linux-386-longtest builds in x/exp and x/mobile

[toothrot]

Creating release branches for Go 1.15 RC1:

# golang.org/x/arch v0.0.0-20200511175325-f7c78586839d
# golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9
# golang.org/x/mod v0.3.0
# golang.org/x/net v0.0.0-20200707034311-ab3426394381
# golang.org/x/sys v0.0.0-20200501145240-bc7a7d42d5c3
# golang.org/x/text v0.3.3-0.20200430171850-afb9336c4530
# golang.org/x/tools v0.0.0-20200616133436-c1934b75d054
# golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543

[toothrot]

Labeling okay-after-beta1, as all repos included in the release are green.