fix an internal error in advanced_tls.test. Previous check is to check against the prefix of the ServerName, which might be different in various environments. We'd better not rely on checking that.
Also, two fixes:
- Fix long-standing `.travis.yml` bug where `VET_SKIP_PROTO` was not `export`ed (so not seen by `vet.sh`).
- Update `vet.sh` to work with new `goimports -l` that does not print a `:` after filenames.
- .travis.yml:
- Download dependencies at install time
- Run race and non-race in separate instances
- Use braces around env var names
- vet.sh:
- Run `go mod tidy` to keep it tidy
- Stop filtering transport errors from go/vet output
- Use braces around env var names
- Makefile:
- Reorder alphabetically
- Add "vetdeps" as a dependency of "vet"
- Add "testappengine" to "all"
ChannelZ is a debugging feature in grpc. It has not had appengine support; this commit
removes its support for 1.6, too. This change is because in a recent sys/unix commit
1.6 support was broken.
This PR removes race testing from all travis runs except go1.10 except during nightlies.
| | Before* | After** | % Change |
| --------------- | ------- | ------- | -------- |
| Total runtime | 65m | 45m | -31% |
| Wall-clock time | 21m | 11.5m | -45% |
*: (one random sample I found in our PRs)
**: (latest run on this PR)
Also includes some organization changes (making proper use of `before_install` and `install`) / cleanups in `.travis.yml` to make it easier to understand. As a result, we now stop unintentionally running `vet` on a second build by accident (go1.10 with retry enabled).
- Define env variable RUN386 and set GOARCH later, instead of assigning to GOARCH directly. Otherwise, gimme will override GOARCH back
- Add misspell check
- Disable -race on 386
* Check "x/net/context" with `go vet` like "context"
Includes a signficant revamp of .travis.yml to execute a separate script for
all of the things we check before allowing a merge.
* fix bug in benchmain
It's too noisy and not worth it. I don't like encouraging the code
coverage game, wherein people write bad tests to get points.
It's annoying to get a red X in the PR status when coverage dropped
by 1 line or 0.4%.
Good code review should be sufficient to verify we have both tests
and good tests.
