Ah yes - that's what I thought you meant in the first place!

Yes, that leads to a long PHP 7.1 build - see https://travis-ci.org/gitlost/wp-cli/builds/244175178?utm_source=github_status&utm_medium=notification for comparative times. (That build comes from gitlost#11)

I'll do that anyway on #4161 and you can see what you think. (It also requires a fix on cli.feature to make it "re-entrant".)

Yes, that leads to a long PHP 7.1 build - see https://travis-ci.org/gitlost/wp-cli/builds/244175178?utm_source=github_status&utm_medium=notification for comparative times. (That build comes from gitlost#11)

Oh. I meant to only run the specific scenario. So I think that's the only tag we need to include, instead of all of them again.

I'll do that anyway on #4161 and you can see what you think. (It also requires a fix on cli.feature to make it "re-entrant".)

Sounds good.

Ok we're still at cross-purposes I think. There is no scenario specific to opcache.save_comments being disabled. There is one scenario "bootstrap.feature:3" (https://github.com/wp-cli/wp-cli/blob/master/features/bootstrap.feature#L3) which can't be run with opcache.save_comments disabled as it installs and runs wp-cli 1.1.0 which then craps out.

There is no scenario specific to opcache.save_comments being disabled.

Oh, I see. I misunderstood previously.

I don't think we need to run the entire test suite with opcache.save_comments disabled. In fact, the PHPUnit tests seem sufficient to me.

Oh ok I'll remove that job from the travis then.