Fix style (space between a .. b) in std/json.d

[wilzbach]

See also: #5541 (comment)

CC @CyberShadow

(will investigate how it managed to get through Circle)

[dlang-bot]

Thanks for your pull request, @wilzbach!

Bugzilla references

Your PR doesn't reference any Bugzilla issue.

If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog.

[wilzbach]

I cherry-picked the commit from #5547 (comment)

[wilzbach]

(will investigate how it managed to get through Circle)

#5511 was never green on CircleCi:

It's simply not required:

[CyberShadow]

Thanks :)

[wilzbach]

Regarding #5547 (comment)

maybe because everyone ignores CircleCI, since it fails quite often

@MartinNowak No AFAICT it simply has never been enforced.
Btw regarding failures - I am not so sure whether this is true anymore. Here are the build statistics for the last month:

Phobos builds over the last month in details:

There were only four failures in the last thirty days and half of them were our fault:

• https://circleci.com/gh/dlang/phobos/3368 -> our fault (old, unrebased PR)
• https://circleci.com/gh/dlang/phobos/3253 -> CodeCov outage
• https://circleci.com/gh/dlang/phobos/3206 -> DUB registry outage
• https://circleci.com/gh/dlang/phobos/3067 -> our fault (old, unrebased PR)

takes ages to load logs.

That might even be an temporary error on the CircleCi side - I have never experienced this (except when being on my phone).

@CyberShadow @MartinNowak can we require CircleCi to pass for the stable branch as well?

[CyberShadow]

@CyberShadow @MartinNowak can we require CircleCi to pass for the stable branch as well?

Sounds good to me but it's ultimately up to Martin.

takes ages to load logs.

The UI is clunky and it does take a bunch of scrolling and several clicks to open the relevant section for no apparent reason.

[MartinNowak]

I'd rather not do that. We want people to be able to merge even with small coverage decreases. Might as well adjust the limits, but I'd rather leave that to common sense.
What needs to be fixed is dlang/dlang-bot#69 instead.

[wilzbach]

I'd rather not do that. We want people to be able to merge even with small coverage decreases. Might as well adjust the limits, but I'd rather leave that to common sense.

How is this related? CircleCi is about the linting, public tests etc. - it doesn't fail if the coverage decreases. That's CodeCov to which CircleCi sends the coverage files.

[MartinNowak]

How is this related? CircleCi is about the linting, public tests etc. - it doesn't fail if the coverage decreases. That's CodeCov to which CircleCi sends the coverage files.

I see, only meant the latter, indeed makes sense for Linting. Any plans to work on dlang/dlang-bot#69 regardless.

[wilzbach]

indeed makes sense for Linting.

Enabled then :)

Any plans to work on dlang/dlang-bot#69 regardless.

It's not on my main priority list, I might get to it on the next bot hacking session in August, but no promises.