updates:
- [github.com/python-jsonschema/check-jsonschema: 0.27.0 → 0.27.2](0.27.0...0.27.2)
- [github.com/psf/black: 23.10.0 → 23.11.0](psf/black@23.10.0...23.11.0)

…onfig

[pre-commit.ci] pre-commit autoupdate

Bumps [responses](https://github.com/getsentry/responses) from 0.23.3 to 0.24.1.
- [Release notes](https://github.com/getsentry/responses/releases)
- [Changelog](https://github.com/getsentry/responses/blob/master/CHANGES)
- [Commits](getsentry/responses@0.23.3...0.24.1)

---
updated-dependencies:
- dependency-name: responses
  dependency-type: direct:development
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Lazy reads are slightly customized to better handle named pipes.

The lazy file reads are combined with an early close immediately after
the instance loader gets data from a file. As a result,
check-jsonschema should only keep one instance file open at a time.

The previous implementation tried to repeatedly read data for a schema
input, even across multiple instances. As a result, if the schema
source was ephemeral (e.g. a memfd), the subsequent reads would fail.

Co-authored-by: sirosen <sirosen@users.noreply.github.com>

Create a directory of 2000 instance files, then iterate through them
and verify that the number of open files never exceeds 2000 at any
point during iteration.

In order for pyjson5 to be satisfied, it must be passed the real file
descriptor (or something satisfying `io.IOBase`). For simplicity, pass
it the underlying open() result.

Also improve the file close behavior using a `finally` block.

Improve file handling with lazy reads