Allow Cache to return a single block

[blt]

What does this PR do?

In some cases it is valuable for the block cache to be used without spin.
We expose next_block to return a Block reference as soon as one is
available, blocking the callers meanwhile.

Motivation

Allow single-threaded applications to read from a block cache.

Ref:
https://datadoghq.atlassian.net/browse/SMPTNG-513

[blt]

• Release lading 0.23.4 #1084
• Add telemetry to logrotate_fs #1083
• Install fuse3 in lading container #1080
• Create mount-point in logrotate-fs start if it does not exist #1079
• Make sure state starts at close to now tick #1078
• Uniformly advance time in logrotate_fs #1077
• Avoid some allocations, hashes that aren't needed #1076
• Simulate all ticks, not just some #1075
• Configuration and logging improvements to logrotate_fs #1074
• Tidy up some clippy allows in logrotate_fs #1071
• Update State::advance_time to be less indented #1070
• Make logrotate_fs a generator #1066
• Introduce FileHandle #1063
• Improve property tests #1058
• Rotate log files #1057
• Move names into a pre-created name pool #1056
• Add component pieces for rotation #1055
• Adjust time advancement to be done by State #1053
• Avoid extra name lookup #1044
• Allow readers to read consistently #1043
• Ensure file size updates due to getattr #1042
• Update file timestamps to reflect real systemtime #1041
• Begin work on a logrotate filesystem #1034
• Allow Cache to return a single block #1037 👈
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

Join @blt and the rest of your teammates on Graphite

[blt]

Merge activity

• Oct 28, 10:40 AM EDT: A user merged this pull request with Graphite.