Eddy Ashton
Eddy Ashton
We will probably do this as a client-side script (which fetches the CA cert, and outputs a proposal containing multiple actions), rather than a fresh action (where the service would...
This doesn't work like I thought it did. Inline `close()` _should_ be ok, and we need to `uv_close()` _at least_ the stdin pipe once we're done writing to it. And...
The failures are so verbose we need to look at the raw logs, but on the first run these are the failing tests: ``` 2024-11-07T15:18:36.9778496Z The following tests FAILED: 2024-11-07T15:18:36.9779075Z...
First change knocks out of a few of those failures already: ``` 2024-11-07T16:20:07.7189192Z 40 - recovery_test_cft_api_0 (Failed) 2024-11-07T16:20:07.7189554Z 41 - recovery_test_cft_api_1 (Failed) 2024-11-07T16:20:07.7189916Z 44 - regression_test_suite (Failed) 2024-11-07T16:20:07.7190245Z 45 -...
Closing this PR, superceded by @maxtropets' work in #7201/#7232.
We need to establish whether `5.latest` and `6.0.2`/`6.0.3` are live-compatible. LTS doesn't test this, the usual guarantee is only to `6.0.0`. This is hard because it is cross-OS. We could...
Confirmed in this branch that 5.latest and 6.0.2 are live-compatible: https://github.com/microsoft/CCF/pull/7026#issuecomment-2943217728 However rather than yank releases or hacking a monotonic-forcing 5.latest, we're opting for option 3: - Clearly document the...
>This is related to but somewhat separate from the discussion around wait for commit. I think we'd need a phantom write _and_ wait-for-commit behaviour to produce a linearizable read? If...
To use the AppendEntries roundtrip, do we need to include a new counter value in AEs (and responses)? I'm specifically worried about delayed arrival of ACKs. You can't currently distinguish...