Philip Jenvey

@fzzzy did most of this in #494 but we're postponing it for now until we decide how to handle Spanner's usage of the encoded offset (or lack thereof)

Dependent on #645

The history collection seems much more susceptible to this issue. I don't think I've yet seen any instance of this other than w/ history (collection id = 4 in the...

This was fixed in #632 The `AND expiry > CURRENT_TIMESTAMP()` was removed from batch `INSERT` & `UPDATE`. This reduced 5xx errors, more closely matches the Python version, and additionally matches...

It looks like one potential option for using the new std `Backtrace` would be to render it via e.g. `format!("{:#?}", backtrace)` and passing that to `parse_stacktrace`. Its own `backtrace_to_stacktrace` impl...

I hadn't noticed `Backtrace::frames` has not stabilized yet, thanks. Agreed that that is a blocker for this.

This style was common with the `failure` crate: it recommended it as one of its patterns as it simplifies handling of extra data like the Optional backtrace here in to...

#488 appears to be the cause of this change

Bumping this from a 3 to at least a 5. syncstorage's own local e2e test suite (`test_storage.py`) can be invoked via `run.py` -- in theory from the existing docker image...

The JIRA sync mistakenly closed this