Refactor getting replication updates from database v2. #7740
Conversation
erikjohnston
force-pushed
the
erikj/move_limited_checks
branch
from
05f520a
to
8cd151b
Compare
There was a problem hiding this comment.
I think this looks OK overall. I think I see why the db_query_to_update_function existed now, as there's a lot of similar, but not quite the same-ish code.
Although this duplicates code a bit, the fewer levels of indirection will be useful, I think! 👍
Note that I kind of skipped some of the comments since they seem to be copied and pasted?
| updates = [(row[0], row[1:]) for row in txn] | ||
| limited = False | ||
| upto_token = current_id | ||
| if len(updates) >= limit: |
There was a problem hiding this comment.
#7636 seems to use a mixture of >, >= and == to determine if the stream was limited. I'm not sure if this inconsistency is on purpose or not. I would have thought > would be the proper choice here to avoid unnecessary work, but maybe that's wrong.
There was a problem hiding this comment.
Ah, yes, sorry. It mostly depends on the SQL query. Some are LIMIT ? and so will only ever return at most limit rows. Some have more fuzzy bounds and so could return more. I don't think anywhere uses >? If they do it sounds like a bug
There was a problem hiding this comment.
Ah, typing uses >, but that is a special case because we have all results in memory and don't query the DB with a limit
There was a problem hiding this comment.
Interesting. Is this worth a comment in any of the locations?
There was a problem hiding this comment.
I've added a comment to the typing one 👍
| @@ -98,77 +98,69 @@ def get_pushers(txn): | |||
| rows = yield self.db.runInteraction("get_all_pushers", get_pushers) | |||
| return rows | |||
|
|
|||
| def get_all_updated_pushers(self, last_id, current_id, limit): | |||
There was a problem hiding this comment.
Was get_all_updated_pushers unused?
There was a problem hiding this comment.
There are still a few streams left that don't have store functions that follow the new scheme, but those are special snowflakes that have more complicated handling in their stream class. We should probably fix those up to, but I don't think they
@erikjohnston What was the rest of this final sentence? 😄
I think that this PR looks good though.
|
How am I so bad at writing? I think that was meant to be:
|
* commit '43726783e': (22 commits) 1.17.0rc1 Fix some spelling mistakes / typos. (#7811) `update_membership` declaration: now always returns an event id. (#7809) Improve stacktraces from exceptions in background processes (#7808) Fix `can only concatenate list (not "tuple") to list` exception (#7810) Pass original request headers from workers to the main process. (#7797) Generate real events when we reject invites (#7804) Add `HomeServer.signing_key` property (#7805) Revert "Update the installation docs on apt-transport-https (#7801)" Do not use simplejson in Synapse. (#7800) Stop passing bytes when dumping JSON (#7799) Update the installation docs on apt-transport-https (#7801) shuffle changelog slightly Change Caddy links (old is deprecated) (#7789) Stop populating unused table `local_invites`. (#7793) Refactor getting replication updates from database v2. (#7740) Add libwebp dependency to Dockerfile (#7791) Add documentation for JWT login type and improve sample config. (#7776) Convert the appservice handler to async/await. (#7775) Don't ignore `set_tweak` actions with no explicit `value`. (#7766) ...
This is a follow on from #7636 to do most of the other streams. There are still a few streams left that don't have store functions that follow the new scheme, but those are special snowflakes that have more complicated handling in their stream class. We should probably fix those up to, but I don't think they
c.f. #7340