fix(dsstore): persist inherited LTS trie #13276
Conversation
| @@ -23,6 +23,8 @@ | |||
| -include_lib("snabbkaffe/include/snabbkaffe.hrl"). | |||
| -include_lib("stdlib/include/assert.hrl"). | |||
|
|
|||
| -import(emqx_utils_conv, [bin/1]). | |||
There was a problem hiding this comment.
Dropped in 10e78ea.
keynslug
force-pushed
the
fix/ds/inherited-lts-lost
branch
from
6f046a5
to
49aca5b
Compare
| || I <- lists:seq(1, 200), | ||
| TS <- Timestamps, | ||
| TS <- lists:seq(0, 1_000, 10), |
There was a problem hiding this comment.
It's pretty confusing that TS is used for both timestamp and as a topic level. I think it's better to use a constant timestamp for all messages (I believe it's irrelevant in this test), and rename TS to J.
There was a problem hiding this comment.
It's actually not used as a topic level, but it indeed should be irrelevant.
| PrevRuntimeData :: term() | ||
| ) -> | ||
| {_Schema, cf_refs()}. |
There was a problem hiding this comment.
| PrevRuntimeData :: term() | |
| ) -> | |
| {_Schema, cf_refs()}. | |
| Schema | undefined | |
| ) -> | |
| {Schema, cf_refs()}. |
?
There was a problem hiding this comment.
The last argument is actually RuntimeData (i.e. the result of open/5), not Schema.
| make_topic(T1, T2, T3, T4) -> | ||
| make_topic([T1, T2, T3, T4]). |
There was a problem hiding this comment.
This function looks very specific to a particular testcase. Please remove it and use make_topic/1 with list argument.
Before this commit, inherited trie was actually only kept in memory cache. Also simplify storage backend behaviour around inheriting previous generation's legacy.
keynslug
force-pushed
the
fix/ds/inherited-lts-lost
branch
from
49aca5b
to
82588fb
Compare
| Topic = emqx_topic:join(["wildcard", integer_to_list(I), "suffix", Suffix]), | ||
| {TS, make_message(TS, Topic, integer_to_binary(TS))} | ||
| end | ||
| {TS, make_message(TS, make_topic([wildcard, I, suffix, Suffix]), bin(TS))} |
There was a problem hiding this comment.
Same here, TS is probably not needed.
There was a problem hiding this comment.
Had a different impression actually: sounds important for a replay test to have messages spread both in time and in "topic space" (here, TS is not part of a topic).
Anyway, it'd probably be better to tackle the logic of unrelated testcase in a separate PR.
Release version: v/e5.7.1
Summary
Before this commit, inherited trie was actually only kept in memory cache.
Also simplify storage backend behaviour around inheriting previous generation's legacy.
PR Checklist
Please convert it to a draft if any of the following conditions are not met. Reviewers may skip over until all the items are checked:
Added property-based tests for code which performs user input validationchanges/(ce|ee)/(feat|perf|fix|breaking)-<PR-id>.en.mdfilesCreated PR to emqx-docs if documentation update is required, or link to a follow-up jira ticket