Matrix-Mirrors/anonymousland-synapse
1
0
Fork 0
mirror of https://git.anonymousland.org/anonymousland/synapse.git synced 2025-05-04 06:54:54 -04:00
Code Issues Projects Releases Packages Wiki Activity

Minor @cachedList enhancements (#9975)

Browse source 
- use a tuple rather than a list for the iterable that is passed into the
  wrapped function, for performance

- test that we can pass an iterable and that keys are correctly deduped.
This commit is contained in:
Richard van der Hoff 2021-05-14 11:12:36 +01:00 • committed by GitHub
parent 52ed9655ed
commit 5090f26b63
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 31 additions and 20 deletions
Download patch file Download diff file Expand all files Collapse all files

14
synapse/util/caches/descriptors.py
View file

@ -322,8 +322,8 @@ class DeferredCacheDescriptor(_CacheDescriptorBase):
class DeferredCacheListDescriptor(_CacheDescriptorBase):
"""Wraps an existing cache to support bulk fetching of keys.
Given a list of keys it looks in the cache to find any hits, then passes
the list of missing keys to the wrapped function.
Given an iterable of keys it looks in the cache to find any hits, then passes
the tuple of missing keys to the wrapped function.
Once wrapped, the function returns a Deferred which resolves to the list
of results.
@ -437,7 +437,9 @@ class DeferredCacheListDescriptor(_CacheDescriptorBase):
return f
args_to_call = dict(arg_dict)
args_to_call[self.list_name] = list(missing)
# copy the missing set before sending it to the callee, to guard against
# modification.
args_to_call[self.list_name] = tuple(missing)
cached_defers.append(
defer.maybeDeferred(
@ -522,14 +524,14 @@ def cachedList(
Used to do batch lookups for an already created cache. A single argument
is specified as a list that is iterated through to lookup keys in the
original cache. A new list consisting of the keys that weren't in the cache
get passed to the original function, the result of which is stored in the
original cache. A new tuple consisting of the (deduplicated) keys that weren't in
the cache gets passed to the original function, the result of which is stored in the
cache.
Args:
cached_method_name: The name of the single-item lookup method.
This is only used to find the cache to use.
list_name: The name of the argument that is the list to use to
list_name: The name of the argument that is the iterable to use to
do batch lookups in the cache.
num_args: Number of arguments to use as the key in the cache
(including list_name). Defaults to all named parameters.