fix(feed/search): parse feed once + content-based search cache - #225
Merged
Merged
Conversation
getEpisodes fetched and parsed the feed twice on a cold call: once via getFeed (to populate channel metadata) and again directly to read the items. This doubled the network round-trips and XML parses on every cold path, e.g. the URIHandler resume-link flow that constructs a FeedParser with no cached feed. Fetch and parse the document a single time and reuse it for both the channel metadata and the episode items. The metadata extraction moves to a private extractFeed(body, url) helper that getFeed and getEpisodes both call, so behavior (metadata population, the "Invalid RSS feed" guard, this.feed caching) is unchanged - only the redundant second fetch is removed. Resolves deepsec finding other-redundant-fetch.
The Fuse search index was cached in a WeakMap keyed by the episodes array and reused whenever the cached size matched the array length. Length is a weak fingerprint: the same array reference mutated in place at the same length (an entry swapped or edited) would return the stale index built from the old contents. Validate the cache with a content signature (a JSON-framed list of each episode's title and streamUrl) instead of length, so any content or ordering change rebuilds the index while an unchanged list keeps reusing it across keystrokes (preserving the #149 optimization). Adds the first unit tests for searchEpisodes, including a Fuse-construction spy proving the index is reused when unchanged and rebuilt when the content changes. Resolves deepsec finding other-stale-cache.
Deploying podnotes with
|
| Latest commit: |
c57fae3
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://de662011.podnotes.pages.dev |
| Branch Preview URL: | https://chhoumann-deepsec-feed-cache.podnotes.pages.dev |
chhoumann
marked this pull request as ready for review
June 29, 2026 07:20
github-actions Bot
pushed a commit
that referenced
this pull request
Jul 9, 2026
## [2.17.3](2.17.2...2.17.3) (2026-07-09) ### Bug Fixes * **feed/search:** parse feed once + content-based search cache ([#225](#225)) ([053d51f](053d51f)), closes [#149](#149) * make episode identity key collision-resistant and prototype-safe ([#226](#226)) ([a5683db](a5683db)) * **opml:** correct import progress math and saved-count reporting ([#221](#221)) ([a79e529](a79e529)) * **security:** validate feed/URI URLs and cap download size ([#223](#223)) ([edef281](edef281)) * **template:** neutralize feed-controlled note injection ([#228](#228)) ([ef4ecbd](ef4ecbd)) * **timestamp:** escape live table-cell pipe after an escaped backslash ([#227](#227)) ([a34dfca](a34dfca)) * **transcription:** resolve three deepsec transcription-pipeline bugs ([#224](#224)) ([83c34e7](83c34e7))
Contributor
|
🎉 This PR is included in version 2.17.3 🎉 The release is available on GitHub release Your semantic-release bot 📦🚀 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes two deepsec-confirmed data-layer bugs in the feed/search path.
getEpisodesfetched the feed twice (slug:other-redundant-fetch)FeedParser.getEpisodes(url)parsed the feed once viagetFeed(to populatechannel metadata) and then parsed it again directly to read the items - two
network round-trips and two XML parses per cold call. This was actively
reachable: the
URIHandlerresume-link path ({{episodelink}}/ share links)constructs
new FeedParser()with no cached feed, so both fetches fired.The fix fetches and parses the document once and reuses it for both the channel
metadata and the items. Metadata extraction moves to a private
extractFeed(body, url)thatgetFeedandgetEpisodesshare, so metadatapopulation, the
Invalid RSS feedguard, andthis.feedcaching are unchanged -only the redundant second fetch is gone.
Search cache returned stale results on in-place mutation (slug:
other-stale-cache)searchEpisodescached a Fuse index in aWeakMapkeyed by the episodes arrayand reused it whenever the cached size equalled the array length. Length is a
weak fingerprint: the same array reference mutated in place at the same length
(an entry swapped or edited) returned the stale index built from the old
contents.
The fix validates the cache with a content signature (a JSON-framed list of each
episode's
titleandstreamUrl) instead of length. Any content or orderingchange rebuilds the index; an unchanged list keeps reusing it across keystrokes,
preserving the index-reuse optimization from #149.
streamUrlis folded in sothe cache is keyed on episode identity rather than coupling correctness to
fuse.js's internal live-reference behavior.
Tests
feedParser.test.ts: thegetEpisodestests now assert a single fetch(
toHaveBeenCalledTimes(1)); a sharedfeedResponsehelper replaces theduplicated mock setup that previously implied two fetches.
searchEpisodes.test.ts(new): covers the fuzzy match, empty/whitespaceshort-circuits, cache reuse for an unchanged list, fresh results after an
in-place same-length mutation, and a rebuild on a same-title identity swap. The
cache tests use a Fuse-construction spy and were verified to fail against the
old implementation.
Gates
lint,format:check,typecheck,build, and the unit suite pass. Note: thepre-existing
PodcastView.integration.test.ts > reopening a show reuses full in-memory cache without refetchingtest is flaky onmaster(timing-based,fails intermittently without these changes); it mocks
getEpisodesand does notexercise the code touched here.
Scope is limited to
src/parser/feedParser.tsandsrc/utility/searchEpisodes.ts(plus their tests).