• Move search form to a dedicated page
• Add Readeck integration
• Add feed option to disable HTTP/2 to avoid fingerprinting
• Add Enter key as a hotkey to open selected item
• Proxify video element poster attribute
• Add a couple of new possible locations for feeds
  • Hugo likes to generate index.xml
  • feed.atom and feed.rss are used by enterprise-scale/old-school gigantic CMS
• Fix categories import from Thunderbird's OPML
• Fix logo misalignment when using languages that are more verbose than English
• Google Reader: Do not return a 500 error when no items is returned
• Handle RDF feeds with duplicated <title> elements
• Sort integrations alphabetically
• Add more URL validation in media proxy
• Add unit test to ensure each translation has the correct number of plurals
• Add missing plurals for some languages
• Makefile: quiet git describe and rev-parse stderr: When building from a tarball instead of a cloned git repo, there would be two fatal: not a git repository errors emitted even though the build succeeds. This is because of how VERSION and COMMIT are set in the Makefile. This PR suppresses the stderr for these variable assignments.
• Makefile: do not force CGO_ENABLED=0 for miniflux target
• Add GitHub Action pipeline to build packages on-demand
• Remove Golint (deprecated), use staticcheck and golangci-lint instead
• Build amd64/arm64 Debian packages with CGO disabled
• Update go.mod and add .exe suffix to Windows binary
• Add a couple of fuzzers
• Fix CodeQL workflow
• Code and performance improvements:
  • Use an io.ReadSeeker instead of an io.Reader to parse feeds
  • Speed up the sanitizer:
    • Allow Youtube URLs to start with www
    • Use strings.Builder instead of a bytes.Buffer
    • Use a strings.NewReader instead of a bytes.NewBufferString
    • Sprinkles a couple of continue to make the code-flow more obvious
    • Inline calls to inList, and put their parameters in the right order
    • Simplify isPixelTracker
    • Simplify isValidIframeSource, by extracting the hostname and comparing it directly, instead of using the full url and checking if it starts with multiple variations of the same one (//, http:, https:// multiplied by /www.)
    • Add a benchmark
    • Instead of having to allocate a ~100 keys map containing possibly dynamic values (at least to the go compiler), allocate it once in a global variable. This significantly speeds things up, by reducing the garbage
    • Use constant time access for maps instead of iterating on them
    • Build a ~large whitelist map inline instead of constructing it item by item (and remove a duplicate key/value pair)
    • Use slices instead of hand-rolled loops
      collector/allocator involvements.
  • Reuse a Reader instead of copying to a buffer when parsing an Atom feed
  • Preallocate memory when exporting to OPML: This should marginally increase performance when exporting a large amount of feeds to OPML
  • Delay call of view.New after logging the user in: There is no need to do extra work like creating a session and its associated view until the user has been properly identified and as many possibly-failing sql request have been successfully run
  • Use constant-time comparison for anti-csrf tokens: This is probably completely overkill, but since anti-csrf tokens are secrets, they should be compared against untrusted inputs in constant time
  • Simplify and optimize genericProxyRewriter
    • Reduce the amount of nested loops: it's preferable to search the whole page once and filter on it (even with filters that should always be false), than searching it again for every element we're looking for.
    • Factorize the proxying conditions into a shouldProxy function to reduce the copy-pasta.
  • Speed up removeUnlikelyCandidates: .Not returns a brand new Selection, copied element by element
  • Improve EstimateReadingTime's speed by a factor 7
    • Refactorise the tests and add some
    • Use 250 signs instead of the whole text
    • Only check for Korean, Chinese and Japanese script
    • Add a benchmark
    • Use a more idiomatic control flow
  • Don't compute reading-time when unused: If the user doesn't display reading times, there is no need to compute them. This should speed things up a bit, since whatlanggo.Detect is abysmally slow.
  • Simplify username generation for the integration tests: No need to generate random numbers 10 times, generate a single big-enough one. A single int64 should be more than enough
  • Add missing regex anchor detected by CodeQL
  • Don't mix up slices capacity and length
  • Use prepared statements for intervals, ArchiveEntries and updateEnclosures
  • Use modern for-loops introduced with Go 1.22
  • Remove a superfluous condition: No need to check if the length of line is positive since we're checking afterwards that it contains the = sign
  • Close resources as soon as possible, instead of using defer() in a loop
  • Remove superfluous escaping in a regex
  • Use strings.ReplaceAll instead of strings.Replace(…, -1)
  • Use strings.EqualFold instead of strings.ToLower(…) ==
  • Use .WriteString( instead of .Write([]byte(…
  • Use %q instead of "%s"
  • Make internal/worker/worker.go read-only
  • Use a switch-case construct in internal/locale/plural.go instead of an avalanche of if
  • Template functions: simplify formatFileSize and duration implementation
  • Inline some templating functions
  • Make use of printer.Print when possible
  • Add a printer.Print to internal/locale/printer.go: No need to use variadic functions with string format interpolation to generate static strings
  • Minor code simplification in internal/ui/view/view.go: No need to create the map item by item when we can create it in one go
  • Build the map inline in CountAllFeeds(): No need to build an empty map to then add more fields in it one by one
  • Miscellaneous improvements to internal/reader/subscription/finder.go:
    • Surface localizedError in FindSubscriptionsFromWellKnownURLs via slog
    • Use an inline declaration for new subscriptions, like done elsewhere in the
      file, if only for consistency's sake
    • Preallocate the subscriptions slice when using an RSS-bridge,
  • Use an update-where for MarkCategoryAsRead instead of a subquery
  • Simplify CleanOldUserSessions' query: No need for a subquery, filtering on created_at directly is enough
  • Simplify cleanupEntries' query
    • NOT (hash=ANY(%4)) can be expressed as hash NOT IN $4
    • There is no need for a subquery operating on the same table, moving the conditions out is equivalent.
  • Reformat ArchiveEntries's query for consistency's sake and replace the =ANY with an IN
  • Reformat the query in GetEntryIDs and GetReadTime's query for consistency's sake
  • Simplify WeeklyFeedEntryCount: No need for a BETWEEN: we want to filter on entries published in the last week, no need to express is as "entries published between now and last week", "entries published after last week" is enough
  • Add some tests for add_image_title
  • Remove github.com/google/uuid dependencies: Replace it with a hand-rolled implementation. Heck, an UUID isn't even a requirement according to Omnivore API docs
  • Simplify internal/reader/icon/finder.go:
    • Use a simple regex to parse data uri instead of a hand-rolled parser, and document what fields are considered mandatory.
    • Use case-insensitive matching to find (fav)icons, instead of doing the same query twice with different letter cases
    • Add apple-touch-icon-precomposed.png as a fallback favicon
    • Reorder the queries to have icon first, since it seems to be the most popular one. It used to be last, meaning that pages had to be parsed completely 4 times, instead of one now.
    • Minor factorisation in findIconURLsFromHTMLDocument
  • Small refactoring of internal/reader/date/parser.go:
    • Split dates formats into those that require local times and those who don't, so that there is no need to have a switch-case in the for loop with around 250 iterations at most.
    • Be more strict when it comes to timezones, previously invalid ones like -13 were accepted. Also add a test for this.
    • Bail out early if the date is an empty string.
  • Make use of Go ≥ 1.21 slices package instead of hand-rolled loops
  • Reorder the fields of the Entry struct to save some memory
• Dependencies update:
  • Bump golang.org/x/oauth2 from 0.17.0 to 0.18.0
  • Bump github.com/prometheus/client_golang from 1.18.0 to 1.19.0
  • Bump github.com/tdewolff/minify/v2 from 2.20.16 to 2.20.18
  • Bump github.com/PuerkitoBio/goquery from 1.8.1 to 1.9.1
  • Bump golang.org/x/crypto from 0.19.0 to 0.20.0
  • Bump github.com/go-jose/go-jose/v3 from 3.0.1 to 3.0.3