Miniflux - Package Updates
-
[1.6.1]
- Update v2 to 2.2.7
- Full Changelog
- refactor: combine feed icon handlers to use only
externalIconID - fix(ui): update share feature to correctly select the title element and handle empty titles
- fix(ui): update entry tags display logic to show links based on user authentication
- fix(ui): remove touch-action style to prevent horizontal scrolling issues
- fix(ui): log a warning for an empty client secret
- fix(ui): change labels from "Read / Unread" to "Mark as Read"
- fix(ui): avoid 500 errors and NaN when marking a deleted entry as read
- fix(subscription): add
/rss/feed.xmlto the list of known feed URLs
-
[1.6.2]
- Update v2 to 2.2.8
- Full Changelog
- fix(api):
hide_globallycategories field should be a boolean - fix(ui): add missing
awaitwhen callingnavigator.share()method - fix(ui): replace share link with a form button for better accessibility
- feat(telegrambot): replace "Go to website" button with "Go to Miniflux"
- feat(locale): update Polish translation
- feat(locale): update German translation
- feat(locale): update Chinese translation
- feat(config): add
SCHEDULER_ROUND_ROBIN_MAX_INTERVALoption - feat(cli): add
-reset-feed-next-check-atargument - feat(api): add
update_contentquery parameter to/entries/{entryID}/fetch-contentendpoint
-
[1.6.3]
- Update v2 to 2.2.9
- Full Changelog
- fix(webauthn): correct argument in debug log
- fix(sanitizer): MathML tags are not fully supported by
golang.org/x/net/html - fix(migrations): prevent failure at version 45 with long entry URLs
- fix(locale): localize Git commit label in about page
- fix(googlereader): return a 400 instead of 500 for invalid edit requests
- fix(googlereader): handle various item ID formats
- fix(googlereader): avoid panic for inexisting feed or category
- fix(googlereader):
/items/contentsshould accept short form item IDs - feat(webauthn): prefer creation of a client-side discoverable credential
- feat(urlcleaner): remove the
refparameter from url
-
[1.6.4]
- Update v2 to 2.2.10
- Full Changelog
- test(sanitizer): add unit test for 0x0 pixel tracker
- fix(readability): do not remove elements within code blocks
- fix(karakeep): correct method name and improve error handling in
SaveURL - fix(filter): skip invalid rules instead of exiting the loop
- feat(ui): display external link in single entry view because the URL was not visible on mobile (no mouse over)
- feat(ui): avoid showing an excessive number of tags
- feat(ui): add user setting to control
target="_blank"on links - feat(sanitizer): validate MathML XML namespace
- feat(sanitizer): consider images of size 0x0 as pixel trackers
- feat(sanitizer): add validation for empty
widthandheightattributes in img tags
-
[1.6.5]
- Update v2 to 2.2.11
- Full Changelog
- TLS support for Unix sockets: Miniflux can now serve TLS over Unix domain sockets using CERT_FILE and KEY_FILE (#fcf86e3).
- RSS fallback: If a feed entry has no URL, Miniflux now uses the enclosure URL as a fallback (#d9de9d1).
- Bearer token for Linkwarden: The Linkwarden integration now uses Bearer token authorization instead of cookies (#1d11623).
- Cookie policy improvement: SameSiteStrictMode is enforced for cookies when OAuth2/OIDC is not used (#135ce1d).
- Readability engine: Avoid removing elements with the content class during readability parsing (#66b269e).
- Fixed an issue with feeds having excessive leading whitespace causing parser buffer issues (#54abd0a).
- Properly preserve UTF-8 when truncating strings for full-text search (#703f113).
- Fixed logic error in enclosure type detection (#50d5cb9).
- Fixed incorrect filter rule parsing of Windows-style newlines (#dc81725).
- Fixed a panic in startAutoCertTLSServer function when using Let's Encrypt automatic certificates (#f7a6b02)
-
[1.6.6]
- Update v2 to 2.2.12
- Full Changelog
- Keep only metadata of removed entries to reduce database size.
- Removed entry status is now immutable and cannot be changed back to
unreadorreadstatus. - SVG favicons are now minified before storing them in the database.
- Added support for resizing WebP images.
- Main menu now includes icons.
- Added Progressive Web App (PWA) shortcuts for quick access to common actions.
- Added direct link to the Apache 2.0 license on the About page.
- Feed-level webhook URLs now take priority when saving entries.
- New option:
POLLING_LIMIT_PER_HOSTto limit concurrent requests per host. - Added a rewrite rule to remove useless heading images on Phoronix articles.
-
[1.6.7]
- Update v2 to 2.2.13
- Full Changelog
- UI: Added a "Back to top" link for easier navigation.
- Integrations: Added support for Wallabag tags.
- Added support for LinkTaco service to save articles.
- API: Fixed issue where
removedentries could still be returned. It was generating an error when fetching entries. - Icons: Improved handling of relative icon URLs within subfolders.
- Timezone: Ensured only current IANA time zones are used. This avoids issues with Debian Trixie where deprecated time zones have been removed.
-
[1.6.8]
- Update v2 to 2.2.14
- Full Changelog
- Go Client: Allow passing a custom
http.Clientand add context support to API methods. - UI: Redirect users back to the original page after logging in.
- Template: Improved Content Security Policy: extracted CSP generation into a function, added systematic nonces, and changed
default-srcto'none'for stronger security. - Integrations: Added tags option for the Karakeep integration.
- Integrations: Added new Archive.org integration.
- Rewrite Rules: Added
remove_img_blur_paramsrule. - Rewrite Rules: Added
add_image_titlerule for explainxkcd.com. - Fixed CSS layout overflow when external links are too long.
- Fixed JSON Feed parser to fallback to
external_urlwhenurlis missing. - Updated scraper rule for Dark Reading.
-
[1.6.9]
- Update v2 to 2.2.15
- Full Changelog
- New configuration option to disable the Miniflux API
- Added option to save entries to a specific Linkwarden collection
- YouTube subscription improvements:
- Allow feed entries with
<i>and<small>tags - URL Cleaner: Remove additional trackers from URLs
- YouTube embeds: Avoid Error 153 (
video player configuration error) in various scenarios - API:
fetchContentendpoint now properly rewrites media URLs when using the media proxy - Security: Only relative paths are now allowed for the
redirectURLparameter
-
[1.6.10]
- Update v2 to 2.2.16
- Full Changelog
- Disallow the media proxy from fetching resources on private networks to mitigate potential SSRF issues. This behavior is configurable at the instance level.
- Disallow fetching feed icons from private networks to reduce the SSRF attack surface. This is also configurable at the instance level.
- Add the
TRUSTED_REVERSE_PROXY_NETWORKSconfiguration option to prevent spoofing of HTTP headers such asX-Forwarded-For,X-Forwarded-Proto, andX-Real-Ip. This option must be configured whenAUTH_PROXY_HEADERis enabled. - Stop logging generated Google Reader API tokens, even when debug mode is enabled.
- Remove the CORS handler from the Google Reader API, as it is not intended to be used by web clients, reducing the overall attack surface.
- Avoid indexing the content of removed entries, significantly reducing database index size after cleanup.
- Add a new API endpoint to import entries into an existing feed.
- Execute the content sanitizer when updating or importing entries through the API to ensure consistent sanitization.
- Improve Google Reader API compatibility by removing unnecessary output parameter checks and aligning behavior with other open-source RSS readers.
- Add smooth page transitions for a more polished navigation experience.