1. 25 Jan, 2019 2 commits
  2. 24 Jan, 2019 16 commits
  3. 23 Jan, 2019 2 commits
  4. 22 Jan, 2019 20 commits
    • Renon's avatar
      fix apache conf file for 0.18 · 8af6b6ce
      Renon authored
    • Eliot Berriot's avatar
    • Jo Vuit's avatar
      Typo · 6f6f6ac2
      Jo Vuit authored
    • Jo Vuit's avatar
      Typo · ed4c3604
      Jo Vuit authored
    • Jo Vuit's avatar
      Missing word · 47ee8c99
      Jo Vuit authored
    • Jo Vuit's avatar
      Typos · e6995dc4
      Jo Vuit authored
    • Jo Vuit's avatar
      Typo · 7488957c
      Jo Vuit authored
    • Jo Vuit's avatar
      Typo · e28f0a89
      Jo Vuit authored
    • Eliot Berriot's avatar
      Merge branch 'master' into develop · cd698d2e
      Eliot Berriot authored
    • Eliot Berriot's avatar
      Fixed missing contributors thanks · d61208f1
      Eliot Berriot authored
    • Eliot Berriot's avatar
      Merge tag '0.18' into develop · 543ae370
      Eliot Berriot authored
      0.18 "Naomi" (2019-01-22)
      This release is dedicated to Naomi, an early contributor and beta tester of Funkwhale.
      Her positivity, love and support have been incredibly helpful and helped shape the project
      as you can enjoy it today. Thank you so much Naomi <3
      Upgrade instructions are available at
      https://docs.funkwhale.audio/index.html, ensure you also execute the intructions
      marked with ``[manual action required]`` and ``[manual action suggested]``.
      See ``Full changelog`` below for an exhaustive list of changes!
      Audio transcoding is back!
      After removal of our first, buggy transcoding implementation, we're proud to announce
      that this feature is back. It is enabled by default, and can be configured/disabled
      in your instance settings!
      This feature works in the browser, with federated/non-federated tracks and using Subsonic clients.
      Transcoded tracks are generated on the fly, and cached for a configurable amount of time,
      to reduce the load on the server.
      Licensing and copyright information
      Funkwhale is now able to parse copyright and license data from file and store
      this information. Apart from displaying it on each track detail page,
      no additional behaviour is currently implemented to use this new data, but this
      will change in future releases.
      License and copyright data is also broadcasted over federation.
      License matching is done on the content of the ``License`` tag in the files,
      with a fallback on the ``Copyright`` tag.
      Funkwhale will successfully extract licensing data for the following licenses:
      - Creative Commons 0 (Public Domain)
      - Creative Commons 1.0 (All declinations)
      - Creative Commons 2.0 (All declinations)
      - Creative Commons 2.5 (All declinations and countries)
      - Creative Commons 3.0 (All declinations and countries)
      - Creative Commons 4.0 (All declinations)
      Support for other licenses such as Art Libre or WTFPL will be added in future releases.
      Instance-level moderation tools
      This release includes a first set of moderation tools that will give more control
      to admins about the way their instance federate with other instance and accounts on the network.
      Using these tools, it's now possible to:
      - Browse known accounts and domains, and associated data (storage size, software version, etc.)
      - Purge data belonging to given accounts and domains
      - Block or partially restrict interactions with any account or domain
      All those features are usable using a brand new "moderation" permission, meaning
      you can appoints one or nultiple moderators to help with this task.
      I'd like to thank all Mastodon contributors, because some of the these tools are heavily
      inspired from what's being done in Mastodon. Thank you so much!
      Iframe widget to embed public tracks and albums [manual action required]
      Funkwhale now support embedding a lightweight audio player on external websites
      for album and tracks that are available in public libraries. Important pages,
      such as artist, album and track pages also include OpenGraph tags that will
      enable previews on compatible apps (like sharing a Funkwhale track link on Mastodon
      or Twitter).
      To achieve that, we had to tweak the way Funkwhale front-end is served. You'll have
      to modify your nginx configuration when upgrading to keep your instance working.
      **On docker setups**, edit your ``/srv/funkwhale/nginx/funkwhale.template`` and replace
      the ``location /api/`` and `location /` blocks by the following snippets::
          location / {
              include /etc/nginx/funkwhale_proxy.conf;
              # this is needed if you have file import via upload enabled
              client_max_body_size ${NGINX_MAX_BODY_SIZE};
              proxy_pass   http://funkwhale-api/;
          location /front/ {
              alias /frontend/;
      The change of configuration will be picked when restarting your nginx container.
      **On non-docker setups**, edit your ``/etc/nginx/sites-available/funkwhale.conf`` file,
      and replace the ``location /api/`` and `location /` blocks by the following snippets::
          location / {
              include /etc/nginx/funkwhale_proxy.conf;
              # this is needed if you have file import via upload enabled
              client_max_body_size ${NGINX_MAX_BODY_SIZE};
              proxy_pass   http://funkwhale-api/;
          location /front/ {
              alias ${FUNKWHALE_FRONTEND_PATH}/;
      Replace ``${FUNKWHALE_FRONTEND_PATH}`` by the corresponding variable from your .env file,
      which should be ``/srv/funkwhale/front/dist`` by default, then reload your nginx process with
      ``sudo systemctl reload nginx``.
      Alternative docker deployment method
      Thanks to the awesome done by @thetarkus at https://github.com/thetarkus/docker-funkwhale,
      we're now able to provide an alternative and easier Docker deployment method!
      In contrast with our current, multi-container offer, this method integrates
      all Funkwhale processes and services (database, redis, etc.) into a single, easier to deploy container.
      Both method will coexist in parallel, as each one has pros and cons. You can learn more
      about this exciting new deployment option by visiting https://docs.funkwhale.audio/installation/docker.html!
      Automatically load .env file
      On non-docker deployments, earlier versions required you to source
      the config/.env file before launching any Funkwhale command, with ``export $(cat config/.env | grep -v ^# | xargs)``
      This led to more complex and error prode deployment / setup.
      This is not the case anymore, and Funkwhale will automatically load this file if it's available.
      Delete pre 0.17 federated tracks [manual action suggested]
      If you were using Funkwhale before the 0.17 release and federated with other instances,
      it's possible that you still have some unplayable federated files in the database.
      To purge the database of those entries, you can run the following command:
      On docker setups::
          docker-compose run --rm api python manage.py script delete_pre_017_federated_uploads --no-input
      On non-docker setups::
          python manage.py script delete_pre_017_federated_uploads --no-input
      Enable gzip compression [manual action suggested]
      Gzip compression will be enabled on new instances by default
      and will reduce the amount of bandwidth consumed by your instance.
      If you with to benefit from gzip compression on your instance,
      edit your reverse proxy virtualhost file (located at ``/etc/nginx/sites-available/funkwhale.conf``) and add the following snippet
      in the server block, then reload your nginx server::
          server {
              # ... exiting configuration
              # compression settings
              gzip on;
              gzip_comp_level    5;
              gzip_min_length    256;
              gzip_proxied       any;
              gzip_vary          on;
              # end of compression settings
      Full changelog
      - Allow embedding of albums and tracks available in public libraries via an <iframe> (#578)
      - Audio transcoding is back! (#272)
      - First set of instance level moderation tools (#580, !521)
      - Store licensing and copyright information from file metadata, if available (#308)
      - Add UI elements for multi-disc albums (#631)
      - Added alternative funkwhale/all-in-one docker image (#614)
      - Broadcast library updates (name, description, visibility) over federation
      - Based Docker image on alpine to have a smaller (and faster to build) image
      - Improved front-end performance by stripping unused dependencies, reducing bundle size
        and enabling gzip compression
      - Improved accessibility by using main/section/nav tags and aria-labels in most critical places (#612)
      - The progress bar in the player now display loading state / buffer loading (#586)
      - Added "type: funkwhale" and "funkwhale-version" in Subsonic responses (#573)
      - Documented keyboard shortcuts, list is now available by pressing "h" or in the footer (#611)
      - Documented which Subsonic endpoints are implemented (#575)
      - Hide invitation code field during signup when it's not required (#410)
      - Importer will now pick embedded images in files with OTHER type if no COVER_FRONT is present
      - Improved keyboard accessibility on player, queue and various controls (#576)
      - Improved performance when listing playable tracks, albums and artists
      - Increased default upload limit from 30 to 100MB (#654)
      - Load env file in config/.env automatically to avoid sourcing it by hand (#626)
      - More resilient date parsing during audio import, will not crash anymore on
        invalid dates (#622)
      - Now start radios immediatly, skipping any existing tracks in queue (#585)
      - Officially support connecting to a password protected redis server, with
        the redis://:password@localhost:6379/0 scheme (#640)
      - Performance improvement when fetching favorites, down to a single, small http request
      - Removed "Activity" page, since all the data is available on the "Browse" page (#600)
      - Removed the need to specify the DJANGO_ALLOWED_HOSTS variable
      - Restructured the footer, added useful links and removed unused content
      - Show short entries first in search results to improve UX
      - Store disc number and order tracks by disc number / position) (#507)
      - Strip EXIF metadata from uploaded avatars to avoid leaking private data (#374)
      - Support blind key rotation in HTTP Signatures (#658)
      - Support setting a server URL in settings.json (#650)
      - Updated default docker postgres version from 9.4 to 11 (#656)
      - Updated lots of dependencies (especially django 2.0->2.1), and removed unused dependencies (#657)
      - Improved test suite speed by reducing / disabling expensive operations (#648)
      - Fixed parsing of embedded file cover for ogg files tagged with MusicBrainz (#469)
      - Upgraded core dependencies to fix websocket/messaging issues and possible memory leaks (#643)
      - Fix ".None" extension when downloading Flac file (#473)
      - Fixed None extension when downloading an in-place imported file (#621)
      - Added a script to prune pre 0.17 federated tracks (#564)
      - Advertise public libraries properly in ActivityPub representations (#553)
      - Allow opus file upload (#598)
      - Do not display "view on MusicBrainz" button if we miss the mbid (#422)
      - Do not try to create unaccent extension if it's already present (#663)
      - Ensure admin links in sidebar are displayed for users with relavant permissions, and only them (#597)
      - Fix broken websocket connexion under Chrome browser (#589)
      - Fix play button not starting playback with empty queue (#632)
      - Fixed a styling inconsistency on about page when instance description was missing (#659)
      - Fixed a UI discrepency in playlist tracks count (#647)
      - Fixed greyed tracks in radio builder and detail page (#637)
      - Fixed inconsistencies in subsonic error responses (#616)
      - Fixed incorrect icon for "next track" in player control (#613)
      - Fixed malformed search string when redirecting to LyricsWiki (#608)
      - Fixed missing track count on various library cards (#581)
      - Fixed skipped track when appending multiple tracks to the queue under certain conditions (#209)
      - Fixed wrong album/track count on artist page (#599)
      - Hide unplayable/emtpy playlists in "Browse playlist" pages (#424)
      - Initial UI render using correct language from browser (#644)
      - Invalid URI for reverse proxy websocket with apache (#617)
      - Properly encode Wikipedia and lyrics search urls (#470)
      - Refresh profile after user settings update to avoid cache issues (#606)
      - Use role=button instead of empty links for player controls (#610)
      - Deploy documentation from the master branch instead of the develop branch to avoid inconsistencies (#642)
      - Document how to find and use library id when importing files in CLI (#562)
      - Fix documentation typos (#645)
    • Eliot Berriot's avatar
      Merge branch 'release/0.18' · 338e1a85
      Eliot Berriot authored
    • Eliot Berriot's avatar
      Version bump and changelog · f44d2b06
      Eliot Berriot authored
    • Eliot Berriot's avatar
      Merge branch 'master' into develop · af2b4965
      Eliot Berriot authored
    • Eliot Berriot's avatar
      Merge branch 'test-install-docs' into 'develop' · 924f44c6
      Eliot Berriot authored
      Test install docs for 0.18
      See merge request funkwhale/funkwhale!551
    • Eliot Berriot's avatar
    • Eliot Berriot's avatar
      Disabled http2 in nginx template · 635865ac
      Eliot Berriot authored
    • Eliot Berriot's avatar
    • Eliot Berriot's avatar
    • Eliot Berriot's avatar