Setting only padding-left made the logo left-shifted on mobile page. Margin-right was old setting for adjusting clearance between old logo and letter "Mastodon".
* Allow domain blocks to reject media without silencing or suspending
* Fix typo
* Hide 'Reject media' button when superfluous, instead of disabling it
* Properly hide 'reject media' checkbox on page load if needed
This may happen when resubmitting the domain block form after an error.
* Don't ask whether undoing a media-only block should be retroactive
* Rename :media_only block to :noop
* Display :noop block as None in frontend, allow blocks that do nothing
* Remove 'coding' line auto-added by emacs
Change the placeholder used in the content warning field from "Content Warning" to "Write your warning here". This change should made it easier to understand what the field is about.
Change le message de substitution utilisé dans le champ d’avertissement de « Avertissement » à « Écrivez ici votre avertissement ». Ce changement devrait rendre plus évidente la fonction du champ.
* add fr-FR locales to the landing page
* moar french locales
* terms of service/privacy policy
* remove un-used locales in the client (yarn manage:translations)
* update french locales in the client
* remove duplicate locales in fr.yml
* fix typos per PR comments in fr.yml
* put back default messages
* translate untranslated keys on the client
* add "push" after notifications
* correctly ident ToS in fr.yml
This PR adds a new notification cleaning mode, super perfectly tuned for accessibility, and removes the previous notification cleaning functionality as it's now redundant.
* w.i.p. notif clearing mode
* Better CSS for selected notification and shorter text if Stretch is off
* wip for rebase ~
* all working in notif clearing mode, except the actual removal
* bulk delete route for piggo
* cleaning + refactor. endpoint gives 422 for some reason
* formatting
* use the right route
* fix broken destroy_multiple
* load more notifs after succ cleaning
* satisfy eslint
* Removed CSS for the old notif delete button
* Tabindex=0 is mandatory
In order to make it possible to tab to this element you must have tab index = 0. Removing this violates WCAG and makes it impossible to use the interface without good eyesight and a mouse. So nobody with certain mobility impairments, vision impairments, or brain injuries would be able to use this feature if you don't have tabindex=0
* Corrected aria-label
Previous label implied a different behavior from what actually happens
* aria role localization & made the overlay behave like a checkbox
* checkboxes css and better contrast
* color tuning for the notif overlay
* fanceh checkboxes etc and nice backgrounds
* SHUT UP TRAVIS
* Use the same emoji data on the frontend and backend
* Move emoji.json to repository, add tests
This way you don't need to install node dependencies if you only
want to run Ruby code
* feat(compose): More space on mobile devices
* feat(compose): Hide navigation when typing on mobile devices
* fix(compose): Make animation faster
* fix(navigation_bar): Remove hardcoded title
* fix(compose): Prevent accidental bluring
* fix(compose): Increase max-height to 600px
* Do not raise unretryable exceptions in ResolveRemoteAccountService
* Removed fatal exceptions from ResolveRemoteAccountService
Exceptions that cannot be retried should not be raised. New exception
class for those that can be retried (Mastodon::UnexpectedResponseError)
* Refactor ResolveRemoteAccountService
* Remove trailing whitespace
* Use redis locks around critical ResolveRemoteAccountService code
* Add test for race condition of lock
* Wrap methods of ProcessFeedService::ProcessEntry in classes
This is a change same with 002ed7dc62, except
that it has the following changes:
* Revert irrelevant change in find_or_create_conversation
* Fix error handling for RemoteActivity
* Introduce Ostatus name space
app/javascript/mastodon/main.js delayed the execution of modules,
but other entry points didn't. That leads to failure in executing
modules, which requires those polyfills.
Strictly enforce the rule to require any modules after loading
polyfill in entry points.
Though size of extended-video-player is already fixed to 80vw*80vh in components.scss, player size was also set to original video size in extended_video_player.js. Video size is fixed to 80vw*80vh, so video player's size must also be fixed to 80vw*80vh.
* Shrink too wide single image modal
Fix too wide react-swipeable-view-container
Fix 0067f80 shrinking all react-swipeable-view-container
Change to apply max-width of react-swipeable-view-container only under media-modal.
Fix b30b03b just a typo
* Centering contents in image-loader
Centering small img, canvas, and video in image-loader.
* Locale script now accepts overrides and new keys from glitch/locales
* Revert glitchsoc changes to mastodon/locales to prevent future merge conflicts
* Add dependency on idn-ruby to speed up URI normalization
* Use normalized_host instead of normalize.host when applicable
When we are only interested in the normalized host, calling normalized_host
avoids normalizing the other components of the URI as well as creating a
new object
* Improve webfinger templates and make tests more flexible
* Clean up AS2 representation of actor
* Refactor outbox
* Create activities representation
* Add representations of followers/following collections, do not redirect /users/:username route if format is empty
* Remove unused translations
* ActivityPub endpoint for single statuses, add ActivityPub::TagManager for better
URL/URI generation
* Add ActivityPub::TagManager#to
* Represent all attachments as Document instead of Image/Video specifically
(Because for remote ones we may not know for sure)
Add mentions and hashtags representation to AP notes
* Add AP-resolvable hashtag URIs
* Use ActiveModelSerializers for ActivityPub
* Clean up unused translations
* Separate route for object and activity
* Adjust cc/to matrices
* Add to/cc to activities, ensure announce activity embeds target status and
not the wrapper status, add "id" to all collections
* Add Request class with HTTP signature generator
Spec: https://tools.ietf.org/html/draft-cavage-http-signatures-06
* Add HTTP signature verification concern
* Add test for SignatureVerification concern
* Add basic test for Request class
* Make PuSH subscribe/unsubscribe requests use new Request class
Accidentally fix lease_seconds not being set and sent properly, and
change the new minimum subscription duration to 1 day
* Make all PuSH workers use new Request class
* Make Salmon sender use new Request class
* Make FetchLinkService use new Request class
* Make FetchAtomService use the new Request class
* Make Remotable use the new Request class
* Make ResolveRemoteAccountService use the new Request class
* Add more tests
* Allow +-30 seconds window for signed request to remain valid
* Disable time window validation for signed requests, restore 7 days
as PuSH subscription duration (which was previous default due to a bug)
* chore(yarn): Install babel-plugin-preval as development dependency
* feat(babel): Add preval as a plugin
* feat(emojione_light): Prevaled module what tree-shaked emojione
* refactor(emoji): Use emojione_light
* feat: Preload emojione_picker bundle
* fix(emojione_light): Do not use Object.entries
* fix(emojify): Update tests
* chore(emojione_light): Remove silly ascii art
- Use unicode when selecting emoji through picker
- Convert shortcodes to unicode when storing text input server-side
- Do not convert shortcodes in JS anymore
Render function for BundleContainer must not be methods.
React doesn't know dependency of the method, so they won't rerender on property updates.
In this case, when you close modal and open another modal immediately,
old modal will be open instead of new one.
* First lame pass at adding optional git commit hash display on /about/more page.
Currently, this is implemented by checking for the existence of a file called CURRENT_RELEASE in the home directory of the user running Mastodon. If the file exists, its contents are added.
I've modified my update process to include the following before precompiling assets:
git log -1 | head -n 1 | cut -d " " -f2 > ~/CURRENT_RELEASE
That puts the current commit hash into the file ~/CURRENT_RELEASE, but you figured that out because you're a smart cookie.
As I am quite sure this is a horrible methodology for implementing this, I look forward to any improvements you have to offer!
* Updated to handle instances that share a user - the CURRENT_RELEASE file now lives in the instance's base directory.
This also requires modifying the update hook to `git log -1 | head -n 1 | cut -d " " -f2 > CURRENT_RELEASE`
* Redesign the landing page, mount public timeline on it
* Adjust the standalone mounted component to the lacking of router
* Adjust auth layout pages to new design
* Fix tests
* Standalone public timeline polling every 5 seconds
* Remove now obsolete translations
* Add responsive design for new landing page
* Address reviews
* Add floating clouds behind frontpage form
* Use access token from public page when available
* Fix mentions and hashtags links, cursor on status content in standalone mode
* Add footer link to source code
* Fix errors on pages that don't embed the component, use classnames
* Fix tests
* Change anonymous autoPlayGif default to false
* When gif autoplay is disabled, hover to play
* Add option to hide the timeline preview
* Slightly improve alt layout
* Add elephant friend to new frontpage
* Display "back to mastodon" in place of "login" when logged in on frontpage
* Change polling time to 3s
* Added a success page to remote following
Includes follow-through links to web (the old redirect target) and back to the remote user's profile
* Use Account.new in spec instead of a fake with only id
(fixes spec)
* Fabricate(:account) over Account.new
* Remove self from the success text
(and all HTML with it)
Since there is little point in retrying so often when a service is down
or does not exist anymore. Subscriptions are renewed 1 day before they
should expire, so retrying in 30 minutes, then 2 hours, then 12 hours
is fine. If even after that, the remote server does not work, there is
little sense in retrying more often than once a day
Also, uniqueness of the job should ensure that failed retries will
not result in multiple retries for the same endpoint when the next
resubscription cycle comes
