Fixes#16515
Not using a router object somehow made `this.history` lag behind the real
browser history whenever pushing a new history item in `replyCompose`.
Not using the context-provided router in this case was an oversight made
when porting glitch-soc changes in #16499.
The auto-linking code basically rewrote the whole string escaping non-ascii
characters in an inefficient way, and building a full character offset map
between the unescaped and escaped texts before sending the contents to
TwitterText's extractor.
Instead of doing that, this commit changes the TwitterText regexps to include
valid IRI characters in addition to valid URI characters.
Conflicts:
- `CONTRIBUTING.md`:
Upstream file has changed and we're quoting it.
Ported the changes.
- `README.md`:
Upstream file has changed but we have a completely different one.
Kept our version.
- `lib/mastodon/version.rb`:
Upstream has changed from `tootsuite/mastodon` to `mastodon/mastodon`,
but we're still `glitch-soc/mastodon`.
Kept our version.
- `spec/presenters/instance_presenter_spec.rb`:
Upstream has changed from `tootsuite/mastodon` to `mastodon/mastodon`,
but we're still `glitch-soc/mastodon`.
Kept our version.
* Change references to tootsuite/mastodon to mastodon/mastodon
* Remove obsolete test fixture
* Replace occurrences of tootsuite/mastodon with mastodon/mastodon in CHANGELOG
And a few other places
* Refactor shouldUpdateScroll passing
So far, shouldUpdateScroll has been manually passed down from the very top of
the React component hierarchy even though it is a static function common to
all ScrollContainer instances, so replaced that with a custom class extending
ScrollContainer.
* Generalize “press back to close modal” to any modal and to public pages
* Fix boost confirmation modal closing media modal
This simplifies the logic to:
- when the last modal gets closed and we're in our history buffer state, go back
- whenever a modal is open, ensure we're in a history buffer state by
potentially pushing one
Conflicts:
- `app/serializers/rest/instance_serializer.rb`:
Upstream changed the fields returned by /api/v1/instance by adding a
`configuration` field holding a lot of useful information making our
`max_toot_chars` and `poll_limits` fields obsolete.
Keeping those around for now for compatibility.
- `app/validators/status_length_validator.rb`:
No real conflict, just URL_PLACEHOLDER_CHARS introduced too close to
MAX_CHARS which is defined differently in glitch-soc.
Ported upstream changes.
Conflicts:
- `app/helpers/accounts_helper.rb`:
Conflict due to upstream changing how followers count is displayed while we
have an option to hide followers count.
Ported upstream change.
- `app/views/accounts/_header.html.haml`:
Conflict due to upstream changing how followers count is displayed while we
have an option to hide followers count.
Ported upstream change.
- `app/views/directories/index.html.haml`:
Conflict due to upstream changing how followers count is displayed while we
have an option to hide followers count.
Ported upstream change.
3.4.1 dropped account_stats.lock_version, but in a way breaking the usual
upgrade path by requiring services to be reloaded after the post-migrations.
Indeed, `self.locking_column = nil` was not enough for Rails to ignore the
`lock_version` column when preparing statements on application load, resulting
in some ActiveRecord queries (typically those involving
`includes(:account_stat)`) erroring out with:
ActiveRecord::StatementInvalid (PG::UndefinedColumn: ERROR: column account_stats.lock_version does not exist
Always mark jemalloc needed if jemalloc is enabled by akihikodaki · Pull Request #4627 · ruby/ruby
https://github.com/ruby/ruby/pull/4627
> Symbols exported by jemalloc is referred by the shared library but not
> by the executables when building Ruby as a shared library with
> jemalloc. It causes shared libraries such as the GNU C++ library
> occasionally rely on the memory allocator provided by the standard C
> library. Worse, the resolved symbols can later be replaced with
> jemalloc, and jemalloc may see pointers from the standard C library,
> which results in various failures.
> e.g. https://github.com/tootsuite/mastodon/issues/15751
As a workaround, do not rely on jemalloc enablement of Ruby, and
preload libjemalloc.so instead.
* Fix anonymous access to outbox not being cached by the reverse proxy
Up until now, anonymous access to outbox was marked as public, but with a
0 duration for caching, which means remote proxies would only serve from cache
when the server was completely overwhelmed.
Changed that cache duration to one minute, so that repeated anonymous access
to one account's outbox can be appropriately cached.
Also added `Signature` to the `Vary` header in case a page is requested, so
that authenticated fetches are never served from cache (which only contains
public toots).
* Remove Vary: Accept header from webfinger controller
Indeed, we have stopped returning xrd, and only ever return jrd, so the
Accept request header does not matter anymore.
* Cache negative webfinger hits for 3 minutes
