Conflicts:
- `README.md`:
Discarded upstream changes: we have our own README
- `app/controllers/follower_accounts_controller.rb`:
Port upstream's minor refactoring
Conflicts:
- `app/models/concerns/domain_materializable.rb`:
Fixed a code style issue upstream in a PR that got merged in glitch-soc
earlier.
Changed the code to match upstream's.
As pointed out by https://github.com/mastodon/mastodon/pull/21297#discussion_r1028372193
at least one of our dependencies already dropped support for Ruby 2.6, and we
had removed Ruby 2.6 tests from the CI over a year ago (#16861).
So stop advertising Ruby 2.6 support, bump targeted version, and drop some
compatibility code.
Conflicts:
- `config/initializers/content_security_policy.rb`:
Our config file is pretty different from upstream.
Upstream changed CSP directive `script-src` to include
`wasm-unsafe-eval` instead of `unsafe-eval`, which we
did not include.
Added `wasm-unsafe-eval` to `script-src` to fix
execution of the OCR web worker.
- `package.json`:
Upstream updated a dependency (`array-includes`) textually
adjacent to a glitch-soc-only dependency (`atrament`).
Updated `array-includes` as upstream did.
Conflicts:
- `app/views/admin/announcements/edit.html.haml`:
Upstream change too close to theming-related glitch-soc change.
Ported upstream changes.
- `app/views/admin/announcements/new.html.haml`
Upstream change too close to theming-related glitch-soc change.
Ported upstream changes.
Conflicts:
- `app/models/account.rb`:
Conflict because we (glitch-soc) have disabled trending of posts without
review.
Discarded that upstream change.
- `app/views/admin/settings/discovery/show.html.haml`:
Just an extra setting in glitch-soc.
Kept that extra setting.
Backport from vanilla: 36bc90e8aa
The Blurhash library used by Mastodon requires an input encoded as 24
bits raw RGB data. The conversion to raw RGB using Imagemagick did not
previously specify the desired bit depth. In some situations, this leads
Imagemagick to output in a pixel format using less bpp than expected.
This then manifested as segfaults of the Sidekiq process due to
out-of-bounds read, or potentially a (highly noisy) memory infoleak.
