* feat(workspaces): delete workspace emit event
* feat(workspaces): move workspace group metrics to the backend
* Removed FE mixpanel group update
* Remove fragment
* test(gatekeeper): add unittest to new gatekeeper service
---------
Co-authored-by: Mike Tasset <mike.tasset@gmail.com>
* feat(regions): repo functions for copying project branches and commits
* chore(regions): wire up move to resolver
* chore(regions): successful basic test of project region change
* fix(regions): sabrina carpenter please please please
* fix(regions): repair multiregion test setup
* chore(regions): appease ts
* chore(multiregion): update test multiregion config
* chore(multiregion): fix test docker config and test
* chore(multiregion): use transaction
* chore(multiregion): maybe this will work
* fix(multiregion): drop subs synchronously
* chore(multiregion): desperate test logs
* chore(multiregion): somehow that worked?
* chore(multiregion): add load-bearing log statement
* chore(multiregion): move services
* fix(multiregion): test drop waits
* chore(regions): fix import
* chore(regions): make test a bit more thorough for good measure
* fix(regions): speed up inserts
* fix(regions): ignore workspace conflict on move
* feat(viewer-lib): Added property to PointQueryResult which tells if the requested point is in frustum when projecting and unprojecting
* feat(frontend-2): Updated the comment bubbles screen location computation step to account for their anchor's inclusion in the camera's frustum
* New SelectMulti with no headless
* Use FormSelectMulti in developer settings
* Fix build
* isMultiItemSelected
* minor
* Tidy up FormSelectBase useage
* feat(viewer-lib): Updates to the PassReader extension along with underlying viewer library updates
Implemented reading framebuffer contents for framebuffers with multiple attachements. The current version of three.js that we are using does not support this. Because we still need to drag along WebGL 1.0 support, only attachement 0 can be read for now, which does not bother us.
DepthNormalPass now specifies it's MRT output target as the outputTarget
PassReader's read function is now overloaded and it can take a pass name as a string or a GPass | GPass[]
Had to add a small type augmentation since the current version of types-three library does a poor job when it comes to WebGLMultipleRenderTargets
Updated PassReader extension in frontend and updated the call to read in order to make sure depth reading works in other view modes that write depth. The only view mode that does not draw to depth is Shaded mode
* fix(viewer-lib): Fixed the classic sandbox compile error
* fix(frontend-2): Updated depth reading to work with MRT depth from our view mode pipelines
* Add ... to all suggested prompts
* Update copy and remove some icons
* Update spacings
* Increase gaps
* Some more polishing from call
* Add submit on enter
* Fix overflow
* Typo
* keyup > keypress
---------
Co-authored-by: Benjamin Ottensten <benjamin.ottensten@gmail.com>
Iain Sproat