- Remove standalone InboxService and move functionality to UserService
- Rename inbox to user notifications for better API consistency
- Add ListUserNotifications, UpdateUserNotification, DeleteUserNotification methods
- Update frontend components to use new notification endpoints
- Update store layer to support new notification model
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Remove extra blank line in memo_service.go (goimports)
- Remove invalid fields from CreateMemoRequest call (validateOnly, requestId)
- Clean up unnecessary comments
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
BREAKING CHANGE: Removed DeleteMemoTag and RenameMemoTag API endpoints
for better API consistency. Tags should now be managed by updating memo
content directly via UpdateMemo endpoint.
Backend changes:
- Remove RenameMemoTag and DeleteMemoTag RPC methods from proto
- Remove backend implementations in memo_service.go
- Regenerate protocol buffers (Go, TypeScript, OpenAPI)
Frontend changes:
- Remove RenameTagDialog component
- Simplify TagsSection to remove rename/delete functionality
- Improve tag styling with active state highlighting
- Add smooth transitions and better hover interactions
- Polish TagTree component for consistency
- Tags now only support click-to-filter (no inline editing)
Style improvements:
- Active tags highlighted with primary color and font-medium
- Consistent hover states across flat and tree views
- Better spacing and visual hierarchy
- Improved empty state styling
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Changed Cache-Control max-age from 7 days to 1 hour with immutable directive.
This prevents users from experiencing blank pages or JS errors when accessing
frequently redeployed instances (e.g., demo environments) where old cached
assets may reference files that no longer exist after redeployment.
Since Vite generates content-hashed filenames, the immutable directive prevents
unnecessary revalidation while the shorter cache duration ensures fresh assets
are served within an hour of redeployment.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add comprehensive inline documentation for auth services
- Document session-based and token-based authentication flows
- Clarify authentication priority and validation logic
- Add detailed comments for JWT token structure and claims
- Fix reactions seed data to use memo UIDs instead of numeric IDs
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Removed the wikilink extension from markdown services in test and API service.
- Deleted the DefaultLink and WikiLink components, simplifying link handling.
- Updated ConditionalComponent to remove wikilink checks.
- Adjusted MemoContent to exclude wikilink handling in markdown rendering.
- Refined markdown styles for compact rendering, enhancing readability.
- Added a Markdown Styling Guide to document the new compact styling approach.
- Removed the `nodes` field from the `Memo` interface in `memo_service.ts`.
- Updated the `createBaseMemo` function and the `Memo` message functions to reflect the removal of `nodes`.
- Cleaned up the serialization and deserialization logic accordingly.
chore: remove code-inspector-plugin from Vite configuration
- Deleted the `codeInspectorPlugin` from the Vite configuration in `vite.config.mts`.
- Simplified the plugins array to include only `react` and `tailwindcss`.
- Update GetUser to accept both numeric IDs and username strings (users/{id} or users/{username})
- Implement CEL filter parsing for username-based lookups
- Update proto documentation to reflect dual lookup capability
- Simplify frontend user store to use GetUser instead of ListUsers filter
- Update ListUsers filter documentation to show current capabilities
- Updated memo and reaction filtering logic to use a unified engine for compiling filter expressions into SQL statements.
- Removed redundant filter parsing and conversion code from ListMemoRelations, ListReactions, and ListAttachments methods.
- Introduced IDList and UIDList fields in FindMemo and FindReaction structs to support filtering by multiple IDs.
- Removed old filter test files for reactions and attachments, as the filtering logic has been centralized.
- Updated tests for memo filtering to reflect the new SQL statement compilation approach.
- Ensured that unsupported user filters return an error in ListUsers method.
- Bump protoc-gen-go version from v1.36.8 to v1.36.9 in multiple generated files.
- Enhance OpenAPI descriptions for various endpoints to improve clarity and readability.
- Fix typo in the description list handling in the markdown service.
- Update HTMLElement component to handle children and self-closing tags correctly.
- Modify types in the markdown service to include children and self-closing properties for HTML elements.
* feat: update ItalicNode to use repeated children instead of content
* chore: update gomark dependency to latest version
---------
Co-authored-by: Anubhav Singh <anubhav@mail.techindika.com>
This removes the content of the <title> element in the RSS feeds that Memo produces.
Why remove? Every RSS client I can find shows the <title> next to the <description> when viewing an item. This creates a duplicate (but often trimmed, so less useful) version of <description> right above the actual text the user wants to read (often in a much larger font). It similarly makes lists of items in some clients extremely tall, as 128 characters is a lot of hard-to-read text — especially when Memos renders links as their URL in titles.
Why an empty tag? The RSS 1.0 and 2.0 specs require that a <title> element is present.
Examples from elsewhere:
- micro.blog uses an empty <title /> element: https://www.manton.org/feed.xml
- Bluesky omits the <title> element: https://bsky.app/profile/did%3Aplc%3Aqvzn322kmcvd7xtnips5xaun/rss
- Mastodon omits the <title> element: https://mastodon.social/@scalzi.rss
* feat: support set global Default memo visibility
* feat: Move to WorkspaceMemoRelatedSetting
* fix: goimports check failure
* chore: change item name to default_visibility
* fix: goimports check failure
* Set max thumbnail width to timeline img max width
* Prevent images less than thumbnail size from being scaled up
* Apply suggestions from code review
---------
Co-authored-by: boojack <24653555+boojack@users.noreply.github.com>
* Reintroducing thumbnails
* Aligned with linting rules
* making recomended code review change
- changed method names to start with lower case as they are not used outside of their package
- made receiver types for struct funcs to be pointers to not need to create copies
Trying to cover all linting issues
- converted slog warning to use attributes when logging warnings
- seperated imports to have package files in their own section
* Update go.mod
---------
Co-authored-by: boojack <24653555+boojack@users.noreply.github.com>
When using the "export to Markdown" feature in Memos, the files included in the ZIP folder don't feature any kind of reference to their alphanumeric UID from the server's database, which completely breaks the point of links inside the files made with the `[[memos/<uid>]]` format, since it's impossible to know which Markdown file this kind of string inside other files refers to.
This pull request modifies the ExportMemos server function, to add the UID of every memo immediately after the date in the filename. For example, an exported memo would now be called: `YYYY-MM-DDThh:mm:ss+hh:mm-AbcDefGhiJklMnoPqrStu1-PUBLIC.md`.
