* fix: Remap internal links when duplicating a document tree
Duplicating a document tree now assigns the identifiers of every duplicate
before any content is written, so links and mentions between the documents
being duplicated point at the copies rather than the originals.
* fix: Remap fully qualified links when duplicating a document tree
Links written as a full url to this installation are now remapped alongside
relative ones, and stay fully qualified. Where a link is displayed as its own
url the text is updated to match.
Also trims surrounding whitespace in sanitizeUrl, which otherwise failed
validation and prepended a second scheme to an already qualified url.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: Remap links across trees when duplicating a collection
The collection duplication task duplicated each root document separately, so
only links within a single tree were remapped. It now duplicates the whole
collection as one unit, with identifiers assigned across every tree before any
content is written.
* fix: Match document links by identifier rather than host when duplicating
An installation can be reached through more than one host, so comparing a
link's host against the configured url left fully qualified links to a
document in the duplicated set unmapped. The document a link identifies now
decides whether it is replaced, and the host it was written with is kept.
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: Add TextPack (.textpack) import support
Adds a single-document importer for TextPack, the zipped variant of
TextBundle used by Bear, Ulysses, iA Writer and others. Bare .textbundle
directories are not supported, as a directory cannot be delivered through
a browser file input.
The bundle's text entry may use any extension, per the spec, with
info.json's type deciding whether it can be read as markdown. Assets are
inlined as data URIs for the existing attachment pipeline to pick up,
bounded by the attachment size limit and a memory ceiling, and only for
media types markdown-it accepts as a link destination.
Also fixes an existing bug where a file embedded in an HTML or email
import as a data URI was stored in the document as base64 rather than
being uploaded as an attachment.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* Refactor to individual converters
* refactor
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: Return a compact success payload from MCP mutation tools
Mutation tools serialized the entire resource — and for documents the
full markdown body — back to the client on every write, which the caller
already has. They now acknowledge the write with the identifying fields
needed to reference the resource.
The one exception is update_document with editMode "patch", which still
echoes the resulting content so the caller can verify what was replaced.
For move_document and restore_document the breadcrumb is now resolved
after the transaction commits, so it reflects the new location rather
than the cached pre-move document structure.
* test: Update MCP scope enforcement tests for the new response shape
* test: Assert the success acknowledgement across all mutation tool variants
Memberships pointing at soft-deleted or archived documents were still
returned by userMemberships.list and groupMemberships.list, consuming
slots in the first page the sidebar requests. Inner join the document so
they are excluded from the query entirely.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Restores field-level picking in presentURL instead of casting the entire
UnfurlData blob, so transient metadata attached by unfurl plugins or
internal callers is never serialized into API responses.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Widen the narrowed instanceof Error checks introduced in the TypeScript 6
upgrade with a shared errToId helper that accepts any object carrying a
string id, restoring the login email-matching recovery path.
* feat: Add support for web page and email import
* fix: Type mailparser result explicitly
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* Combine HTML and Web Page
* fix: Validate mime type before inlining MIME archive parts
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: Replace import document file picker with a modal
Choosing "Import document" now opens a dialog that lists the supported
file formats with an icon for each, and accepts files either dropped onto
it or chosen with the system file picker.
Also closes the gaps between the formats the dialog advertises and those
the importer actually accepts:
- `.tsv` was offered by the file picker but rejected by the server, it is
now converted to a table like CSV
- `.csv` and `.htm` are now recognised by extension, so they still import
when the browser reports an unhelpful mime type
- `.markdown` is now offered by the file picker, it was already supported
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Dj1UnjnRX336CQcdzw3ott
* Add intermediate dialog for file import
* tweaks, polish
* Update failure message
* test
---------
Co-authored-by: Claude <noreply@anthropic.com>
sequelize-strict-attributes defines per-instance accessors for attributes
omitted from a query's selection, which trips Sequelize's public class field
detection. The removed `Fix` decorator used to suppress this per model; do it
once on the base model instead.
* perf: Reduce data loading on revisions.list endpoint
* test: Cover includeContent option on Document.findByPk
Also fixes a JSDoc typo on the includeViews option.
---------
Co-authored-by: Claude <noreply@anthropic.com>
collectDocumentsAndAttachments merged a folder/file sharing a title by scanning
the sibling list with out.find(d => d.title === child.title) once per child —
O(n²) over the number of siblings in one directory, which for a flat import can
be large. Index the siblings by title in a Map and look up in O(1). First
occurrence wins, matching find; the merge still mutates the same object.
* Move socket management from client -> server
* feedback
* fix: Reconcile collection rooms on visibility change
Handle collections.update transitions between private and team-visible by
joining or rebuilding the collection channel, route collection events through
the members channel so guests are excluded, and use removeData for the
collection ids cache invalidation.
---------
Co-authored-by: Claude <noreply@anthropic.com>
* wip
* generalize
* refactor
* Extract SearchResultIcon for command bar results
Both search action builders rendered the same document icon inline. Move it
into a shared component alongside the other command bar pieces, forwarding
size so the command bar can continue to size icons via cloneElement.
Also drop the redundant charAt(0).toUpperCase() on the initial, as the Icon
component already normalizes it.
Co-Authored-By: Claude <noreply@anthropic.com>
* PR feedback
Prevent same doc showing twice in cmd k
Improve result ordering in main app command bar
Refactor to flattenTree
* Upgrade Fuse to 7.5.0, use workers
* fix: Re-register search actions when index is enriched
Include result contexts in the command bar registration key so that
snippets refresh when the same documents are enriched, and only mark a
server search as cached once its results are actually fed to the index.
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
