Conflict differ UI #42
Labels
No labels
area:docs
area:identity
area:ops
area:plugin
area:server
channel:community
channel:direct
channel:owned
channel:press
channel:social
e2ee-constrained
gate:at-ga
gate:pre-ga
marketing
parity
relay:absent
relay:planned
relay:requested
relay:supported
risk:additive
risk:contract
risk:none
usability
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
Nectenda/nectenda#42
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Part of #37.
What Relay does
A diff/merge resolution view since 0.3.3, opened proactively from a "stack" icon,
themed to match Obsidian, with guides for resolving conflicts and working through
a backlog of them. Note their own issue #134 complains the differ mislabels which
side is which — theirs is not exemplary, it is merely present.
What we do today
We write a conflict copy and show a Notice —
"<name> (conflicted copy <stamp>)<ext>"(
file-sync.ts:446,blob-sync.ts:365). That is the safer behaviour andCLAUDE.mdis right to prefer it: disagreement is visible and recoverable wheresilent convergence is not. But nothing helps the user afterwards.
Why this is more than polish
.nectenda-backups/is a dot-directory, and Obsidian's vault index excludesit — the same fact recorded in
CLAUDE.mdas having once made backups lookmissing.
docs/sync-limitations.mdwarns testers about it. Users have the sameproblem and no workaround: first-sync backups exist and are effectively
unreachable from inside Obsidian.
So this is a live hole in the claim the product leads with. Both inputs are
already on disk; nothing surfaces them.
What changes
A differ over the conflict copy and its original, plus a browser for
.nectenda-backups/that reads through the adapter rather than the vault index.Risk
risk:none.Verification
Assert through an API that can actually observe dot-directories —
CLAUDE.mdrecords that using the vault index here produced a wrong conclusion once already.
Moved to the Vikunja board as NEC-19: https://projectron.nerchure.com/tasks/19