Pearl Edits and Reflects Linear — manage Linear.app issues as org-mode in Emacs https://git.cjennings.net/pearl/atom?h=main 2026-06-05T11:01:13+00:00 2026-06-05T11:01:13+00:00 Craig Jennings c@cjennings.net 2026-06-05T11:01:13+00:00 urn:sha1:ecfc66294e57788df715c98297ef425fd9dbd9ef The full-replace render saved the Org file, but the in-place paths didn't, so a refresh or a comment fetch left the buffer modified-but-unsaved and the file on disk stale. Five paths now save through pearl--save-if-visiting: refresh-current-view, refresh-current-issue, fetch-all-comments, create-comment, and the merge branch of a dirty source-switch. The helper saves only when the buffer visits a file, so it's a no-op in a test buffer. Saving the Org file never pushes to Linear, so an unpushed edit still waits for pearl-save-issue. This only keeps the disk copy in step with what's on screen. fetch-all-comments surfaced the gap first. An audit turned up the other four. 2026-06-05T06:06:18+00:00 Craig Jennings c@cjennings.net 2026-06-05T06:06:18+00:00 urn:sha1:3f491e73cdbdc03d652bce81fdfc68eabb313213 pearl-fetch-all-comments pages the whole comment connection for the issue at point and replaces its Comments subtree with the complete set. It's bound to C-; L f c and in the transient. The bulk list and view fetch showed only the newest pearl-list-comments-shown comments with an N/M+ overflow marker, so anything past the cap couldn't surface in the buffer. The replacement is surgical: pearl--replace-comments-subtree swaps only the Comments subtree, so the issue title, drawer, and description body keep any unsaved edits, and the new heading carries an exact N/N count. It refuses when the issue has unpushed comment edits (reusing the save model's dirty scanner) so the fetch can't clobber an unsent change.