diff options
| author | Craig Jennings <c@cjennings.net> | 2026-05-28 10:45:03 -0500 |
|---|---|---|
| committer | Craig Jennings <c@cjennings.net> | 2026-05-28 10:45:03 -0500 |
| commit | eb5629570b26b84610ec9ed1a27d40ae5407fca1 (patch) | |
| tree | c174b192b54984724323ec5d01d9adff7642dcd9 /pearl.el | |
| parent | b1df299c164bd688ce4d0887603e6b71ea51bd22 (diff) | |
| download | pearl-eb5629570b26b84610ec9ed1a27d40ae5407fca1.tar.gz pearl-eb5629570b26b84610ec9ed1a27d40ae5407fca1.zip | |
feat(merge): preserve the user's view across a merge refresh
A merge refresh (the "merge keeping edits" choice on a dirty buffer, or pearl-refresh-current-view on a clean one) used to call pearl--restore-page-visibility at the tail, re-folding the whole buffer to its #+STARTUP overview. That collapsed every subtree the user had expanded, including the comment they were just editing, and disturbed point. The edit-then-merge flow felt jarring.
The fix: pearl--merge-issues-into-buffer now returns a marker for every subtree it re-rendered or appended, alongside the existing counts (:touched-markers (M1 M2 ...)). The merge call sites fold just those subtrees via a new pearl--fold-touched-subtrees helper. Kept (locally-edited) and untouched subtrees stay exactly as the user had them, and point is preserved because the localized fold doesn't touch the rest of the page.
The full-rebuild paths (Branch A no-buffer, Branch B clean-buffer replace, Branch C-discard) still call pearl--restore-page-visibility. They re-render the whole buffer so the global re-fold is the right behavior.
Pre-commit review caught a critical correctness bug in the first draft: the alist marker from pearl--issue-subtree-markers is insertion-type t (it needs to advance past replaces of EARLIER subtrees so it stays anchored to its own heading), and reusing it after pearl--replace-issue-subtree-at-point captured the post-advance position, which is the NEXT subtree's heading, so the fold collapsed the wrong subtree. The fix copies the marker as type-nil BEFORE the replace, so it stays anchored at the deletion start and lands on the new heading after the re-insert. I strengthened the marker test to verify the IDs at the touched markers match the expected set ({a, b, c} for 2 updated + 1 added), not just that they're some valid headings. The original test would have passed against the bug.
Five new tests in test-pearl-merge.el cover: merge returns markers with the correct IDs at their headings, skipped (locally-edited) issues don't add to touched-markers, fold-subtree-at-marker hides body content while leaving the heading visible, fold-touched-subtrees is a no-op when pearl-fold-after-update is nil, and fold-subtree-at-marker tolerates a nil marker silently. All 674 ert tests pass. make compile and make lint are clean.
Diffstat (limited to 'pearl.el')
| -rw-r--r-- | pearl.el | 68 |
1 files changed, 59 insertions, 9 deletions
@@ -298,10 +298,41 @@ folds shut like the rest of the page, rather than rendering expanded." (narrow-to-region start end) (pearl--hide-or-tidy-drawers))) +(defun pearl--fold-subtree-at-marker (marker) + "Fold the subtree at MARKER to outline-only and tidy its drawers. +The subtree's heading stays visible; everything under it (body, drawers, +comment children) is collapsed. A merge refresh re-renders updated +subtrees and inserts new ones expanded; this folds just those so they +match what the buffer's `#+STARTUP' visibility would have produced, +without touching subtrees the user left expanded. A nil or dead MARKER +is a silent no-op." + (when (and marker (marker-buffer marker)) + (save-excursion + (goto-char marker) + (org-back-to-heading t) + (let ((beg (point)) + (end (save-excursion (org-end-of-subtree t t) (point)))) + (cond ((fboundp 'org-fold-subtree) (org-fold-subtree t)) + ((fboundp 'outline-hide-subtree) (outline-hide-subtree))) + (pearl--hide-drawers-in-region beg end))))) + +(defun pearl--fold-touched-subtrees (markers) + "Fold each subtree at the markers in MARKERS. +Used after `pearl--merge-issues-into-buffer' so the page stays as the +user had it. Untouched and locally-edited subtrees are left exactly as +they were; only the subtrees the merge actually re-rendered (updated) +or appended (added) come in folded. A no-op when +`pearl-fold-after-update' is nil." + (when pearl-fold-after-update + (dolist (m markers) + (pearl--fold-subtree-at-marker m)))) + (defun pearl--restore-page-visibility () "Re-fold the whole current buffer to its `#+STARTUP' visibility and hide drawers. -Used after a full repopulation (list / view / merge refresh) so the page does -not sprawl open. A no-op when `pearl-fold-after-update' is nil." +Used after a full repopulation (list / view replace path) so the page does +not sprawl open. A no-op when `pearl-fold-after-update' is nil. The merge +path uses `pearl--fold-touched-subtrees' instead so an edit-then-merge flow +keeps the user's expanded subtrees visible." (when pearl-fold-after-update ;; A full in-place rebuild (Branch B / merge) leaves Org's parsed startup ;; options stale -- the new `#+STARTUP:' text is in the buffer but @@ -3256,7 +3287,11 @@ just writing the file and leaving it off-screen." (pearl--update-source-header (length issues) truncated) (pearl--update-derived-todo-header (pearl--gather-header-states issues)) (pearl-highlight-comments) - (pearl--restore-page-visibility) + ;; Fold only the subtrees the merge re-rendered or appended; leave + ;; the rest of the page (and point) as the user had them, so an + ;; edit-then-merge flow doesn't collapse the subtree they were + ;; editing. + (pearl--fold-touched-subtrees (plist-get counts :touched-markers)) (message "Merged into %s: %d updated, %d added, %d dropped%s" (file-name-nondirectory org-file-path) (plist-get counts :updated) @@ -5153,11 +5188,18 @@ in place; an issue new to ISSUES is appended after the last one; an issue no longer in ISSUES is dropped. A subtree with any unpushed local edit (see `pearl--subtree-has-local-edits-p' -- title, state, priority, assignee, labels, description, or a comment) is never overwritten and never dropped; it is kept -and counted, so a refresh can't lose un-synced work. Returns a plist of counts -\(:updated :added :dropped :skipped)." +and counted, so a refresh can't lose un-synced work. + +Returns a plist of counts plus markers: +\(:updated N :added N :dropped N :skipped N :touched-markers (M1 M2 ...)) +where `:touched-markers' carries one marker per subtree the merge re-rendered +(updated) or appended (added). Callers use the marker list to fold just +those subtrees rather than re-folding the whole buffer, so an +edit-then-merge flow keeps the user's expanded subtrees expanded." (let ((existing (pearl--issue-subtree-markers)) (fetched-ids (mapcar (lambda (i) (plist-get i :id)) issues)) - (updated 0) (added 0) (dropped 0) (skipped 0)) + (updated 0) (added 0) (dropped 0) (skipped 0) + (touched-markers '())) ;; Existing issues still in the result: re-render in place, unless dirty. (dolist (issue issues) (let ((marker (cdr (assoc (plist-get issue :id) existing)))) @@ -5167,6 +5209,9 @@ and counted, so a refresh can't lose un-synced work. Returns a plist of counts (if (pearl--subtree-has-local-edits-p) (setq skipped (1+ skipped)) (pearl--replace-issue-subtree-at-point issue) + ;; `marker' still points at the new heading -- markers track + ;; the buffer position across the replace's delete-then-insert. + (push (copy-marker marker) touched-markers) (setq updated (1+ updated))))))) ;; Existing issues absent from the result: drop them, but keep dirty ones. (dolist (cell existing) @@ -5184,9 +5229,12 @@ and counted, so a refresh can't lose un-synced work. Returns a plist of counts (save-excursion (goto-char (point-max)) (unless (bolp) (insert "\n")) - (insert (pearl--format-issue-as-org-entry issue))) + (let ((insertion-start (point-marker))) + (insert (pearl--format-issue-as-org-entry issue)) + (push insertion-start touched-markers))) (setq added (1+ added)))) - (list :updated updated :added added :dropped dropped :skipped skipped))) + (list :updated updated :added added :dropped dropped :skipped skipped + :touched-markers (nreverse touched-markers)))) (defun pearl--update-source-header (issue-count truncated) "Refresh the active file's run-at, count, and truncation header lines. @@ -5258,7 +5306,9 @@ of the replace path." (pearl--update-source-header (length issues) truncated) (pearl--update-derived-todo-header (pearl--gather-header-states issues)) (pearl-highlight-comments) - (pearl--restore-page-visibility) + ;; Same as the dirty-merge path: fold only the touched subtrees, leave + ;; the rest of the page and point alone. + (pearl--fold-touched-subtrees (plist-get counts :touched-markers)) (pearl--surface-buffer (current-buffer)) (message "Refreshed %s: %d updated, %d added, %d dropped%s" (pearl--source-name source) |
