feat: show all icons alphabetically on empty search by nanoraptor · Pull Request #210 · vicinaehq/extensions

nanoraptor · 2026-03-23T09:29:25Z

Modified filterIconIndex to return all icons sorted alphabetically when search is empty
Updated display logic to show all icons instead of 'Start searching' message
Updated test to verify new behavior
All 11 tests passing

- Modified filterIconIndex to return all icons sorted alphabetically when search is empty - Updated display logic to show all icons instead of 'Start searching' message - Updated test to verify new behavior - All 11 tests passing Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

This PR changes Nerd Font Search so that when the search box is empty (and “All icon packs” is selected), the UI shows an alphabetical icon list instead of an empty “Start searching” state, and updates tests accordingly.

Changes:

Updated filterIconIndex to return an alphabetically sorted list for short/empty search in the “all packs” mode.
Updated the grid display/empty-state logic to show icons (or a loading message) rather than a “Start searching” prompt.
Updated the pack filter behavior test to validate the new empty-search behavior.

Reviewed changes

Copilot reviewed 3 out of 4 changed files in this pull request and generated 4 comments.

File	Description
`extensions/nerdfont-search/src/filtering.ts`	Changes short/empty-search behavior to return a sorted icon list for all packs.
`extensions/nerdfont-search/src/nerdfonts-search.tsx`	Adjusts which icons are displayed when search is empty and updates empty-state messaging.
`extensions/nerdfont-search/test/pack-filter-behavior.test.ts`	Updates assertions for the new empty-search behavior and sorting expectation.
`extensions/nerdfont-search/package-lock.json`	Lockfile metadata changes.

Files not reviewed (1)

extensions/nerdfont-search/package-lock.json: Language not supported

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-03-23T09:32:39Z

 	if (searchText.length < 3) {
 		if (selectedPack === PACK_FILTER_ALL) {
-			return [];
+			return iconIndex


iconIndex.sort(...) mutates the shared iconIndex array in-place. Since the same array is cached (cachedIconIndex) and also passed into the Fuse instance (new Fuse(decodedIndex, ...)), this can reorder the underlying list after Fuse has been constructed and can lead to incorrect search results or hard-to-debug state issues. Use a non-mutating approach (e.g., sort a shallow copy) so the cached index and Fuse’s backing array aren’t modified.

Suggested change

return iconIndex

return [...iconIndex]

Copilot · 2026-03-23T09:32:40Z

 	if (searchText.length < 3) {
 		if (selectedPack === PACK_FILTER_ALL) {
-			return [];
+			return iconIndex
+				.sort((a, b) => a.displayName.localeCompare(b.displayName))
+				.slice(0, SEARCH_RESULT_LIMIT);
 		}


The new behavior is described as “show all icons when search is empty”, but this branch runs for any searchText.length < 3 and also applies SEARCH_RESULT_LIMIT via slice(0, SEARCH_RESULT_LIMIT). That means 1–2 character queries will ignore the query and just show the default alphabetical list, and empty search will only show the first 200 icons—not “all” icons. Consider handling searchText.length === 0 separately from 1–2 characters, and clarify whether the limit should apply for the empty-search listing.

Copilot · 2026-03-23T09:32:40Z

+	assert.ok(noSearchAll.length > 0, "all-pack short search should return all icons");
+	assert.ok(
+		noSearchAll.every((a, i, arr) => i === 0 || a.displayName >= arr[i - 1].displayName),


The alphabetical-order assertion uses a.displayName >= arr[i - 1].displayName, which is a different ordering than the production code’s localeCompare and can behave differently with case/locale-specific characters. Also, the message says “return all icons”, but the implementation applies SEARCH_RESULT_LIMIT, so the assertion text is misleading. Align the test’s ordering check and expected wording with the production sorting and any intentional result limiting.

Suggested change

assert.ok(noSearchAll.length > 0, "all-pack short search should return all icons");

assert.ok(

noSearchAll.every((a, i, arr) => i === 0 || a.displayName >= arr[i - 1].displayName),

assert.ok(noSearchAll.length > 0, "all-pack short search should return icons");

assert.ok(

noSearchAll.every(

(icon, i, arr) =>

i === 0 || arr[i - 1].displayName.localeCompare(icon.displayName) <= 0

),

Copilot · 2026-03-23T09:32:40Z

-                    : "Enter at least 3 characters to search for icons"
+              : searchText.length >= 3
+                ? "Try a different search term or pick another icon pack"
+                : "Your icon library is loading. Please wait..."


The empty-state copy now defaults to “Loading icons...” / “Your icon library is loading...” whenever displayIcons.length === 0 and searchText.length < 3. But isLoading can be false while displayIcons is still empty (e.g., the index load failed and iconIndex stayed empty in useIconData’s catch). In that case the UI would misleadingly claim it’s still loading. Consider keying the empty-state messaging off isLoading (and possibly adding an explicit error/failed-to-load state) instead of assuming empty results always mean loading.

SiriusCrain · 2026-03-25T08:55:52Z

@nanoraptor is it ok for you if I will push some perfomance fixes to your PR?

nanoraptor · 2026-03-25T09:04:41Z

@nanoraptor is it ok for you if I will push some perfomance fixes to your PR?

Okay

SiriusCrain · 2026-03-25T09:34:15Z

I cannot push since it's your fork, here is patch, apply it please:
nerdfont-presort-optimization.patch

What was fixed:

Icons are pre-sorted, before they were sorted on each filterIconIndex call
map lookup + slice instead of filtering 10K items
removed array mutation
fixed search

5 runs median result

Before fix:

Scenario	Median	Mean	P95	Min	Max
empty-search-all	3.42ms	3.45ms	3.77ms	3.38ms	3.93ms
empty-search-pack	0.44ms	0.44ms	0.45ms	0.43ms	0.56ms
short-search-all	3.28ms	3.28ms	3.44ms	3.25ms	3.68ms
fuse-search-all	113.20ms	113.58ms	116.92ms	110.56ms	129.55ms
fuse-search-pack	144.04ms	144.88ms	147.92ms	141.80ms	167.30ms

After fix:

Scenario	Median	Mean	P95	Min	Max
empty-search-all	0.00ms	0.01ms	0.01ms	0.00ms	0.03ms
empty-search-pack	0.13ms	0.13ms	0.15ms	0.12ms	0.30ms
short-search-all	0.00ms	0.00ms	0.01ms	0.00ms	0.01ms
fuse-search-all	114.86ms	116.00ms	125.70ms	112.07ms	135.63ms
fuse-search-pack	145.82ms	146.05ms	147.83ms	143.79ms	164.66ms

fbosch · 2026-03-25T10:50:26Z

I cannot push since it's your fork, here is patch, apply it please: nerdfont-presort-optimization.patch

What was fixed:
* Icons are pre-sorted, before they were sorted on each `filterIconIndex` call

* map lookup + slice instead of filtering 10K items

* removed array mutation

* fixed search
5 runs median result

Before fix:
Scenario Median Mean P95 Min Max
empty-search-all 3.42ms 3.45ms 3.77ms 3.38ms 3.93ms
empty-search-pack 0.44ms 0.44ms 0.45ms 0.43ms 0.56ms
short-search-all 3.28ms 3.28ms 3.44ms 3.25ms 3.68ms
fuse-search-all 113.20ms 113.58ms 116.92ms 110.56ms 129.55ms
fuse-search-pack 144.04ms 144.88ms 147.92ms 141.80ms 167.30ms

After fix:
Scenario Median Mean P95 Min Max
empty-search-all 0.00ms 0.01ms 0.01ms 0.00ms 0.03ms
empty-search-pack 0.13ms 0.13ms 0.15ms 0.12ms 0.30ms
short-search-all 0.00ms 0.00ms 0.01ms 0.00ms 0.01ms
fuse-search-all 114.86ms 116.00ms 125.70ms 112.07ms 135.63ms
fuse-search-pack 145.82ms 146.05ms 147.83ms 143.79ms 164.66ms

The decision of whether or not this is something we wanna go with still remains:
#209 (comment)

Maybe making it opt-in could be a solution?

SiriusCrain · 2026-03-25T12:30:07Z

I cannot push since it's your fork, here is patch, apply it please: nerdfont-presort-optimization.patch

What was fixed:
* Icons are pre-sorted, before they were sorted on each `filterIconIndex` call

* map lookup + slice instead of filtering 10K items

* removed array mutation

* fixed search
5 runs median result

Before fix:
Scenario Median Mean P95 Min Max
empty-search-all 3.42ms 3.45ms 3.77ms 3.38ms 3.93ms
empty-search-pack 0.44ms 0.44ms 0.45ms 0.43ms 0.56ms
short-search-all 3.28ms 3.28ms 3.44ms 3.25ms 3.68ms
fuse-search-all 113.20ms 113.58ms 116.92ms 110.56ms 129.55ms
fuse-search-pack 144.04ms 144.88ms 147.92ms 141.80ms 167.30ms

After fix:
Scenario Median Mean P95 Min Max
empty-search-all 0.00ms 0.01ms 0.01ms 0.00ms 0.03ms
empty-search-pack 0.13ms 0.13ms 0.15ms 0.12ms 0.30ms
short-search-all 0.00ms 0.00ms 0.01ms 0.00ms 0.01ms
fuse-search-all 114.86ms 116.00ms 125.70ms 112.07ms 135.63ms
fuse-search-pack 145.82ms 146.05ms 147.83ms 143.79ms 164.66ms
The decision of whether or not this is something we wanna go with still remains:
#209 (comment)

Maybe making it opt-in could be a solution?

I just fixed issues in current PR only following it's logic)
Of course if behavior will change we can adapt as well

Performance fixes for NerdFont search extension including: - Add packIndex parameter for optimized pack filtering - Update search configuration and Fuse options - Enhance icon data handling and search functionality - Add pack indexing for better performance - Update UI logic for empty states and icon display - Update tests Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

nanoraptor · 2026-03-27T22:35:44Z

I cannot push since it's your fork, here is patch, apply it please: nerdfont-presort-optimization.patch

What was fixed:

Icons are pre-sorted, before they were sorted on each filterIconIndex call

map lookup + slice instead of filtering 10K items

removed array mutation

fixed search

5 runs median result

Before fix:

Scenario Median Mean P95 Min Max
empty-search-all 3.42ms 3.45ms 3.77ms 3.38ms 3.93ms
empty-search-pack 0.44ms 0.44ms 0.45ms 0.43ms 0.56ms
short-search-all 3.28ms 3.28ms 3.44ms 3.25ms 3.68ms
fuse-search-all 113.20ms 113.58ms 116.92ms 110.56ms 129.55ms
fuse-search-pack 144.04ms 144.88ms 147.92ms 141.80ms 167.30ms
After fix:

Scenario Median Mean P95 Min Max
empty-search-all 0.00ms 0.01ms 0.01ms 0.00ms 0.03ms
empty-search-pack 0.13ms 0.13ms 0.15ms 0.12ms 0.30ms
short-search-all 0.00ms 0.00ms 0.01ms 0.00ms 0.01ms
fuse-search-all 114.86ms 116.00ms 125.70ms 112.07ms 135.63ms
fuse-search-pack 145.82ms 146.05ms 147.83ms 143.79ms 164.66ms

applied

aurelleb · 2026-04-24T12:32:19Z

there are conflicts atm. besides, is this ready or not?

Copilot AI review requested due to automatic review settings March 23, 2026 09:29

Copilot started reviewing on behalf of nanoraptor March 23, 2026 09:30 View session

nanoraptor mentioned this pull request Mar 23, 2026

[nerdfont-search] Feature Request: display all nerd fonts when no input is given #209

Open

Copilot AI reviewed Mar 23, 2026

View reviewed changes

nanoraptor and others added 2 commits March 28, 2026 03:59

Merge main branch with patch changes

e1e0df1

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: show all icons alphabetically on empty search#210

feat: show all icons alphabetically on empty search#210
nanoraptor wants to merge 3 commits intovicinaehq:mainfrom
nanoraptor:feat/show-all-icons-on-empty-search

nanoraptor commented Mar 23, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Mar 23, 2026

Uh oh!

Copilot AI Mar 23, 2026

Uh oh!

Copilot AI Mar 23, 2026

Uh oh!

Copilot AI Mar 23, 2026

Uh oh!

SiriusCrain commented Mar 25, 2026

Uh oh!

nanoraptor commented Mar 25, 2026

Uh oh!

SiriusCrain commented Mar 25, 2026 •

edited

Loading

Uh oh!

fbosch commented Mar 25, 2026

Uh oh!

SiriusCrain commented Mar 25, 2026

Uh oh!

nanoraptor commented Mar 27, 2026

Uh oh!

aurelleb commented Apr 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Conversation

nanoraptor commented Mar 23, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 23, 2026

Choose a reason for hiding this comment

Uh oh!

SiriusCrain commented Mar 25, 2026

Uh oh!

nanoraptor commented Mar 25, 2026

Uh oh!

SiriusCrain commented Mar 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

fbosch commented Mar 25, 2026

Uh oh!

SiriusCrain commented Mar 25, 2026

Uh oh!

nanoraptor commented Mar 27, 2026

Uh oh!

aurelleb commented Apr 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

SiriusCrain commented Mar 25, 2026 •

edited

Loading