XUtils: ignore empty needles in a multi-term filter

[rawsun007]

String_contains_i() splits the filter on | and hands each needle to strcasestr(), which reports a match against anything for an empty needle. String_split() emits an empty field for a leading or repeated separator but drops a trailing one, so the same typo behaved differently depending on where it landed:

filter	before	after
nginx|	nothing extra	unchanged
|nginx	every row	only nginx rows
nginx||postgres	every row	nginx or postgres rows

So while composing a multi-term filter, a stray or half-typed | silently stopped the filter from filtering — and the same mistake at the end of the line was harmless. That inconsistency is what convinced me it wasn't intentional: if an empty alternative were meant to match everything, a trailing | would do it too.

Empty needles are now skipped. A filter consisting only of separators carries no search term and still matches everything, the same as an empty filter — the incremental-search callers already treat a null filter that way.

Fixed in String_contains_i() rather than String_split(), since the empty-field behaviour there is relied on elsewhere and only the filter's semantics are wrong.

Verification, on darwin: linked a small driver against the tree's own XUtils.o and ran 14 cases covering a leading, trailing, doubled and interior separator on both the matching and non-matching side, plus case-insensitivity and separators alone. All pass; reverting the change fails exactly |nginx and nginx||postgres. Full make clean with no new warnings under the tree's flags (-Wall -Wextra -pedantic and the rest).

I could not add a regression test: the only harness here is test_spec.lua, which drives a terminal rather than calling into C, so there is nowhere for a unit test of this to live. Happy to write a test_spec.lua case that types a filter into a live htop if you would rather have coverage that way — it seemed worse than the C-level evidence above, but it is your call.

AI disclosure per docs/ai-contributions-policy.md: found and written with Claude Code (Claude Opus 5), recorded in the commit's Assisted-by: trailer. @rawsun007 authorised the Signed-off-by: and reviewed the change.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: ASSERTIVE

Plan: Team

Run ID: 2da47536-28fc-4d88-84e2-fda12d31a8db

📥 Commits

Reviewing files that changed from the base of the PR and between a3dd8ad and 0dd0e9b.

📒 Files selected for processing (1)

• XUtils.c

Included review availability: Your plan provides up to 8 included reviews per hour; 7 remain after this review.

---

📝 Walkthrough

Walkthrough

String_contains_i now skips empty needles caused by leading or repeated | separators. It tracks the number of non-empty needles searched. A separator-only pattern returns true, which makes it behave like an empty filter. Non-empty needles continue to use case-insensitive substring matching.

Poem

Pipes part, but blanks stand still
Searches skip the empty sill
Needles count what they inspect
Empty filters now connect
Rows await a clearer sign

Merge Risk: ⚪ Minimal · up to 0dd0e

Filters with leading, repeated, or separator-only terms now behave as intended without changing non-empty term matching. The change is ready to merge.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[fasterit]

I'll merge this without the first AI comment. Today's AIs tend to work around their limited memory and context window by adding rather useless, verbose (to a human) commentary. So I'll cull that one. Otherwise thank you for the PR.

[fasterit]

Merged as b9a2a63