justintsus-experiment Public
tsus-web: no .tsus files▶ dates: absolute ✔ a bare day is that UTC day, both ends exact (1.81475ms) ✔ > a day starts the next day; >= a day includes the day (0.516208ms) ✔ < a day ends before it; <= a day includes all of it (0.220375ms) ✔ a range includes both days whole (0.248667ms) ✔ open-ended ranges with * (0.16875ms) ✔ a leap day exists in 2024 (0.112584ms) ✔ two date terms are ANDed (0.105708ms) ✔ pulls filter on created_at and updated_at too (0.178083ms)✔ dates: absolute (3.962625ms)▶ dates: relative ages ✔ <7d: younger than 7 days (exactly 7 days old is not); <=7d: at most 7 days old (0.2335ms) ✔ >7d: older than 7 days; >=7d: at least 7 days old (0.14675ms) ✔ 1w and 168h are 7d (0.122667ms) ✔ hours: <2h and <1h (exactly an hour old is not younger than 1h) (0.083084ms) ✔ relative dates use ctx.now, not the clock (0.125583ms) ✔ relative and absolute on created: together (0.072042ms)✔ dates: relative ages (0.913459ms)▶ dates: validation ✔ impossible or malformed days are bad_date (0.233875ms) ✔ operators: = is not one, and an operator needs a date (0.057417ms) ✔ ranges: reversed, *..*, missing sides and relative sides are bad_date (0.051667ms) ✔ relative values: an operator is required, N is 1 to 9999, units are h, d, w (0.080416ms)✔ dates: validation (0.483459ms)▶ parse: basics ✔ the empty input and whitespace-only input are valid, empty queries (0.720458ms) ✔ every supported qualifier parses for issues (0.937417ms) ✔ every supported qualifier parses for pulls (0.251458ms) ✔ parseQuery never throws on odd strings (0.33375ms)✔ parse: basics (2.820708ms)▶ parse: kinds of terms ✔ text that is not a qualifier stays free text: 12:30, a1:b, :x (0.20625ms) ✔ #n with 1 to 9 digits and no leading zero is a number; anything else is text (0.076166ms) ✔ a lone - is the free-text word - (0.062917ms) ✔ a quoted qualifier is a phrase, not a qualifier (0.087375ms) ✔ OR is a word, not an operator (0.08525ms) ✔ qualifier keys are case-insensitive (0.102416ms) ✔ negation applies to qualifiers and to free text (0.07ms)✔ parse: kinds of terms (0.87225ms)▶ parse: quotes ✔ a qualifier takes a quoted value (0.13575ms) ✔ \" and \\ are escapes inside quotes; another backslash stays (0.077166ms) ✔ an unterminated quote is one error from the term to the end, and parsing stops (0.474042ms) ✔ an unterminated qualifier quote spans the whole term (0.078041ms) ✔ text right after a closing quote is unexpected_quote (0.0615ms) ✔ a quote inside bare text is unexpected_quote, spanning to the next whitespace (0.054417ms) ✔ an empty phrase is empty_value: "" and " " (0.065833ms)✔ parse: quotes (1.038875ms)▶ parse: errors ✔ unknown qualifiers are errors, assignee: included (0.148875ms) ✔ qualifiers of the other target are unsupported_qualifier (0.051875ms) ✔ an unknown key is reported before an empty value, and unsupported before empty (0.033625ms) ✔ empty values: label:, label:"", author:, is: (0.050208ms) ✔ negation is not allowed on no:, sort:, created:, updated: and #n (an empty value is reported first) (0.045916ms) ✔ negation_not_allowed comes before bad_value (0.027917ms) ✔ is: values depend on the target (0.085834ms) ✔ sort: values depend on the target (0.059375ms) ✔ labels follow the label rules: characters and 32 at most (0.078083ms) ✔ user references: @me, a UUID or an agent slug (0.0675ms) ✔ a second sort: with another value is a conflict on the second term (0.057208ms) ✔ error spans cover the whole term, the - included (0.069667ms) ✔ spans are UTF-16 offsets: an emoji before the term counts two (0.053375ms) ✔ every invalid term gets one error, in input order (0.035375ms) ✔ a label value cannot carry SQL: it is a bad_value (0.030791ms)✔ parse: errors (1.031833ms)▶ parse: unicode and whitespace ✔ no-break and ideographic spaces separate terms (0.062625ms) ✔ unicode free text and emoji are kept exactly (0.041042ms)✔ parse: unicode and whitespace (0.135ms)▶ parse: limits ✔ 1024 code units are allowed (0.109833ms) ✔ 1025 code units are too_long, alone, spanning the input (9.070542ms) ✔ an emoji counts as two code units against the limit (0.248292ms) ✔ a very long input is rejected quickly (0.999458ms) ✔ 32 terms are allowed; the 33rd is too_many_terms from its start to the end (0.795667ms) ✔ terms are counted as written: 33 copies of one term are too many, invalid ones included (0.104375ms) ✔ many terms of every kind parse fast (4.409542ms)✔ parse: limits (15.947084ms)▶ scope ✔ no issues query returns a row of another repository or organization (4.335208ms) ✔ no pulls query returns a row of another repository or organization (1.243833ms) ✔ text that only the other repositories' rows match finds nothing (0.565333ms) ✔ #1 is this repository's #1 (0.1605ms) ✔ the scope follows ctx: another repository of the org sees its own rows only (0.137833ms) ✔ the organization is part of the scope: the right repo id under another org matches nothing (0.109ms) ✔ the empty query is still scoped, with repoId and orgId as params (0.077625ms) ✔ repo:, org: and is:public cannot widen the scope: they are errors (0.19325ms) ✔ an agent slug resolves in the repository's organization only (0.271ms) ✔ generated inputs: whatever parses stays in scope and round-trips (35.048542ms)✔ scope (42.985125ms)▶ issues: state, labels, order ✔ the empty query lists every issue in scope, newest created first (1.492042ms) ✔ is:open, is:closed and -is:open (0.861917ms) ✔ label:bug, and label:bug -label:wontfix (0.476375ms) ✔ label:"needs review" matches the whole label, and label:BUG equals label:bug (0.446167ms) ✔ no:label (0.143209ms) ✔ sort:comments-desc and -asc, ties broken by number in the same direction (0.138458ms) ✔ sort:created-asc, sort:updated-asc and sort:updated-desc (0.135083ms) ✔ #n picks one issue; two numbers match nothing; a missing number matches nothing (0.155625ms)✔ issues: state, labels, order (4.490042ms)▶ issues: people ✔ author:@me is the viewer (0.191834ms) ✔ author:@me with no viewer matches nothing; -author:@me then matches everything (0.145542ms) ✔ author:<uuid>, in any case (0.223042ms) ✔ author:<slug> is the agent of the repository's organization only (0.155667ms)✔ issues: people (0.835958ms)▶ issues: free text ✔ a word matches the title or the body, ASCII case-insensitively (0.218333ms) ✔ a phrase is one substring; bare words are separate terms (0.174208ms) ✔ OR is an ordinary word (0.095916ms) ✔ % and _ are matched literally (0.095375ms) ✔ a backslash is matched literally (0.189916ms) ✔ a double quote inside a phrase matches (0.066292ms) ✔ -word excludes issues whose title or body has it (0.095792ms) ✔ text and qualifiers combine (0.174ms)✔ issues: free text (1.210208ms)▶ pulls ✔ the empty query lists every pull request in scope, newest created first (0.121083ms) ✔ is:open includes drafts; is:closed includes merged (as on GitHub) (0.121709ms) ✔ is:merged and -is:merged (0.086833ms) ✔ is:draft in any state, -is:draft, and is:open -is:draft (0.1095ms) ✔ author: matches the author or the agent's owner (0.176958ms) ✔ -author: keeps pull requests whose owner_id is NULL (0.086791ms) ✔ review-requested: looks at this repository's requests only (0.173958ms) ✔ review-requested:@me with no viewer matches nothing; its negation matches everything (0.073417ms) ✔ free text matches the title, with % and _ literal (0.106583ms) ✔ sort:updated-desc and #n (0.124917ms)✔ pulls (1.287458ms)▶ serialize: canonical form ✔ normalization, quoting and duplicates follow the canonical rules (2.4565ms) ✔ pull request qualifiers serialize the same way (0.132791ms) ✔ the canonical form of the empty query is the empty string (0.069ms)✔ serialize: canonical form (3.15925ms)▶ serialize: round trips ✔ serialize is idempotent: the canonical form parses back to itself (0.484958ms) ✔ the canonical form evaluates to the same SQL and params as the input (1.023084ms) ✔ pulls: idempotent and the same SQL (0.196542ms) ✔ the canonical form survives a URL's query string (0.48325ms) ✔ terms keep their order (the canonical form is not sorted) (0.097083ms) ✔ differently written equal terms collapse into the first (0.085292ms)✔ serialize: round trips (2.568625ms)▶ sql: injection attempts are bound, never interpolated ✔ free text and phrases carrying SQL (2.952042ms) ✔ qualifier values and numbers carrying markers (0.604625ms) ✔ pull request queries carrying SQL (0.237583ms)✔ sql: injection attempts are bound, never interpolated (4.265375ms)▶ sql: shape ✔ no date or number from the input appears in the SQL text (0.440709ms) ✔ ctx values are bound too, even when they contain quotes (0.587041ms) ✔ where keeps its meaning when the caller ANDs more conditions after it (0.411208ms) ✔ every query of a mixed corpus has as many placeholders as params, and a fixed orderBy (0.351042ms) ✔ toSql does not throw for odd but valid queries (0.4575ms) ✔ a ? or :name in the text is a value, not a placeholder (0.157125ms)✔ sql: shape (2.594458ms)ℹ tests 118ℹ suites 18ℹ pass 118ℹ fail 0ℹ cancelled 0ℹ skipped 0ℹ todo 0ℹ duration_ms 657.221167hidden tests: pass 118 / total 118 (worktree <rebasely checkout>/.claude/worktrees/agent-a7895120081998e14)