Aquesta és la versió actual. · Data de publicació:
Fitxers d’instal·lació 0.10.4
| Arquitectura | Tipus | Àmbit | Mida del fitxer | Baixa | SHA256 |
|---|---|---|---|---|---|
| 64 bits | ZIP → Portable | — | — | github.com | VirusTotal |
L’enllaç i l’SHA256 provenen del manifest comunitari del Gestor de paquets del Windows. Convé comparar el hash del fitxer baixat abans d’executar-lo.
Notes de la versió
### Added
• `IS NULL` and `IS NOT NULL` operators.
• `depth` field.
• `CONTAINS_CYRILLIC` function; better Japanese text detection.
• `STRPOS` function, split from `LOCATE`.
• `rlike` as a regex operator alias.
• Encrypted entries in zip archives; more fields available for archived files, with archive paths included in their values.
• `include:` directives in `.hgignore` and `.dockerignore`.
• Function argument counts and literals compared with typed expressions are checked when the query is parsed.
• Integration test suite covering traversal, ignore files, formats, grouping, dates, archives and CLI flags.
### Changed
• Expressions carry static types; caches, sorting, grouping and aggregates work on typed values instead of rendered text.
• `fsize` is numeric:it compares, groups, sorts and aggregates by byte count while printing the formatted size.
• `CURRENT_DATE`, `NOW`, `DATE_ADD`, `DATE_SUB`, `FROM_UNIXTIME`, `LAST_DAY` and `DATE_TRUNC` return datetimes.
• `LEAST` and `GREATEST` operate on strings as well as numbers.
• `xattr()` returns values that are not UTF-8 as `0x..` hex instead of an empty string ([#187](https://github.com/jhspetersson/fselect/issues/187)).
• Unknown columns in `GROUP BY` and `ORDER BY`, ungrouped columns in grouped queries, an empty `IN ()` and an incomplete `WHERE` clause are rejected.
• A literal that can never match a typed expression (`length(name) > 'abc'`) is rejected before the search starts.
• `LIKE ''` matches the empty string.
• Bare words are treated as function calls only for functions that take no arguments.
• Root aliases are case-insensitive; JSON column names are lowercase.
• Search roots defined by subqueries inherit the root options of the parent query; a grouped subquery in `FROM` yields one entry per group.
• Faster sorting, aggregates, subquery caching, BFS traversal with `LIMIT`, deferred evaluation of heavy columns.
### Fixed
• `GROUP BY` merged distinct values that render alike, such as sizes of 10,241 and 10,242 bytes both shown as `10.00KiB`.
• Missing boolean values sorted as `false` instead of last.
• `HAVING`, aggregate aliases and `MIN`/`MAX`/`COALESCE` corner cases.
• Correlated subqueries:caching, nested contexts and the group values they see.
• Symlink handling and ignore rules applied across symlinks.
• SVG dimension detection, MP4 duration and `exif_alt` for negative altitudes.
• Errors no longer break JSON and HTML output or the row count.
• `--help` and `--version` after other arguments; paths containing whitespace; floating point literals and unary minus.
• Windows:datetime formatting, Unix permission fields, extended attributes on musl builds.
• Subquery `OFFSET` applied twice; depth calculation; path under symlinks.