diff --git a/crates/bashkit/src/builtins/rg/mod.rs b/crates/bashkit/src/builtins/rg/mod.rs index 534170b86..f12ded94a 100644 --- a/crates/bashkit/src/builtins/rg/mod.rs +++ b/crates/bashkit/src/builtins/rg/mod.rs @@ -52,6 +52,7 @@ struct RgOptions { text: bool, binary: bool, search_zip: bool, + preprocessor: Option, crlf: bool, multiline: bool, multiline_dotall: bool, @@ -187,6 +188,7 @@ impl RgOptions { text: false, binary: false, search_zip: false, + preprocessor: None, crlf: false, multiline: false, multiline_dotall: false, @@ -398,6 +400,12 @@ impl RgOptions { opts.search_zip = true; } else if p.flag("--no-search-zip") { opts.search_zip = false; + } else if let Some(val) = long_value(&mut p, "--pre")? { + opts.preprocessor = Some(val); + } else if p.flag("--no-pre") { + opts.preprocessor = None; + } else if let Some(val) = long_value(&mut p, "--pre-glob")? { + let _ = RgGlobRule::parse(&val, false, opts.glob_case_insensitive)?; } else if p.flag("--crlf") { opts.crlf = true; opts.null_data = false; @@ -913,6 +921,14 @@ impl RgOptions { } opts.paths = positional; + if let Some(preprocessor) = &opts.preprocessor + && !preprocessor.is_empty() + && preprocessor != "cat" + { + return Err(Error::Execution(format!( + "rg: unsupported --pre command: {preprocessor}" + ))); + } Ok(opts) } @@ -2622,7 +2638,7 @@ fn append_rg_stats(output: &mut String, stats: &RgSearchStats, bytes_printed: us #[async_trait] impl Builtin for Rg { async fn execute(&self, ctx: Context<'_>) -> Result { - let help_text = "Usage: rg [OPTIONS] PATTERN [PATH...]\nRecursively search for a pattern.\n\n -i, --ignore-case\tcase insensitive\n -S, --smart-case\tcase insensitive if pattern is lowercase\n -s, --case-sensitive\tcase sensitive\n -n, --line-number\tshow line numbers\n -N, --no-line-number\tsuppress line numbers\n --column\tshow column numbers\n -b, --byte-offset\tshow byte offsets\n --vimgrep\tshow file:line:column:match lines\n --json\tshow JSON Lines events\n --stats\tshow search statistics\n --null\tterminate path fields with NUL\n -c, --count\tcount matching lines\n --count-matches\tcount individual matches\n --include-zero\tinclude zero counts\n -l, --files-with-matches\tfiles with matches\n --files-without-match\tfiles without matches\n --files\tprint files that would be searched\n -v, --invert-match\tinvert match\n -w, --word-regexp\tword boundary\n -x, --line-regexp\tmatch whole lines\n -F, --fixed-strings\tfixed strings (literal)\n -a, --text\tsearch binary files as text\n --binary\tsearch binary files and print binary-match summaries\n -z, --search-zip\tsearch gzip-compressed files\n --no-search-zip\tdisable compressed file search\n --crlf\ttreat CRLF as line terminators for $ anchors\n --no-crlf\tdisable CRLF line terminator mode\n -U, --multiline\tenable matching across line boundaries\n --no-multiline\tdisable multiline matching\n --multiline-dotall\tmake . match line terminators in multiline mode\n --no-multiline-dotall\tdisable multiline dotall mode\n -o, --only-matching\tshow only matching text\n -q, --quiet\tsuppress output; exit status only\n -e, --regexp PATTERN\tuse PATTERN for matching\n -f, --file PATTERNFILE\tread patterns from file\n -E, --encoding ENCODING\tdecode searched files using ENCODING\n -r, --replace REPLACEMENT\treplace matches in output\n --passthru\tprint matching and non-matching lines\n --trim\ttrim whitespace from output lines\n -m, --max-count NUM\tmax count per file\n -M, --max-columns NUM\tomit lines longer than NUM columns\n --max-columns-preview\tshow prefixes of long lines\n -j, --threads NUM\tset number of search threads (no-op)\n --regex-size-limit NUM\tset regex size limit (no-op)\n --dfa-size-limit NUM\tset DFA size limit (no-op)\n --max-depth NUM\tlimit recursive directory depth\n -A, --after-context NUM\tshow trailing context\n -B, --before-context NUM\tshow leading context\n -C, --context NUM\tshow leading and trailing context\n --context-separator SEP\tset context group separator\n --field-match-separator SEP\tset match field separator\n --field-context-separator SEP\tset context field separator\n -p, --pretty\talias for heading plus line numbers\n --heading\tgroup matches by file\n --no-heading\tdisable heading output\n --sort SORTBY\tsort paths (path only)\n --sortr SORTBY\tsort paths in reverse (path only)\n --sort-files\tsort --files output\n --path-separator SEP\tset displayed path separator\n -g, --glob GLOB\tinclude/exclude paths by glob (!GLOB excludes)\n -t, --type TYPE\tinclude files matching TYPE\n -T, --type-not TYPE\texclude files matching TYPE\n --type-add TYPE:GLOB\tadd a file type glob\n --type-clear TYPE\tclear a file type definition\n --type-list\tshow file type definitions\n --ignore-file FILE\tuse additional ignore file\n --no-ignore\tdo not use ignore files\n --no-ignore-dot\tdo not use .ignore files\n --no-ignore-vcs\tdo not use .gitignore files\n --no-require-git\tuse .gitignore outside git repositories\n --require-git\trequire a git repository for .gitignore files\n -u, --unrestricted\treduce filtering (repeatable)\n --messages\tshow file read diagnostics\n --no-messages\tsuppress file read diagnostics\n --hidden\tsearch hidden files and directories\n --no-hidden\tdo not search hidden files and directories\n -H, --with-filename\tshow filename\n -I, --no-filename\tsuppress filename\n --line-buffered\tforce line buffering (no-op)\n --block-buffered\tforce block buffering (no-op)\n --no-config\tdo not read config files (no-op)\n --mmap\tsearch using memory maps when possible (no-op)\n --no-mmap\tdisable memory maps (no-op)\n -P, --pcre2\tuse PCRE2 regex engine for supported patterns (no-op)\n --no-pcre2\tdisable PCRE2 regex engine (no-op)\n --pcre2-version\tshow PCRE2 version information\n --unicode\tenable Unicode mode (no-op)\n --no-unicode\tdisable Unicode mode (no-op)\n --pcre2-unicode\tenable PCRE2 Unicode mode (no-op)\n --no-pcre2-unicode\tdisable PCRE2 Unicode mode (no-op)\n --engine ENGINE\tselect regex engine: default, auto, pcre2 (no-op)\n --auto-hybrid-regex\tuse PCRE2 when needed (no-op)\n --no-auto-hybrid-regex\tdisable auto hybrid regex (no-op)\n --color MODE\tcolor output (no-op)\n --colors SPEC\tconfigure colors (no-op)\n -h, --help\tdisplay this help and exit\n -V, --version\toutput version information and exit\n"; + let help_text = "Usage: rg [OPTIONS] PATTERN [PATH...]\nRecursively search for a pattern.\n\n -i, --ignore-case\tcase insensitive\n -S, --smart-case\tcase insensitive if pattern is lowercase\n -s, --case-sensitive\tcase sensitive\n -n, --line-number\tshow line numbers\n -N, --no-line-number\tsuppress line numbers\n --column\tshow column numbers\n -b, --byte-offset\tshow byte offsets\n --vimgrep\tshow file:line:column:match lines\n --json\tshow JSON Lines events\n --stats\tshow search statistics\n --null\tterminate path fields with NUL\n -c, --count\tcount matching lines\n --count-matches\tcount individual matches\n --include-zero\tinclude zero counts\n -l, --files-with-matches\tfiles with matches\n --files-without-match\tfiles without matches\n --files\tprint files that would be searched\n -v, --invert-match\tinvert match\n -w, --word-regexp\tword boundary\n -x, --line-regexp\tmatch whole lines\n -F, --fixed-strings\tfixed strings (literal)\n -a, --text\tsearch binary files as text\n --binary\tsearch binary files and print binary-match summaries\n -z, --search-zip\tsearch gzip-compressed files\n --no-search-zip\tdisable compressed file search\n --pre COMMAND\trun a preprocessor before searching (cat/empty supported)\n --pre-glob GLOB\tlimit preprocessor paths by glob\n --no-pre\tdisable preprocessing\n --crlf\ttreat CRLF as line terminators for $ anchors\n --no-crlf\tdisable CRLF line terminator mode\n -U, --multiline\tenable matching across line boundaries\n --no-multiline\tdisable multiline matching\n --multiline-dotall\tmake . match line terminators in multiline mode\n --no-multiline-dotall\tdisable multiline dotall mode\n -o, --only-matching\tshow only matching text\n -q, --quiet\tsuppress output; exit status only\n -e, --regexp PATTERN\tuse PATTERN for matching\n -f, --file PATTERNFILE\tread patterns from file\n -E, --encoding ENCODING\tdecode searched files using ENCODING\n -r, --replace REPLACEMENT\treplace matches in output\n --passthru\tprint matching and non-matching lines\n --trim\ttrim whitespace from output lines\n -m, --max-count NUM\tmax count per file\n -M, --max-columns NUM\tomit lines longer than NUM columns\n --max-columns-preview\tshow prefixes of long lines\n -j, --threads NUM\tset number of search threads (no-op)\n --regex-size-limit NUM\tset regex size limit (no-op)\n --dfa-size-limit NUM\tset DFA size limit (no-op)\n --max-depth NUM\tlimit recursive directory depth\n -A, --after-context NUM\tshow trailing context\n -B, --before-context NUM\tshow leading context\n -C, --context NUM\tshow leading and trailing context\n --context-separator SEP\tset context group separator\n --field-match-separator SEP\tset match field separator\n --field-context-separator SEP\tset context field separator\n -p, --pretty\talias for heading plus line numbers\n --heading\tgroup matches by file\n --no-heading\tdisable heading output\n --sort SORTBY\tsort paths (path only)\n --sortr SORTBY\tsort paths in reverse (path only)\n --sort-files\tsort --files output\n --path-separator SEP\tset displayed path separator\n -g, --glob GLOB\tinclude/exclude paths by glob (!GLOB excludes)\n -t, --type TYPE\tinclude files matching TYPE\n -T, --type-not TYPE\texclude files matching TYPE\n --type-add TYPE:GLOB\tadd a file type glob\n --type-clear TYPE\tclear a file type definition\n --type-list\tshow file type definitions\n --ignore-file FILE\tuse additional ignore file\n --no-ignore\tdo not use ignore files\n --no-ignore-dot\tdo not use .ignore files\n --no-ignore-vcs\tdo not use .gitignore files\n --no-require-git\tuse .gitignore outside git repositories\n --require-git\trequire a git repository for .gitignore files\n -u, --unrestricted\treduce filtering (repeatable)\n --messages\tshow file read diagnostics\n --no-messages\tsuppress file read diagnostics\n --hidden\tsearch hidden files and directories\n --no-hidden\tdo not search hidden files and directories\n -H, --with-filename\tshow filename\n -I, --no-filename\tsuppress filename\n --line-buffered\tforce line buffering (no-op)\n --block-buffered\tforce block buffering (no-op)\n --no-config\tdo not read config files (no-op)\n --mmap\tsearch using memory maps when possible (no-op)\n --no-mmap\tdisable memory maps (no-op)\n -P, --pcre2\tuse PCRE2 regex engine for supported patterns (no-op)\n --no-pcre2\tdisable PCRE2 regex engine (no-op)\n --pcre2-version\tshow PCRE2 version information\n --unicode\tenable Unicode mode (no-op)\n --no-unicode\tdisable Unicode mode (no-op)\n --pcre2-unicode\tenable PCRE2 Unicode mode (no-op)\n --no-pcre2-unicode\tdisable PCRE2 Unicode mode (no-op)\n --engine ENGINE\tselect regex engine: default, auto, pcre2 (no-op)\n --auto-hybrid-regex\tuse PCRE2 when needed (no-op)\n --no-auto-hybrid-regex\tdisable auto hybrid regex (no-op)\n --color MODE\tcolor output (no-op)\n --colors SPEC\tconfigure colors (no-op)\n -h, --help\tdisplay this help and exit\n -V, --version\toutput version information and exit\n"; let help_text = help_text.replace( " --max-depth NUM\tlimit recursive directory depth\n", " -d, --max-depth NUM\tlimit recursive directory depth\n --maxdepth NUM\talias for --max-depth\n", @@ -5406,6 +5422,38 @@ mod tests { cwd: "/", output: RgDiffOutput::Exact, }, + RgDiffCase { + name: "pre cat explicit file", + args: &["--pre", "cat", "needle", "proj/a.txt"], + stdin: None, + files: DIFF_BASIC_FILES, + cwd: "/", + output: RgDiffOutput::Exact, + }, + RgDiffCase { + name: "pre empty explicit file", + args: &["--pre=", "needle", "proj/a.txt"], + stdin: None, + files: DIFF_BASIC_FILES, + cwd: "/", + output: RgDiffOutput::Exact, + }, + RgDiffCase { + name: "pre glob with cat recursive", + args: &["--pre=cat", "--pre-glob", "*.txt", "needle", "proj"], + stdin: None, + files: DIFF_BASIC_FILES, + cwd: "/", + output: RgDiffOutput::UnorderedLines, + }, + RgDiffCase { + name: "no pre disables pre command", + args: &["--pre=false", "--no-pre", "needle", "proj/a.txt"], + stdin: None, + files: DIFF_BASIC_FILES, + cwd: "/", + output: RgDiffOutput::Exact, + }, RgDiffCase { name: "missing file keeps stdout and exits 2", args: &["needle", "proj/a.txt", "proj/missing.txt"], @@ -5805,6 +5853,9 @@ mod tests { assert!(long_help.stdout.contains("--multiline")); assert!(long_help.stdout.contains("--multiline-dotall")); assert!(long_help.stdout.contains("-z, --search-zip")); + assert!(long_help.stdout.contains("--pre COMMAND")); + assert!(long_help.stdout.contains("--pre-glob")); + assert!(long_help.stdout.contains("--no-pre")); assert!(long_help.stdout.contains("--no-ignore-files")); assert!(long_help.stdout.contains("--ignore-file-case-insensitive")); assert!(long_help.stdout.contains("--ignore-dot")); @@ -6076,6 +6127,21 @@ mod tests { assert!(result.is_err()); } + #[tokio::test] + async fn test_rg_unsupported_preprocessor_errors() { + let args: Vec = vec![ + "--pre".to_string(), + "sed s/x/y/".to_string(), + "needle".to_string(), + "file.txt".to_string(), + ]; + let result = RgOptions::parse(&args); + assert!(matches!( + result, + Err(Error::Execution(msg)) if msg == "rg: unsupported --pre command: sed s/x/y/" + )); + } + #[tokio::test] async fn test_rg_basic_match() { let result = run_rg(