the --stats flag should note that it is unconditionally enabled when --json is used

[cstyles]

What version of ripgrep are you using?

ripgrep 13.0.0
-SIMD -AVX (compiled)

How did you install ripgrep?

Homebrew

What operating system are you using ripgrep on?

macOS 12.6

Describe your bug.

When ripgrep is invoked with --json, the final line is always a stats summary. It seems like this is included implicitly whenever --json is used. However, I would expect that explicitly including --no-stats would disable the summary.

What are the steps to reproduce the behavior?

Invoke ripgrep with both the --json and --no-stats flags.

What is the actual behavior?

$ rg --debug --json --no-stats pattern file.txt
DEBUG|rg::config|crates/core/config.rs:40: /Users/cstyles/.ripgreprc: arguments loaded from config file: ["--smart-case", "--max-columns=300", "--max-columns-preview"]
DEBUG|rg::args|crates/core/args.rs:543: final argv: ["rg", "--smart-case", "--max-columns=300", "--max-columns-preview", "--debug", "--json", "--no-stats", "pattern", "file.txt"]
DEBUG|grep_regex::literal|crates/regex/src/literal.rs:109: required literals found: [Cut(PATTE), Cut(pATTE), Cut(PaTTE), Cut(paTTE), Cut(PAtTE), Cut(pAtTE), Cut(PatTE), Cut(patTE), Cut(PATtE), Cut(pATtE), Cut(PaTtE), Cut(paTtE), Cut(PAttE), Cut(pAttE), Cut(PattE), Cut(pattE), Cut(PATTe), Cut(pATTe), Cut(PaTTe), Cut(paTTe), Cut(PAtTe), Cut(pAtTe), Cut(PatTe), Cut(patTe), Cut(PATte), Cut(pATte), Cut(PaTte), Cut(paTte), Cut(PAtte), Cut(pAtte), Cut(Patte), Cut(patte)]
DEBUG|grep_regex::matcher|crates/regex/src/matcher.rs:50: extracted fast line regex: (?-u:PATTE|pATTE|PaTTE|paTTE|PAtTE|pAtTE|PatTE|patTE|PATtE|pATtE|PaTtE|paTtE|PAttE|pAttE|PattE|pattE|PATTe|pATTe|PaTTe|paTTe|PAtTe|pAtTe|PatTe|patTe|PATte|pATte|PaTte|paTte|PAtte|pAtte|Patte|patte)
DEBUG|globset|crates/globset/src/lib.rs:421: built glob set; 0 literals, 0 basenames, 12 extensions, 0 prefixes, 0 suffixes, 0 required extensions, 0 regexes
DEBUG|globset|crates/globset/src/lib.rs:421: built glob set; 0 literals, 4 basenames, 2 extensions, 0 prefixes, 0 suffixes, 0 required extensions, 0 regexes
{"type":"begin","data":{"path":{"text":"file.txt"}}}
{"type":"match","data":{"path":{"text":"file.txt"},"lines":{"text":"pattern\n"},"line_number":2,"absolute_offset":7,"submatches":[{"match":{"text":"pattern"},"start":0,"end":7}]}}
{"type":"end","data":{"path":{"text":"file.txt"},"binary_offset":null,"stats":{"elapsed":{"secs":0,"nanos":42958,"human":"0.000043s"},"searches":1,"searches_with_match":1,"bytes_searched":21,"bytes_printed":233,"matched_lines":1,"matches":1}}}
{"data":{"elapsed_total":{"human":"0.001521s","nanos":1521416,"secs":0},"stats":{"bytes_printed":233,"bytes_searched":21,"elapsed":{"human":"0.000043s","nanos":42958,"secs":0},"matched_lines":1,"matches":1,"searches":1,"searches_with_match":1}},"type":"summary"}

What is the expected behavior?

Including the --no-stats flag should disable the stats summary.

[BurntSushi]

This is intended behavior at the moment:

ripgrep/crates/core/args.rs

Lines 1558 to 1566 in eab044d

	/// Returns true if and only if aggregate statistics for a search should
	/// be tracked.
	///
	/// Generally, this is only enabled when explicitly requested by in the
	/// command line arguments via the --stats flag, but this can also be
	/// enabled implicitly via the output format, e.g., for JSON Lines.
	fn stats(&self) -> bool {
	self.output_kind() == OutputKind::JSON || self.is_present("stats")
	}

I'm not inclined to change this. But the --stats flag could use an extra sentence noting that it appears in the JSON output unconditionally and can't be disabled. There is no real benefit to disabling it. It only uses one line, and the JSON format itself requires the work needed to compute the stats anyway. So there's really no extra cost.

[cstyles]

So there's really no extra cost.

The cost comes in the form of complexity in whatever consumes the JSON output. Even if I don't care about the statistics, I still need to add logic to either handle or ignore them. It's not a huge deal but being able to exclude them would be nice and, IMO, more intuitive and consistent. Ultimately it's your call though.

[BurntSushi]

I acknowledge there is an increase in complexity, but one that is extremely small IMO. Small enough to keep it unconditionally enabled.

[cstyles]

True. My thought was that the required change to ripgrep / grep-printer would also be extremely small. That said, if you disagree I won't be offended. Feel free to close this.