Stop on first non-match after positive match

[sveniu]

Feature request summary

Stop search on first non-matching line, but only after having seen matching lines.

Motivation

On sorted or semi-sorted files, it is often useful to extract consecutive lines that match a given pattern. In some cases involving large files, time and IO can be saved by stopping the search after having matched such a run of consecutive lines.

I'm fairly convinced this feature can save countless CPU cycles and IO in big-ish data situations, particularly where timestamp matching is involved.

Example

Input:

2021-01-29T07:08:09+0100 foo
2021-01-29T08:08:09+0100 bar
2021-01-29T09:08:09+0100 baz
2021-01-29T10:08:09+0100 boz

Behaviour when searching for pattern ^2021-01-29T08 with this option enabled:

1. First line does not match; continue to next.
2. Second line matches; print it and continue to next.
3. Third line does not match; break and exit; line 4 is never scanned.

Pseudo code

have_seen_matches = null
for each line:
  if line matches:
    print line
    have_seen_matches = true
  else:
    if have_seen_matches == true:
      break

Suggested documentation

--stop-on-nonmatch
    Stop reading a file on the first non-matching line that follows a matching
    line. At least one line must have matched for this to take effect.

[BurntSushi]

Why doesn't the --max-count flag achieve what you need here?

[sveniu]

Why doesn't the --max-count flag achieve what you need here?

Because the number of matching lines is not known when starting the search. I might be misunderstanding what you actually mean, or my description or use case was unclear.