--max-count has precedence over --after-context and --context

[chelmertz]

I think --after-context (-A) and --context (-C) should have the same precedence as --before-context (-B), when being combined with --max-count (-m), but they do not:

$ cat > rgtest.txt <<EOF
> a
> b
> c
> d
> e
> EOF
$ rg -B2 c rgtest.txt
1-a
2-b
3:c
$ rg -C2 c rgtest.txt
1-a
2-b
3:c
4-d
5-e
$ rg -A2 c rgtest.txt
3:c
4-d
5-e
$ # and now with --max-count:
$ rg -B2 c -m1 rgtest.txt
1-a
2-b
3:c
$ rg -C2 c -m1 rgtest.txt
1-a
2-b
3:c
$ rg -A2 c -m1 rgtest.txt
3:c

From my expectations, -B works correctly, -C works like -B which is unexpected, and -A has no effect at all.

My use case is to show a colleague one instance of many repeated matches, but also with some context. If -A combined with -m would yield the same result as with -B, then my use case would have been fulfilled.

I am using ripgrep-0.3.2-1.fc24.x86_64, packaged for Fedora through copr (by carlgeorge-ripgrep).

Thanks for a great program! (edit: removed trailing whitespace in code block)

[BurntSushi]

Thanks for the bug report. This is definitely a bug. But I don't think it has anything to do with precedence specifically. For example:

[andrew@Cheetah ripgrep-402] cat > rgtest.txt <<EOF
> a
> b
> c
> d
> e
> a
> b
> c
> d
> e
> EOF
[andrew@Cheetah ripgrep-402] rg -C2 c -m1 rgtest.txt
1-a
2-b
3:c
[andrew@Cheetah ripgrep-402] rg -C2 c -m2 rgtest.txt
1-a
2-b
3:c
4-d
5-e
6-a
7-b
8:c

Notice that in the last example, the "after" context for the first match is printed correctly. Only the last match loses it.

This is probably just a bug in the search code where it neglects to print the context after the last match if it decides to quit early. I'll note that grep handles this correctly:

[andrew@Cheetah ripgrep-402] grep -C2 c -m2 rgtest.txt
a
b
c
d
e
a
b
c
d
e

This could be a little tricky to fix. Not sure.