Some characters placed just after capture groups break replaced output

[tamwile]

What version of ripgrep are you using?

`rg --version

ripgrep 13.0.0
-SIMD -AVX (compiled)
+SIMD +AVX (runtime)`

How did you install ripgrep?

cargo install

What operating system are you using ripgrep on?

Archlinux 5.17.4-arch1-1

Describe your bug.

When we use the replace flag -r, some characters seem to break the replaced result
if they are placed just after a capture group like $1.
For example underscore (_).

What are the steps to reproduce the behavior?

Run
ls file.mkv | rg '(.*).mkv' -r '$1_en.ass'
Here, _ is just after the 1 in '$1', and it will break the output.
This also happen if we have any alphabet letter after the 1.

What is the actual behavior?

Here are some commands i tried.

ls file.mkv | rg '(.*).mkv' -r '$1_en.ass'
.ass
The $1 is ignored and _ has apparently eaten the "en" part of the pattern.

ls file.mkv | rg '(.*).mkv' -r '$1$_en.ass'
file.ass
Since $$ is used to escape $, i tried to write $_,
and while it still eats the "en" part, the $1 is correctly replaced this time.

ls file.mkv | rg '(.*).mkv' -r '$1 en.ass'
file en.ass
Here we can see that the output is correct if a space follow the 1, this is also the case for ., !, and probably others.

What is the expected behavior?

The expected behavior to the first command listed above is :
file_en.ass

I think this is a bug, but if this is intentional, then please show me where in the doc i could have find this, because i have not.
Is there a way to escape the capture group and/or find a way for underscore to not break the output ?

[BurntSushi]

The behavior is correct. The docs sadly do not mention the ${1} syntax. It's documented in the regex engine, but you shouldn't need to dig into that. Specifically, this bit:

The longest possible name consisting of the characters [_0-9A-Za-z] is used. e.g., $1a looks up the capture group named 1a and not the capture group at index 1. To exert more precise control over the name, or to refer to a capture group name that uses characters outside of [_0-9A-Za-z], use braces, e.g., ${1}a or ${foo[bar].baz}. When using braces, any sequence of characters is permitted. If the sequence does not refer to a capture group name in the corresponding regex, then it is replaced with an empty string.

[tamwile]

Thanks, this totally answers my question.
I leave the issue open until the doc is updated.

[brunetton]

Sorry for my duplicate in #2455

It came to my mind that this might be this kind of problem; and I've to admit: I didn't go deep enough into the doc to find this section. But, IMHO (very humble), it doesn't seem very logical to me that the priority is given to users that use named groups over users (97% I suppose) that don't.

I expected that to match $1_en.ass one should use ${1_en.ass}, and $1_en.ass to be equivalent to ${1}_en.ass

[BurntSushi]

Yes, if I could go back in time, I might tweak these rules. But I don't feel comfortable making a breaking change of this sort at this point. The problem is that it would be a silent breaking change. It would silently change the semantics of replacement strings that previously worked.