Could you include the full command you are running?

I'm running just rg café.
The literal 'TopÑAPA', mentioned in the error string, is indeed within some of my files.

I'm just having a hard time understanding, because that error message to me implies that TopÑAPA was in your pattern...

Is there anyway you can provide a full reproduction? I.e., the full command and a file that you are searching that causes the problem?

The error message also implies that the pattern is over 6,000 characters.

Yes, I also had difficulties to try to understand the error message.
I'll try to narrow this down and provide you something reproducible.
But can't do it now. I'll come back soon.
Thx.

Thanks! To be clear, I can run rg café without any issues and see correct results.

To provide more context, if I run rg '(?-u)café', where that (?-u) disables Unicode support, then I get your error message:

Error parsing regex near ')café' at character offset 9: Unicode features are not allowed when the Unicode (u) flag is not set.

In this case, it's expected behavior, because when Unicode mode is disabled, you lose the ability to search for Unicode string literals. (I can go into more gruesome detail here if anyone is curious, but it's actually a design decision in the regex engine, not ripgrep itself.)

Anyway, I kind of expect that there's some weirdness going on, so I shall look forward to more details. :-) Thank you for looking into it!

It is something with .gitignore parsing - I can repro by running c:\Users\samh\bin\rg.exe --files-with-matches "(?-u)." "c:\foo" on a folder with only the following .gitignore file:

Offset(h) 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F

00000000  EF BB BF 6E 6F 64 65 5F 6D 6F 64 75 6C 65 73 0D  node_modules.
00000010  0A 61 7A 75 72 65 5F 65 72 72 6F 72              .azure_error

yielding:

Error parsing regex near '\\])node_' at character offset 28: Unicode features are not allowed when the Unicode (u) flag is not set.

Ah, I got it now. The globber translates patterns to regexes and disables Unicode support. But this will fail if any literal is not an ASCII codepoint.

Yes! I can confirm that.
'TopÑapa' literal is present in one of my .gitignore files.
Removing that makes it all work again.

I get a similar error, I guess is my .gitignore file, but don't know how to search for '])# Cr'

rg --debug hola                                                                                                                  Unity/kleptocat develop
DEBUG:grep::search: regex ast:
Literal {
    chars: [
        'h',
        'o',
        'l',
        'a'
    ],
    casei: false
}
DEBUG:grep::literals: literal prefixes detected: Literals { lits: [Complete(hola)], limit_size: 250, limit_class: 10 }
Error parsing regex near '\\])\# Cr' at character offset 28: Unicode features are not allowed when the Unicode (u) flag is not set.
No files were searched, which means ripgrep probably applied a filter you didn't expect. Try running again with --debug.

fixed in e96e930