Unrecognised escape sequence (\/)

[wolf99]

I am trying to add a .lycheeignore file that will ignore https://www.godaddy.com/. (Because it returns infinite HTTP 302s as it tries to redirect to a localized page).

I'd like to be able to ignore even if the trailing sforward slash is present or not.
So I construct the following regex:

godaddy\.com(\/)?$

git-ecosystem/git-credential-manager@e848ba8

However, lychee appears to barf on this regex:

+ eval lychee --format markdown --output /tmp/lychee/out.md --user-agent '"Mozilla/5.0' '(Windows' NT '10.0;' 'Win64;' 'x64)' AppleWebKit/[53](https://github.com/GitCredentialManager/git-credential-manager/runs/7055448925?check_suite_focus=true#step:3:56)7.36 '(KHTML,' like 'Gecko)' Chrome/100.0.4896.75 'Safari/537.36"' --no-progress .
++ lychee --format markdown --output /tmp/lychee/out.md --user-agent 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/100.0.4896.75 Safari/537.36' --no-progress .
Error: regex parse error:
    godaddy\.com(\/)?$
                 ^^
error: unrecognized escape sequence
+ exit_code=1

I've stared at this for about 30 mins on regex101.com, but I cannot see how this is wrong regex.

See: https://regex101.com/r/P1B4wT/1

I tried with a non-capturing group also, but no change: git-ecosystem/git-credential-manager@bc76c6e

[mre]

You also need to escape the escape character:

godaddy\.com(\\/)?$

Explanation: the first backslash is because of the extraction of the backslash character from the text file (it's a special character for text), the second is the actual regex escape.

BTW if you want to ignore all godaddy domains you can just use godaddy\.com as a pattern. Your regex currently just includes the homepage because of the $ at the end.

[mre]

Moved to the main lychee repo, because the error is reproducible without the action.

[mre]

I'm gonna close this as resolved. Feel free to reopen in case it doesn't work. 😊

[wolf99]

Not sure I grok the explanation, but it works 🤷
Thanks!

[mre]

Backslash is commonly associated with escaping the next character in a string.
I thought it fails because it should be escaped once for the string literal and once for the regex as described here.

However I think in reality it's because \/ is a no-op escape in the Rust regex crate (it does nothing as / doesn't need to be escaped) and that is not allowed. See the discussion here.
That means you can also just use

godaddy\.com/?$