diff options
| author | adisbladis <adisbladis@gmail.com> | 2024-09-03 01:11:13 +1200 |
|---|---|---|
| committer | adisbladis <adisbladis@gmail.com> | 2024-09-03 01:15:51 +1200 |
| commit | 8a207815796726eea82c690ef6d8c52ad9d87707 (patch) | |
| tree | 6b6f7edc20c34fe0735f36ffa543ed6e59c7b38c | |
| parent | lib.systems: add rustTarget for riscv32 (diff) | |
| download | nixpkgs-8a207815796726eea82c690ef6d8c52ad9d87707.tar.gz | |
lib.strings.trimWith: Refactor scoping
I wanted to use this instead of [my ownimplementation](https://github.com/nix-community/pyproject.nix/blob/65d4134/lib/util.nix#L18-L23) and got a small performance regression.
| -rw-r--r-- | lib/strings.nix | 7 |
1 files changed, 4 insertions, 3 deletions
diff --git a/lib/strings.nix b/lib/strings.nix index 54a5b9c67434..ffde541c6958 100644 --- a/lib/strings.nix +++ b/lib/strings.nix @@ -408,7 +408,6 @@ rec { start ? false, end ? false, }: - s: let # Define our own whitespace character class instead of using # `[:space:]`, which is not well-defined. @@ -425,12 +424,14 @@ rec { "(.*[^${chars}])[${chars}]*" else "(.*)"; - + in + s: + let # If the string was empty or entirely whitespace, # then the regex may not match and `res` will be `null`. res = match regex s; in - optionalString (res != null) (head res); + optionalString (res != null) (head res); /** Construct a Unix-style, colon-separated search path consisting of |