fix(ci): remove invalid --version=14 clang-format argument

[kou]

Fixes #1461.

[kou]

We may want to use the official(?) clang-format plugin https://github.com/pre-commit/mirrors-clang-format :

diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
index c13ad362..3cfd2924 100644
--- a/.pre-commit-config.yaml
+++ b/.pre-commit-config.yaml
@@ -39,12 +39,13 @@ repos:
       files: '\.(bat|sln)$'
     - id: trailing-whitespace
       exclude: "^r/.*?/_snaps/.*?.md$"
-  - repo: https://github.com/pocc/pre-commit-hooks
-    rev: v1.3.5
+  - repo: https://github.com/pre-commit/mirrors-clang-format
+    rev: "v14.0.6"
     hooks:
       - id: clang-format
-        args: [-i]
-        types_or: [c, c++]
+        types_or:
+          - c++
+          - c
   - repo: https://github.com/cheshirekow/cmake-format-precommit
     rev: v0.6.13
     hooks:

[kou]

https://github.com/pocc/pre-commit-hooks doesn't work if clang-format in $PATH isn't clang-format 14. But https://github.com/pre-commit/mirrors-clang-format works even when clang-format in $PATH isn't clang-format 14 because it installs clang-format from PyPI internally.

[kou]

This is for avoiding name conflict on Windows.
windef.h defines min macro. It conflicts with std::numeric_limits::min(). If we use (std::numeric_limits::min)(), we can avoid the min macro expansion.

[kou]

ditto.

[lidavidm]

Thanks Kou!