RFC: Make "disable i18n" logic optional

docs/config.md

@@ -101,6 +101,10 @@ Callables or other Python objects have to be passed in `setup.py` (via the `use_
         Setuptools will still normalize it to create the final distribution,
         so as to stay compliant with the python packaging standards.
 
+`disable_i18n`
+:   A boolean flag indicating if the values `LC_ALL="C" LANGUAGE=""` should be
+    injected into the environment when running commands such as `git` and `hg`.
+    Defaults to `True`.
 
 ## environment variables
 

src/setuptools_scm/_config.py

@@ -103,6 +103,7 @@ class Configuration:
     dist_name: str | None = None
     version_cls: type[_VersionT] = _Version
     search_parent_directories: bool = False
+    disable_i18n: bool = True
 
     parent: _t.PathT | None = None
 

src/setuptools_scm/_run_cmd.py

@@ -7,13 +7,15 @@
 import warnings
 from typing import Callable
 from typing import Mapping
+from typing import Optional
 from typing import overload
 from typing import Sequence
 from typing import TYPE_CHECKING
 from typing import TypeVar
 
 from . import _log
 from . import _types as _t
+from . import Configuration
 
 if TYPE_CHECKING:
     BaseCompletedProcess = subprocess.CompletedProcess[str]
@@ -128,25 +130,35 @@ def run(
     trace: bool = True,
     timeout: int = 20,
     check: bool = False,
+    config: Configuration | None = None,
 ) -> CompletedProcess:
     if isinstance(cmd, str):
         cmd = shlex.split(cmd)
     else:
         cmd = [os.fspath(x) for x in cmd]
     cmd_4_trace = " ".join(map(_unsafe_quote_for_display, cmd))
     log.debug("at %s\n    $ %s ", cwd, cmd_4_trace)
+
+    run_env = dict(
+        avoid_pip_isolation(no_git_env(os.environ)),
+        # os.environ,
+        HGPLAIN="1",
+    )
+    if not config or (config and config.disable_i18n):
+        run_env.update(
+            {
+                # try to disable i18n
+                "LC_ALL": "C",
+                "LANGUAGE": "",
+            }
+        )
+    log.debug("Shell environment to be used for command: %s", run_env)
+
     res = subprocess.run(
         cmd,
         capture_output=True,
         cwd=os.fspath(cwd),
-        env=dict(
-            avoid_pip_isolation(no_git_env(os.environ)),
-            # os.environ,
-            # try to disable i18n
-            LC_ALL="C",
-            LANGUAGE="",
-            HGPLAIN="1",
-        ),
+        env=run_env,
         text=True,
         timeout=timeout,
     )
@@ -171,10 +183,13 @@ def _unsafe_quote_for_display(item: _t.PathT) -> str:
 
 
 def has_command(
-    name: str, args: Sequence[str] = ["version"], warn: bool = True
+    name: str,
+    args: Sequence[str] = ["version"],
+    warn: bool = True,
+    config: Configuration | None = None,
 ) -> bool:
     try:
-        p = run([name, *args], cwd=".", timeout=5)
+        p = run([name, *args], cwd=".", timeout=5, config=config)
     except OSError as e:
         log.warning("command %s missing: %s", name, e)
         res = False
@@ -189,6 +204,6 @@ def has_command(
     return res
 
 
-def require_command(name: str) -> None:
-    if not has_command(name, warn=False):
+def require_command(name: str, config: Configuration | None = None) -> None:
+    if not has_command(name, warn=False, config=config):
         raise OSError(f"{name!r} was not found")

src/setuptools_scm/hg.py

@@ -3,7 +3,9 @@
 import datetime
 import logging
 import os
+from dataclasses import dataclass
 from pathlib import Path
+from typing import Optional
 from typing import TYPE_CHECKING
 
 from . import Configuration
@@ -22,13 +24,18 @@
 log = logging.getLogger(__name__)
 
 
+@dataclass()
 class HgWorkdir(Workdir):
+    config: Configuration | None = None
+
     @classmethod
-    def from_potential_worktree(cls, wd: _t.PathT) -> HgWorkdir | None:
-        res = _run(["hg", "root"], wd)
+    def from_potential_worktree(
+        cls, wd: _t.PathT, config: Configuration
+    ) -> HgWorkdir | None:
+        res = _run(["hg", "root"], wd, config=config)
         if res.returncode:
             return None
-        return cls(Path(res.stdout))
+        return cls(path=Path(res.stdout), config=config)
 
     def get_meta(self, config: Configuration) -> ScmVersion | None:
         node: str
@@ -47,6 +54,7 @@ def get_meta(self, config: Configuration) -> ScmVersion | None:
             ["hg", "id", "-T", "{branch}\n{if(dirty, 1, 0)}\n{date|shortdate}"],
             cwd=self.path,
             check=True,
+            config=config,
         ).stdout.split("\n")
         dirty = bool(int(dirty_str))
         node_date = datetime.date.fromisoformat(dirty_date if dirty else node_date_str)
@@ -109,7 +117,7 @@ def get_meta(self, config: Configuration) -> ScmVersion | None:
     def hg_log(self, revset: str, template: str) -> str:
         cmd = ["hg", "log", "-r", revset, "-T", template]
 
-        return _run(cmd, cwd=self.path, check=True).stdout
+        return _run(cmd, cwd=self.path, check=True, config=self.config).stdout
 
     def get_latest_normalizable_tag(self) -> str | None:
         # Gets all tags containing a '.' (see #229) from oldest to newest
@@ -143,9 +151,9 @@ def check_changes_since_tag(self, tag: str | None) -> bool:
 
 
 def parse(root: _t.PathT, config: Configuration) -> ScmVersion | None:
-    _require_command("hg")
+    _require_command("hg", config=config)
     if os.path.exists(os.path.join(root, ".hg/git")):
-        res = _run(["hg", "path"], root)
+        res = _run(["hg", "path"], root, config=config)
         if not res.returncode:
             for line in res.stdout.split("\n"):
                 if line.startswith("default ="):
@@ -158,7 +166,7 @@ def parse(root: _t.PathT, config: Configuration) -> ScmVersion | None:
                         if wd_hggit:
                             return _git_parse_inner(config, wd_hggit)
 
-    wd = HgWorkdir.from_potential_worktree(config.absolute_root)
+    wd = HgWorkdir.from_potential_worktree(config.absolute_root, config=config)
 
     if wd is None:
         return None