Add a warning for windows users wrt stdout piping

Fix #905

Author: Yomguithereal

minet/cli/run.py

@@ -15,7 +15,6 @@
 import multiprocessing
 import casanova
 import shlex
-import platform
 from casanova.exceptions import MissingColumnError
 from contextlib import ExitStack
 
@@ -35,7 +34,6 @@
 from minet.cli.argparse import resolve_arg_dependencies, build_parser, get_subparser
 from minet.cli.exceptions import NotResumableError, InvalidArgumentsError, FatalError
 
-WINDOWS = "windows" in platform.system().lower()
 
 GLOBAL_SETUP_IS_DONE = False
 
@@ -48,11 +46,6 @@ def global_setup() -> None:
 
     GLOBAL_SETUP_IS_DONE = True
 
-    # Issue #497, utf-8 encoding for windows stdout
-    if WINDOWS:
-        sys.__stdout__.reconfigure(encoding="utf-8")
-        sys.__stderr__.reconfigure(encoding="utf-8")
-
     # Freezing multiprocessing support for pyinstaller etc.
     multiprocessing.freeze_support()
 

minet/cli/utils.py

@@ -139,7 +139,28 @@ def emit(self, record):
 def acquire_cross_platform_stdout():
     # As per #254: stdout need to be wrapped so that windows get a correct csv
     # stream output
+    # As per #905: powershell sometimes writes a UTF-16-LE BOM in the stdout
+    # before one has any chance to reconfigure the stream as UTF-8
+    # Reference: https://stackoverflow.com/questions/68487529/how-to-ensure-python-prints-utf-8-and-not-utf-16-le-when-piped-in-powershell
     if "windows" in platform.system().lower():
+        console.vprint(
+            [
+                "You piped the result of your command into stdout.",
+                'Stdout piping (">") is known to be unreliable on Windows shells (cmd.exe, PowerShell)',
+                "and can sometimes result in corrupted data.",
+                "",
+                "Are you sure you know what you are doing?",
+                "If you are not sure, please use the --output flag instead.",
+                "",
+            ],
+            style="warning",
+        )
+
+        # As per #497: we reconfigure the stdout to be UTF-8 on
+        # windows
+        sys.__stdout__.reconfigure(encoding="utf-8")
+        sys.__stderr__.reconfigure(encoding="utf-8")
+
         return open(
             sys.__stdout__.fileno(),
             mode=sys.__stdout__.mode,