swidth: use cross platform implementation, fixes #7493

This implementation should be good enough for our usecase (paths) and has no external dependencies.

There is also a wcwidth library which might be even better, but would add another dependency.

Author: ThomasWaldmann

src/borg/platform/__init__.py

@@ -10,7 +10,7 @@
 
 from .base import ENOATTR, API_VERSION
 from .base import SaveFile, sync_dir, fdatasync, safe_fadvise
-from .base import get_process_id, fqdn, hostname, hostid
+from .base import get_process_id, fqdn, hostname, hostid, swidth
 
 # work around pyinstaller "forgetting" to include the xattr module
 from . import xattr  # noqa: F401
@@ -24,7 +24,6 @@
     from .linux import set_flags, get_flags
     from .linux import SyncFile
     from .posix import process_alive, local_pid_alive
-    from .posix import swidth
     from .posix import get_errno
     from .posix import getosusername
     from . import posix_ug as platform_ug
@@ -36,7 +35,6 @@
     from .base import get_flags
     from .base import SyncFile
     from .posix import process_alive, local_pid_alive
-    from .posix import swidth
     from .posix import get_errno
     from .posix import getosusername
     from . import posix_ug as platform_ug
@@ -47,7 +45,6 @@
     from .base import set_flags, get_flags
     from .base import SyncFile
     from .posix import process_alive, local_pid_alive
-    from .posix import swidth
     from .posix import get_errno
     from .posix import getosusername
     from . import posix_ug as platform_ug
@@ -60,7 +57,6 @@
     from .base import get_flags
     from .base import SyncFile
     from .posix import process_alive, local_pid_alive
-    from .posix import swidth
     from .posix import get_errno
     from .posix import getosusername
     from . import posix_ug as platform_ug
@@ -72,7 +68,6 @@
     from .base import set_flags, get_flags
     from .base import SyncFile
     from .posix import process_alive, local_pid_alive
-    from .posix import swidth
     from .posix import get_errno
     from .posix import getosusername
     from . import posix_ug as platform_ug
@@ -84,7 +79,6 @@
     from .base import set_flags, get_flags
     from .base import SyncFile
     from .windows import process_alive, local_pid_alive
-    from .base import swidth
     from .windows import getosusername
     from . import windows_ug as platform_ug
 

src/borg/platform/base.py

@@ -1,6 +1,7 @@
 import errno
 import os
 import socket
+import unicodedata
 import uuid
 from pathlib import Path
 
@@ -266,7 +267,20 @@ def swidth(s):
 
     For western scripts, this is just len(s), but for cjk glyphs, 2 cells are used.
     """
-    return len(s)
+    width = 0
+    for char in s:
+        # Get the East Asian Width property
+        ea_width = unicodedata.east_asian_width(char)
+
+        # Wide (W) and Fullwidth (F) characters take 2 cells
+        if ea_width in ("W", "F"):
+            width += 2
+        # Not a zero-width characters (combining marks, format characters)
+        elif unicodedata.category(char) not in ("Mn", "Me", "Cf"):
+            # Normal characters take 1 cell
+            width += 1
+
+    return width
 
 
 # patched socket.getfqdn() - see https://bugs.python.org/issue5004

src/borg/platform/posix.pyx

@@ -5,34 +5,11 @@ from . import posix_ug
 
 from libc.errno cimport errno as c_errno
 
-from cpython.mem cimport PyMem_Free
-from libc.stddef cimport wchar_t
-
-cdef extern from "wchar.h":
-    # https://www.man7.org/linux/man-pages/man3/wcswidth.3.html
-    cdef int wcswidth(const wchar_t *s, size_t n)
-
-
-cdef extern from "Python.h":
-    # https://docs.python.org/3/c-api/unicode.html#c.PyUnicode_AsWideCharString
-    wchar_t* PyUnicode_AsWideCharString(object, Py_ssize_t*) except NULL
-
 
 def get_errno():
     return c_errno
 
 
-def swidth(s):
-    cdef Py_ssize_t size
-    cdef wchar_t *as_wchar = PyUnicode_AsWideCharString(s, &size)
-    terminal_width = wcswidth(as_wchar, <size_t>size)
-    PyMem_Free(as_wchar)
-    if terminal_width >= 0:
-        return terminal_width
-    else:
-        return len(s)
-
-
 def process_alive(host, pid, thread):
     """
     Check whether the (host, pid, thread_id) combination corresponds to a process potentially alive.

src/borg/testsuite/platform/all_test.py

@@ -0,0 +1,13 @@
+from ...platform import swidth
+
+
+def test_swidth_ascii():
+    assert swidth("borg") == 4
+
+
+def test_swidth_cjk():
+    assert swidth("バックアップ") == 6 * 2
+
+
+def test_swidth_mixed():
+    assert swidth("borgバックアップ") == 4 + 6 * 2