gh-113317: Move global utility functions into libclinic (#113986)

Establish Tools/clinic/libclinic/utils.py and move the following
functions over there:

- compute_checksum()
- create_regex()
- write_file()

diff --git a/Tools/clinic/clinic.py b/Tools/clinic/clinic.py
index f6f95580f1a177ab0f9322d9cd85244af2ccb52e..d45159f47a09893979806035d4ce161682b55da3 100755 (executable)
--- a/Tools/clinic/clinic.py
+++ b/Tools/clinic/clinic.py
@@ -16,7 +16,6 @@ import copy
 import dataclasses as dc
 import enum
 import functools
-import hashlib
 import inspect
 import io
 import itertools
@@ -1792,21 +1791,6 @@ class CLanguage(Language):
         return clinic.get_destination('block').dump()
 
 
-def create_regex(
-        before: str,
-        after: str,
-        word: bool = True,
-        whole_line: bool = True
-) -> re.Pattern[str]:
-    """Create an re object for matching marker lines."""
-    group_re = r"\w+" if word else ".+"
-    pattern = r'{}({}){}'
-    if whole_line:
-        pattern = '^' + pattern + '$'
-    pattern = pattern.format(re.escape(before), group_re, re.escape(after))
-    return re.compile(pattern)
-
-
 @dc.dataclass(slots=True, repr=False)
 class Block:
     r"""
@@ -1905,8 +1889,9 @@ class BlockParser:
         self.language = language
         before, _, after = language.start_line.partition('{dsl_name}')
         assert _ == '{dsl_name}'
-        self.find_start_re = create_regex(before, after, whole_line=False)
-        self.start_re = create_regex(before, after)
+        self.find_start_re = libclinic.create_regex(before, after,
+                                                    whole_line=False)
+        self.start_re = libclinic.create_regex(before, after)
         self.verify = verify
         self.last_checksum_re: re.Pattern[str] | None = None
         self.last_dsl_name: str | None = None
@@ -1995,7 +1980,7 @@ class BlockParser:
         else:
             before, _, after = self.language.checksum_line.format(dsl_name=dsl_name, arguments='{arguments}').partition('{arguments}')
             assert _ == '{arguments}'
-            checksum_re = create_regex(before, after, word=False)
+            checksum_re = libclinic.create_regex(before, after, word=False)
             self.last_dsl_name = dsl_name
             self.last_checksum_re = checksum_re
         assert checksum_re is not None
@@ -2029,7 +2014,7 @@ class BlockParser:
                 else:
                     checksum = d['checksum']
 
-                computed = compute_checksum(output, len(checksum))
+                computed = libclinic.compute_checksum(output, len(checksum))
                 if checksum != computed:
                     fail("Checksum mismatch! "
                          f"Expected {checksum!r}, computed {computed!r}. "
@@ -2142,8 +2127,8 @@ class BlockPrinter:
             write(output)
 
         arguments = "output={output} input={input}".format(
-            output=compute_checksum(output, 16),
-            input=compute_checksum(input, 16)
+            output=libclinic.compute_checksum(output, 16),
+            input=libclinic.compute_checksum(input, 16)
         )
         write(self.language.checksum_line.format(dsl_name=dsl_name, arguments=arguments))
         write("\n")
@@ -2245,27 +2230,6 @@ extensions: LangDict = { name: CLanguage for name in "c cc cpp cxx h hh hpp hxx"
 extensions['py'] = PythonLanguage
 
 
-def write_file(filename: str, new_contents: str) -> None:
-    try:
-        with open(filename, encoding="utf-8") as fp:
-            old_contents = fp.read()
-
-        if old_contents == new_contents:
-            # no change: avoid modifying the file modification time
-            return
-    except FileNotFoundError:
-        pass
-    # Atomic write using a temporary file and os.replace()
-    filename_new = f"{filename}.new"
-    with open(filename_new, "w", encoding="utf-8") as fp:
-        fp.write(new_contents)
-    try:
-        os.replace(filename_new, filename)
-    except:
-        os.unlink(filename_new)
-        raise
-
-
 ClassDict = dict[str, "Class"]
 DestinationDict = dict[str, Destination]
 ModuleDict = dict[str, "Module"]
@@ -2505,7 +2469,8 @@ impl_definition block
                                           core_includes=True,
                                           limited_capi=self.limited_capi,
                                           header_includes=self.includes)
-                    write_file(destination.filename, printer_2.f.getvalue())
+                    libclinic.write_file(destination.filename,
+                                         printer_2.f.getvalue())
                     continue
 
         return printer.f.getvalue()
@@ -2578,18 +2543,7 @@ def parse_file(
                     limited_capi=limited_capi)
     cooked = clinic.parse(raw)
 
-    write_file(output, cooked)
-
-
-def compute_checksum(
-        input: str | None,
-        length: int | None = None
-) -> str:
-    input = input or ''
-    s = hashlib.sha1(input.encode('utf-8')).hexdigest()
-    if length:
-        s = s[:length]
-    return s
+    libclinic.write_file(output, cooked)
 
 
 class PythonParser:

diff --git a/Tools/clinic/libclinic/__init__.py b/Tools/clinic/libclinic/__init__.py
index d4e7a0c5cf7b7656a50ce246c573ef2e6aa7b1e0..f26bf3ef6d55fe663737e4fe8613ba9088218074 100644 (file)
--- a/Tools/clinic/libclinic/__init__.py
+++ b/Tools/clinic/libclinic/__init__.py
@@ -15,6 +15,11 @@ from .formatting import (
     wrap_declarations,
     wrapped_c_string_literal,
 )
+from .utils import (
+    create_regex,
+    compute_checksum,
+    write_file,
+)
 
 
 __all__ = [
@@ -32,6 +37,11 @@ __all__ = [
     "suffix_all_lines",
     "wrap_declarations",
     "wrapped_c_string_literal",
+
+    # Utility functions
+    "create_regex",
+    "compute_checksum",
+    "write_file",
 ]
 
 

diff --git a/Tools/clinic/libclinic/utils.py b/Tools/clinic/libclinic/utils.py
new file mode 100644 (file)
index 0000000..1514558
--- /dev/null
+++ b/Tools/clinic/libclinic/utils.py
@@ -0,0 +1,45 @@
+import hashlib
+import re
+import os
+
+
+def write_file(filename: str, new_contents: str) -> None:
+    """Write new content to file, iff the content changed."""
+    try:
+        with open(filename, encoding="utf-8") as fp:
+            old_contents = fp.read()
+
+        if old_contents == new_contents:
+            # no change: avoid modifying the file modification time
+            return
+    except FileNotFoundError:
+        pass
+    # Atomic write using a temporary file and os.replace()
+    filename_new = f"{filename}.new"
+    with open(filename_new, "w", encoding="utf-8") as fp:
+        fp.write(new_contents)
+    try:
+        os.replace(filename_new, filename)
+    except:
+        os.unlink(filename_new)
+        raise
+
+
+def compute_checksum(input_: str, length: int | None = None) -> str:
+    checksum = hashlib.sha1(input_.encode("utf-8")).hexdigest()
+    if length:
+        checksum = checksum[:length]
+    return checksum
+
+
+def create_regex(
+    before: str, after: str, word: bool = True, whole_line: bool = True
+) -> re.Pattern[str]:
+    """Create a regex object for matching marker lines."""
+    group_re = r"\w+" if word else ".+"
+    before = re.escape(before)
+    after = re.escape(after)
+    pattern = fr"{before}({group_re}){after}"
+    if whole_line:
+        pattern = fr"^{pattern}$"
+    return re.compile(pattern)