Merge 13ad58351a into 3afb7e1a21

docs/CONTRIBUTORS

@@ -284,6 +284,7 @@ Thomas Brown
 Thomas Dybdahl Ahle
 Thomas Santerre
 Tim Hutt
+Tim Paine
 Tim Snyder
 Tobias Jensen
 Tobias Rosenkranz

src/flexfix

@@ -12,10 +12,48 @@
 ######################################################################
 # DESCRIPTION: Edits flex output to get around various broken flex issues.
 
+import os
 import re
+import subprocess
 import sys
 import platform
 
+
+def _flexlexer_uses_size_t() -> bool:
+    """Return True if the FlexLexer.h that will be used declares LexerInput with size_t."""
+    candidates = []
+    env_paths = os.environ.get("CPATH", "") + os.pathsep + os.environ.get("CPLUS_INCLUDE_PATH", "")
+    for d in env_paths.split(os.pathsep):
+        if d:
+            candidates.append(os.path.join(d, "FlexLexer.h"))
+    if platform.system() == "Darwin":
+        try:
+            sdk = subprocess.check_output(["xcrun", "--show-sdk-path"],
+                                          stderr=subprocess.DEVNULL).decode().strip()
+            if sdk:
+                candidates.append(os.path.join(sdk, "usr", "include", "FlexLexer.h"))
+        except (OSError, subprocess.CalledProcessError):
+            pass
+        candidates.extend([
+            "/opt/homebrew/include/FlexLexer.h",
+            "/usr/local/include/FlexLexer.h",
+        ])
+    candidates.extend(["/usr/include/FlexLexer.h"])
+    for path in candidates:
+        try:
+            with open(path, "r", encoding="utf-8", errors="ignore") as fh:
+                text = fh.read()
+        except OSError:
+            continue
+        if re.search(r"\bLexerInput\s*\(\s*char\s*\*\s*\w*\s*,\s*size_t\b", text):
+            return True
+        if re.search(r"\bLexerInput\s*\(\s*char\s*\*\s*\w*\s*,\s*int\b", text):
+            return False
+    return False
+
+
+_LEXER_USES_SIZE_T = _flexlexer_uses_size_t()
+
 for line in sys.stdin:
     # Fix flex 2.6.1 warning
     line = re.sub(r'for \( i = 0; i < _yybytes_len; \+\+i \)',
@@ -48,14 +86,12 @@ for line in sys.stdin:
     line = re.sub(r'(#line \d+ ".*)_pretmp', r'\1', line)
     # Fix 'register' storage class specifier is deprecated and incompatible with C++17
     line = re.sub(r'register ', '', line)
-    # Change int to size_t on LexerInput/LexerOutput when using macOS Tahoe
+    # Change int to size_t on LexerInput/LexerOutput when the FlexLexer.h in use
-    if platform.system() == "Darwin":
+    # declares them with size_t (e.g. macOS Tahoe's system flex). Skip when the
-        try:
+    # actual FlexLexer.h still uses int (e.g. Homebrew flex 2.6.4) to avoid an
-            osv = platform.mac_ver()[0]  # Example "26.2"
+    # out-of-line-definition mismatch.
-            if len(osv) > 2 and osv[2] == "." and int(osv[:2]) >= 26:
+    if _LEXER_USES_SIZE_T:
-                if "::LexerInput(" in line or "::LexerOutput(" in line:
+        if "::LexerInput(" in line or "::LexerOutput(" in line:
-                    line = re.sub(r'int ', r'size_t ', line)
+            line = re.sub(r'int ', r'size_t ', line)
-        except IndexError as e:
-            print(f"Unexpected index error {e}")
 
     print(line, end='')