catch build files that cannot be opened in utf8 mode and emit useful error

Previously, if a junked meson.build or native.ini was used we got a
lengthy traceback ending in UnicodeDecodeError.

Fixes: #13154
Fixes: #13156

5 files changed, 28 insertions, 17 deletions

diff --git a/mesonbuild/ast/interpreter.py b/mesonbuild/ast/interpreter.py
index 53ddc10..15d2793 100644
--- a/mesonbuild/ast/interpreter.py
+++ b/mesonbuild/ast/interpreter.py
@@ -189,9 +189,7 @@ class AstInterpreter(InterpreterBase):
         if not os.path.isfile(absname):
             sys.stderr.write(f'Unable to find build file {buildfilename} --> Skipping\n')
             return
-        with open(absname, encoding='utf-8') as f:
-            code = f.read()
-        assert isinstance(code, str)
+        code = self.read_buildfile(absname, buildfilename)
         try:
             codeblock = mparser.Parser(code, absname).parse()
         except mesonlib.MesonException as me:
diff --git a/mesonbuild/coredata.py b/mesonbuild/coredata.py
index 3cf873f..7213115 100644
--- a/mesonbuild/coredata.py
+++ b/mesonbuild/coredata.py
@@ -1102,14 +1102,18 @@ class MachineFileParser():
         self.sections: T.Dict[str, T.Dict[str, T.Union[str, bool, int, T.List[str]]]] = {}
 
         for fname in filenames:
-            with open(fname, encoding='utf-8') as f:
-                content = f.read()
-                content = content.replace('@GLOBAL_SOURCE_ROOT@', sourcedir)
-                content = content.replace('@DIRNAME@', os.path.dirname(fname))
-                try:
-                    self.parser.read_string(content, fname)
-                except configparser.Error as e:
-                    raise EnvironmentException(f'Malformed machine file: {e}')
+            try:
+                with open(fname, encoding='utf-8') as f:
+                    content = f.read()
+            except UnicodeDecodeError as e:
+                raise EnvironmentException(f'Malformed machine file {fname!r} failed to parse as unicode: {e}')
+
+            content = content.replace('@GLOBAL_SOURCE_ROOT@', sourcedir)
+            content = content.replace('@DIRNAME@', os.path.dirname(fname))
+            try:
+                self.parser.read_string(content, fname)
+            except configparser.Error as e:
+                raise EnvironmentException(f'Malformed machine file: {e}')
 
         # Parse [constants] first so they can be used in other sections
         if self.parser.has_section('constants'):
diff --git a/mesonbuild/interpreter/interpreter.py b/mesonbuild/interpreter/interpreter.py
index de8f24d..f6b4cf4 100644
--- a/mesonbuild/interpreter/interpreter.py
+++ b/mesonbuild/interpreter/interpreter.py
@@ -2425,9 +2425,7 @@ class Interpreter(InterpreterBase, HoldableObject):
         if not os.path.isfile(absname):
             self.subdir = prev_subdir
             raise InterpreterException(f"Nonexistent build file '{buildfilename!s}'")
-        with open(absname, encoding='utf-8') as f:
-            code = f.read()
-        assert isinstance(code, str)
+        code = self.read_buildfile(absname, buildfilename)
         try:
             codeblock = mparser.Parser(code, absname).parse()
         except mesonlib.MesonException as me:
diff --git a/mesonbuild/interpreterbase/interpreterbase.py b/mesonbuild/interpreterbase/interpreterbase.py
index ccc3349..525d5d6 100644
--- a/mesonbuild/interpreterbase/interpreterbase.py
+++ b/mesonbuild/interpreterbase/interpreterbase.py
@@ -93,12 +93,19 @@ class InterpreterBase:
         # do nothing in an AST interpreter
         return
 
+    def read_buildfile(self, fname: str, errname: str) -> str:
+        try:
+            with open(fname, encoding='utf-8') as f:
+                return f.read()
+        except UnicodeDecodeError as e:
+            node = mparser.BaseNode(1, 1, errname)
+            raise InvalidCode.from_node(f'Build file failed to parse as unicode: {e}', node=node)
+
     def load_root_meson_file(self) -> None:
         mesonfile = os.path.join(self.source_root, self.subdir, environment.build_filename)
         if not os.path.isfile(mesonfile):
             raise InvalidArguments(f'Missing Meson file in {mesonfile}')
-        with open(mesonfile, encoding='utf-8') as mf:
-            code = mf.read()
+        code = self.read_buildfile(mesonfile, mesonfile)
         if code.isspace():
             raise InvalidCode('Builder file is empty.')
         assert isinstance(code, str)
diff --git a/mesonbuild/optinterpreter.py b/mesonbuild/optinterpreter.py
index 980dadd..599da65 100644
--- a/mesonbuild/optinterpreter.py
+++ b/mesonbuild/optinterpreter.py
@@ -78,7 +78,11 @@ class OptionInterpreter:
     def process(self, option_file: str) -> None:
         try:
             with open(option_file, encoding='utf-8') as f:
-                ast = mparser.Parser(f.read(), option_file).parse()
+                code = f.read()
+        except UnicodeDecodeError as e:
+            raise mesonlib.MesonException(f'Malformed option file {option_file!r} failed to parse as unicode: {e}')
+        try:
+            ast = mparser.Parser(code, option_file).parse()
         except mesonlib.MesonException as me:
             me.file = option_file
             raise me