parser: preserve escape chars in strings

use separate Node for multiline strings

13 files changed, 74 insertions, 49 deletions

diff --git a/mesonbuild/ast/interpreter.py b/mesonbuild/ast/interpreter.py
index 9e098d0..501be7d 100644
--- a/mesonbuild/ast/interpreter.py
+++ b/mesonbuild/ast/interpreter.py
@@ -239,7 +239,7 @@ class AstInterpreter(InterpreterBase):
 
     def evaluate_dictstatement(self, node: mparser.DictNode) -> TYPE_nkwargs:
         def resolve_key(node: mparser.BaseNode) -> str:
-            if isinstance(node, mparser.StringNode):
+            if isinstance(node, mparser.BaseStringNode):
                 return node.value
             return '__AST_UNKNOWN__'
         arguments, kwargs = self.reduce_arguments(node.args, key_resolver=resolve_key)
diff --git a/mesonbuild/ast/introspection.py b/mesonbuild/ast/introspection.py
index e8055c5..50aa744 100644
--- a/mesonbuild/ast/introspection.py
+++ b/mesonbuild/ast/introspection.py
@@ -27,7 +27,7 @@ from ..build import Executable, Jar, SharedLibrary, SharedModule, StaticLibrary
 from ..compilers import detect_compiler_for
 from ..interpreterbase import InvalidArguments
 from ..mesonlib import MachineChoice, OptionKey
-from ..mparser import BaseNode, ArithmeticNode, ArrayNode, ElementaryNode, IdNode, FunctionNode, StringNode
+from ..mparser import BaseNode, ArithmeticNode, ArrayNode, ElementaryNode, IdNode, FunctionNode, BaseStringNode
 from .interpreter import AstInterpreter
 
 if T.TYPE_CHECKING:
@@ -128,7 +128,7 @@ class IntrospectionInterpreter(AstInterpreter):
 
         if not self.is_subproject() and 'subproject_dir' in kwargs:
             spdirname = kwargs['subproject_dir']
-            if isinstance(spdirname, StringNode):
+            if isinstance(spdirname, BaseStringNode):
                 assert isinstance(spdirname.value, str)
                 self.subproject_dir = spdirname.value
         if not self.is_subproject():
@@ -174,7 +174,7 @@ class IntrospectionInterpreter(AstInterpreter):
         for l in self.flatten_args(raw_langs):
             if isinstance(l, str):
                 langs.append(l)
-            elif isinstance(l, StringNode):
+            elif isinstance(l, BaseStringNode):
                 langs.append(l.value)
 
         for lang in sorted(langs, key=compilers.sort_clink):
@@ -263,7 +263,7 @@ class IntrospectionInterpreter(AstInterpreter):
                 # Pop the first element if the function is a build target function
                 if isinstance(curr, FunctionNode) and curr.func_name in BUILD_TARGET_FUNCTIONS:
                     arg_nodes.pop(0)
-                elementary_nodes = [x for x in arg_nodes if isinstance(x, (str, StringNode))]
+                elementary_nodes = [x for x in arg_nodes if isinstance(x, (str, BaseStringNode))]
                 inqueue += [x for x in arg_nodes if isinstance(x, (FunctionNode, ArrayNode, IdNode, ArithmeticNode))]
                 if elementary_nodes:
                     res += [curr]
diff --git a/mesonbuild/ast/printer.py b/mesonbuild/ast/printer.py
index ebf63af..4626e5e 100644
--- a/mesonbuild/ast/printer.py
+++ b/mesonbuild/ast/printer.py
@@ -84,7 +84,17 @@ class AstPrinter(AstVisitor):
 
     def visit_FormatStringNode(self, node: mparser.FormatStringNode) -> None:
         assert isinstance(node.value, str)
-        self.append("f'" + node.value + "'", node)
+        self.append("f'" + self.escape(node.value) + "'", node)
+        node.lineno = self.curr_line or node.lineno
+
+    def visit_MultilineStringNode(self, node: mparser.StringNode) -> None:
+        assert isinstance(node.value, str)
+        self.append("'''" + node.value + "'''", node)
+        node.lineno = self.curr_line or node.lineno
+
+    def visit_FormatMultilineStringNode(self, node: mparser.FormatStringNode) -> None:
+        assert isinstance(node.value, str)
+        self.append("f'''" + node.value + "'''", node)
         node.lineno = self.curr_line or node.lineno
 
     def visit_ContinueNode(self, node: mparser.ContinueNode) -> None:
diff --git a/mesonbuild/ast/visitor.py b/mesonbuild/ast/visitor.py
index 51a6620..69b93c7 100644
--- a/mesonbuild/ast/visitor.py
+++ b/mesonbuild/ast/visitor.py
@@ -43,6 +43,12 @@ class AstVisitor:
     def visit_FormatStringNode(self, node: mparser.FormatStringNode) -> None:
         self.visit_default_func(node)
 
+    def visit_MultilineStringNode(self, node: mparser.StringNode) -> None:
+        self.visit_default_func(node)
+
+    def visit_FormatMultilineStringNode(self, node: mparser.FormatStringNode) -> None:
+        self.visit_default_func(node)
+
     def visit_ContinueNode(self, node: mparser.ContinueNode) -> None:
         self.visit_default_func(node)
 
diff --git a/mesonbuild/cmake/interpreter.py b/mesonbuild/cmake/interpreter.py
index f577ba7..80dfff5 100644
--- a/mesonbuild/cmake/interpreter.py
+++ b/mesonbuild/cmake/interpreter.py
@@ -960,7 +960,7 @@ class CMakeInterpreter:
             return Token(tid, self.subdir.as_posix(), 0, 0, 0, None, val)
 
         def string(value: str) -> StringNode:
-            return StringNode(token(val=value))
+            return StringNode(token(val=value), escape=False)
 
         def id_node(value: str) -> IdNode:
             return IdNode(token(val=value))
diff --git a/mesonbuild/coredata.py b/mesonbuild/coredata.py
index f151c7b..f91c583 100644
--- a/mesonbuild/coredata.py
+++ b/mesonbuild/coredata.py
@@ -1098,7 +1098,7 @@ class MachineFileParser():
         return section
 
     def _evaluate_statement(self, node: mparser.BaseNode) -> T.Union[str, bool, int, T.List[str]]:
-        if isinstance(node, (mparser.StringNode)):
+        if isinstance(node, (mparser.BaseStringNode)):
             return node.value
         elif isinstance(node, mparser.BooleanNode):
             return node.value
diff --git a/mesonbuild/interpreter/interpreter.py b/mesonbuild/interpreter/interpreter.py
index 9b005cb..0814739 100644
--- a/mesonbuild/interpreter/interpreter.py
+++ b/mesonbuild/interpreter/interpreter.py
@@ -536,7 +536,7 @@ class Interpreter(InterpreterBase, HoldableObject):
             assert isinstance(kw, mparser.IdNode), 'for mypy'
             if kw.value == 'meson_version':
                 # mypy does not understand "and isinstance"
-                if isinstance(val, mparser.StringNode):
+                if isinstance(val, mparser.BaseStringNode):
                     self.handle_meson_version(val.value, val)
 
     def get_build_def_files(self) -> mesonlib.OrderedSet[str]:
diff --git a/mesonbuild/interpreterbase/helpers.py b/mesonbuild/interpreterbase/helpers.py
index f2ee1b1..917969b 100644
--- a/mesonbuild/interpreterbase/helpers.py
+++ b/mesonbuild/interpreterbase/helpers.py
@@ -25,7 +25,7 @@ if T.TYPE_CHECKING:
     from .baseobjects import TYPE_var, TYPE_kwargs, SubProject
 
 def flatten(args: T.Union['TYPE_var', T.List['TYPE_var']]) -> T.List['TYPE_var']:
-    if isinstance(args, mparser.StringNode):
+    if isinstance(args, mparser.BaseStringNode):
         assert isinstance(args.value, str)
         return [args.value]
     if not isinstance(args, collections.abc.Sequence):
@@ -35,7 +35,7 @@ def flatten(args: T.Union['TYPE_var', T.List['TYPE_var']]) -> T.List['TYPE_var']
         if isinstance(a, list):
             rest = flatten(a)
             result = result + rest
-        elif isinstance(a, mparser.StringNode):
+        elif isinstance(a, mparser.BaseStringNode):
             result.append(a.value)
         else:
             result.append(a)
diff --git a/mesonbuild/interpreterbase/interpreterbase.py b/mesonbuild/interpreterbase/interpreterbase.py
index f35da33..eab96ef 100644
--- a/mesonbuild/interpreterbase/interpreterbase.py
+++ b/mesonbuild/interpreterbase/interpreterbase.py
@@ -196,8 +196,13 @@ class InterpreterBase:
             self.assignment(cur)
         elif isinstance(cur, mparser.MethodNode):
             return self.method_call(cur)
-        elif isinstance(cur, mparser.StringNode):
-            return self._holderify(cur.value)
+        elif isinstance(cur, mparser.BaseStringNode):
+            if isinstance(cur, mparser.MultilineFormatStringNode):
+                return self.evaluate_multiline_fstring(cur)
+            elif isinstance(cur, mparser.FormatStringNode):
+                return self.evaluate_fstring(cur)
+            else:
+                return self._holderify(cur.value)
         elif isinstance(cur, mparser.BooleanNode):
             return self._holderify(cur.value)
         elif isinstance(cur, mparser.IfClauseNode):
@@ -230,11 +235,6 @@ class InterpreterBase:
             return self.evaluate_indexing(cur)
         elif isinstance(cur, mparser.TernaryNode):
             return self.evaluate_ternary(cur)
-        elif isinstance(cur, mparser.FormatStringNode):
-            if isinstance(cur, mparser.MultilineFormatStringNode):
-                return self.evaluate_multiline_fstring(cur)
-            else:
-                return self.evaluate_fstring(cur)
         elif isinstance(cur, mparser.ContinueNode):
             raise ContinueRequest()
         elif isinstance(cur, mparser.BreakNode):
@@ -256,7 +256,7 @@ class InterpreterBase:
     @FeatureNew('dict', '0.47.0')
     def evaluate_dictstatement(self, cur: mparser.DictNode) -> InterpreterObject:
         def resolve_key(key: mparser.BaseNode) -> str:
-            if not isinstance(key, mparser.StringNode):
+            if not isinstance(key, mparser.BaseStringNode):
                 FeatureNew.single_use('Dictionary entry using non literal key', '0.53.0', self.subproject)
             key_holder = self.evaluate_statement(key)
             if key_holder is None:
@@ -428,9 +428,7 @@ class InterpreterBase:
         return self.evaluate_fstring(node)
 
     @FeatureNew('format strings', '0.58.0')
-    def evaluate_fstring(self, node: mparser.FormatStringNode) -> InterpreterObject:
-        assert isinstance(node, mparser.FormatStringNode)
-
+    def evaluate_fstring(self, node: T.Union[mparser.FormatStringNode, mparser.MultilineFormatStringNode]) -> InterpreterObject:
         def replace(match: T.Match[str]) -> str:
             var = str(match.group(1))
             try:
diff --git a/mesonbuild/mintro.py b/mesonbuild/mintro.py
index 0391535..cf4a53a 100644
--- a/mesonbuild/mintro.py
+++ b/mesonbuild/mintro.py
@@ -36,7 +36,7 @@ from .dependencies import Dependency
 from . import environment
 from .interpreterbase import ObjectHolder
 from .mesonlib import OptionKey
-from .mparser import FunctionNode, ArrayNode, ArgumentNode, StringNode
+from .mparser import FunctionNode, ArrayNode, ArgumentNode, BaseStringNode
 
 if T.TYPE_CHECKING:
     import argparse
@@ -194,7 +194,7 @@ def list_targets_from_source(intr: IntrospectionInterpreter) -> T.List[T.Dict[st
             elif isinstance(n, ArgumentNode):
                 args = n.arguments
             for j in args:
-                if isinstance(j, StringNode):
+                if isinstance(j, BaseStringNode):
                     assert isinstance(j.value, str)
                     res += [Path(j.value)]
                 elif isinstance(j, str):
diff --git a/mesonbuild/mparser.py b/mesonbuild/mparser.py
index f9b8967..6144794 100644
--- a/mesonbuild/mparser.py
+++ b/mesonbuild/mparser.py
@@ -193,23 +193,15 @@ class Lexer:
                     elif tid == 'dblquote':
                         raise ParseException('Double quotes are not supported. Use single quotes.', self.getline(line_start), lineno, col)
                     elif tid in {'string', 'fstring'}:
-                        # Handle here and not on the regexp to give a better error message.
                         if match_text.find("\n") != -1:
                             msg = ("Newline character in a string detected, use ''' (three single quotes) "
                                    "for multiline strings instead.\n"
                                    "This will become a hard error in a future Meson release.")
                             mlog.warning(mlog.code_line(msg, self.getline(line_start), col), location=BaseNode(lineno, col, filename))
                         value = match_text[2 if tid == 'fstring' else 1:-1]
-                        value = ESCAPE_SEQUENCE_SINGLE_RE.sub(decode_match, value)
                     elif tid in {'multiline_string', 'multiline_fstring'}:
-                        # For multiline strings, parse out the value and pass
-                        # through the normal string logic.
-                        # For multiline format strings, we have to emit a
-                        # different AST node so we can add a feature check,
-                        # but otherwise, it follows the normal fstring logic.
                         if tid == 'multiline_string':
                             value = match_text[3:-3]
-                            tid = 'string'
                         else:
                             value = match_text[4:-3]
                         lines = match_text.split('\n')
@@ -295,13 +287,30 @@ class NumberNode(ElementaryNode[int]):
         self.value = int(token.value, base=0)
         self.bytespan = token.bytespan
 
-class StringNode(ElementaryNode[str]):
+class BaseStringNode(ElementaryNode[str]):
     pass
 
-class FormatStringNode(ElementaryNode[str]):
+@dataclass(unsafe_hash=True)
+class StringNode(BaseStringNode):
+
+    raw_value: str = field(hash=False)
+
+    def __init__(self, token: Token[str], escape: bool = True):
+        super().__init__(token)
+        self.value = ESCAPE_SEQUENCE_SINGLE_RE.sub(decode_match, token.value) if escape else token.value
+        self.raw_value = token.value
+
+class FormatStringNode(StringNode):
     pass
 
-class MultilineFormatStringNode(FormatStringNode):
+@dataclass(unsafe_hash=True)
+class MultilineStringNode(BaseStringNode):
+
+    def __init__(self, token: Token[str]):
+        super().__init__(token)
+        self.value = token.value
+
+class MultilineFormatStringNode(MultilineStringNode):
     pass
 
 class ContinueNode(ElementaryNode):
@@ -819,6 +828,8 @@ class Parser:
             return StringNode(t)
         if self.accept('fstring'):
             return FormatStringNode(t)
+        if self.accept('multiline_string'):
+            return MultilineStringNode(t)
         if self.accept('multiline_fstring'):
             return MultilineFormatStringNode(t)
         return EmptyNode(self.current.lineno, self.current.colno, self.current.filename)
diff --git a/mesonbuild/optinterpreter.py b/mesonbuild/optinterpreter.py
index 37ef55a..9eeb3fd 100644
--- a/mesonbuild/optinterpreter.py
+++ b/mesonbuild/optinterpreter.py
@@ -115,7 +115,7 @@ class OptionInterpreter:
             return arg
         if isinstance(arg, mparser.ParenthesizedNode):
             return self.reduce_single(arg.inner)
-        elif isinstance(arg, (mparser.StringNode, mparser.BooleanNode,
+        elif isinstance(arg, (mparser.BaseStringNode, mparser.BooleanNode,
                               mparser.NumberNode)):
             return arg.value
         elif isinstance(arg, mparser.ArrayNode):
@@ -123,7 +123,7 @@ class OptionInterpreter:
         elif isinstance(arg, mparser.DictNode):
             d = {}
             for k, v in arg.args.kwargs.items():
-                if not isinstance(k, mparser.StringNode):
+                if not isinstance(k, mparser.BaseStringNode):
                     raise OptionException('Dictionary keys must be a string literal')
                 d[k.value] = self.reduce_single(v)
             return d
diff --git a/mesonbuild/rewriter.py b/mesonbuild/rewriter.py
index a9b2e88..0d7b111 100644
--- a/mesonbuild/rewriter.py
+++ b/mesonbuild/rewriter.py
@@ -28,7 +28,7 @@ from .ast import IntrospectionInterpreter, BUILD_TARGET_FUNCTIONS, AstConditionL
 from mesonbuild.mesonlib import MesonException, setup_vsenv
 from . import mlog, environment
 from functools import wraps
-from .mparser import Token, ArrayNode, ArgumentNode, AssignmentNode, BooleanNode, ElementaryNode, IdNode, FunctionNode, StringNode
+from .mparser import Token, ArrayNode, ArgumentNode, AssignmentNode, BaseStringNode, BooleanNode, ElementaryNode, IdNode, FunctionNode, StringNode
 import json, os, re, sys
 import typing as T
 
@@ -267,12 +267,12 @@ class MTypeStrList(MTypeList):
         return StringNode(Token('', '', 0, 0, 0, None, str(value)))
 
     def _check_is_equal(self, node, value) -> bool:
-        if isinstance(node, StringNode):
+        if isinstance(node, BaseStringNode):
             return node.value == value
         return False
 
     def _check_regex_matches(self, node, regex: str) -> bool:
-        if isinstance(node, StringNode):
+        if isinstance(node, BaseStringNode):
             return re.match(regex, node.value) is not None
         return False
 
@@ -292,7 +292,7 @@ class MTypeIDList(MTypeList):
         return False
 
     def _check_regex_matches(self, node, regex: str) -> bool:
-        if isinstance(node, StringNode):
+        if isinstance(node, BaseStringNode):
             return re.match(regex, node.value) is not None
         return False
 
@@ -652,7 +652,7 @@ class Rewriter:
             src_list = []
             for i in target['sources']:
                 for j in arg_list_from_node(i):
-                    if isinstance(j, StringNode):
+                    if isinstance(j, BaseStringNode):
                         src_list += [j.value]
 
             # Generate the new String nodes
@@ -686,7 +686,7 @@ class Rewriter:
             def find_node(src):
                 for i in target['sources']:
                     for j in arg_list_from_node(i):
-                        if isinstance(j, StringNode):
+                        if isinstance(j, BaseStringNode):
                             if j.value == src:
                                 return i, j
                 return None, None
@@ -745,7 +745,7 @@ class Rewriter:
             extra_files_list = []
             for i in target['extra_files']:
                 for j in arg_list_from_node(i):
-                    if isinstance(j, StringNode):
+                    if isinstance(j, BaseStringNode):
                         extra_files_list += [j.value]
 
             # Generate the new String nodes
@@ -776,7 +776,7 @@ class Rewriter:
             def find_node(src):
                 for i in target['extra_files']:
                     for j in arg_list_from_node(i):
-                        if isinstance(j, StringNode):
+                        if isinstance(j, BaseStringNode):
                             if j.value == src:
                                 return i, j
                 return None, None
@@ -845,12 +845,12 @@ class Rewriter:
             src_list = []
             for i in target['sources']:
                 for j in arg_list_from_node(i):
-                    if isinstance(j, StringNode):
+                    if isinstance(j, BaseStringNode):
                         src_list += [j.value]
             extra_files_list = []
             for i in target['extra_files']:
                 for j in arg_list_from_node(i):
-                    if isinstance(j, StringNode):
+                    if isinstance(j, BaseStringNode):
                         extra_files_list += [j.value]
             test_data = {
                 'name': target['name'],
@@ -865,8 +865,8 @@ class Rewriter:
             alphanum_key = lambda key: [convert(c) for c in re.split('([0-9]+)', key)]
             path_sorter = lambda key: ([(key.count('/') <= idx, alphanum_key(x)) for idx, x in enumerate(key.split('/'))])
 
-            unknown = [x for x in i.arguments if not isinstance(x, StringNode)]
-            sources = [x for x in i.arguments if isinstance(x, StringNode)]
+            unknown = [x for x in i.arguments if not isinstance(x, BaseStringNode)]
+            sources = [x for x in i.arguments if isinstance(x, BaseStringNode)]
             sources = sorted(sources, key=lambda x: path_sorter(x.value))
             i.arguments = unknown + sources