v1.3.1: Some frontend bug fixes

CHANGELOG.md

@@ -5,9 +5,12 @@ This file contains the notable changes in the ISLa project since version 0.2a1
 
 ## [unreleased]
 
+## [1.3.1] - 2022-09-30
+
 ### Changed
 
 - Fixed bug in `unparse_grammar`: Backslashes are correctly escaped now.
+- Fixed bug in parsing of match expressions: Escaped symbols are not handled correctly.
 - Made order of terminals in REST_GRAMMAR deterministic.
 
 ## [1.3.0] - 2022-09-27

pyproject.toml

@@ -8,7 +8,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "isla-solver"
-version = "1.3.0"
+version = "1.3.1"
 authors = [
     { name = "Dominic Steinhöfel", email = "dominic.steinhoefel@cispa.de" },
 ]

src/isla/__init__.py

@@ -1 +1 @@
-__version__ = "1.3.0"
+__version__ = "1.3.1"

src/isla/helpers.py

@@ -798,3 +798,22 @@ def result(inp: T) -> T:
         return inp
 
     return result
+
+
+def instantiate_escaped_symbols(text: str) -> str:
+    backslash_escape_placeholder = "$$BESC$$"
+    assert backslash_escape_placeholder not in text
+
+    repl_map = {
+        "\\b": "\b",
+        "\\t": "\t",
+        "\\n": "\n",
+        "\\r": "\r",
+        '\\"': '"',
+    }
+
+    text = text.replace("\\\\", backslash_escape_placeholder)
+    for escaped_char in repl_map:
+        text = text.replace(escaped_char, repl_map[escaped_char])
+
+    return text.replace(backslash_escape_placeholder, "\\")

src/isla/language.py

@@ -53,6 +53,7 @@
     chain_functions,
     eassert,
     Exceptional,
+    instantiate_escaped_symbols,
 )
 from isla.helpers import (
     replace_line_breaks,
@@ -2681,10 +2682,8 @@ def exitMatchExprOptional(self, ctx: MexprParser.MatchExprOptionalContext):
 
     def exitMatchExprChars(self, ctx: MexprParser.MatchExprCharsContext):
         text = antlr_get_text_with_whitespace(ctx)
-        text = text.replace("{{", "{")
-        text = text.replace("}}", "}")
-        text = text.replace('\\"', '"')
-        self.result.append(text)
+        text = text.replace("{{", "{").replace("}}", "}")
+        self.result.append(instantiate_escaped_symbols(text))
 
     def exitMatchExprVar(self, ctx: MexprParser.MatchExprVarContext):
         self.result.append(
@@ -3809,22 +3808,7 @@ def exitAlternative(self, ctx: bnfParser.AlternativeContext):
                 assert child_text[-1] == '"'
                 child_text = child_text[1:-1]
 
-            # Instantiate escaped characters
-            backslash_escape_placeholder = "$$BESC$$"
-            assert backslash_escape_placeholder not in child_text
-            child_text = child_text.replace("\\\\", backslash_escape_placeholder)
-            repl_map = {
-                "\\b": "\b",
-                "\\t": "\t",
-                "\\n": "\n",
-                "\\r": "\r",
-                '\\"': '"',
-            }
-            for escaped_char in repl_map:
-                child_text = child_text.replace(escaped_char, repl_map[escaped_char])
-            child_text = child_text.replace(backslash_escape_placeholder, "\\")
-
-            elems.append(child_text)
+            elems.append(instantiate_escaped_symbols(child_text))
         self.partial_results[ctx] = "".join(elems)
 
 

tests/test_concrete_syntax.py

@@ -24,6 +24,7 @@
     unparse_isla,
     parse_bnf,
     unparse_grammar,
+    ISLaEmitter,
 )
 from isla.z3_helpers import z3_eq
 from isla_formalizations import scriptsizec
@@ -601,8 +602,8 @@ def test_unparse_grammar(self):
     def test_parse_bnf_with_escape_chars(self):
         grammar_str = rf'''
 <start> ::= <A>
-<A> ::= "\r" | "\n" | "\"" | "\\t" | "\\\\\\"'''
-        expected = {"<start>": ["<A>"], "<A>": ["\r", "\n", '"', "\\t", "\\\\\\"]}
+<A> ::= "\\a\\" | "\r" | "\n" | "\"" | "\\t" | "\\\\\\"'''
+        expected = {"<start>": ["<A>"], "<A>": ["\\a\\", "\r", "\n", '"', "\\t", "\\\\\\"]}
 
         self.assertEqual(expected, parse_bnf(grammar_str))
 
@@ -773,8 +774,27 @@ def test_unbound_variable(self):
     def test_parse_rest_bnf(self):
         unparsed = unparse_grammar(REST_GRAMMAR)
         parsed = parse_bnf(unparsed)
+        print(parsed)
         gg.GrammarGraph.from_grammar(parsed)
 
+    def test_parse_match_expression_with_escape_char(self):
+        isla_emitter = ISLaEmitter(None)
+
+        DummyVariable.cnt = 0
+        parsed = isla_emitter.parse_mexpr(
+            r"{<a> a}a\\n\n\r{<b> b}<c>", VariableManager()
+        )
+
+        DummyVariable.cnt = 0
+        expected = language.BindExpression(
+            language.BoundVariable("a", "<a>"),
+            DummyVariable("a\\n\n\r"),
+            language.BoundVariable("b", "<b>"),
+            language.DummyVariable("<c>"),
+        )
+
+        self.assertEqual(expected, parsed)
+
 
 if __name__ == "__main__":
     unittest.main()