Merge pull request #16 from AntonLydike/anton/fix-capture-syntax

regex: Syntax fixes for a bunch of stuff (fixes #4, #12)
AntonLydike · Jul 12, 2024 · 716e954 · 716e954
2 parents acaaae0 + d3d6b07
commit 716e954
Show file tree

Hide file tree

Showing 4 changed files with 24 additions and 14 deletions.
diff --git a/filecheck/parser.py b/filecheck/parser.py
@@ -47,7 +47,7 @@ def pattern_for_opts(opts: Options) -> tuple[re.Pattern[str], re.Pattern[str]]:
 )
 
 
-LINE_SPLIT_RE = split = re.compile(r"(\{\{|\[\[|]]|}})")
+LINE_SPLIT_RE = split = re.compile(r"(\{\{|\[\[\$?[#a-zA-Z_]|]|})")
 
 
 @dataclass
@@ -150,17 +150,21 @@ def parse_args(self, arg: str, line: str) -> list[UOp]:
         offset = len(line) - len(arg)
         while parts:
             part = parts.pop(0)
-            if part == "[[":
+            if part.startswith("[["):
+                brackets = 2
                 # grab parts greedily until we hit a ]]
-                while not part.endswith("]]"):
+                while brackets > 0:
                     if not parts:
                         raise ParseError(
                             "Invalid substitution block, no ]]",
                             self.line_no,
                             offset,
                             line,
                         )
-                    part += parts.pop(0)
+                    addition = parts.pop(0)
+                    brackets += addition.count("[") - addition.count("\\[")
+                    brackets -= addition.count("]") + addition.count("\\]")
+                    part += addition
                 # check if we are a simple capture pattern [[<name>:<regex>]]
                 if match := VAR_CAPTURE_PATTERN.fullmatch(part):
                     re_expr = posix_to_python_regex(match.group(2))

diff --git a/filecheck/regex.py b/filecheck/regex.py
@@ -5,15 +5,14 @@
     r"\[:(alpha|upper|lower|digit|alnum|xdigit|space|blank|print|punct|graph|word|ascii|cntrl):]"
 )
 POSIX_REGEXP_REPLACEMENTS = {
-    "alpha": "[A-Za-z]",
-    "upper": "[A-Z]",
-    "lower": "[a-z]",
-    "digit": "[0-9]",
-    "alnum": "[A-Za-z0-9]",
-    "xdigit": "[A-Fa-f0-9]",
+    "alpha": "A-Za-z",
+    "upper": "A-Z",
+    "lower": "a-z",
+    "digit": "0-9",
+    "alnum": "A-Za-z0-9",
+    "xdigit": "A-Fa-f0-9",
     "space": r"\s",
-    "blank": r"[ \t]",
-    "word": r"\w+",
+    "blank": r" \t",
 }
 
 NEGATED_SET_WITHOUT_NEWLINES = re.compile(r"([^\\]|^)\[\^((?!\\n))")

diff --git a/tests/filecheck/regex.test b/tests/filecheck/regex.test
@@ -4,7 +4,7 @@ sample text with a number: 144
 // CHECK: sample text with a number: {{\d+}}
 
 sample text with another number: 12*12
-// CHECK: sample text with another number: {{([:digit:]{2})}}*{{([:digit:]{2})}}
+// CHECK: sample text with another number: {{([[:digit:]]{2})}}*{{([[:digit:]]{2})}}
 
 // make sure that negative capturing groups don't capture newlines
 test 123

diff --git a/tests/filecheck/variables.test b/tests/filecheck/variables.test
@@ -24,7 +24,7 @@ print z
 
 // check for posix style regex
 same FF77 FF77
-// CHECK: [[VAR:[:xdigit:]+]] [[VAR]]
+// CHECK: [[VAR:[[:xdigit:]]+]] [[VAR]]
 
 // numeric captures:
 print 0xFF00FF00
@@ -42,3 +42,10 @@ commutativity states that (x * y) = (y * x)
 // CHECK: ([[a:(\w+)(x?)]] * [[b:(\w+)(x?)]]) = ([[b]] * [[a]])
 again: (x * y) = (y * x)
 // CHECK: ([[a]] * [[b]]) = ([[b]] * [[a]])
+
+
+// check that we parse ambigous syntax correctly:
+test %arg1
+// CHECK:       test [[ARG:%[[:alnum:]]+]]
+test [%arg1][0]
+// CHECK-NEXT:  test [[[ARG]]][0]