Add style checker

Ref. #799

ide/gnatstudio/recordflux.py

@@ -221,18 +221,14 @@ def __on_gps_started():
     r"(?P<filename>[^:]+):"
     r"(?P<line>\d+):"
     r"(?P<column>\d+): "
-    r"(?P<subsystem>core|parser|model|cli|internal|graph): "
+    r"(?P<subsystem>\w+): "
     r"(?P<severity>info|warning|error): "
     r"(?P<message>.*)"
     r"$"
 )
 
 generic_message_re = re.compile(
-    r"^"
-    r"(?P<subsystem>core|parser|model|cli|internal|graph): "
-    r"(?P<severity>info|warning|error): "
-    r"(?P<message>.*)"
-    r"$"
+    r"^(?P<subsystem>\w+): (?P<severity>info|warning|error): (?P<message>.*)$"
 )
 
 

rflx/error.py

@@ -62,6 +62,7 @@ class Subsystem(Enum):
     PYRFLX = auto()
     ID = auto()
     VALIDATOR = auto()
+    STYLE = auto()
 
     def __str__(self) -> str:
         return str.lower(self.name)

rflx/expression.py

@@ -928,7 +928,7 @@ def simplified(self) -> Expr:
 
     @property
     def symbol(self) -> str:
-        return "**"
+        return " ** "
 
     def ada_expr(self) -> ada.Expr:
         return ada.Pow(self.left.ada_expr(), self.right.ada_expr())

rflx/specification/parser.py

@@ -15,6 +15,7 @@
 from rflx.model import declaration as decl, statement as stmt
 from rflx.specification.const import RESERVED_WORDS
 
+from . import style
 from .cache import Cache
 
 log = logging.getLogger(__name__)
@@ -1541,6 +1542,7 @@ def parse(self, *specfiles: Path) -> None:
 
         for f in specfiles:
             error.extend(self.__parse_specfile(f))
+            error.extend(style.check(f))
         self.__sort_specs_topologically()
         error.propagate()
 

rflx/specification/style.py

@@ -0,0 +1,173 @@
+from __future__ import annotations
+
+import re
+from pathlib import Path
+
+from rflx.error import Location, RecordFluxError, Severity, Subsystem
+
+INCORRECT_LINE_TERMINATORS = "\r"
+ILLEGAL_WHITESPACE_CHARACTERS = "\t\x0b\x0c"
+KEYWORD_INDENTATION = [
+    ("begin", [3, 6]),
+    ("end message", [6]),
+    ("end", [0, 3, 6]),
+    ("for", [3, 12, 15]),
+    ("generic", [3]),
+    ("goto", [9]),
+    ("is", [3, 6]),
+    ("message", [6]),
+    ("package", [0]),
+    ("session", [3]),
+    ("state", [6]),
+    ("then", [12]),
+    ("transition", [6]),
+    ("type", [3, 6]),
+    ("with", [0, 3, 6, 9, 12, 15]),
+]
+
+
+def check(spec_file: Path) -> RecordFluxError:
+    error = RecordFluxError()
+
+    with open(spec_file, encoding="utf-8", newline="") as f:
+        specification = f.read()
+
+    if not specification:
+        return error
+
+    blank_lines = 0
+    lines = specification.split("\n")
+
+    for i, l in enumerate(lines, start=1):
+        strings = [range(m.start(), m.end()) for m in re.finditer(r'"[^"]*"', l.split("--")[0])]
+
+        blank_lines = _check_blank_lines(error, l, i, spec_file, blank_lines, len(lines))
+        _check_characters(error, l, i, spec_file, strings)
+        _check_indentation(error, l, i, spec_file)
+        _check_line_length(error, l, i, spec_file)
+        _check_token_spacing(error, l, i, spec_file, strings)
+        _check_trailing_spaces(error, l, i, spec_file)
+
+    return error
+
+
+def _append(error: RecordFluxError, message: str, row: int, col: int, spec_file: Path) -> None:
+    error.extend(
+        [
+            (
+                message,
+                Subsystem.STYLE,
+                Severity.ERROR,
+                Location((row, col), spec_file),
+            )
+        ]
+    )
+
+
+def _check_blank_lines(
+    error: RecordFluxError, line: str, row: int, spec_file: Path, blank_lines: int, row_count: int
+) -> int:
+    if line == "":
+        if row == 1:
+            _append(error, "leading blank line", row, 1, spec_file)
+        if blank_lines > 0 and row == row_count:
+            _append(error, "trailing blank line", row - 1, 1, spec_file)
+        blank_lines += 1
+    else:
+        if blank_lines > 1:
+            _append(error, "multiple blank lines", row - 1, 1, spec_file)
+        blank_lines = 0
+
+    return blank_lines
+
+
+def _check_characters(
+    error: RecordFluxError, line: str, row: int, spec_file: Path, strings: list[range]
+) -> None:
+    for j, c in enumerate(line, start=1):
+        if c == INCORRECT_LINE_TERMINATORS:
+            s = repr(c).replace("'", '"')
+            _append(error, f"incorrect line terminator {s}", row, j, spec_file)
+        if c in ILLEGAL_WHITESPACE_CHARACTERS:
+            s = repr(c).replace("'", '"')
+            _append(error, f"illegal whitespace character {s}", row, j, spec_file)
+
+    for j, c in enumerate(line.split("--")[0], start=1):
+        if any(j in r for r in strings):
+            continue
+
+
+def _check_indentation(error: RecordFluxError, line: str, row: int, spec_file: Path) -> None:
+    match = re.match(r" *", line)
+    assert match
+    indentation = match.end()
+
+    for keyword, indentations in KEYWORD_INDENTATION:
+        if re.match(rf" *{keyword}", line) and indentation not in indentations:
+            expected_indentation = (
+                ", ".join(str(row) for row in indentations[:-1]) + f" or {indentations[-1]}"
+                if len(indentations) > 1
+                else str(indentations[0])
+            )
+            _append(
+                error,
+                f"unexpected keyword indentation (expected {expected_indentation})",
+                row,
+                match.end(),
+                spec_file,
+            )
+
+
+def _check_token_spacing(
+    error: RecordFluxError, line: str, row: int, spec_file: Path, strings: list[range]
+) -> None:
+    for match in re.finditer(
+        r"--+|/?=(?!>)|<=?|>=?|=>|:=|\+|[(]+\-\d+[)]+|\-|/|\*\*?|::?|'|;|,",
+        line,
+    ):
+        if "--" in line and match.start() > line.find("--"):
+            continue
+
+        if any(match.start() in r for r in strings) or any(match.end() in r for r in strings):
+            # ISSUE: nedbat/coveragepy#772
+            # A dummy statement is needed to disable the peephole optimizer, so that the continue
+            # statement is detected during coverage analysis.
+            # CPython 3.8 and 3.9 are affected. The issue is fixed in CPython 3.10.
+            dummy = 0  # noqa: F841
+            continue
+
+        token = match.group(0)
+
+        if token in [";", ","]:
+            space_before = False
+            space_after = True
+        elif token in ["::", "'"]:
+            space_before = False
+            space_after = False
+        else:
+            space_before = True
+            space_after = True
+
+        if space_before:
+            assert token != ";"
+            if match.start() > 1 and line[match.start() - 1] not in " (":
+                _append(error, f'missing space before "{token}"', row, match.start() + 1, spec_file)
+        else:
+            if match.start() > 1 and line[match.start() - 1] == " ":
+                _append(error, f'space before "{token}"', row, match.start() + 1, spec_file)
+        if space_after:
+            if match.end() < len(line) and line[match.end()] not in " ;\n":
+                _append(error, f'missing space after "{token}"', row, match.end() + 1, spec_file)
+        else:
+            if match.end() < len(line) and line[match.end()] == " ":
+                _append(error, f'space after "{token}"', row, match.end() + 2, spec_file)
+
+
+def _check_trailing_spaces(error: RecordFluxError, line: str, row: int, spec_file: Path) -> None:
+    if line.endswith(" "):
+        _append(error, "trailing whitespace", row, len(line), spec_file)
+
+
+def _check_line_length(error: RecordFluxError, line: str, row: int, spec_file: Path) -> None:
+    if len(line) > 120:
+        _append(error, f"line too long ({len(line)}/120)", row, 121, spec_file)

tests/data/specs/always_valid_aspect.rflx

@@ -12,7 +12,7 @@ package Always_Valid_Aspect is
        V6 => 6)
    with Size => 8, Always_Valid;
 
-   type T is mod 2**8;
+   type T is mod 2 ** 8;
 
    type Message is
       message

tests/data/specs/checksum_message.rflx

@@ -1,8 +1,8 @@
 package Checksum_Message is
 
-   type A is mod 2**8;
-   type B is mod 2**8;
-   type C is mod 2**8;
+   type A is mod 2 ** 8;
+   type B is mod 2 ** 8;
+   type C is mod 2 ** 8;
 
    type Message is
       message

tests/data/specs/ethernet.rflx

@@ -1,9 +1,9 @@
 package Ethernet is
 
-   type Address is mod 2**48;
-   type Type_Length is range 46 .. 2**16 - 1 with Size => 16;
+   type Address is mod 2 ** 48;
+   type Type_Length is range 46 .. 2 ** 16 - 1 with Size => 16;
    type TPID is range 16#8100# .. 16#8100# with Size => 16;
-   type TCI is mod 2**16;
+   type TCI is mod 2 ** 16;
 
    type Frame is
       message

tests/data/specs/icmp.rflx

@@ -36,12 +36,12 @@ package ICMP is
    ) with Size => 8;
 
    type Code_Zero is range 0 .. 0 with Size => 8;
-   type Checksum is mod 2**16;
-   type Identifier is mod 2**16;
-   type Sequence_Number is mod 2**16;
-   type Pointer is mod 2**8;
-   type Timestamp is mod 2**32;
-   type Gateway_Internet_Address is mod 2**32;
+   type Checksum is mod 2 ** 16;
+   type Identifier is mod 2 ** 16;
+   type Sequence_Number is mod 2 ** 16;
+   type Pointer is mod 2 ** 8;
+   type Timestamp is mod 2 ** 32;
+   type Gateway_Internet_Address is mod 2 ** 32;
    type Unused_32 is range 0 .. 0 with Size => 32;
    type Unused_24 is range 0 .. 0 with Size => 24;
 

tests/data/specs/ipv4.rflx

@@ -2,19 +2,19 @@ package IPv4 is
 
    type Version is range 4 .. 4 with Size => 4;
    type IHL is range 5 .. 15 with Size => 4;
-   type DCSP is mod 2**6;
-   type ECN is mod 2**2;
-   type Total_Length is mod 2**16;
-   type Identification is mod 2**16;
-   type Fragment_Offset is mod 2**13;
-   type TTL is mod 2**8;
+   type DCSP is mod 2 ** 6;
+   type ECN is mod 2 ** 2;
+   type Total_Length is mod 2 ** 16;
+   type Identification is mod 2 ** 16;
+   type Fragment_Offset is mod 2 ** 13;
+   type TTL is mod 2 ** 8;
    type Protocol is (P_ICMP => 1, P_UDP => 17) with Size => 8, Always_Valid;
-   type Header_Checksum is mod 2**16;
-   type Address is mod 2**32;
+   type Header_Checksum is mod 2 ** 16;
+   type Address is mod 2 ** 32;
 
    type Option_Class is (Control => 0, Debugging_And_Measurement => 2) with Size => 2;
-   type Option_Number is mod 2**5;
-   type Option_Length is range 2 .. 2**8 - 1 with Size => 8;
+   type Option_Number is mod 2 ** 5;
+   type Option_Length is range 2 .. 2 ** 8 - 1 with Size => 8;
 
    type Option is
       message

tests/data/specs/message_in_message.rflx

@@ -1,6 +1,6 @@
 package Message_In_Message is
 
-   type Length is mod 2**16;
+   type Length is mod 2 ** 16;
 
    type Length_Value is
       message

tests/data/specs/message_size.rflx

@@ -1,6 +1,6 @@
 package Message_Size is
 
-   type T is mod 2**8;
+   type T is mod 2 ** 8;
 
    type Msg is
       message

tests/data/specs/message_type_size_condition.rflx

@@ -1,6 +1,6 @@
 package Message_Type_Size_Condition is
 
-   type T1 is mod 2**8;
+   type T1 is mod 2 ** 8;
    type Message is
       message
          F1 : T1

tests/data/specs/multiple_errors.rflx

@@ -1,12 +1,12 @@
 package Multiple_Errors is
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
-   type INVALID is mod 2**8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
+   type INVALID is mod 2 ** 8;
 end Multiple_Errors;

tests/data/specs/no_conditionals.rflx

@@ -1,6 +1,6 @@
 package No_Conditionals is
 
-   type Tag is mod 2**16;
+   type Tag is mod 2 ** 16;
 
    type Message is
       message

tests/data/specs/p1.rflx

@@ -1,6 +1,6 @@
 package P1 is
 
-   type Kind is mod 2**16;
+   type Kind is mod 2 ** 16;
 
    type Frame is
       message

tests/data/specs/p2.rflx

@@ -2,7 +2,7 @@ with P4;
 
 package P2 is
 
-   type Length is mod 2**16;
+   type Length is mod 2 ** 16;
 
    type Packet is
       message

tests/data/specs/p3.rflx

@@ -2,7 +2,7 @@ with P2;
 
 package P3 is
 
-   type T is mod 2**32;
+   type T is mod 2 ** 32;
 
    type Packet is
       message

tests/data/specs/p4.rflx

@@ -1,5 +1,5 @@
 package P4 is
 
-   type Protocol_Number is mod 2**16;
+   type Protocol_Number is mod 2 ** 16;
 
 end P4;