- parser.py, syntax.py: bug concerning collection of errors fixed

DHParser/error.py

@@ -66,8 +66,8 @@ class Error:
         return prefix + "%s: %s" % (self.level_str, self.message)
 
     def __repr__(self):
-        return 'Error("%s", %i, %s, %i, %i, %i)' \
-               % (self.message, self.level, repr(self.code), self.pos, self.line, self.column)
+        return 'Error("%s", %s, %i, %i, %i)' \
+               % (self.message, repr(self.code), self.pos, self.line, self.column)
 
     @property
     def level_str(self):

DHParser/parser.py

@@ -2039,7 +2039,7 @@ class Compiler:
         else:
             self._dirty_flag = True
         result = self.compile(node)
-        self.propagate_error_flags(node)
+        self.propagate_error_flags(node, lazy=True)
         return result
 
     def set_grammar_name(self, grammar_name="", grammar_source=""):
@@ -2058,12 +2058,13 @@ class Compiler:
         self.grammar_source = load_if_file(grammar_source)
 
     @staticmethod
-    def propagate_error_flags(node: Node) -> None:
-        if node.error_flag < Error.HIGHEST:
+    def propagate_error_flags(node: Node, lazy: bool = True) -> None:
+        # See test_parser.TestCompilerClass.test_propagate_error()..
+        if not lazy or node.error_flag < Error.HIGHEST:
             for child in node.children:
                 Compiler.propagate_error_flags(child)
                 node.error_flag = max(node.error_flag, child.error_flag)
-                if node.error_flag >= Error.HIGHEST:
+                if lazy and node.error_flag >= Error.HIGHEST:
                     return
 
     @staticmethod
@@ -2164,6 +2165,7 @@ def compile_source(source: str,
             syntax_tree.log(log_file_name + '.ast')
         if not is_error(syntax_tree.error_flag):
             result = compiler(syntax_tree)
+        # print(syntax_tree.as_sxpr())
         messages.extend(syntax_tree.collect_errors(source_text))
         syntax_tree.error_flag = max(syntax_tree.error_flag, efl)
     return result, messages, syntax_tree

DHParser/syntaxtree.py

@@ -398,21 +398,25 @@ class Node(collections.abc.Sized):
             return []
 
     def _collect_errors(self, lbreaks: List[int] = [], clear_errors=False) -> List[Error]:
-        if self.error_flag:
-            errors = self.errors
-            if lbreaks:
-                for err in errors:
-                    err.pos = self.pos
-                    err.line, err.column = line_col(lbreaks, err.pos)
-            if clear_errors:
-                self._errors = []
-                self.error_flag = 0
-            if self.children:
-                for child in self.children:
-                    errors.extend(child._collect_errors(lbreaks, clear_errors))
-            return errors
+        errors = self.errors
+        if errors and lbreaks:
+            for err in errors:
+                err.pos = self.pos
+                err.line, err.column = line_col(lbreaks, err.pos)
+        if self.children:
+            for child in self.children:
+                errors.extend(child._collect_errors(lbreaks, clear_errors))
+        if clear_errors:
+            self._errors = []
+            self.error_flag = 0
         else:
-            return []
+            if self._errors:
+                self.error_flag = max(err.code for err in self.errors)
+            if self.children:
+                max_child_error = max(child.error_flag for child in self.children)
+                self.error_flag = max(self.error_flag, max_child_error)
+        return errors
+
 
 
     def _tree_repr(self, tab, open_fn, close_fn, data_fn=identity, density=0) -> str:
@@ -482,6 +486,8 @@ class Node(collections.abc.Sized):
             # s += " '(pos %i)" % node.pos
             if src:
                 txt += " '(pos %i " % node.pos  # + " %i %i)" % line_col(src, node.pos)
+            if node.error_flag:
+                txt += " HAS ERRORS"
             if node.errors:
                 txt += " '(err '(%s))" % ' '.join(str(err).replace('"', r'\"')
                                                   for err in node.errors)

build_cython-modules.bat

+python.exe setup.py build_ext --inplace

buildpackages.bat

+python3.exe setup.py sdist bdist
\ No newline at end of file

examples/MLW/Beispiele/imperium/imperium.mlw

@@ -12,12 +12,15 @@ SCHREIBWEISE
 		ym-: Chart. Sangall. A 194.
 		impir-: v. ibi.
 		
-        form. sing.:
-            gen.:
-                -ri : v. ibi. adde Annal. Plac. a. 1266 p. 516,21.
-                -iae : Chron. Fred. 2,33. p. 56,22. 2,35.
-            abl.:
-                -um : Chron. Fred. 2,15. 2,35sqq. capit. p. 43. confunditur c. imperitus : v. ibi.
+    form. sing.:
+        gen.:
+            -ri: v. ibi. adde Annal. Plac. a. 1266 p. 516,21.
+            -iae: Chron. Fred. 2,33. p. 56,22. 2,35.
+        abl.:
+            -um: Chron. Fred. 2,15. 2,35sqq. capit. p. 43.
+
+    confunditur c.:
+        imperitus: v. ibi.
 
 
 BEDEUTUNG

examples/MLW/MLW.ebnf

@@ -17,7 +17,7 @@ Artikel           = [LZ]
                     [ArtikelKopf]
                     BedeutungsPosition
                     [VerweisPosition]
-                    { SubArtikel }
+                    { UnterArtikel }
                     ArtikelVerfasser
                     [LZ]  DATEI_ENDE
 
@@ -41,9 +41,9 @@ LemmaVariante     = LAT_WORT_TEIL { "-" LAT_WORT_TEIL }
 
 ## GRAMMATIK-POSITION ##
 
-GrammatikPosition = "GRAMMATIK" [LZ] Grammatik §ABS { GrammatikVariante §ABS }
+GrammatikPosition = "GRAMMATIK" [LZ] §Grammatik ABS { GrammatikVariante §ABS }
 
-Grammatik        = §wortart ABS flexion [genus]
+Grammatik        = wortart §ABS flexion [genus]
 
 wortart          = "nomen"  | "n."
                  | "verb"   | "v."
@@ -51,14 +51,16 @@ wortart          = "nomen"  | "n."
                  | "adjektiv" | "adj."
                  | "praeposition" | "praep."
 
-flexion          = FLEX { "," §FLEX }
+flexion          = deklination | konjugation
+deklination      = FLEX §"," FLEX
+konjugation      = FLEX
 FLEX             = /-?[a-z]+/~
 
 genus            = "maskulinum" | "m."
                  | "femininum" | "f."
                  | "neutrum" | "n."
 
-GrammatikVariante  = [wortart ABS] flexion [genus] DPP  { Beleg }+
+GrammatikVariante  = [wortart ABS] flexion [genus] DPP { Beleg }+
 
 
 
@@ -73,14 +75,32 @@ Etymologie         = FREITEXT
 
 #### ARTIKEL-KOPF ############################################################
 
-ArtikelKopf     = { SchreibweisenPosition | StrukturPosition
-                  | GebrauchPosition | MetrikPosition | VerwechselungPosition }+
+ArtikelKopf     = <   SchreibweisenPosition | StrukturPosition | GebrauchPosition
+                    | MetrikPosition | VerwechselungPosition >
 
-SchreibweisenPosition =  "SCHREIBWEISE" [LZ] §SWTyp DPP [LZ]
-                         SWVariante { ABS SWVariante} [LZ]
-SWTyp           = "script. fat-" | "script."
-SWVariante      = Schreibweise DPP Beleg
-Schreibweise    = ZEICHENFOLGE
+## Schreibweisen-Position ##
+
+# TODO: Ggf. noch zu ergänzen um: Zusatz, Mehrere Tyen innerhalb der Schreibweisen-Position.
+
+SchreibweisenPosition =  "SCHREIBWEISE" [LZ] { SWKategorie }+
+
+SWKategorie      = SWTyp DPP [LZ] §{SWUnterkategorie | SWVariante { ABS SWVariante }}+ [LZ]
+SWUnterkategorie = SWUnterTyp DPP [LZ] §SWVariante { ABS SWVariante }+ [LZ]
+SWTyp            = scriptfat | scriptform | script | form | OFFEN
+SWVariante       = !KATEGORIENZEILE Schreibweise DPP Beleg
+Schreibweise     = ZEICHENFOLGE
+
+scriptfat  = "script." "fat-"
+scriptform = "script. " "form"
+script     = "srcipt."
+form       = "form"
+
+
+#### STRUKTUR-POSITION #######################################################
+
+StrukturPosition = "STRUKTUR" [LZ] { STVariante }+
+
+STVariante = "TODO"
 
 
 
@@ -96,7 +116,16 @@ DeutscheBedeutung = SW_DEU /(?:(?![A-ZÄÖÜ][A-ZÄÖÜ]).)+/~
 
 Belege          = "BELEGE" [LZ] (EinBeleg | { "*" EinBeleg }) ABS
 EinBeleg        = { !([LZ] "*" | SCHLUESSELWORT) /\s*[^\n]*/~ [ZW] }+ [Zusatz]
-Zusatz          = "ZUSATZ" /\s*.*/
+
+
+#### VERWEIS-POSITION #####################################################
+
+VerweisPosition = "VERWEISE"
+
+
+#### UNTER-ARTIKEL ########################################################
+
+UnterArtikel = "UNTER-ARTIKEL"
 
 
 #### AUTOR/AUTORIN ###########################################################
@@ -171,6 +200,7 @@ RZS              = /\s*?\n|$/               # Rückwärtiger Zeilensprung oder T
 
 ZEILENSPRUNG     = /[ \t]*\n/~
 KOMMENTARZEILEN  = { /[ \t]*\n?[ \t]*/ COMMENT__ }  # echte Kommentarzeilen
+KATEGORIENZEILE  = /[^:\n]+[:][ \t]*\n/     # Kategorienzeilen enthalten genau einen Doppelpunkt am Ende der Zeile
 
 DATEI_ENDE       = !/./
 NIEMALS          = /(?!.)/

examples/MLW/MLW_ebnf_ERRORS.txt

-Warning: Rule "SW_GRIECH" is not connected to parser root "Artikel" !
-Warning: Rule "VerweisZiel" is not connected to parser root "Artikel" !
-Warning: Rule "DEU_WORT" is not connected to parser root "Artikel" !
-Warning: Rule "DEU_GROSS" is not connected to parser root "Artikel" !
-Warning: Rule "DEU_KLEIN" is not connected to parser root "Artikel" !
-Warning: Rule "LAT_WORT" is not connected to parser root "Artikel" !
-Warning: Rule "GROSSSCHRIFT" is not connected to parser root "Artikel" !
-Warning: Rule "ZW" is not connected to parser root "Artikel" !
-Warning: Rule "LÜCKE" is not connected to parser root "Artikel" !
-Warning: Rule "LEERZEILE" is not connected to parser root "Artikel" !
-Warning: Rule "RZS" is not connected to parser root "Artikel" !
-Warning: Rule "KOMMENTARZEILEN" is not connected to parser root "Artikel" !
-Warning: Rule "NIEMALS" is not connected to parser root "Artikel" !
+line:  31, column: 49, Error: Missing definition for symbol 'LemmaWort'
+line:  70, column: 22, Error: Missing definition for symbol 'LAT'
+line:  70, column: 28, Error: Missing definition for symbol 'GRI'
+line:  78, column: 66, Error: Missing definition for symbol 'GebrauchPosition'
+line:  79, column: 23, Error: Missing definition for symbol 'MetrikPosition'
+line:  79, column: 40, Error: Missing definition for symbol 'VerwechselungPosition'
+line:  88, column: 20, Error: Missing definition for symbol 'SWUnterTyp'
+line:  89, column: 61, Error: Missing definition for symbol 'OFFEN'
+line: 142, column:  1, Warning: Rule "SW_GRIECH" is not connected to parser root "Artikel" !
+line: 158, column: 20, Error: Missing definition for symbol 'Autor'
+line: 158, column: 30, Error: Missing definition for symbol 'Werk'
+line: 158, column: 39, Error: Missing definition for symbol 'Stelle'
+line: 158, column: 51, Error: Missing definition for symbol 'Datierung'
+line: 158, column: 67, Error: Missing definition for symbol 'Edition'
+line: 162, column:  1, Warning: Rule "VerweisZiel" is not connected to parser root "Artikel" !
+line: 173, column:  1, Warning: Rule "DEU_GROSS" is not connected to parser root "Artikel" !
+line: 174, column:  1, Warning: Rule "DEU_KLEIN" is not connected to parser root "Artikel" !
+line: 175, column:  1, Warning: Rule "LAT_WORT" is not connected to parser root "Artikel" !
+line: 177, column:  1, Warning: Rule "GROSSSCHRIFT" is not connected to parser root "Artikel" !
+line: 199, column:  1, Warning: Rule "RZS" is not connected to parser root "Artikel" !
+line: 206, column:  1, Warning: Rule "NIEMALS" is not connected to parser root "Artikel" !
+line: 208, column:  1, Warning: Rule "DUMMY" is not connected to parser root "Artikel" !

test/test_parser.py

@@ -27,8 +27,9 @@ sys.path.extend(['../', './'])
 from DHParser.toolkit import is_logging, logging, compile_python_object
 from DHParser.stringview import StringView
 from DHParser.error import Error
+from DHParser.syntaxtree import mock_syntax_tree
 from DHParser.parser import compile_source, Retrieve, Grammar, Forward, Token, ZeroOrMore, RE, \
-    RegExp, Lookbehind, NegativeLookahead, OneOrMore, Series, Alternative, AllOf, SomeOf
+    RegExp, Lookbehind, NegativeLookahead, OneOrMore, Series, Alternative, AllOf, SomeOf, Compiler
 from DHParser.ebnf import get_ebnf_grammar, get_ebnf_transformer, get_ebnf_compiler
 from DHParser.dsl import grammar_provider, DHPARSER_IMPORTS
 
@@ -542,6 +543,22 @@ class TestBorderlineCases:
         assert not cst.error_flag
 
 
+class TestCompilerClass:
+    def test_error_propagations(self):
+        tree = mock_syntax_tree('(A (B 1) (C (D (E 2) (F 3))))')
+        A = tree
+        B = next(tree.find(lambda node: str(node) == "1"))
+        D = next(tree.find(lambda node: node.parser.name == "D"))
+        F = next(tree.find(lambda node: str(node) == "3"))
+        B.add_error("Error in child node")
+        F.add_error("Error in child's child node")
+        Compiler.propagate_error_flags(tree, lazy=True)
+        assert A.error_flag
+        assert not D.error_flag
+        Compiler.propagate_error_flags(tree, lazy=False)
+        assert D.error_flag
+
+
 if __name__ == "__main__":
     from DHParser.testing import runner
     with logging(False):

test/test_syntaxtree.py

@@ -23,6 +23,7 @@ import copy
 import sys
 sys.path.extend(['../', './'])
 
+from DHParser.error import Error
 from DHParser.syntaxtree import Node, mock_syntax_tree, TOKEN_PTYPE
 from DHParser.transform import traverse, reduce_single_child, \
     replace_by_single_child, flatten, remove_expendables
@@ -124,6 +125,23 @@ class TestNode:
         assert nd1.pos == 0, "Expected Node.pos == 0, got %i" % nd1.pos
         assert nd2.pos == 3, "Expected Node.pos == 3, got %i" % nd2.pos
 
+    def test_collect_errors(self):
+        tree = mock_syntax_tree('(A (B 1) (C (D (E 2) (F 3))))')
+        A = tree
+        B = next(tree.find(lambda node: str(node) == "1"))
+        D = next(tree.find(lambda node: node.parser.name == "D"))
+        F = next(tree.find(lambda node: str(node) == "3"))
+        B.add_error("Error in child node")
+        F.add_error("Error in child's child node")
+        tree.error_flag = Error.ERROR
+        errors = tree.collect_errors()
+        assert len(errors) == 2, str(errors)
+        assert A.error_flag
+        assert D.error_flag
+        errors = tree.collect_errors(clear_errors=True)
+        assert len(errors) == 2
+        assert not D.error_flag
+
 
 class TestErrorHandling:
     def test_error_flag_propagation(self):