Fix bug in parsing of hex code points constants.

This happened in both: echo -e '_\u03bc_' and echo $'_\u03bc_'
oilshell · Jun 8, 2018 · f238a6d37fc8390a4c5245f7a7352d6ae4b5aceb · f238a6d
1 parent d5cbdc5
commit f238a6d37fc8390a4c5245f7a7352d6ae4b5aceb
Unified Split

Showing with 5 additions and 3 deletions.

+1 −1 asdl/asdl_.py

+2 −0 gold/echo-e.sh

+2 −2 osh/lex.py
diff --git a/asdl/asdl_.py b/asdl/asdl_.py
@@ -301,7 +301,7 @@ def Print(self, f, indent):

 class Sum(AST):
    def __init__(self, types, attributes=None):
-        self.types = types
+        self.types = types  # List[Constructor]
        self.attributes = attributes or []

    def Print(self, f, indent):
diff --git a/gold/echo-e.sh b/gold/echo-e.sh
@@ -17,6 +17,8 @@ echo -e 'abcd\x65f'
 echo -e 'abcd\044e'
 echo -e 'abcd\u0065f'
 echo -e 'abcd\U00000065f'
+echo -e '_\u03bc_'
+echo -e '_\U000003bc_'
 echo -en '\03777' | od -A n -t x1 | sed 's/ \+/ /g'
 echo -en '\04000' | od -A n -t x1 | sed 's/ \+/ /g'
 echo -e '\0777' | od -A n -t x1 | sed 's/ \+/ /g'
diff --git a/osh/lex.py b/osh/lex.py
@@ -395,8 +395,8 @@ def IsKeyword(name):

  # \x6 is valid in bash
  R(r'\\x[0-9a-fA-F]{1,2}', Id.Char_Hex),
-  R(r'\\u[0-9]{1,4}', Id.Char_Unicode4),
-  R(r'\\U[0-9]{1,8}', Id.Char_Unicode8),
+  R(r'\\u[0-9a-fA-F]{1,4}', Id.Char_Unicode4),
+  R(r'\\U[0-9a-fA-F]{1,8}', Id.Char_Unicode8),

  R(r'\\[0abeEfrtnv\\]', Id.Char_OneChar),