From c8058c162f4adcd1c11fa90ab151aeaf79aa0c14 Mon Sep 17 00:00:00 2001
From: WillForan <willforan@gmail.com>
Date: Sun, 29 Dec 2024 14:46:06 -0500
Subject: [PATCH 1/2] fix #99: implement commented string_quote_removal

---
 bashlex/heredoc.py   | 18 ++++++++++++++++--
 tests/test_parser.py | 17 +++++++++++++++++
 2 files changed, 33 insertions(+), 2 deletions(-)

diff --git a/bashlex/heredoc.py b/bashlex/heredoc.py
index c6ee2e4f..c1da11a7 100644
--- a/bashlex/heredoc.py
+++ b/bashlex/heredoc.py
@@ -1,4 +1,5 @@
 from bashlex import ast, errors
+import re
 
 def gatherheredocuments(tokenizer):
     # if we're at the end of the input and we're not strict, allow skipping
@@ -11,9 +12,22 @@ def gatherheredocuments(tokenizer):
         redirnode, killleading = tokenizer.redirstack.pop(0)
         makeheredoc(tokenizer, redirnode, 0, killleading)
 
+
+def string_quote_removal(word):
+    """remove single quotes for heredoc
+    >>> string_quote_removal("'EOF'")
+    'EOF'
+    >>> string_quote_removal("EOF")
+    'EOF'
+    """
+    quote_match = re.search("^'(.*)'$", word)
+    if quote_match:
+        word = quote_match.group(1)
+    return word
+
+
 def makeheredoc(tokenizer, redirnode, lineno, killleading):
-    # redirword = string_quote_removal(redirectnode.word)
-    redirword = redirnode.output.word
+    redirword = string_quote_removal(redirnode.output.word)
     document = []
 
     startpos = tokenizer._shell_input_line_index
diff --git a/tests/test_parser.py b/tests/test_parser.py
index 10809c79..f04e83c8 100644
--- a/tests/test_parser.py
+++ b/tests/test_parser.py
@@ -779,6 +779,23 @@ def test_heredoc_with_actual_doc(self):
                                 "delimited by end-of-file \\(wanted 'EOF'",
                                 parse, s)
 
+    def test_heredoc_singlequotes(self):
+        doc = 'foo\nbar\nEOF'
+        s = '''a <<'EOF'
+%s''' % doc
+
+        self.assertASTEquals(s,
+                commandnode("a <<'EOF'",
+                  wordnode('a'),
+                  redirectnode("<<'EOF'\n%s" % doc, None, '<<', wordnode("'EOF'"),
+                      heredocnode(doc))
+                ))
+
+        s = "a <<'EOF'\nb"
+        self.assertRaisesRegex(errors.ParsingError,
+                                "delimited by end-of-file \\(wanted 'EOF'",
+                                parse, s)
+
     def test_herestring(self):
         s = 'a <<<"b\nc"'
         self.assertASTEquals(s,

From fff4df03615a86519236762f8e7a32457bac00c5 Mon Sep 17 00:00:00 2001
From: WillForan <willforan@gmail.com>
Date: Sat, 18 Jan 2025 20:59:33 -0500
Subject: [PATCH 2/2] string_quote_removal: rm escaped and double quotes (#99)

---
 bashlex/heredoc.py | 29 +++++++++++++++++++++++++----
 1 file changed, 25 insertions(+), 4 deletions(-)

diff --git a/bashlex/heredoc.py b/bashlex/heredoc.py
index c1da11a7..d043c16a 100644
--- a/bashlex/heredoc.py
+++ b/bashlex/heredoc.py
@@ -14,15 +14,36 @@ def gatherheredocuments(tokenizer):
 
 
 def string_quote_removal(word):
-    """remove single quotes for heredoc
+    """
+    Remove surrounding quotes for heredoc token.
+    Other quotes are removed from the token unless escaped.
+
+    See definition in bash's source
+    https://github.com/bminor/bash/blob/master/subst.c#L11892
+
+    >>> string_quote_removal("EOF")
+    'EOF'
     >>> string_quote_removal("'EOF'")
     'EOF'
-    >>> string_quote_removal("EOF")
+    >>> string_quote_removal('"EOF"')
     'EOF'
+    >>> string_quote_removal('HERE\\\\"DOC\\\\"')
+    'HERE"DOC"'
+    >>> string_quote_removal('"HERE"x"DOC"')
+    'HERExDOC'
+    
     """
-    quote_match = re.search("^'(.*)'$", word)
+    # remove paired quote from start and end
+    quote_match = re.search("^([\"'])(.*)\\1$", word)
     if quote_match:
-        word = quote_match.group(1)
+        word = quote_match.group(2)
+
+    # removing unescaped quotes. assumes matching quote pairs
+    # bash code would fail to parse otherwise (?)
+    word = re.sub("(?<=[^\\\\])[\"']","", word)
+
+    # escaped quotes in input become literal in returned token
+    word = word.replace('\\"','"').replace("\\'","'")
     return word