Merge github.com:mathiasertl/pygments

fixes #1645
pygments · Dec 28, 2020 · 00a31bc · 00a31bc
2 parents a3d4100 + 456ac59
commit 00a31bc
Show file tree

Hide file tree

Showing 3 changed files with 27 additions and 5 deletions.
diff --git a/CHANGES b/CHANGES
@@ -22,6 +22,7 @@ Version 2.7.4
 - Limit recursion with nesting Ruby heredocs (#1638)
 - Fixed a few inefficient regexes for guessing lexers
 - Fixed the raw token lexer handling of Unicode (#1616)
+- Lex trailing whitespace as part of the prompt in shell lexers (#1645)
 
 Thanks to Google's OSS-Fuzz project for finding many of these bugs.
 

diff --git a/pygments/lexers/shell.py b/pygments/lexers/shell.py
@@ -232,7 +232,7 @@ class BashSessionLexer(ShellSessionBaseLexer):
     _innerLexerCls = BashLexer
     _ps1rgx = re.compile(
         r'^((?:(?:\[.*?\])|(?:\(\S+\))?(?:| |sh\S*?|\w+\S+[@:]\S+(?:\s+\S+)' \
-        r'?|\[\S+[@:][^\n]+\].+))\s*[$#%])(.*\n?)')
+        r'?|\[\S+[@:][^\n]+\].+))\s*[$#%]\s*)(.*\n?)')
     _ps2 = '>'
 
 

diff --git a/tests/test_shell.py b/tests/test_shell.py
@@ -143,8 +143,7 @@ def test_end_of_line_nums(lexer_bash):
 def test_newline_in_echo(lexer_session):
     fragment = '$ echo \\\nhi\nhi\n'
     tokens = [
-        (Token.Generic.Prompt, '$'),
-        (Token.Text, ' '),
+        (Token.Generic.Prompt, '$ '),
         (Token.Name.Builtin, 'echo'),
         (Token.Text, ' '),
         (Token.Literal.String.Escape, '\\\n'),
@@ -155,6 +154,29 @@ def test_newline_in_echo(lexer_session):
     assert list(lexer_session.get_tokens(fragment)) == tokens
 
 
+def test_newline_in_ls(lexer_session):
+    fragment = '$ ls \\\nhi\nhi\n'
+    tokens = [
+        (Token.Generic.Prompt, '$ '),
+        (Token.Text, 'ls'),
+        (Token.Text, ' '),
+        (Token.Literal.String.Escape, '\\\n'),
+        (Token.Text, 'hi'),
+        (Token.Text, '\n'),
+        (Token.Generic.Output, 'hi\n'),
+    ]
+    assert list(lexer_session.get_tokens(fragment)) == tokens
+
+
+def test_comment_after_prompt(lexer_session):
+    fragment = '$# comment'
+    tokens = [
+        (Token.Generic.Prompt, '$'),
+        (Token.Comment.Single, '# comment\n'),
+    ]
+    assert list(lexer_session.get_tokens(fragment)) == tokens
+
+
 def test_msdos_gt_only(lexer_msdos):
     fragment = '> py\nhi\n'
     tokens = [
@@ -208,8 +230,7 @@ def test_virtualenv(lexer_session):
     tokens = [
         (Token.Generic.Prompt.VirtualEnv, '(env)'),
         (Token.Text, ' '),
-        (Token.Generic.Prompt, '[~/project]$'),
-        (Token.Text, ' '),
+        (Token.Generic.Prompt, '[~/project]$ '),
         (Token.Text, 'foo'),
         (Token.Text, ' '),
         (Token.Text, '-h'),