Fix bug: _is_binary_stream should recognize TextIOWrapper as non-binary, escaped \r\n should be removed (#616)

* detect TextIOWrapper as non-binary

* I don't understand the CHANGELOG.md format, hope this is good enough

* Delete \\\r\n in Literal Strings (ref. section 7.3.4.2 of PDF32000_2008)

* Keep Travis CI happy

* Added test

* Remove pdfminer/Changelog

* Prettify _parse_string_1

* Add CHANGELOG.md

* Satisfy flake8

* Update CHANGELOG.md

* Use logging.Logger.warning instead of warning.warn in most cases, following
 the Python official guidance that warning.warn is directed at _developers_,
 not users

 * (pdfdocument.py) remove declarations of PDFTextExtractionNotAllowedWarning,
			PDFNoValidXRefWarning

 * (pdfpage.py) Don't import warning, don't use PDFTextExtractionNotAllowedWarning

 * (tools/dumppdf.py) Don't import warning, don't use PDFNoValidXRefWarning

 * (tests/test_tools_dumppdf.py) Don't import warning, check for logging.WARN rather
				  than PDFNoValidXRefWarning

* get name right

* make flake8 happy

* Revert "make flake8 happy"

This reverts commit 4592769686.

* Revert "get name right"

This reverts commit 80091ea211.

* Revert "Use logging.Logger.warning instead of warning.warn in most cases, following"

This reverts commit 3c1e3d6606.

* Revert "Merge branch 'preferLoggingToWarning' into hst"

This reverts commit 9d9d139921, reversing
changes made to 80091ea211.

* Revert "Revert "Merge branch 'preferLoggingToWarning' into hst""

This reverts commit b3da21934d.

Co-authored-by: Henry S. Thompson <ht@home.hst.name>
Co-authored-by: Pieter Marsman <pietermarsman@gmail.com>

CHANGELOG.md

@@ -20,6 +20,8 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
 - Fix `.paint_path` logic for handling single line segments and extracting point-on-curve positions of Beziér path commands ([#530](https://github.com/pdfminer/pdfminer.six/pull/530))
 - Raising `UnboundLocalError` when a bad `--output-type`  is used ([#610](https://github.com/pdfminer/pdfminer.six/pull/610))
 - `TypeError` when using `TagExtractor` with non-string or non-bytes tag values ([#610](https://github.com/pdfminer/pdfminer.six/pull/610))
+- Using `io.TextIOBase` as the file to write to ([#616](https://github.com/pdfminer/pdfminer.six/pull/616))
+- Parsing \r\n after the escape character in a literal string ([#616](https://github.com/pdfminer/pdfminer.six/pull/616))
 
 ## Removed
 - Support for Python 3.4 and 3.5 ([#522](https://github.com/pdfminer/pdfminer.six/pull/522))

pdfminer/converter.py

@@ -181,6 +181,8 @@ class PDFConverter(PDFLayoutAnalyzer):
             return True
         elif isinstance(outfp, io.StringIO):
             return False
+        elif isinstance(outfp, io.TextIOBase):
+            return False
 
         return True
 

pdfminer/psparser.py

@@ -444,16 +444,29 @@ class PSBaseParser:
         return j+1
 
     def _parse_string_1(self, s, i):
+        """Parse literal strings
+
+        PDF Reference 3.2.3
+        """
         c = s[i:i+1]
         if OCT_STRING.match(c) and len(self.oct) < 3:
             self.oct += c
             return i+1
-        if self.oct:
+
+        elif self.oct:
             self._curtoken += bytes((int(self.oct, 8),))
             self._parse1 = self._parse_string
             return i
-        if c in ESC_STRING:
+
+        elif c in ESC_STRING:
             self._curtoken += bytes((ESC_STRING[c],))
+
+        elif c == b'\r' and len(s) > i+1 and s[i+1:i+2] == b'\n':
+            # If current and next character is \r\n skip both because enters
+            # after a \ are ignored
+            i += 1
+
+        # default action
         self._parse1 = self._parse_string
         return i+1
 

tests/test_converter.py

@@ -207,3 +207,6 @@ class TestBinaryDetector():
 
     def test_non_file_like_object_defaults_to_binary(self):
         assert_true(PDFConverter._is_binary_stream(object()))
+
+    def test_textiowrapper(self):
+        assert_false(PDFConverter._is_binary_stream(io.TextIOBase()))