Merging upstream version 1.5.5.

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

6 files changed, 110 insertions, 13 deletions

diff --git a/README.md b/README.md
index 7a58a1f..17aabd2 100644
--- a/README.md
+++ b/README.md
@@ -21,6 +21,7 @@ If you have an actual file on disk, you can get the most information possible
 (a superset of all other methods):
 
 ```python
+>>> from identify import identify
 >>> identify.tags_from_path('/path/to/file.py')
 {'file', 'text', 'python', 'non-executable'}
 >>> identify.tags_from_path('/path/to/file-with-shebang')
diff --git a/identify/extensions.py b/identify/extensions.py
index 4475a41..1110ccb 100644
--- a/identify/extensions.py
+++ b/identify/extensions.py
@@ -10,11 +10,11 @@ EXTENSIONS = {
     'asar': {'binary', 'asar'},
     'bash': {'text', 'shell', 'bash'},
     'bat': {'text', 'batch'},
+    'bib': {'text', 'bib'},
     'bmp': {'binary', 'image', 'bitmap'},
     'bz2': {'binary', 'bzip2'},
     'c': {'text', 'c'},
     'cc': {'text', 'c++'},
-    'cu': {'text', 'cuda'},
     'cfg': {'text'},
     'chs': {'text', 'c2hs'},
     'clj': {'text', 'clojure'},
@@ -31,6 +31,7 @@ EXTENSIONS = {
     'cson': {'text', 'cson'},
     'css': {'text', 'css'},
     'csv': {'text', 'csv'},
+    'cu': {'text', 'cuda'},
     'cxx': {'text', 'c++'},
     'dart': {'text', 'dart'},
     'def': {'text', 'def'},
@@ -89,6 +90,7 @@ EXTENSIONS = {
     'key': {'text', 'pem'},
     'kml': {'text', 'kml', 'xml'},
     'kt': {'text', 'kotlin'},
+    'lean': {'text', 'lean'},
     'less': {'text', 'less'},
     'lhs': {'text', 'literate-haskell'},
     'libsonnet': {'text', 'jsonnet'},
@@ -130,31 +132,32 @@ EXTENSIONS = {
     'proto': {'text', 'proto'},
     'puml': {'text', 'plantuml'},
     'purs': {'text', 'purescript'},
+    'pxd': {'text', 'cython'},
+    'pxi': {'text', 'cython'},
     'py': {'text', 'python'},
     'pyi': {'text', 'pyi'},
     'pyx': {'text', 'cython'},
     'pyz': {'binary', 'pyz'},
     'pyzw': {'binary', 'pyz'},
-    'pxd': {'text', 'cython'},
-    'pxi': {'text', 'cython'},
     'r': {'text', 'r'},
     'rb': {'text', 'ruby'},
     'rs': {'text', 'rust'},
     'rst': {'text', 'rst'},
     's': {'text', 'asm'},
+    'sass': {'text', 'sass'},
     'sbt': {'text', 'sbt', 'scala'},
     'sc': {'text', 'scala'},
     'scala': {'text', 'scala'},
-    'scss': {'text', 'scss'},
     'scm': {'text', 'scheme'},
+    'scss': {'text', 'scss'},
     'sh': {'text', 'shell'},
     'sls': {'text', 'salt'},
     'so': {'binary'},
     'sol': {'text', 'solidity'},
     'spec': {'text', 'spec'},
+    'sql': {'text', 'sql'},
     'ss': {'text', 'scheme'},
     'styl': {'text', 'stylus'},
-    'sql': {'text', 'sql'},
     'sv': {'text', 'system-verilog'},
     'svg': {'text', 'image', 'svg', 'xml'},
     'svh': {'text', 'system-verilog'},
@@ -163,15 +166,17 @@ EXTENSIONS = {
     'swiftdeps': {'text', 'swiftdeps'},
     'tac': {'text', 'twisted', 'python'},
     'tar': {'binary', 'tar'},
+    'tex': {'text', 'tex'},
+    'tf': {'text', 'terraform'},
+    'tfvars': {'text', 'terraform'},
     'tgz': {'binary', 'gzip'},
     'thrift': {'text', 'thrift'},
     'tiff': {'binary', 'image', 'tiff'},
     'toml': {'text', 'toml'},
-    'tf': {'text', 'terraform'},
-    'tfvars': {'text', 'terraform'},
     'ts': {'text', 'ts'},
     'tsx': {'text', 'tsx'},
     'ttf': {'binary', 'ttf'},
+    'txsprofile': {'text', 'ini', 'txsprofile'},
     'txt': {'text', 'plain-text'},
     'v': {'text', 'verilog'},
     'vdx': {'text', 'vdx'},
@@ -181,11 +186,12 @@ EXTENSIONS = {
     'vue': {'text', 'vue'},
     'war': {'binary', 'zip', 'jar'},
     'wav': {'binary', 'audio', 'wav'},
-    'wkt': {'text', 'wkt'},
     'whl': {'binary', 'wheel', 'zip'},
+    'wkt': {'text', 'wkt'},
     'woff': {'binary', 'woff'},
     'woff2': {'binary', 'woff2'},
     'wsgi': {'text', 'wsgi', 'python'},
+    'xhtml': {'text', 'xml', 'html', 'xhtml'},
     'xml': {'text', 'xml'},
     'xq': {'text', 'xquery'},
     'xql': {'text', 'xquery'},
@@ -209,30 +215,32 @@ EXTENSIONS_NEED_BINARY_CHECK = {
 
 NAMES = {
     '.babelrc': EXTENSIONS['json'] | {'babelrc'},
-    '.bashrc': EXTENSIONS['bash'],
     '.bash_aliases': EXTENSIONS['bash'],
     '.bash_profile': EXTENSIONS['bash'],
+    '.bashrc': EXTENSIONS['bash'],
     '.bowerrc': EXTENSIONS['json'] | {'bowerrc'},
     '.coveragerc': EXTENSIONS['ini'] | {'coveragerc'},
     '.cshrc': EXTENSIONS['csh'],
     '.dockerignore': {'text', 'dockerignore'},
     '.editorconfig': {'text', 'editorconfig'},
-    '.gitconfig': EXTENSIONS['ini'] | {'gitconfig'},
-    '.hgrc': EXTENSIONS['ini'] | {'hgrc'},
+    '.flake8': EXTENSIONS['ini'] | {'flake8'},
     '.gitattributes': {'text', 'gitattributes'},
+    '.gitconfig': EXTENSIONS['ini'] | {'gitconfig'},
     '.gitignore': {'text', 'gitignore'},
     '.gitmodules': {'text', 'gitmodules'},
+    '.hgrc': EXTENSIONS['ini'] | {'hgrc'},
     '.jshintrc': EXTENSIONS['json'] | {'jshintrc'},
     '.mailmap': {'text', 'mailmap'},
     '.mention-bot': EXTENSIONS['json'] | {'mention-bot'},
     '.npmignore': {'text', 'npmignore'},
     '.pdbrc': EXTENSIONS['py'] | {'pdbrc'},
     '.pypirc': EXTENSIONS['ini'] | {'pypirc'},
+    '.rstcheck.cfg': EXTENSIONS['ini'],
     '.yamllint': EXTENSIONS['yaml'] | {'yamllint'},
     '.zshrc': EXTENSIONS['zsh'],
     'AUTHORS': EXTENSIONS['txt'],
-    'BUILD.bazel': {'text', 'bazel'},
     'BUILD': {'text', 'bazel'},
+    'BUILD.bazel': {'text', 'bazel'},
     'CMakeLists.txt': EXTENSIONS['cmake'],
     'COPYING': EXTENSIONS['txt'],
     'Dockerfile': {'text', 'dockerfile'},
diff --git a/identify/identify.py b/identify/identify.py
index 8a21d8b..1c0e677 100644
--- a/identify/identify.py
+++ b/identify/identify.py
@@ -141,6 +141,27 @@ def _shebang_split(line):
         return line.split()
 
 
+def _parse_nix_shebang(bytesio, cmd):
+    while bytesio.read(2) == b'#!':
+        next_line = bytesio.readline()
+        try:
+            next_line = next_line.decode('UTF-8')
+        except UnicodeDecodeError:
+            return cmd
+
+        for c in next_line:
+            if c not in printable:
+                return cmd
+
+        line_tokens = tuple(_shebang_split(next_line.strip()))
+        for i, token in enumerate(line_tokens[:-1]):
+            if token != '-i':
+                continue
+            # the argument to -i flag
+            cmd = (line_tokens[i + 1],)
+    return cmd
+
+
 def parse_shebang(bytesio):
     """Parse the shebang from a file opened for reading binary."""
     if bytesio.read(2) != b'#!':
@@ -159,6 +180,8 @@ def parse_shebang(bytesio):
     cmd = tuple(_shebang_split(first_line.strip()))
     if cmd and cmd[0] == '/usr/bin/env':
         cmd = cmd[1:]
+        if cmd == ('nix-shell',):
+            return _parse_nix_shebang(bytesio, cmd)
     return cmd
 
 
diff --git a/identify/interpreters.py b/identify/interpreters.py
index bcffb5a..7feb4b1 100644
--- a/identify/interpreters.py
+++ b/identify/interpreters.py
@@ -3,9 +3,14 @@ from __future__ import absolute_import
 from __future__ import unicode_literals
 
 INTERPRETERS = {
+    'ash': {'shell', 'ash'},
+    'awk': {'awk'},
     'bash': {'shell', 'bash'},
+    'bats': {'shell', 'bash', 'bats'},
     'csh': {'shell', 'csh'},
     'dash': {'shell', 'dash'},
+    'expect': {'expect'},
+    'ksh': {'shell', 'ksh'},
     'node': {'javascript'},
     'nodejs': {'javascript'},
     'perl': {'perl'},
diff --git a/setup.cfg b/setup.cfg
index ea0b137..c742f5a 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -1,6 +1,6 @@
 [metadata]
 name = identify
-version = 1.4.29
+version = 1.5.5
 description = File identification library for Python
 long_description = file: README.md
 long_description_content_type = text/markdown
diff --git a/tests/identify_test.py b/tests/identify_test.py
index 44406a1..a80c401 100644
--- a/tests/identify_test.py
+++ b/tests/identify_test.py
@@ -217,6 +217,66 @@ def test_file_is_text_does_not_exist(tmpdir):
         (b"#!/path'with/quotes    y", ("/path'with/quotes", 'y')),
         # Don't regress on leading/trailing ws
         (b"#! /path'with/quotes y ", ("/path'with/quotes", 'y')),
+        # Test nix-shell specialites with shebang on second line
+        (
+            b'#! /usr/bin/env nix-shell\n'
+            b'#! nix-shell -i bash -p python',
+            ('bash',),
+        ),
+        (
+            b'#! /usr/bin/env nix-shell\n'
+            b'#! nix-shell -i python -p coreutils',
+            ('python',),
+        ),
+        (
+            b'#! /usr/bin/env nix-shell\n'
+            b'#! nix-shell -p coreutils -i python',
+            ('python',),
+        ),
+        # multi-line and no whitespace variation
+        (
+            b'#! /usr/bin/env nix-shell\n'
+            b'#! nix-shell -p coreutils\n'
+            b'#! nix-shell -i python',
+            ('python',),
+        ),
+        (
+            b'#! /usr/bin/env nix-shell\n'
+            b'#!nix-shell -p coreutils\n'
+            b'#!nix-shell -i python',
+            ('python',),
+        ),
+        (
+            b'#! /usr/bin/env nix-shell\n'
+            b'#!\xf9\x93\x01\x42\xcd',
+            ('nix-shell',),
+        ),
+        (
+            b'#! /usr/bin/env nix-shell\n'
+            b'#!\x00\x00\x00\x00',
+            ('nix-shell',),
+        ),
+        # non-proper nix-shell
+        (b'#! /usr/bin/nix-shell', ('/usr/bin/nix-shell',)),
+        (b'#! /usr/bin/env nix-shell', ('nix-shell',)),
+        (
+            b'#! /usr/bin/env nix-shell non-portable-argument',
+            ('nix-shell', 'non-portable-argument'),
+        ),
+        (
+            b'#! /usr/bin/env nix-shell\n'
+            b'#! nix-shell -i',
+            ('nix-shell',),   # guard against index error
+        ),
+        # interpret quotes correctly
+        (
+            b'#!/usr/bin/env nix-shell\n'
+            b'#!nix-shell --argstr x "a -i python3 p"\n'
+            b'#!nix-shell -p hello\n'
+            b'#!nix-shell -i bash\n'
+            b'#!nix-shell --argstr y "b -i runhaskell q"',
+            ('bash',),
+        ),
         (b'\xf9\x93\x01\x42\xcd', ()),
         (b'#!\xf9\x93\x01\x42\xcd', ()),
         (b'#!\x00\x00\x00\x00', ()),