Update Dockerfile lexer

Add Dockerfiles new keyword in the lexer
Improve lexer to parse json arrays on specific commands
Handle line breaks as a bash syntax

2 files changed, 45 insertions, 10 deletions

diff --git a/pygments/lexers/configs.py b/pygments/lexers/configs.py
index c39b1a52..4af2adb6 100644
--- a/pygments/lexers/configs.py
+++ b/pygments/lexers/configs.py
@@ -15,6 +15,7 @@ from pygments.lexer import RegexLexer, default, words, bygroups, include, using
 from pygments.token import Text, Comment, Operator, Keyword, Name, String, \
     Number, Punctuation, Whitespace, Literal
 from pygments.lexers.shell import BashLexer
+from pygments.lexers.data import JsonLexer
 
 __all__ = ['IniLexer', 'RegeditLexer', 'PropertiesLexer', 'KconfigLexer',
            'Cfengine3Lexer', 'ApacheConfLexer', 'SquidConfLexer',
@@ -539,20 +540,25 @@ class DockerLexer(RegexLexer):
     filenames = ['Dockerfile', '*.docker']
     mimetypes = ['text/x-dockerfile-config']
 
-    _keywords = (r'(?:FROM|MAINTAINER|CMD|EXPOSE|ENV|ADD|ENTRYPOINT|'
-                 r'VOLUME|WORKDIR)')
-
+    _keywords = (r'(?:FROM|MAINTAINER|EXPOSE|WORKDIR|USER|STOPSIGNAL)')
+    _bash_keywords = (r'(?:RUN|CMD|ENTRYPOINT|ENV|ARG|LABEL|ADD|COPY)')
+    _lb = r'(?:\s*\\?\s*)' # dockerfile line break regex
     flags = re.IGNORECASE | re.MULTILINE
 
     tokens = {
         'root': [
-            (r'^(ONBUILD)(\s+)(%s)\b' % (_keywords,),
-             bygroups(Name.Keyword, Whitespace, Keyword)),
-            (r'^(%s)\b(.*)' % (_keywords,), bygroups(Keyword, String)),
             (r'#.*', Comment),
-            (r'RUN', Keyword),  # Rest of line falls through
+            (r'(ONBUILD)(%s)' % (_lb,), bygroups(Keyword, using(BashLexer))),
+            (r'(HEALTHCHECK)((%s--\w+=\w+%s)*)' % (_lb, _lb),
+                bygroups(Keyword, using(BashLexer))),
+            (r'(VOLUME|ENTRYPOINT|CMD|SHELL)(%s)(\[.*?\])' % (_lb,),
+                bygroups(Keyword, using(BashLexer), using(JsonLexer))),
+            (r'(LABEL|ENV|ARG)((%s\w+=\w+%s)*)' % (_lb, _lb),
+                bygroups(Keyword, using(BashLexer))),
+            (r'(%s|VOLUME)\b(.*)' % (_keywords), bygroups(Keyword, String)),
+            (r'(%s)' % (_bash_keywords,), Keyword),
             (r'(.*\\\n)*.+', using(BashLexer)),
-        ],
+        ]
     }
 
 
diff --git a/tests/examplefiles/docker.docker b/tests/examplefiles/docker.docker
index d65385b6..1ae3c3a1 100644
--- a/tests/examplefiles/docker.docker
+++ b/tests/examplefiles/docker.docker
@@ -1,5 +1,34 @@
-maintainer First O'Last
+FROM alpine:3.5
+MAINTAINER First O'Last
 
+# comment
 run echo \
   123 $bar
-# comment
+RUN apk --update add rsync dumb-init
+
+# Test env with both syntax
+ENV FOO = "BAR"
+ENV FOO \
+	"BAR"
+
+COPY foo "bar"
+COPY foo \
+	"bar"
+
+HEALTHCHECK \
+    --interval=5m --timeout=3s \
+    CMD curl -f http://localhost/ || exit 1
+
+# ONBUILD keyword, then with linebreak
+ONBUILD ADD . /app/src
+ONBUILD \
+	RUN echo 123 $bar
+
+# Potential JSON array parsing, mixed with linebreaks
+VOLUME \
+    /foo
+VOLUME \
+    ["/bar"]
+VOLUME ["/bar"]
+VOLUME /foo
+CMD ["foo", "bar"]