Merge pull request #175 from mukundans91/validateImageName

Added regex pattern based validation for image name argument

repo2docker/app.py

@@ -30,7 +30,7 @@ from .buildpacks import (
     PythonBuildPack, DockerBuildPack, LegacyBinderDockerBuildPack,
     CondaBuildPack, JuliaBuildPack, Python2BuildPack, BaseImage
 )
-from .utils import execute_cmd, ByteSpecification, maybe_cleanup
+from .utils import execute_cmd, ByteSpecification, maybe_cleanup, is_valid_docker_image_name
 from . import __version__
 
 
@@ -180,6 +180,25 @@ class Repo2Docker(Application):
                                extra=dict(phase='failed'))
                 sys.exit(1)
 
+    def validate_image_name(self, image_name):
+        """
+        Validate image_name read by argparse contains only lowercase characters
+
+        Args:
+            image_name (string): argument read by the argument parser
+
+        Returns:
+            unmodified image_name
+
+        Raises:
+            ArgumentTypeError: if image_name contains characters that are not lowercase
+        """
+
+        if not is_valid_docker_image_name(image_name):
+            msg = "%r is not a valid docker image name. Image name can contain only lowercase characters." % image_name
+            raise argparse.ArgumentTypeError(msg)
+        return image_name
+
     def get_argparser(self):
         argparser = argparse.ArgumentParser()
         argparser.add_argument(
@@ -204,7 +223,8 @@ class Repo2Docker(Application):
         argparser.add_argument(
             '--image-name',
             help=('Name of image to be built. If unspecified will be '
-                  'autogenerated')
+                  'autogenerated'),
+            type=self.validate_image_name
         )
 
         argparser.add_argument(

repo2docker/utils.py

@@ -2,6 +2,7 @@ from contextlib import contextmanager
 from functools import partial
 import shutil
 import subprocess
+import re
 
 from traitlets import Integer
 
@@ -56,6 +57,81 @@ def maybe_cleanup(path, cleanup=False):
         shutil.rmtree(path, ignore_errors=True)
 
 
+def is_valid_docker_image_name(image_name):
+    """
+    Function that constructs a regex representing the docker image name and tests it against the given image_name
+    Reference Regex definition in https://github.com/docker/distribution/blob/master/reference/regexp.go
+
+    Args:
+        image_name: string representing a docker image name
+
+    Returns:
+        True if image_name is valid else False
+
+    Example:
+
+        'test.Com/name:latest' is a valid tag
+
+        'Test/name:latest' is not a valid tag
+
+        Note:
+
+        This function has a stricter pattern than https://github.com/docker/distribution/blob/master/reference/regexp.go
+
+        This pattern will not allow cases like
+        'TEST.com/name:latest' though docker considers it a valid tag
+    """
+    reference_regex = re.compile(r"""^ # Anchored at start and end of string
+
+                        ( # Start capturing name
+
+                        (?: # start grouping the optional registry domain name part
+
+                        (?:[a-z0-9]|[a-z0-9][a-z0-9-]*[a-z0-9]) # lowercase only '<domain-name-component>'
+
+                        (?: # start optional group
+
+                        (?:\.(?:[a-zA-Z0-9]|[a-zA-Z0-9][a-zA-Z0-9-]*[a-zA-Z0-9]))+ # multiple repetitions of pattern '.<domain-name-component>'
+
+                        )? # end optional grouping part of the '.' separated domain name
+
+                        (?::[0-9]+)?/ # '<domain-name>' followed by an optional '<port>' component followed by '/' literal
+
+                        )? # end grouping the optional registry domain part
+
+                        # start <name-pattern>
+                        [a-z0-9]+   # must have a <name-component>
+                        (?:
+                        (?:(?:[\._]|__|[-]*)[a-z0-9]+)+ # repeat the pattern '<separator><name-component>'
+                        )? # optionally have multiple repetitions of the above line
+                        # end <name-pattern>
+
+                        (?: # start optional name components
+
+                        (?: # start multiple repetitions
+
+                        /   # separate multiple name components by /
+                        # start <name-pattern>
+                        [a-z0-9]+ # must have a <name-component>
+                        (?:
+                        (?:(?:[\._]|__|[-]*)[a-z0-9]+)+ # repeat the pattern '<separator><name-component>'
+                        )? # optionally have multiple repetitions of the above line
+                        # end <name-pattern>
+
+                        )+ # multiple repetitions of the pattern '/<name-component><separator><name-component>'
+
+                        )? # optionally have  the above group
+
+                        ) # end capturing name
+
+                        (?::([\w][\w.-]{0,127}))? # optional capture <tag-pattern>=':<tag>'
+                        (?:@[A-Za-z][A-Za-z0-9]*(?:[-_+.][A-Za-z][A-Za-z0-9]*)*[:][[:xdigit:]]{32,})? # optionally capture <digest-pattern>='@<digest>'
+                        $""",
+                                 re.VERBOSE)
+
+    return reference_regex.match(image_name) is not None
+
+
 class ByteSpecification(Integer):
     """
     Allow easily specifying bytes in units of 1024 with suffixes

tests/argumentvalidation.py

@@ -0,0 +1,90 @@
+"""
+Tests that runs validity checks on arguments passed in from shell
+"""
+
+import os
+import subprocess
+
+def does_validate_image_name(builddir, image_name):
+    try:
+        output = subprocess.check_output(
+            [
+                'repo2docker',
+                '--no-run',
+                '--no-build',
+                '--image-name',
+                str(image_name),
+                builddir
+            ],
+            stderr=subprocess.STDOUT,
+        ).decode()
+        return True
+    except subprocess.CalledProcessError as e:
+        output = e.output.decode()
+        if "error: argument --image-name: %r is not a valid docker image name. " \
+           "Image name can contain only lowercase characters." % image_name in output:
+            return False
+        else:
+            raise
+
+
+def test_image_name_fail():
+    """
+    Test to check if repo2docker throws image_name validation error on --image-name argument containing
+    uppercase characters and _ characters in incorrect positions.
+    """
+
+    builddir = os.path.dirname(__file__)
+
+    assert not does_validate_image_name(builddir, 'Test/Invalid_name:1.0.0')
+
+
+def test_image_name_underscore_fail():
+    """
+    Test to check if repo2docker throws image_name validation error on --image-name argument starts with _.
+    """
+
+    builddir = os.path.dirname(__file__)
+
+    assert not does_validate_image_name(builddir, '_test/invalid_name:1.0.0')
+
+
+def test_image_name_double_dot_fail():
+    """
+    Test to check if repo2docker throws image_name validation error on --image-name argument contains consecutive dots.
+    """
+
+    builddir = os.path.dirname(__file__)
+
+    assert not does_validate_image_name(builddir, 'test..com/invalid_name:1.0.0')
+
+
+def test_image_name_valid_restircted_registry_domain_name_fail():
+    """
+    Test to check if repo2docker throws image_name validation error on -image-name argument being invalid. Based on the
+    regex definitions first part of registry domain cannot contain uppercase characters
+    """
+
+    builddir = os.path.dirname(__file__)
+
+    assert not does_validate_image_name(builddir, 'Test.com/valid_name:1.0.0')
+
+
+def test_image_name_valid_registry_domain_name_success():
+    """
+    Test to check if repo2docker runs with a valid --image-name argument.
+    """
+
+    builddir = os.path.dirname(__file__) + '/dockerfile/simple/'
+
+    assert does_validate_image_name(builddir, 'test.COM/valid_name:1.0.0')
+
+
+def test_image_name_valid_name_success():
+    """
+    Test to check if repo2docker runs with a valid --image-name argument.
+    """
+
+    builddir = os.path.dirname(__file__) + '/dockerfile/simple/'
+
+    assert does_validate_image_name(builddir, 'test.com/valid_name:1.0.0')