The Maven plugins that will expose your Git passwords and how Docker helps prove it

A few days ago, I noticed something very troubling after using the Maven release plugin to publish a release artifact to an internal Maven repository: my git password was exposed in the Maven build output as well as a git.properties file that the Maven Git commit ID plugin generated. These files are is now sitting in Artifactory for all to read.

screenshot-example

Not cool. For Maven-based projects, I typically use the the Maven Release Plugin. Because we’d also like to track some of the git metadata about how the build was produced, we also use the Maven Git commit ID plugin as well, which plays quite nicely SpringBoot. So I was very disturbed to see my password all over the place.

What is happening here?

First of all, I should be clear that the project in question is using the Maven command line wrapper and pulling down Maven 3.3.9, which is the latest at the time of this writing. I’m also using Git 2.7.4 and the current release of the Maven Git commit ID plugin, which is was 2.2.0. For the most part, everything is current.

This issue here is not specific plugin any single plugin (but it looks like the Maven Release Plugin is the core offender), but rather the issues only manifest themselves in certain conditions when the group of plugins interact with one another during the release process. The Maven plugins in question are:

The combination of these plugins will expose your Git passwords when using Git over either HTTP or HTTPS when the Maven Release plugins release:prepare and release:perform plugins are invoked, but curiously not when the package,install, or deploy goals are invoked. Additionally if you’re using the Maven Git Commit ID Plugin to capture commit information in your build, the generated git.properties will contain your user name and password when using the default setting and this file will be visible in the Maven repository your artifact is published to. What appears to be happening is that the Maven release plugin in rewriting the git origin URI and including the credentials in the URI. Thus, when the git commit ID plugin goes to resolve the git.remote.origin.url value, it now includes the username and password as well.

Demonstrating the issue

Reproducing this issue was kind of a pain in the ass, but with tools like Docker and Docker Compose, it’s a little less painless. I have created such an environment which does the following:

  • Creates an instance of Artifactory OSS for our Maven repository
  • Creates an instance of Gitbucket for the git repo manager
  • Creates a user named John Yaya who has a username of jyaya and a password of password
  • Creates a container with Maven and Git which also mounts a sample Maven project to demo the issue.

The project is here:

https://github.com/damnhandy/maven-publish-issue

Check the README.md for details on how to run it, but it works fine under Docker Machine on OS X and Linux. This project will startup both Artifactory and Gitbucket and startup a “workspace” container and dump you into the container where you’ll be able to execute the Maven and git commands. The test is 100% repeatable whenever you perform a release.

How you can prevent this?

There are a few ways you can prevent your passwords from being exposed:

Use SSH instead of HTTP/HTTPS in your CI setups

SSH is doesn’t have this issue and won’t expose your passwords. SSH avoids this problem all together, unless of course you use SSH with usernames and passwords. Granted, SSH may not be appropriate in all cases. If you’re in an enterprisey environment, this may be more complicated. SSH is also kind of a pain in the ass on Windows. If SSH isn’t an option, there’s a few more options.

Use Maven Release plugin 2.4.2 or higher

By default, the super pom in Maven 3.3.9 uses version 2.3.2 of the Maven Release plugin. This is fixed as of 2.4.2, but if you don’t define the version to use, you’ll be exposed. You have to explicitly define the version of the Maven Release plugin to use:

<plugin>
   <groupId>org.apache.maven.plugins</groupId>
   <artifactId>maven-release-plugin</artifactId>
   <version>2.5.3</version>
</plugin>

This will keep your passwords out of yours logs in Jenkins, TravisCI, or other CI environment, but it doesn’t address the fact that the Git commit ID plugin, and probably others, will still render the username and password in the URI.

Exclude the git.remote.origin.url property from your build

If you’re using the Git Commit ID Plugin, exclude the git.remote.origin.url property from your build:

<configuration>
	<excludeProperties>
	  <excludeProperty>git.remote.origin.url</excludeProperty>
	</excludeProperties>
</configuration>

This will completely remove the origin URI from the properties file. This can be annoying if you want to track where the code came from, which handy if you have more than one Git repo manager hosting the same code (i.e. you have a mirror repo that is local to one of your global offices). I have submitted PR #241 which attempts to strip out the password if it’s found in the URI.

Use Git Commit ID Plugin 2.2.1 or higher

Update: as of 3/26/2016, version 2.2.1 of the Git Commit ID Plugin 2.2.1 plugin was released which includes PR #241. And that PR fixes the issue altogether.

Advertisement

Adobe XMP Packet Extraction for the Aperture Framework

When it comes to manipulating photographs, I live in Photoshop. One feature of all Adobe products that I like is the ability to annotate images and other documents using their eXtensible Metadata Platform, or XMP. XMP is a collection of RDF statements that get embedded into a document that describe many facets of the document. I’ve always wanted to be able to somehow get that data out of these files and doing something with it for application purposes.

There are projects like Jempbox, which work on manipulating the XMP data but offers no facilities to extract the XMP packet from image files. The Apache XML Graphics Commons is more the ticket I was looking for. The library includes and XMP parser that performs by scanning a files for the XMP header. The approach works quite well and supports pretty much every format supported by the XMP specification. The downside of XML Graphics Commons is that it doesn’t property read all of the RDF statements. Some of the data is skipped or missed completely. To top it off, neither framework allows you to get at the raw RDF data.

What I really wanted to do was to get the XMP packet in its entirety and load it into a triples store like Sesame or Virtuoso. This of course means that you want to have the data available as RDF. Rather than inventing my own framework to do all of this, I found the Aperture Framework. Aperture is simply amazing framework that can extract RDF statements from just about anything. Of course, the one thing that is missing is XMP support. So, I set out on implementing my own Extractor that can suck out the entire XMP packet as RDF. It’s based on the work started in the XML Graphics Commons project, but modified significantly so that it pulls out the RDF data. Once extracted, it’s very easy to store the statements into a triple store and execute SPARQL queries on it.

Right now the, this  XMPExtractor can read XMP from the following formats:

  • JPEG Images (image/jpeg)
  • TIFF Images (image/tiff)
  • Adobe DNG (image/x-adobe-dng)
  • Portable Network Graphic (image/png)
  • PDF (application/pdf)
  • EPS, Postscipt, and Adobe Illustrator files (application/postscript)
  • Quicktime (video/quicktime)
  • AVI (video/x-msvideo)
  • MPEG-4 (video/mp4)
  • MPEG-2 (video/mpeg)
  • MP3 (audio/mpeg)
  • WAV Audio (audio/x-wav)

On the downside, I’ve found that if you use the XMPExtractor with a Crawler, you’ll run into some problems with Adobe Illustrator files. The problem is that the PDFExtractor mistakes these files for PDFs and then fails. But as long as you’re not using Illustrator files, you should be ok. There’s also a few nitpicks with JPEG files and the JpgExtractor in that the sample files included in the XMP SDK are flagged as invalid JPEG files. However, every JPEG file I created from Photoshop and iPhoto seem to work fine. But after a little more testing, I’ll look at offering it up as a contribution to the project.

Eclipse on Mac Java 6 Reveals More SWT Shortcomings

Two years ago, I raised a few points about some of the short comings of SWT. Because of it’s native bindings, SWT makes the Java mantra of “write once, run anywhere” quite a bit more daunting. For the most part, SWT’s cross-platform support is actually quite good and it is a decent in terms of performance. And, if weren’t for SWT’s existence, we probably wouldn’t have seen Sun address Swing’s performance issues like they did in Java 6. Unfortunately, when a minority platform like OS X makes some steep architectural changes, SWT-based applications end up with more work on thier hands.

As most folks know, Java 6 on Mac OS X 10.5 was a long time coming. It took Apple over a full year after the initial release of Java 6 to get it running on Mac OS X. Now that it’s here and working pretty much “ok”, I decided it was time to start running Java 6 as my default JVM. Then the surprise: Eclipse won’t run under Java 6 on the Mac. Why? Because Java 6 under Leopard is 64-bit. The current version of SWT on OS X relies on Carbon, which is 32-bit and we won’t be seeing 64-bit Carbon anytime soon. Support for 32-bit Cocoa is planned for later next year, but I didn’t see word on when 64-bit Cocoa or even just Java 6 support might arrive.

Eclipse is still a great IDE even if I have to continue to run it under Java 5. However, this is one of those things that is annoying each time a platform needs to make significant changes. But this time, you can’t put all of the blame on the Eclipse crew. Apple did an absolutely terrible job keeping the Java community abreast of what thier plans were with Java 6. In fact, it almost seemed that Java 6 would never appear on Leaopard. Coupled with the fact that Java 6 was now going to be only 64-bit and Carbon was not goning to see 64-bit support. But long story short, SWT and therefore Eclipse is always going to be hindered by OS changes to a greater degree than say NetBeans or IDEA.