I am empirically testing OSGi Bundles and their relationships for this I need lots of bundles.
Making these datasets is a difficult task. I already have Eclipse update (1700 Bundles) sites and Spring Enterprise bundle repository for testing, however I want more, anyone out there got massive amounts of bundles.
I don't even need the code, just the manifests would be fine.
Cheers
Google Code Search returns 15800 results for query 'Bundle-SymbolicName file:MANIFEST.MF'. If you find some way to automatically download them and remove duplicates you might get a sizable dataset.
Another similar idea is to find some way to search the Maven central repository for artifacts that include OSGi bundle manifest.
Related
I am working with 10.5.4. I would like to use SVN to keep versioning of my site contents so as editors to be able to commit their changes. I have read a lot about mounting the VFS and that stuff, however with that way everyone coulld access and commit others changes
I would appreciate any further information about using versioning not only for module developments but also for site contents
I know, this is probably not exactly what you are looking for, but you may have a look at the OpenCms Maven plugin.
The main focus of this plugin is, to ease the development process of OpenCms projects. But with a good setup and Maven configuration you might use it to put the contents of your production site under version control. If you are interested, you'll find more information on the official website of the plugin.
I am currently working in a WEB Java project and I have the following requirement: I have to make sure our customers don't have access to the packaged files in order to change them. Actually, I'd like to do that with only a few of those classes but without using any obfuscators due to the size of my project and the resources it uses to accomplish some tasks (reflection, annotations, interceptors, etc).
Does anybody have any tips?
Thanks,
Luan
I deployed a portlet in liferay 7 and it got deployed successfully and was available for use. I want to replace the jsp file, in earlier version I could see my application in tomcat/webapps folder and replace it quickly.
Now I am unable to locate the exploded war in liferay 7. I can only see the war in osgi/war folder.
Can someone help me with that.
Thanks in advance.
While I mostly agree with what Olaf wrote, I do understand the need to be able to make changes in JSP files and try them quickly during development. I'm afraid I don't have the solution for that yet.
However, let me answer the question you asked:
where is war exploded in Liferay 7 tomcat after getting copied in osgi folder
It is NOT (at least not the way it was done by application servers)! When you deploy a WAR file in Liferay 7, it will automatically (on the fly) convert it into OSGi bundle and install it in OSGi runtime. This way now Liferay is fully in charge of deploying plugins and does not need to rely on various application servers.
PLEASE NOTE: Every bundle has it's own state folder. In Liferay those are in <LIFERAY_HOME>/osgi/state. If you know the bundle ID you can easily find it. It may be (I haven't checked) that you'll find some JSP files there. The reason I'm writing this is to warn you (in case you figured it yourself) to NEVER modify bundle's state folder manually. Doing so may brake the whole environment. In worse case scenario you may have to redeploy everything in clean environment.
You should not rely on behavior like this. In previous versions it was the task of the application server to compile changed JSPs at runtime. However, this is bad practice in production systems and totally screws up your maintainability. If you need to update some UI code frequently, I'm suggesting you change your implementation to utilize ADT (Application Display Templates), e.g. through Freemarker or Velocity. Those are meant to be updated at runtime, where the JSP updates were a side effect of Tomcat's default (development friendly, production hostile) configuration
I am looking for a solution to implement security-scanning of the application code-base at the time of a build. The idea is to capture a list of security vulnerabilities early in the software development life cycle.
I have a simple java project which uses a maven build. The java project specifies a number of .jar dependencies and comes up with a .war file as a build output.
I came across (and was able to configure) the dependency-check maven plugin (http://jeremylong.github.io/DependencyCheck/dependency-check-maven/index.html). However, though it scans the dependency jars and comes up with a vulnerability report, it doesn't seem to scan the final artifact - which in my case is the .war file.
How do I ensure that the .war is scanned as well? Is the dependency-check plugin the right tool for this?
dependency-check isn't the right tool for checking your own code. It uses a list of known vulnerability reports to determine if any of your dependancies have known flaws. It does not do an active scan of the code. see Plugin wiki
For checking your own code, HP's Fortify is a decent commercial solution, but if you are working in more of a DIY software setting, I would recommend Sonar. There are certainly many static code analysis tools out there. All have advantages and disadvantages.
I'd like to ask you for best practices with developing with Liferay SDK.
I have the SDK downloaded, I have Eclipse ready, it works, I can create new portlets and run local Liferay instance to test it.
Here is my situation - all the source code I have is in the Eclipse workspace, currently it is only portlets what I'm working on.
Liferay SDK I have in completely different location than workspace. Let's say ~/dev/liferay_sdk.
Eclipse workspace is located in ~/workspace.
At the beggining, it was not working like that. Eclipse from some reason can't find or use Liferay SDK. When I changed "Project validation" in Eclipse/Liferay configuration to "Ignore" the "Liferay Plugin SDK is not valid", it started to work without problems.
Next problem happend when it comes to need to build a WAR for example.
In the portlet directory in the workspace is present "build.xml" file. But inside it refers to another xml file, which should be located one directory up, and this one refers to more thing in relatively location and so on.
In short, it assumes that you have the portlets etc, inside the Liferay SDK.
Like "~/dev/liferay_sdk/portlets".
My question is, Am I wrong completely, or could you suggest me the best practices with this?
I don't want to mix SDK and the code, it sounds wrong to me.
Thanks for help!
I think, the best practice is still when your portlet projects are located inside the Liferay Plugins SDK directory. That way you can take all the advantages of the Liferay IDE plugin for Eclipse, for example. Because as far as I understand Liferay IDE will not allowed you to have portlet projects in another location. It's pretty easy to import projects to Eclipse from inside the Liferay SDK directory, and that's not problem.
But I also faced the same sort of problem when tried to save portlet project to the Git repository. Possible solutions with symbolic links didn't work correctly on every system. Thus I slightly modified the build.xml file to be able to run ant tasks from any directory. For portlets it was something like that:
<project name="your-portlet" basedir="." default="deploy">
<property file="build.properties" />
<property name="project.dir" value="${liferay.sdk.home}" />
<import file="${project.dir}/build-common-plugin.xml" />
</project>
Notice that you should define property "liferay.sdk.home" in build.properties and it should be path to the Liferay Plugins SDK.
As for other types of Liferay plugins (themes, hooks, etc.) you should import another build file for building that type of plugin. For example, for themes it will be:
<import file="${project.dir}/themes/build-common-theme.xml" />
Hope you'll get the idea. :) But think twice before doing something like that.
Liferay plugins are developed inside the Liferay Plugins SDK, its called SDK for a very good reason.
I don't find anything wrong with the plugins-SDK and the code tied togather, below are few reasons why:
If you see the liferay repository of plugins on github, you would find all the sample portlets and other plugins are stored in their respective folders inside plugins-SDK.
So if you want to develop liferay plugins (with or without IDE), the best practice (the only efficient way I think) is to have the projects created inside the respective folders of plugins SDK like portlet projects inside portlets folder, hook project inside hooks folder etc.
If you have used Liferay IDE when you create a plugin project (Liferay project) in this IDE you specify the SDK and the server runtime and what it does is it creates the project inside your Plugins SDK and copies the .settings, .classpath & .project file inside the project created. It does not create the project inside your workspace as eclipse normally does for other projects.
Hope I have managed explain it clearly and this was what you wanted.
I'm already quite happy with the other answers, this could have been distributed through comments at those, but a separate answer gives some more structuring options:
As Prakash says, it's not really bad to do that. In addition to his answer, you do not need to have your code in the workspace directory. Eclipse is happy to put it anywhere in the filesystem - thus while you work with Eclipse you don't even care where exactly your code is (and as you check it into version control - right? - you actually never need to care.
If you want to use Liferay's OOTB ant scripts: They are geared towards exactly the setup you describe: Work in the SDK directory. It's actually not bad, but if you don't like it, you just have to accept that you can't work with build.xml without changing it (like Artem suggests).
Another option is to use maven - this also bypasses the sdk (and the Liferay IDE integration), so you're again free to put your sourcecode whereever you like and let maven do the rest.
I can imagine some rather esoteric and rare issues with Artem's suggestion (like referring to custom parent themes when you imply some relative position) but I consider that as extremely minor, so if that works for you: Go ahead. Just keep in mind that you don't fulfill the basic assumptions that the SDK makes, so you might have to change things that violate the assumptions. I can't imagine this being too hard if you keep this in mind.
Of course, what you miss with that solution is the neat handling of including build.${username}.properties - you'll have to have your own build.properties that define ${liferay.sdk.home}. If you're not working in a team, that's ok. Otherwise you'll have to invent this yourself (and code it) or rely on global parameters to be configured with every team member.