I need to turn Eclipse fragments into plug-ins. Haven't found any wizard to support this.
So I've created a plugin.xml and edited the MANIFEST.MF to look exactly like a plug-in project does. Still, Eclipse somehow detects it's not a plug-in: for example it asks for a Host-plugin.
Also checked .project file, nothing seems to indicate "fragmentness".
Why I need this:
Fragments were used for Unit-testing; but they're rather inconvenient: any modification in them triggers a chain of builds, and they block the exporting of individual plug-ins, unless their project is closed.
Solution
As Greg suggested it was the 'Fragment-Host:' option that did the trick. And as Rüdiger wrote; the project had to be closed/reopened for eclipse to acknowledge the change.
The fragment host is specified in the fragment's MANIFEST.MF:
Fragment-Host: host.plug.in
you will need to remove that.
However many fragments rely on code from their host plugin so it may not be possible to just convert the fragment.
Related
I am an Eclipse/STS user/developer, now trying to use IntelliJ Idea (CE)
2020.2.(1,2,3)
For a project based on Gradle, how spring-integration, when I open the IDE it happens the following
Ok, let the IDE load the project ... but
From above, that is the problem, I don't want that the IDE starts automatically to build/rebuild the project. I just need, open the project and that's all.
Observation: for example in Eclipse/STS exists the option to disable Build Automatically
I did do a research in the Web and I read the following posts and questions:
How to disable automatic gradle builds?
IntelliJ IDEA “Build project automatically” apparently not working
Intellij IDEA Java classes not auto compiling on save
Sadly the dialog options were changed but ...
Therefore:
From above, seems nothing to do.
Observation: from above observe the Build project automatically option is disabled
Even with that disabled and after to restart the IDE, I must always stop manually the build process
So what is missing? or Do I need a special extra plugin to accomplish my goal?
The images that you show indicate that you are building with Gradle, but the Compiler option that you disable is relevant for building projects with Idea not with Gradle.
For the 2020.2 version, you need to do the following:
Open the Setting > Build Tools page.
Disable the "Reload changes in build scripts" option.
This way you can manually control the reload. When you change the build script, you will see a small gradle icon in the right side of the editor.
For more info, refer to the IntelliJ IDEA help > Gradle section.
https://www.jetbrains.com/help/idea/work-with-gradle-projects.html#auto_reload
There are two different things in IntelliJ's Gradle support that sometimes confused: sync and build. Your pictures demonstrate sync process (note caption on the toolwindow). Word build is kind of misleading here.
What is sync? In gradle we use Groovy to define the build procedure. Groovy is an imperative programming language, so it's hard to predict resulting dependencies graph without actually executing the script. During the sync Idea executes configuration phase of gradle build (one that builds dependency graph), and obtains configured objects from the Gradle daemon. This information is used to setup project in the IDE: modules, libraries, dependencies, which sources are test, which are prod, etc.
Actual build is not happening during sync. You can convince yourself by adding syntax error to any source file, and observe that the sync succeeds. But build will fail if you invoke it.
In answer to the original question: you can't disable automatic build, because it is not enabled.
Is it possible to disable sync in Gradle project? Short answer - no. If you need a code browser, which is not required to understand all the cross-references in the source code, IDEA is not the best choice probably.
TL;DR;
Without sync IDE does not know which files are sources, and which are not. IDEA cannot open folders. It only can open projects. Good thing is that module can contain folder. So you can do the following: File | New | Project. Select Empty project, Next, select some random folder outside the source folder you want to open, Finish.
Then add new module:
Select Java in the left panel, everything else keep default, Next, Finish. Then in new module remove existing content root, and add folder with sources as new content root
Resulting project is mostly useless. Tons of red code (at least, unresolved symbols from external libraries), no inspections, no navigation, no sense. But it might be useful in some rare situations indeed.
I have a solution that is using an hybrid .csproj and project.json combination (for nuget management purposes). So basically the "project.json" file is working as a "packages.config" file with a floating version capability.
This solution is using a custom RuleSet that is being distributed via Package, and is imported automatically. On the dev machine, works without a problem.
At the build machine (that is, inside the machine itself, working as an user) the solution also compiles without a problem.
However, when a vNext build (is this the name for the new build system?) is queued, it ignores completely the custom ruleset and just uses the StyleCop one (that is also included), which gives a bunch of warnings. Said warnings should not appear as the Custom RuleSet basically suppresses those warnings (ie: Warning SA1404: Code analysis suppression must have justification,
Warning SA1124: Do not use regions, etc)
As far as I have checked, there is no setting to specify the ruleset, and this works with XAML Builds. What is different in this new build system that is causing this? Is there a way to force/specify the Code Analysis Rule Set from the definition?
Thanks in advance for any help or advice on the matter.
Update/Edit
After debugging back and forth with the wonderful help of jessehouwing I must include the following detail on my initial report (that I ignored as I did not know that it was influential):
I am using SonarQube Analysis on my build definition.
I initially did not mention it as I did not know that it replaces the Code Analysis at Build Time (and not only when it "analyzes", as I thought).
If you are using the SonarQube tasks
The SonarQube tasks generate a new Code Analysis Ruleset file on the fly and will overwrite the one configured for the projects. These rulesets will be used regardless of what you've previously specified.
There is a trick to the naming of the rulesets through which you can include your own overrides.
More information on the structure can be found in the blog post from the SonarQube/Visual Studio team. Basically when you Bind your solution to SonarQube it will generate 2 ruleset files. One which will be overwritten during build, the other containing your customizations.
There is a toolkit/SDK to generate a SonarQube plugin for custom analyzers which allow you to import your rules into SonarQube, so it will know what rules to activate for your project(s).
If you're not using SonarQube
Yes you can specify the ruleset you want to use and force Code Analysis to run. It requires a couple of MsBuild arguments:
/p:RunCodeAnalysis=true /p:CodeAnalysisRuleset="PathToRuleset"
Or you can use my MsBuild helper extension to configure these settings with the help of a UI template:
I have a main plugin project which does not depend on any of Eclipse APIs. But I do want to use Eclipse API's in one of its fragment plug-in. Will it cause any problem for the main plugin?
If you add a dependency (to a plug-in or a package) in a fragment, then you effectively modify the class path of the host project as well. Whether that will cause any changes to the semantics of the host project or any other fragment for the same host project, depends on the specific use in the project.
Having said all that, the normal answer is: no, it should not cause problems, unless you have code that depends on the class path - e.g. if you are using Class.forName(...) or similar...
One final note: when you test this, use -clean argument in the launch configuration to force OSGi to accept the changed dependency. Otherwise, it will be ignored.
I'm happily using the Maven bundle-plugin to create OSGi manifest headers for my modules. However, when there are configuration files that pull in classes which aren't referenced directly in the code, the plugin can't tell which packages it's going to need.
One example is a bundle with domain models that constitute a Persistence Unit for JPA. The driver class is part of the PU configuration and either set in an XML file or at runtime when the EntityManager is instantiated. I have to manually add an Import-Package header for the driver class that I want to load, or I get CNF errors.
Another example is a Struts war, where the web.xml pulls in the Struts dispatcher that's otherwise not found anywhere in the code and has to be manually added to the headers.
How can I avoid this?
I tried adding the required packages as dependencies with a provided scope, but that didn't help.
In the plug-in section of the bnd configuration you can specify plug-ins to analyze these files and contribute to the import-package header. For spring it looks like this:
<_plugin>aQute.lib.spring.SpringComponent</_plugin>
I am not sure, what descriptors are supported on top of spring. Just take a look at the source (it's in the Apache Felix SVN) and see for yourself. In the worst case you have to write your own plug-in, but at least it is possible! Also peter kriens site about the bnd explains the usage and some internals.
Other then that I am not aware of any simple solution.
I currently have an existing plugin, which references a class from a required plugin. I have replaced this code with a reference to a class which is part of my fragment.
I am facing two issues.
If I import my fragment as a jar file, I am not able to see the changes I have made as the plugin running as an eclipse application results in a ClassNotFoundException
To overcome this, I link an additional source (of fragment) to the existing plugin project. However, the link uses an absolute path, and makes it unfit for deployment.
I want to be able to package the plugin with the code modification and be able to "depend" on my fragment code. Is there a way I can add my fragment as a dependency?
For example:
Plugin Project I am changing : org.eclipse.*.editor
it depends on org.eclipse.*.edit
I have a fragment mydomain.*.edit which has org.eclipse.*.edit as host plugin
I want org.eclipse.*.editor to pick up mydomain.*.edit
instead of org.eclipse.*.edit
ps: I have also tried packaging the jar file for the mydomain.*.edit in the plugins directory and try and pick it up from there, it doesnt show up on the list when I click add required plugins on the dependency tab on the plugin.xml file of the org.eclipse.*.editor
Please let me know if I am not clear enough, I will try and rephrase it.
Thanks in advance!
If I understand correctly what you want to do, I don't think that it's possible. You will have to try some other way.
Plugins have dependencies on other plugins. Fragments don't exist as separate runtime entities, but only as extensions of a plugin. So your plugin can only refer to the 'editor' plugin.
Classes provided by a fragment can't (and shouldn't) be accessed directly. They can be returned by the original plugin (A) if they are implementing an executable extension provided by plugin A.
If you refer to the fragment's code from another plugin (B), the classes will be loaded by plugin B's classloader and be different from the ones that are loaded by plugin A.
What is the purpose of your fragment? Do you want to get access to internal code in plugin A? Do you want to extend an eclipse editor?
If you want to extend functionality that the original plugin is not exposing as extensible, I think the only way is to write a plugin, extend the editor class from the original plugin, register it alongside the original one and use it instead.
[Edit] Maybe this link will explain better: Eclipse FAQ
Hope this helps,
Vlad
Thanks Vlad,
Your explanation was very helpful. Unlike the extension based architecture that is truly intended for fragments, I had to modify a certain component in the editor that was not exposed as part of the extension. This modification referred to an external project I created as an fragment but could have been a normal java project packaged a jar file that I could place in the classpath of the editor.
I was able to resolve the dependency issues by placing the jar file in class path, however when I export the plugins and related plugins as jar files and place it in the dropin directory, it does not install correctly. (Nor does placing the jar files in the plugins directory)
The eclipse editor that I am trying to modify uses the EMF project. I have kept the EMF project in the workspace inorder to resolve dependencies of the editor. However when I replace the EMF jar files bundled with eclipse with the one in the workspace, the files that I want to edit are not correctly recognized.
Is there another way of doing this?