I am considering switching a Maven project that I manage to Apache-Ant/Ivy. I need more control over the build process and am getting very frustrated with Maven. Please no comments about how great Maven is. My question is about Ivy.
I would like to set up a "standard" Ant build template that can later be used for other projects with minimal changes.
I will set up a central "enterprise" repository where we can place third-party libraries that are not available in the public Maven repositories (e.g. commercial libraries, Sun libraries, proprietary libraries, etc.). This enterprise repository will be available on our local LAN, but may not be available from outside the office.
Each developer will have a private repository in ~/.ivy/repository. I would like the Ant build to automatically update this private repository with changed versions of libraries from the enterprise repository.
In ~/.ivy/ant, I plan on placing "standard" modules for including in the individual project build.xml files, using the include task in Ant 1.8. These modules will provide things like Scala and Clojure compilation targets with different versions for different Scala and Clojure versions (e.g.: scala-compile-2.9.1.xml, clojure-compile-1.3.xml, etc.) The build modules will be available in the enterprise repository and should be updated automatically in the private repositories if they change.
Each project will follow a standard Maven directory structure: ${project}/src/main/java, ${project}/target/classes, etc.
In the past, I tried using Ivy but the Ant build files got to be very large (> 500 lines for the template build file) and hard to manage/edit. I am hoping that by putting standard targets in their own build modules in the ~/.ivy/ant directory, I can avoid that code bloat.
Can this be done? Am I way off base? The only documentation I can find on Ivy is at the Apache web site (http://ant.apache.org/ivy). Is there any other documentation available, including books?
Rather sensible idea about dividing template build file into includable helper files. Personally, now i'm switchin' a really large project from ant (no dependency managment at all - only copying files from ftp) to ant/ivy solution. So i've done this way - i have a file with milestones targets - i.e ready-to-compile, compiled, ready-to-archiving, archived - so on. I think u got the idea. I've configured dependencies of all this targets ( dependencies in terms of ant, do't get me wrong). In that way - compiled depends on ready-to-compile, ready-to-compilede depends on initialized - smth like this. This targets don't have body - they are for including in every build-file of every module of your multi-module project. The sole purpose of this targets for maintaining the STATE of build, because of this import stuff things become rather tricky and it's hard time to know what target was overriden, and when this target would be run. But with this file i can easily change state of vy build on every sensible milestone. I want in one module to compile help files with exteran exe. No problem - in this project i just do this - ready-to-archiving depends on the target for compiling help. And as this milestones targets are included - i can override only some of them - all others would presere the desired way of building project.
Another part of my strategy - mixins build files - for every specific area. So i have a file for ivy. There i put initializing, resolving, publishing and so on. When i want to use ivy - i just include this file and manage depdendencies through my milestones targets. If the build is typical - i only include this file and have a convention-over-configuration functionality. All out of the box. How?? Just combining with other mixins. Mixins may include other mixins to depend on them. So each mixin is a reusable part of my build strategy. The stuff from OOP - single-concerned unit. In your case it's scala mixin with targets specific for scala stuff.
Then i have delegate.xml that delegates child projects common build activities. I have dist, all, test and whatever u want for multimodule project. The build order is evaluated with ant-ivy task buildlist.
There also some other files - but this are the strategically basic files that helped me to have a reusable and maintanable build with this BIG and VERY Conservative project. So, if u are interested about details, don't be shy and contact me. I will be very pleased to help you, because ivy docs are really comlicated and incomplete.
EDIT: About books - Ant in Action may help you, i took several ideas from this book, and i really highly recommend it everyone to read. There u can find ivy stuff, also. And about ivy docs - sorry, it's all that is available. But when i was in trouble with this cumbersome ivy+ant - i found several interesting articles on private blogs. So ... that may fill the gap some way.
Related
So, at work, I frequently have to create virtually identical ant scripts. Basically the application we provide to our clients is designed to be easily extensible, and we offer a service of designing and creating custom modules for it. Because of the complexity of our application, with lots of cross dependencies, I tend to develop the module within our core dev environment, compile it using IntelliJ, and then run a basic ant script that does the following tasks:
1) Clean build directory
2) Create build directory and directory hierarchy based on package paths.
3) Copy class files (and source files to a separate sources directory).
4) Jar it up.
The thing is, to do this I need to go through the script line by line and change a bunch of property names, so it works for the new use case. I also save all the scripts in case I need to go back to them.
This isn't the worst thing in the world, but I'm always looking for a better way to do things. Hence my idea:
For each specific implementation I would provide an ant script (or other file) of just properties. Key-value pairs, which would have specific prefixes for each key based on what it's used for. I would then want my ant script to run the various tasks, executing each one for the key-value pairs that are appropriate.
For example, copying the class files. I would have a property with a name like "classFile.filePath". I would want the script to call the task for every property it detects that starts with "classFile...".
Honestly, from my current research so far, I'm not confident that this is possible. But... I'm super stubborn, and always looking for new creative options. So, what options do I have? Or are there none?
It's possible to dynamically generate ANT scripts, for example the following does this using an XML input file:
Use pure Ant to search if list of files exists and take action based on condition
Personally I would always try and avoid this level of complexity. Ant is not a programming language.
Looking at what you're trying to achieve it does appear you could benefit from packaging your dependencies as jars and using a Maven repository manager like Nexus or Artifactory for storage. This would simplify each sub-project build. When building projects that depend on these published libraries you can use a dependency management tool like Apache ivy to download them.
Hope that helps your question is fairly broad.
Being a Maven newbie, I want to know if its possible to use multiple classifiers at once; in my case it would be for generating different jars in a single run. I use this command to build my project:
mvn -Dclassifier=bootstrap package
Logically I would think that this is possible:
mvn -Dclassifier=bootstrap,api package
I am using Maven 3.0.4
Your project seems like a candidate for refactoring into a couple of what Maven calls "modules". This involves splitting the code into separate projects within a single directory tree, where the topmost level is normally a parent or aggregator POM with <packaging>pom</packaging> and a <modules/> list containing the sub-project directory names.
Then, I'd advise putting the API interfaces/exceptions/whatnot into an api/ subdirectory with its own pom.xml, and putting the bootstrap classes into a bootstrap/ subdirectory with its own pom.xml. The top-level pom.xml would then list the modules like this:
<modules>
<module>api</module>
<module>bootstrap</module>
</module>
Once you've refactored the project, you will probably want to add a dependency from the bootstrap module to the api module, since I'm guessing the bootstrap will depend on interfaces/etc. from the api.
Now, you should be able to go into the top level of the directory structure and simply call:
mvn clean install
This approach is good because it forces you to think about how different use cases are supported in your code, and it makes dependency cycles between classes harder to miss.
If you want an example to follow, have a look at one of my github projects: Aprox.
NOTE: If you have many modules dependent on the api module, you might want to list it in the top-level pom.xml in the <dependencyManagement/> section, so you can leave off the version in submodule dependency declarations (see Introduction to the Dependency Mechanism).
UPDATE: Legacy Considerations
If you can't refactor the codebase for legacy reasons, etc. then you basically have two options:
Construct a series of pom.xml files in an empty multimodule structure, and use the build-helper-maven-plugin along with source includes/excludes to fragment the codebase and allocate the classes to different modules out of a single source tree.
Maybe use a plugin like the assembly plugin to carve up the target/classes directory (${project.build.directory}) and allocate classes to the different jars. In this scenario, each assembly descriptor requires an <id/> and by default this value becomes the classifier for that assembly jar. Under this plan, the "main" jar output will still be the monolithic one created by the Maven build. If you don't want this, you can use a separate execution of the assembly plugin, and in the configuration use <appendAssemblyId>false</appendAssemblyId>. If the output of that assembly is a jar, then it will effectively replace the old output from the jar plugin. If you decide to pursue this approach, you might want to read the assembly plugin documents to get as much exposure to different examples as you can.
Also, I should note that in both cases you would be stuck with manipulating the list of things produced by using a set of profiles in the pom in order to turn on/off different parts of the build. I'd highly recommend making the default, un-qualified build one that produces everything. This makes it more likely for things like the release plugin to catch everything you want to release, and up-rev versions, etc. appropriately.
These solutions are usually what I promote as migration steps when you can't refactor the codebase all at once. They are especially useful when migrating from something like an Ant build that produces multiple jars out of a single source tree.
I've been developing with Orchard CMS for a few months now (and love it) and the time for launching my website is fast approaching. So far I've just been developing solo out of my BitBucket repo, forking where necessary, not doing anything too fancy. Once I have released though I really need to have a handle on exactly what versions of my modules and themes are in use in production. I figure versioned packages out of my build server is the best way to achieve this.
Currently my repo consists of the source for the entire Orchard instance (minus the App_Data folder), with a solution file that includes the projects that are my modules and themes. My modules take on dependencies of other modules from the App_Data/Dependencies folder.
My question is, is this the best approach to achieve Continuous Integration?
I have my solution building under TeamCity, but I don't include the App_Data folder in my repo, so I need to at least load the setup page so the Dependencies directory gets populated (which doesn't happen on my build server for some reason? Seems only dynamic compilation kicks in for everything?)
Any thoughts or assistance would be greatly appreciated.
UPDATE:
I have decided I will add a lib folder to my solution and store all dependent assemblies there. I will then have my repo consist of only the projects required for my modules / themes. The CI server will then have no problems building the solution, and I can just clone the repo into an Orchard instance for easy development (this means my solution will have to contain a Modules and Themes directory).
I used the following approach.
add Orchard binaries to the repository without any source codes, just in the for they are distributed at the orchard project site
create my solution and all related projects in separate directory, so at the moment dir structure looks like this:
then place your modules projects under orchard/modules folder with all sources and .proj files
add references from your module projects to orchard/bin for any
orchard-specific stuff
add module binaries manually to App_Data/Dependencies folder to be able to reference them
One of the improvements of this approach is to turn off dynamic compilation and store only module binaries, but this will require configuration of the output bin path and additional actions in the build script.
Benefits
You don't have any orchard sources in your repository except of the
modules (but this is solvable by turning off dynamic compilation).
You can easily upgrade orchard binaries and modules almost independently
The build takes less time
TFS 2008 and Common Libraries
I have created a Team Project called "Common Library" that will host code used in numerous different Team Projects throughout TFS. For sake of argument, lets say we have 2 distinct Librarys under the "Common Library" Team Projects, MailProject and LoggingProject. Other projects throughout TFS will be using the binary representation of these projects via branching and not the actual source code.
What is the best way to set up the folder structure for this Team Project? Do I add the project to the "Common Library" and simply "include" the bin/release folder as part of the project?
I have seen some examples of people creating a seperate "Deploy" folder. I assume this is synonamous with the bin/release folder?
We do not want the source code available in other solutions.
Currently, each project has the dll included in the project. Using a mailing module as an example, many projects need the ability to mail. The common module is very stable and mostly static.
However, what if there is a change in the mail module. It seems there would be a better way, than to check out each project and update the dll. Is it possible to allow TFS to grab the latest mail module any time a 'get latest' is called? Either explicitly or implicitly.
Unless you really require the source code for the libraries to be available in the other solutions my advice would be to include the binaries for the libraries in the projects that would use them not really having any explicit link between the two in TFS. Custom labelling of the library builds could be helpful to easily return and rebuild any chosen version of the shared libraries.
If the shared libraries require different versions for different projects then the obvious solution would to create a separate branch for every version of the libraries that need to be customised to a particular project.
TFS does not have a concept similar to SVN's 'externals' though - so if you include a branch from the shared libraries in a project and than branch that project it is very difficult to propagate changes correctly.
I suppose you could also use the Get task in the build and get the latest version of DDLs into the current project from another one, but verify if you can point of Workspace of another project (I have not tired it and MSDN is somewhat vague here). You might need to have a separate workspace for the shared project.
Yet another alternative would be to publish the DLLs for common components to a known location on every build of the shared libs and for individual builds to get whatever version is available from that common location (network share) even via the Copy task. This is simplistic and may cause problems with versioning of the common components but should work well enough in simple case.
For those of you that use Ant with multiple projects, where do you put the build.xml files? Do you put one in each project, or do you put them in a separate project that contains all your Ant-related files?
The usual recommendation is to put a build.xml in each project. But this has a few drawbacks:
It makes it hard to reuse common targets in multiple projects.
Sometimes you want to use Ant to export a project from source control and deploy it. Obviously you can't do this if the build file is in the project itself.
But if you put them all in a common location:
People need to be aware of their location to use them; they can't just use "ant -find" to find the current project's file.
You can't have different build instructions for different branches of the project.
What do you guys do?
EDIT: Thanks for the good suggestions so far. As far Maven, these aren't Java projects, and I get the impression that Maven is only meant for Java.
Place the Ant files with the project. That is the de facto standard and recommended by the creator of Ant. I will try to address some of the issues you have brought up:
Reuse of common targets should be done using techniques as described by Eric Hatcher in his book Java Development with Ant. Basically, you extract all commonality into a some top level files that all other Ant files "inherit" from.
Using Ant to export a project from source control seems odd to me, but if you want to do this, use a different Ant file :-) You can make a target like ant export -Dproject=foo/bar.
For Ant, I recommend you grab that book - it has a ton of helpful techniques.
The real recommendation I would make though is to drop Ant and convert over to Maven - like the Apache Software Foundation did (they maintain both Ant and Maven).
If you're working with independent projects, you can:
put your build.xml at the top level
place common Ant definitions (Antlib) into another project (e.g. config)
use svn:externals to import the common Antlib definition (from 'config') into your project
EDIT The trick with svn:externals is that if you link to the HEAD of some common files, it may happen that they will change after a couple of months/years. So each time you tag, you should change the svn:externals to point to a fix version of the included project. This may come handy when a project has to be rebuild years after it was last built.
My rule of thumb is to put the build.xml file in the directory under which all files are referenced. In other words, no relative paths should start with "../". Where I live, that usually means putting it in the "trunk" directory, which has src, lib, build, docs, etc underneath it.
Doing this makes the paths much cleaner in the file, and it makes it obvious how to build the project.
Where I have multiple projects that need to build, I will create a separate build.xml for each project, and a central build.xml in the directory all the project are in that calls those other build.xml files. That gives you a lot of flexibility with very little work.
I'd expect an Ant build file to be located at the top of a project (it's already a pain to have to look at a the build file to "discover" how to build the project, so if I have to locate it first, it'll drive me totally crazy). Now, regarding all the drawbacks you mentioned, I'm tempted to say: why don't you use Maven?
The way I have done this is in the past (Now I just use Maven):
Have a build.xml in the root of each project
Create an overarching build.xml
for all projects and place it in
the trunk of my repository
The overarching buid.xml has
checkout tasks for each project.
I am guessing when you mentioned
export from repository, you
actually meant import.
The overarching build file also
defines the dependencies, if any
You may update individual projects using each project's individual build file
If you do have common tasks defined, you may inherit from a common build file as well as someone else suggested.
Looks like your set of projects might be a good candidate for migration to Maven, I realize it is not always possible but if you have time, you might want to look into it.