While reading about Maven I came across the line below.
A very significant aspect of Maven is the use of repositories to
manage jar files across different projects.
But, to be frank, I completely disagree on this: Maven requires a lot of knowledge of it, a lot of time to resolve issues related to it, requires a repository, etc.
Please tell me how Maven manages Jar files. If it's the responsibility of the developer themself to specify what jars (with versions) to use, then how does Maven makes it easy to work with, or manage, Jar files?
Maven has a lot of benefits with respect to managing dependencies.
It downloads dependencies you specify from a trusted location (the
Maven Central repository) Dependencies don't need to be stored in
revision control, which saves space
It helps managing different versions of dependencies, which makes builds reliable and reproducible (specifying the same dependency in
different projects guarantees that both projects use the same
dependency)
It manages the inclusion of transitive dependencies for you (adding
a dependency to your project also adds all of that dependency's
dependencies, and their dependencies, and so on)
It manages the exclusion of conflicting versions of transitive
dependencies (if two of your dependencies rely on two different
versions of another dependency, only one version is used. You can
specify which one)
It makes your projects (local and remote) available to one another
to use as dependencies (if a module that you're developing on your
machine depends on another module you're developing on your machine,
you can declare that dependency, and Maven will include the built
version of one module when building the other)
In my opinion, there may be 2 scenarios where the overhead of Maven (yes, there is a steep learning curve compared to e.g. Ant) pays out:
Manage the versions of libraries in use
Manage the resulting libraries of your development
Manage versions of libraries
With Ant, you normally include the used libraries somewhere in a lib folder, and it depends on the good work of your architect, if the source of the libraries is known 3 or 5 years later. Maven helps managing that by providing a place for dependent libraries. There you may exactly fix the version you want to use, and the whole setup and retrieval will be done for you.
Manage resulting libraries
If you want to share your work with others in your company, a repository of shared libraries is the best thing then to have. By using Maven, you are able to deploy new versions to that repository, and others may use them then.
Caveat: You may use Ivy to do the dependency management only, if that is what you want to do.
I recently started to use Maven2 in one of my Java web application projects. Now I had many issues with it, some times project fails to build for no apparent reason and then it suddenly starts to work when nothing was done at all to project. Or some times our project members must delete project from their harddrive and download project again from SVN. There seems to be many very odd bugs in Maven in eclipse, but there some issues I would like know if it is possible to solve this issues.
1) I have understood that Maven2 should be able to get dependencies for added jars, but when I add a new dependency in Eclipse, it fails when I build it, it says dependecies are missing. How can I make maven to download those missing dependecies automatically?
2) I have Tuckey UrlRewrite Filter in use, but public repositories have only old version of this dependecy, so when I use this old version (3.1 when I need 3.2). How can I include this to project? We have many programmers in this project, so setting up local repository would mean that all our programmers would have to install that local repository.
Now I had many issues with it, some times project fails to build for no apparent reason and then it suddenly starts to work when nothing was done at all to project. (...)
Ok and what is the point of this free rant? I use Maven and my builds are 100% reproducible, there are well known practices to follow to achieve this. Maybe you're just not following them. Anyway if you're not happy with it, what can I say, don't use it.
I have understood that Maven2 should be able to get dependencies for added jars, but when I add a new dependency in Eclipse, it fails when I build it, it says dependencies are missing. How can I make maven to download those missing dependencies automatically?
I think you misunderstood, Eclipse won't guess what Maven coordinates to add if you don't provide the required informations for them. Dependencies must be declared in the POM, either by editing the POM manually or by using m2eclipse wizards.
And if this is what you did (and if I misunderstood the question) then please provide the <dependency> declaration and the exact error trace.
I have Tuckey UrlRewrite Filter in use, but public repositories have only old version of this dependecy, so when I use this old version (3.1 when I need 3.2). How can I include this to project? We have many programmers in this project, so setting up local repository would mean that all our programmers would have to install that local repository.
This question has already been asked several times, see for example Maven, how to add additional libs not available in repo where I suggest two possible solutions (use a corporate repository like Nexus or a "file-based" repository, the former suggestion being the preferred one for a long term solution).
Is there a way to tell maven (when doing mvn package, mvn site or ...) not to resolve the dependencies from the local repository?
Background of this question: Sometimes I get into problems, when previously cached dependencies (e.g. SomeProject-0.7-ALPHA) are no longer available in the remote repository. In my local build everything still works fine as the dependency has been cached before. As soon as I share my pom with others, they may get into trouble, as they dont have a cached version of that dependency and the dependency can no longer be resolved from the remote repository.
Any help will be appreciated. Thanks in advance!
Is there a way to tell maven (when doing mvn package, mvn site or ...) not to resolve the dependencies from the local repository?
No, that's how the whole dependency resolution works (via the local repository).
Sometimes I get into problems, when previously cached dependencies (e.g. SomeProject-0.7-ALPHA) are no longer available in the remote repository.
Sorry for stating the obvious but removing dependencies from a remote repository is an horrible practice and leads to... well the kind of troubles you're facing. If possible, avoid doing that.
As soon as I share my pom with others, they may get into trouble, as they dont have a cached version of that dependency and the dependency can no longer be resolved from the remote repository.
One way to check that things would work for others would be to purge the dependencies of the project you're going to share from your local repository and to re-resolve them. Of course, doing this manually would be really painful but the good news is that the Maven Dependency Plugin has a purge-local-repository goal for that.
mvn org.apache.maven.plugins:maven-dependency-plugin:2.1:purge-local-repository \
-DreResolve=true \
-DactTransitively=true \
-Dverbose=true
If this fails, then you know that something is missing and can't be resolved anymore which is basically what you want to know. But you've also lost it so make a backup of your repository if this matters to you.
The basic test of a maven build is to delete the local repository .m2/repository and try to build (mvn package) and see if it is working. If not you have other problems here. As mentioned the problem seemed to be the process you are working with. If a dependency is not available anymore from a remote repository there is an other problem. Are you using a Repository Manager for example Nexus, Archiva, Artifactory ?
I want to start using maven in code I distribute but I can't find an efficient way to work with dependencies. Every new dependency takes me far too long to add.
As a simple example, I need to add Tomcat for compilation. Do I really have to manually trawl the repo in my browser to find the group-id, artifact-id and version number?
In every case it seems easier to find the non-maven downloads. I hope I've missed something obvious.
Do I really have to manually trawl the repo in my browser to find the group-id, artifact-id and version number?
No. Use a repository search engine. Some are mentioned here, more here. Some IDE also offer facilities to search and add dependencies (e.g. Eclipse with m2eclipse).
First of all, using some IDE really helps. Eclipse has m2eclipse plugin and if you define locations of remote maven repositories in your pom definition, you can add dependencies in its pom editor and it will automatically search the proper artifacts for you.
What is really helpful is using repository manager like Nexus / Artifactory, where you gather the most important maven repositories and then you don't need to declare particular repository locations in your pom definition for various artifacts, but everything will be proxied via Nexus.
I am newbie with maven.
Other than its use for managing dependencies, I am finding little use for it.
It was getting so hard to write up a pom.xml, that I generated a ant build.xml from one of maven's tasks (which is a nice handy task...)
I had to tweak the build.xml that was generated by maven. And now all my compiling, testing, etc., is being done with this build.xml..
Is such a combination common? I am thinking of making it permanent in my project.
Other than its use for managing dependencies, I am finding little use for it.
That's because you don't get it :) Dependencies management is only a small part of Maven, Maven has really much more. Quoting Maven: The Definitive Guide:
Maven is a project management tool which encompasses a project object model, a set of standards, a project lifecycle, a dependency management system, and logic for executing plugin goals at defined phases in a lifecycle. When you use Maven, you describe your project using a well-defined project object model, Maven can then apply cross-cutting logic from a set of shared (or custom) plugins.
Maven uses convention over configuration with lots of useful defaults (directory locations, a defined life-cycle, a set of common plugins that know how to build and assemble software), Maven provides a common interface to build project (unlike Ant, you know how to do things like running tests, packaging, etc with every project, no need to open the build script to find out how it's done), Maven implements reuse through maven plugins (build logic is embedded into plugins for DRY purpose, you don't have to repeat yourself over and over, you don't have to copy/paste parts of your build scripts), Maven has a Project Object Model that allows you to describe your project through meta-data (this enables dependency management, remote repositories, reuse of build logic, tool integration, artifacts search...).
So, because Maven provides a lingua franca or shared language for project management, comparing Maven vs. Ant (+ Ivy if you want), Maven vs. Buildr, Maven vs. Gradle is like comparing apples to oranges, the comparison is just irrelevant.
Is such a combination common? I am thinking of making it permanent in my project.
Well, no, that's really not the maven way of doing things. This might seem tempting (because you have the feeling that you regain control because you understand what is happening with Ant) but you are actually repeating yourself again and losing all advantages of Maven. Sure, there is some learning curve with Maven and I'm not saying you'll learn it in one night but once you'll get it, you'll feel the power. I'd thus recommend to keep trying, to ask questions on the mailing list or here on SO, to read the Maven Book, etc. But don't give up.
So you re-did what Maven gives you for free by writing Ant Tasks within a pom.xml?
Besides doing the Dep-Mgmt, Maven will compile the sources, run all test cases and package the whole thing as jar file for you with no extra configuration. That's the default.
It is not common to enrich a pom.xml with Ant clutter. However, some special tasks or legacy Ant tasks are sometimes embedded into the pom.xml lifecycle, but these are exceptions and not the common case.
What exactly is hard when writing a pom.xml?
I wonder, because most of the time you will do this once for a project and not struggle with it all the time. Also, most IDEs have support for creating the minimum pom.xml, which is just a few lines anyway.
I've tried Maven, personally for home projects and professionally at my workplace, and... I hate it. I must admit I don't have a lot of experience, but it doesn't feel good. I get the idea Maven is a less-than-perfect implementation of a good idea. I'll probably get flamed by Maven enthusiasts, but this is my personal opinion.
I think Maven comes into its own in organizations that deal with a network of interrelated projects, like Apache does; where dependencies tend to change a lot and need to be quite explicitly specified to avoid "jar version hell". For isolated projects dependent on a few seldom changing jars, I find it overly intrusive.
To answer your question: I've read forum and blog posts on the Internet of other people doing exactly what you propagate. They use Maven for dependency management and then do their building with Ant. This undermines some of the benefits Maven is supposed to bring, such as the fact that a "normal" build is simpler to specify in Maven than Ant. However, I think you can be encouraged by the fact that you're not the only person with this idea, and it is indeed working for some other people.
I'd like to give you links to quotes, but I came across this stuff in the past few weeks and didn't collect references.
I'm a Maven fan, but it's not without its problems. Some of the issues I remember (and still fight):
Just like Ant, it has a magical syntax that can be hard to understand. If you're familiar with Any you may forget that, but lots of Ant tasks are terribly documented. The same is true for Maven. One of the reasons I eventually switched to Maven, though, is that for many of the mojos (similar to Ant tasks), you don't have to understand how to configure them. You just have to put the various pieces in the right place (which can be as hard as configuring a task...).
The automatic dependency management is amazing!... when it works. When you have to use non-Maven dependencies (like Hadoop) it becomes a problem. You either have to reference them as system scope dependencies, find somebody else who has packaged them, or package them yourself. And you eventually need to setup your own Maven proxy, like Nexus. And that's a whole extra hassle.
Maven is a lot of trouble on non-network or isolated LANs. The automagic is great, as long as you're networked.
It was getting so hard to write up a pom.xml, that I generated a ant build.xml from one of maven's tasks (which is a nice handy task...) I had to tweak the build.xml that was generated by maven. And now all my compiling, testing, etc., is being done with this build.xml..
Well. you can use maven archetype plugin to generate pom :)
Is such a combination common? I am thinking of making it permanent in my project.
JBoss Seam uses Maven internally to handle dependencies and do some targets in Maven. It's a big project that grew up with Ant, and now is difficult to build entire project solely in Maven, but that is going to happen in near future.
If you just need dependency management with Ant, you could give Ivy a try. Maven is a tool for managing the whole lifecycle of the build process.
Personally, I find Maven a nice tool once you get over the learning curve since you can standardize the build process for a lot of projects and there are a lot of great add-ons (particularly for code analysis tools). However, it's very possible that you do enough custom stuff with your build that the Ivy + Ant combo makes more sense.
Honestly, I would love to see a dependency management tool that implemented that part of Maven, specifically as a command line tool. For everything BUT dependency management, I find Maven to be absolutely awful if you're doing anything but exactly what the most general case is. Every time I try to do something that isn't "normal" (system/acceptance tests, etc), I run into a brick wall due to
either horrid documentation,
being told it's "not the maven way" (when it's a task that needs to happen, and "the way" shouldn't be a factor), or
being told to wait for the next version of maven, because maybe it'll be supported then.
I'd love to have a command line tool that can implement the "I need this as a dependency, go get it" functionality of Maven, possibly even using the pom.xml files of various packages. Then I could just use that in a Makefile and happy :)
Simple answer to the question posed: YES <-- click the link for details and reasoning.