I have a task in my build.xml
:
<target name="init" depends="init-ivy">
...
<ivy:cachepath
inline="true"
module="jersey-container-servlet"
organisation="org.glassfish.jersey.containers"
pathid="jersey.classpath"
revision="2.23.2" />
...
</target>
This task downloads ivy if necessary (init-ivy
does that actually) and then invokes ivy to download dependencies. It sets jersey.classpath
as the result.
Right now my build
task depends on the init
task. So every time I build, it checks if dependencies need to be installed. I want to avoid checking dependencies every time and have build
separate from init
. But init
sets jersey.classpath
and build
uses it.
Is there a way to obtain jersey.classpath
from ivy without asking it to check dependencies? Is it even a good practice to not check dependencies in this case?
As explained in this answer, ivy doesn't download jars every time it runs. It caches them locally under "~/.ivy2/cache":
Secondly, you're using ivy in inline mode, presumably to avoid creating an ivy file? ivy cachepath is classified as post resolve task, which means it will automatically call the resolve task, in the background. What the inline mode does is tell ivy to perform a fresh resolve each time, which is wasteful if you have more than one classpath to manage.
Finally have you considered using an ivy file? A single call to the resolve task can work thru all your project's dependencies, cache this information locally and then determine if files need to be downloaded or not. I would recommend always resolving dependencies. It doesn't cost much and it's possible that things have changed between builds (for example if you're using dynamic dependencies, or Maven snapshots).
Here are my standard Ant targets for ivy look like:
<available classname="org.apache.ivy.Main" property="ivy.installed"/>
<target name="resolve" depends="install-ivy">
<ivy:resolve/>
<ivy:report todir='${ivy.reports.dir}' graph='false' xml='false'/>
<ivy:cachepath pathid="compile.path" conf="compile"/>
<ivy:cachepath pathid="runtime.path" conf="runtime"/>
<ivy:cachepath pathid="test.path" conf="test"/>
</target>
<target name="install-ivy" unless="ivy.installed">
<mkdir dir="${user.home}/.ant/lib"/>
<get dest="${user.home}/.ant/lib/ivy.jar" src="http://search.maven.org/remotecontent?filepath=org/apache/ivy/ivy/2.4.0/ivy-2.4.0.jar"/>
<fail message="Ivy has been installed. Run the build again"/>
</target>
Notes:
- Single target called "resolve" that calls ivy to manage my project classpaths. It also generates a useful report for documentation
- The cachepath task is using ivy configurations to group or classify dependencies within the ivy file. This is really the efficiency magic that enables to save time on a single resolution.
- Notice how the install-ivy task has a conditional check to determine if ivy is installed. This trick is not feasible with the rest of your project's dependencies due to complexity. My recommendation is ensure ivy exists and then use it to manage everything else. (Under the hood it'll do its best to be efficient). I don't really understand why Apache Ant doesn't bundle the ivy jar.
来源:https://stackoverflow.com/questions/39313778/how-to-get-ivycachepath-location-without-checking-if-dependencies-downloaded