ant-ivy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alan Chaney <>
Subject Re: 2 Questions for ivyIDE
Date Tue, 20 Mar 2012 17:12:06 GMT

I can't comment on the first part of your answer, but I have commented 
in line on the sources issue.

On 3/20/2012 9:56 AM, David Sills wrote:
> All:
> I love working with Eclipse and ivyIDE with 2 exceptions, and I wonder
> if I am just missing something.
> My environment consists of a company-wide repository, where I place the
> results of building my projects, and of course a local repository, which
> sometimes though not always has folders for some of the same projects.
> As an example, let's assume that my project, say, "xxx-encryption"
> depends on "xxx-common". All is well and Ivy does well with it.
> However, I add some functionality to "xxx-common", up the build number,
> and rebuild. Great. Using Eclipse, I resolve "xxx-encryption" and it
> gets the correct updated value for the version number. In order to
> update the "xxx-encryption" Ivy files, I build "xxx-encryption" even
> though there have not been any changes to the project itself, only its
> dependencies. The company-wide repository's version of ivy.xml is
> correctly updated with the appropriate dependency on the new version of
> "xxx-common". However, the local repository's version of ivy.xml is not!
> This means that when I have a third project, say "xxx-web", that depends
> on "xxx-encryption" and try to resolve that project, the old (now
> incorrect) version of "xxx-common" is taken, since the local
> repository's version is read first if it exists.
> What am I missing? I find myself constantly manually updating the files
> in the local repository in order to get the transitive dependencies
> right.
> Also, can someone explain to me why my Ivy container in ivyIDE doesn't
> allow me to put source code locations on the JARs? That would be hugely
> helpful in debugging, especially in a situation where I have a lot of
> smaller projects that depend on each other and I assemble them as needed
> for a specific requirement. Or do I have to create source JAR files for
> each project using Ivy to make that happen? And will it happen if I do
> make source JAR files?
I don't quite understand what you mean by "source code locations on the 
JARs". Its possible to publish  Ivy artifacts for sources and javadocs  
(or anything else for that matter.) You need to add them to the 
publications section of your Ivy module.

Also, see this thread:

With the current released version you need to have the source and 
javadocs artifacts named with a pattern like:



However, there are a lot of ivy files out there which define the sources 
as source.jar or The next release of ivyde will, hopefully, 
rectify this by adding in some additional metadata.

In the mean time, if you have a corporate repository, you can simply 
arrange a folder structure that complies with existing mechanism and 
update your third party ivy files correspondingly. I've done this for 
one client using XSLT and an ant script.



> Many thanks for anyone who can answer either or both of these questions.
> David Sills

View raw message