[
https://issues.apache.org/jira/browse/IVY-641?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12680130#action_12680130
]
Daniel Dekany edited comment on IVY-641 at 3/9/09 4:54 AM:
-----------------------------------------------------------
I have just looked for a such feature. What I would like to add is that this
would be useful not only for the "work at home" case, but for using Ivy offline
in general (think about laptops). I found that regardless of the TTL I set Ivy
will reach the repository for entries where changing="true", or
rev="latest.integration", etc., and if it can't do that, the Ant task will
fail. So for a more open project (contribs come and go, etc.) currently I have
to simulate the old-school way of dependency management, by ivy:retrieve-ing
into a <projecthome>/lib (with an "updatelib" Ant task that is never called
automatically), and then don't use Ivy anywhere else, just traditional Ant
path-s (instead of the ivy:cachepath). It's not only not elegant this way, but
I practically lose the multiple-configurations feature of Ivy! Now, if I could
build an Ivy *repository* in the lib directory (as easily as I can ivy:retrieve
into it), then I could use ivy:cachepath too. Furthermore, just be replacing
the ivysettings.xml one could switch to true online mode.
Note: I know both the above cases can be addressed by tar-ing the corporate
repo and giving it to the employers. While it certainly works in a
well-controlled corporate environment, for the more open projects (think about
random sf.net stuff, or even Jakarta stuff) it just wouldn't work in practice
that way.
was (Author: [email protected]):
I have just looked for a such feature. What I would like to add is that
this would be useful not only for the "work at home" case, but for using Ivy
offline in general (think about laptops). I found that regardless of the TTL I
set Ivy will reach the repository for entries where changing="true", or
rev="latest.integration", etc., and if it can't do that, the Ant task will
fail. So for a more open project (contribs come and go, etc.) currently I have
to use the simulate old-school way of dependency management, by
ivy:retrieve-ing into a <projecthome>/lib (with an "updatelib" Ant task that is
never called automatically), and then don't use Ivy anywhere else, just
traditional Ant path-s (instead of the ivy:cachepath). It's not only not
elegant this way, but I practically lose the multiple-configurations feature of
Ivy! Now, if I could build an Ivy *repository* in the lib directory (as easily
as I can ivy:retrieve into it), then I could use ivy:cachepath too.
Furthermore, just be replacing the ivysettings.xml one could switch to true
online mode.
Note: I know both the above cases can be addressed by tar-ing the corporate
repo and giving it to the employers. While it certainly works in a
well-controlled corporate environment, for the more open projects (think about
random sf.net stuff, or even Jakarta stuff) it just wouldn't work in practice
that way.
> Need task to clone repository
> -----------------------------
>
> Key: IVY-641
> URL: https://issues.apache.org/jira/browse/IVY-641
> Project: Ivy
> Issue Type: New Feature
> Components: Ant
> Reporter: Pavel Krupets
>
> I installed repository at work (it's not visible from outside due to security
> issues) so I need means to clone repository to local drive so people can work
> remotely or at home. <repreport> can go through repository so I think it
> won't be to difficult to add.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
