SCM Plugin for ClioSoft's SOS?

[Marty]

Hi all,
Anyone have a plugin to talk to SOS repos that they'd be willing to
share?

TIA,
Marty

[Jonathan]

Sorry for dupe in Jenkins Developers, but figured this was a more appropriate place to reply and hopefully re-start a discussion...

I'm also interested in a Cliosoft SOS plugin for Jenkins as I'm exploring it's use in our ASIC flow.  I'm curious what you or anybody else are doing for Cliosoft integration?  I'm actually not really an SOS user (yet), but we do have some existing infrastructure around it and know it'd be vital to getting something like this going.  I was thinking about using the file system scm or sh scripts to pull 'design' things like library and netlists (eventually from Cliosoft), and them maybe maintaining other scripts like tool tcl, perl, etc in a local git repo.  Except, I just found that the multiple scms plugin doesn't seem to work with the file system scm (https://issues.jenkins-ci.org/browse/JENKINS-34895).

Thanks.

-Jonathan

[Martin d'Anjou]

I agree this discussion should be here and not on the dev mailing list.

Have you approached Cliosoft to ask if they would write a Jenkins plugin for you? It surely can help them as well as help you.

For SCM, I am switching to Git. For storing generated assets (pre-compiled libraries, netlists, etc), I am switching to a binary repository (e.g. artifactory).

For managing hardware generated outputs in Jenkins (numerous and large), there is the External Workspace Manager Plugin currently being developed as part of the Jenkins Google Summer of Code. We are discussing the development here, and the design document is here. You are welcome to participate in the discussions, provide feedback, join the on-line meetings, etc.

Martin

[Jonathan]

I intend to follow up with Cliosoft and ask.  I'll review the GSOC doc more closely, and see if there's anything I can add.  I noticed the existing "use custom workspace" option in Jenkins but this seems like a more full featured implementation (including pipeline code support).

I agree about the large amounts of binary data, and I guess I haven't gotten far enough in my investigation of Jenkins to realize that as a problem.   All of our data and machines are on a shared network filesystem, so I don't have to copy between different nodes.  A lot of data would be in Cliosoft, but otherwise I was hoping I could point to an upstream workspace or of a wrapper job.  Also, I think a proper 'populate' from Cliosoft would create links to the data and avoid copying.  This is all relevant to what I'm trying to figure out, just overwhelmed with trying to determine a 'best' approach to start with for a variety of different non-native tasks I'd like to automate.

Thanks.

-Jonathan