Building application from multiple streams
One answer
I am assuming it is possible to create one WAR file from multiple (Eclipse) projects. Looking at RTC, it has one WAR file, but is created from multiple projects.
Multiple Streams basically requires one repository workspace per stream to load the data (as far as I know).
* The Jazz SCM tab built into the Jazz build Engine, can only load one (!) repository workspace or stream. So you can't use this to load the data on disk.
* You can use the SCM command line to load more complex scenarios.
* You can use SCM/LSCM to load multiple repository workspaces to disk. There are some rules with respect to overlap that you need to follow. You don't want the same projects in the two streams. However, if you avoid the scenarios not supported and overlap in general, you will be able to load data for multiple repository workspaces to disk.
* After doing so, you would be able to run a build across the data you loaded.
* Later, if you wanted to, you could create separate snapshots on each stream/repository workspace.
So, I think this is possible,however it requires some work to get the loading done.
Comments
see the discussion here
https://jazz.net/forum/questions/156194/default-or-current-flow-target-used-for-builds
Ralph,
>So, I think this is possible,however it requires some work to get the loading done.
its really easy.. create a workspace dedicated to build.
add all the streams of content as flow targets.
poof... works..no commandline, ... just a build def, use this workspace, and load
I thought it was a fault too, til I thought about it.. really cool capability.
This blows me away. I have to try that myself. Thanks Sam!!!
and, if you do NOT set 'create folder per component', then all the files will be loaded starting with the root of the sandbox workspace folder.. one big directory tree
( the parent POM could/would/should be here if using Maven)
Comments
sam detweiler
Jul 07 '14, 10:40 a.m.I think this goes back to a prior question. if the workspace has multiple flow targets set, then build will pull from all of them, and assuming no negative overlaps, will give you the combined results on the build machine..
then the build process should work from all the files present.
an example would be to take components A&V from stream 1
components X & Q from stream 2
and component Y from stream 3.
then the result in the workspace (and sandbox) will be
components A, V, X, Q and Y