How should I manage multiple development streams on the same workstation?
I am working on a large source tree. I am implementing several features at the same time. Each feature only takes a few lines of code, but they interact with each other so I want to develop them separately.
On git what I'd do is check out a single copy of the source tree and create multiples branches in it, one for each feature. Because the features all differ from the master branch by small text deltas, git is able to store them all efficiently and switch between them instantaneously.
I want to do the same thing with RTC. The only technique I and other members of my team have discovered so far is this:
- Check out the master stream into an Eclipse workspace.
- Every time we start work on a new feature, make a copy of the master workspace directory and open it in an different instance of Eclipse as a new workspace.
With this I have to keep multiple copies of a huge source tree on my machine even when they are identical except for a few files. I run out of disk space in my home directory if I work on more than three or four features at a time.
This is so clearly the wrong way to work that I assume I'm missing something, but neither I nor anyone on my team has figure out the right way to manage development branches. Can someone explain it to us or point to the appropriate place in the documentation?
Accepted answer
Hello William,
one approach may be to use the "Suspend/Resume" operation on your changes to switch between feature development, for example:
1. load workspace.
2. make changes for Feature A and check-in changes.
3. start work on Feature B: suspend the change set for Feature A, start work on Feature B.
4. revisit work on Feature A: suspend work on Feature B, resume the change sets for Feature A.
For more information, see https://pic.dhe.ibm.com/infocenter/clmhelp/v4r0m5/index.jsp?re=1&topic=/com.ibm.team.scm.doc/topics/t_suspending_change_set.html and
https://jazz.net/forum/questions/141517/use-case-for-suspending-a-change-set.
Sonia
one approach may be to use the "Suspend/Resume" operation on your changes to switch between feature development, for example:
1. load workspace.
2. make changes for Feature A and check-in changes.
3. start work on Feature B: suspend the change set for Feature A, start work on Feature B.
4. revisit work on Feature A: suspend work on Feature B, resume the change sets for Feature A.
For more information, see https://pic.dhe.ibm.com/infocenter/clmhelp/v4r0m5/index.jsp?re=1&topic=/com.ibm.team.scm.doc/topics/t_suspending_change_set.html and
https://jazz.net/forum/questions/141517/use-case-for-suspending-a-change-set.
Sonia
One other answer
Hi William,
please find the link below to work with multiple streams.
https://jazz.net/library/article/40
Create a StreamA ( Master) and Create a StreamB ( Feature Stream)
you can set the flow target from Stream B to StreamA ( it flows the changesets)
if you dont want to flow the changes you can always isolate the deopment activity and create a snapshot to record the status of the development activity.
You can record the information by creating a new Snapshot and you dont have to create a new eclipsepse workspace and use the file systems.
Please let me know if you need more information.
Regards,
Arun.