I know that this is not the default way to work in VSTS, but I think that when there is a big change private branch should be used. that way:
1. changes of others will not be a problem
2. at the end of the work merge to the main branch will be done
this is an option - instead of working on a normal workspace.
I wanted to asked if someone works that way, and if it is recommended
thanks
private branch for developer
MrHayes
Shelveset is not good enough because I'm talking about a big change, that can take a month of development. I want to work seperately.
I want to checkin mid versions (concerning my change, and I don't want them to be in the integration). I don't want to sheve my changes after each workday (I'll get lost with all the shelves). that's why I don't see shelving as a part of a version control methodology.
what are the problems in private branch can you list them
joru
Ori,
If you are talking about an item of work performed by a single developer, I would recommend using a Shelveset rather than cluttering up the repository with another branch.
Rizak the Really Horrible
After creating the "private" branch, can it be hidden so other users won't be able to see it
I don't won't anyone to see my developing status, so they wouldn't, accidentally, try to use it instead of the correct one.
mepatuhoo
You could certainly set the permissions so that only you have access to it (from the Source Control Explorer, right click on it to get the Properties dialog and go to the Security tab, or you can use the command line, tf perm). Otherwise, you could simply create a directory under a team project where you and others can have branches that are simply work in progress.
Buck
Soul
Hi
The problem is that each developer doesn't want his others to be able to view his work (and even the fact that he's working on a new branch...)
Is there anyway to hide a branch
Thanks,
Guy
amandeep
I won't offer this up as the recomended development model using TFS ... however when I'm in a situtation where I need to spend an extended period of time on a project or where I expect to create a large amoung of churn - I go off into a branch to do it.
In doing this I only branch the absolute minimum necessary to get the job done. I branch the specific feature into a private branch and then map that into my workspace at the same level as the origional feature code (but as a new name). This way I can hook up the build system more easily and it lets me have the old and new code side-by-side (which can make doing directory diffing easier - especially when merging changes between the two branches).
Robert.
Jie Chen