Best way to version control a WCF application with Git?
Posted
by Sam
on Stack Overflow
See other posts from Stack Overflow
or by Sam
Published on 2010-05-11T23:49:27Z
Indexed on
2010/05/11
23:54 UTC
Read the original article
Hit count: 228
Suppose I have the following projects. The format is [ProjectName] : [ProjectDependency1, ProjectDependency2, etc.]
// Service
CoolLibrary
WcfApp.Core
WcfApp.Contracts
WcfApp.Services : CoolLibrary, WcfApp.Core, WcfApp.Contracts
// Clients
CustomerX.App : WcfApp.Contracts
CustomerY.App : WcfApp.Contracts
CustomerZ.App : WcfApp.Contracts
(On a side note, WcfApp.Contracts
should not depend on WcfApp.Core
, right? Else CustomerX.App would also depend on and thus be exposed to the service domain model?)
(CoolLibrary
is shared with other applications, so I can't just put it inside of WcfApp.Services
.)
All of this code is in-house. I was thinking of having 6 repositories for this. The format is [repository folder name] : [Projects included in repository.]
1. CoolLibrary.git : CoolLibrary
2. WcfApp.Contracts.git : WcfApp.Contracts
3. WcfApp.git : WcfApp.Core, WcfApp.Services
4. CustomerX.App.git : CustomerX.App
5. CustomerY.App.git : CustomerY.App
6. CustomerZ.App.git : CustomerZ.App
How should I manage my project dependencies? I see three options:
I could use binaries which I have to manually copy to each dependent repository. This would be easiest at the start, but my repositories would be a little bloated, and it'd become more tedious as I add more client apps for customers.
I could import dependent code as submodules. This is what I will probably end up doing, although I keep reading on the web that submodules are a hassle.
I also read that I can use something called the subtree merge strategy, but I am not sure how it is different from just cloning the repo into a subdirectory and adding the subdirectory to .gitignore. Is the difference that the subtree is recorded in the master repository, so (for example) cloning it from a different location will also pull the subtree?
I know I asked a lot of questions in this post, but the most important two questions I have are:
1. Am I using the right number and layout of repositories? Should I use less or more?
2. Which of the three dependency management strategies would you recommend? Is there another strategy I haven't considered?
© Stack Overflow or respective owner