Versioning

[Matthew_Kelton]

You know, you don't have to HAVE integration with LabVIEW to use git.  I have used git, svn, Hg, and bzr with no integration and have not seen any great need to have it.  I guess it dpeends on how you do your check ins.  If you think every change should have its own check in, then I could see the integration is nice.  For me, I see check-ins as a problem solved or new functionailty.  Not necessarily huge chunks of code, but it may be several modified files.

It is just as easy to switch to explorer (or your command line) and do a check-in.  For git, I could see some advantages if you want to stage each VI as you go, allowing for easier partial check ins.

[Jonas_Stein]

do you know how to run a

git commit -a -m "one more step"

git push

after each compilation or at some other frequent event?

I think its a bit anoying to switch programs to get kind of backup.

[smercurio_fc]

Why would you perform a commit after each "compilation", and what do you mean by compilation? Is it related to your other query? I think you may be "overusing" the source code control system. Even if we we talking about a text-based program it makes no sense to me to commit after I compile. What if I have bugs? I commit at the end of the work day, or sometimes in the middle if I made a lot of changes.

[Matthew_Kelton]

If that's what you want, then you don't want version control.  Simply doing a commit at no logical state and saying the message should be "one more step" means nothing.  You can use something like PureSync to create backups for you automatically.

commit messaging should make sense and be descriptive of what has changed since the last commit.  How are you going to track down any specific version of a file when all your commit messages are meaningless?

[Jonas_Stein]

i can make a manual commit, when i have done a nice step... 

[smercurio_fc]

Then what's the issue?