tags:

views:

77

answers:

3

Every coworker works on a copy of the same VPC and uses Tortoise to commit files. Some questions have arisen due the fact all VPC's have the same host-name and commit to the same repository. Questions about whether this will or not lead to corruption or conflicts of any sort in the repository.

I initially believed that this would not pose any problem since commits are atomic and subversion does not care about the host-name of the committing machine. But I just can't find any source stating this as a true fact.

So the question is, are there any concerns that need to be considered when committing files from multiple VPC's that all have the same host-name?

+6  A: 

No, this will not be a problem. Commits are done on a per-connection basis, it does not rely on the host being unique.

As an example, it's perfectly legal, and supported, to commit to the same repository from two different directories on the same machine, at the same time, at least if they do not lead to any normal conflicts due to editing the same files, etc.

Lasse V. Karlsen
Any chance on providing a link to some article or other source to back this up?
Mez
SVN was originally architected to use Berkely DB, and it used its transaction support, one transaction per connection. Later it was changed away from Berkeley DB since it made the repositories bound to their creation-platform, but the rest of the architect was kept. Internally, SVN still uses a transaction system, one transaction per connection.
Lasse V. Karlsen
But no, I don't have an article handy that points this out, but I *have* done the multiple simultaneous commits from a build machine, where multiple sub-projects was commited through a parallell branch in my FinalBuilder script.
Lasse V. Karlsen
+4  A: 

It should not pose a problem, because the author (committer) (not the host machine) is what is logged in SVN. THe fact that all the VPCs have the same host-name should do nothing to the repository

phsr
+2  A: 

This shouldn't be a problem : I'm often working with two different checkouts on the same machine, and never had nay kind of problem with it -- actually, on a project I'm currently working on, we are several developpers, some of us using several checkouts on the same machine -- and no problem at all : it works exactly as we each had only one checkout on distinct machines.

SVN is working on a "per commit" basis : the machine, host, hostname, user, whatever, ... don't matter.

Each commit will have its different revision number, and, in your case, different user name.

Pascal MARTIN