Sourcetree has performed phenomenally for me so far, but I've noticed a case where the application can slow down to a near crawl.

Our office has two networks: One for connecting to the company domain and another for high bandwidth. Since our company's domain network is slow, I tend to stay connected to the high bandwidth network. However, if my computer starts connected to the company domain network, I am automatically granted access to some very slow network drives.

Note I do not have any repositories stored on the network drives.

If Windows does not detect the slow network drives (started computer on high bandwidth network), Sourcetree runs spectacularly.

If Windows has access to the slow network drives (started computer on company domain network), Sourcetree sees a significant, but still workable slowdown.

If Windows detects the slow network drives but no longer has access (started computer on company domain network then switched to high bandwidth network), Sourcetree slows to a crawl and individual components see various delays. A diff of a file may pop up before the staging area has refreshed. It may take half a minute for a small repo to open. If I then manually disconnect from the network drives, performance is restored.

I assume this has to do with the way Sourcetree or Windows scans the file system. The mere presence of a missing or slow drive is enough to grind work to a near halt. It is not so much an issue for me, but it is a performance issue. Is this a flaw in Windows or could Sourcetree exclude looking at drives that have no repository?