Code review comment for lp:~parthm/bzr/81689-win-symlink-warning

Parth, you are a star for tackling this.

One thing this branch really needs is some tests that cover the code paths for where symlinks are unsupported. I think the basic plan should be:

* Create a branch in memory that contains symlinks
* Patch out os.symlink, with the both the "doesn't exist" and "raises error" cases covered
* Run command of some sort on the branch

With the different interesting combinations exercised.

I'm failing to find a good example at the moment, but I think you can use bzrlib.branchbuilder for the first step, and can check the testcase log for your warning message and the treeshape on disk as assertions. Yell if you need any help from me.