I was unable to reproduce either of these behaviors using VBP 5.6b and Vault 3.0.1 client (against the sourcegear test server). I was able to reproduce a crash in the Vault client if the current directory wasn't set back to a valid drive+path after deleting the subst drive (this sequence has a consistent although slightly different behavior from a batch file: the shell complains about the directory being invalid and won't execute the vault command).
I've attached a modified version of the Vault.bld sample that follows the described steps below. Does the 2nd setworkingfolder step fail in your case? I've also attached a equivalent batch file also test to see if the behavior differs there vs VBP. Both work successfully here.
|