Re: High level Registry API

With respect to DeleteKey() allowing recursive deletes. Either require using SHDeleteKey() to do it correctly, or don't allow recursive deletes at all. The problem being is that when deleting registry keys recursively on NT the entire operation should completly succeed or completly fail. If you naively recursively delete using RegDeleteKey() you can end up with a partially completed delete operation which is completly annoying to go about fixing. (bad, bad, bad) Bill

With respect to DeleteKey() allowing recursive deletes. Either require using SHDeleteKey() to do it correctly, or don't allow recursive deletes at all. The problem being is that when deleting registry keys recursively on NT the entire operation should completly succeed or completly fail. If you naively recursively delete using RegDeleteKey() you can end up with a partially completed delete operation which is completly annoying to go about fixing. (bad, bad, bad)
Bill This is a usefull hint.
BTW: win32com has the same problem, see win32com\server\register.py recurse_delete_key() Thanks, Thomas

In article <01bd01bf6f07$5d432940$4500a8c0@thomasnotebook>, Thomas Heller <thomas.heller@ion-tof.com> writes
With respect to DeleteKey() allowing recursive deletes. Either require using SHDeleteKey() to do it correctly, or don't allow recursive deletes at all. The problem being is that when deleting registry keys recursively on NT the entire operation should completly succeed or completly fail. If you naively recursively delete using RegDeleteKey() you can end up with a partially completed delete operation which is completly annoying to go about fixing. (bad, bad, bad)
Bill This is a usefull hint.
BTW: win32com has the same problem, see win32com\server\register.py recurse_delete_key()
Thanks,
Thomas ...
As a matter of interest is the Win32 registry single writer? -- Robin Becker
participants (3)
-
Bill Tutt
-
Robin Becker
-
Thomas Heller