Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.
--- Tony Hoyle wrote: > rrankin1424-cvsnt at yahoo dotcom wrote: > > OK, is this related to or causing the rename problem all of a sudden? I've > been > > using Cygwin with CYGWIN='...ntsec...' and CVSNT on this server for about > three > > years with no problems until now. > > No, it's been doing that for a while. If you haven't noticed it the > it's not been an issue for you. > > > > > Another thought... What happens when CVSNT sees the CYGWIN environment > variable > > *without* ntsec? Does CVSNT actually parse the contents of CYGWIN to see if > it > > contains 'ntsec'? If so, will it correctly handle (and ignore) the > 'smbntsec' > > setting? Ntsec is now the default in Cygwin. If you want to use ntea, you > have > > to set it explicitly. If I can remove 'ntsec', but leave 'smbntsec', might > that > > help? > > smbntsec must be new... haven't seen that one. The default for CVSNT is > always ntea, unless you tell it otherwise by explicitly setting CYGWIN. > It understands ntsec and nontsec - the parsing code is actually very > similar to the code cygwin itself uses (or used to use - if they've > added a new option that'll be separate). > > There's also the CVSNT environment variable for emergencies (not even > documented!) So if you set CVSNT=ntea it'll override the cygwin. > IIRC, smbntsec has been around for 6 months to a year. I've had it set for at least 6 months. The CVSNT variable was suggested to me by John McNamee (thanks John!). I set CVSNT=nontsec nontea on the server per his suggestion and now my permissions in the repository are unchanged after a commit. Actually, the CYGWIN env var interaction probably explains some odd permission problems I've had for a while on the client box I use. I just set CVSNT there as well. We'll see. Unfortunately, this had no effect on the rename problem in 58c. I'm about as sure as I can reasonably be that this is *not* an A/V issue, so it looks like I'm going to have to get the cvsnt source, build a debug version and try to debug this. Could be a while, though... Rick