[cvsnt] Re: cannot stat on commit/release woes

[John Kinson]

We get it here all the time too.  Our production repositories are hosted 
on Win2k server machines, our development PCs run Win2k (workstation). 
AFAIK no unusual timezone settings are in effect (we're in the UK in 
what's optimistically called summertime).  All filesystems are NTFS. 
Even testing locally (i.e. sspi access to a cvsnt server running on the 
local PC) has the same problem.  All other cvsnt commands work fine, 
it's just release that misbehaves.

I would disagree about the release command being pointless though; I 
frequently check out sandboxes from many different repositories, and 
when done, want a safe way to delete the directory without accidently 
discarding uncommitted changes.  "cvs release -d" is *exactly* the right 
command to use for this.  Or is there an easier way?

JK

Tony Hoyle wrote:
> David Somers wrote:
> 
>> Also, when releasing a directory, I'm told my files are altered, even 
>> though
>> status says they're up-to-date.
> 
> That's caused by a wierd combination of using a FAT drive and timezone 
> settings... not sure it's fixable as I've not been able to duplicate it 
> despite trying hard to.  Few people get it, and since release is almost 
> entirely pointless I've not bothered too much with it.
> 
> Tony

-- 
http://www.yellowradio.com/

If technology doesn't seem like magic,
it's probably obsolete.