On Thu, Nov 4, 2010 at 6:47 AM, David Aldrich <david.aldr...@eu.nec.com> wrote: > Hi Nico > >> Hard to tell. Did you compile it yourself, or are you using your >> distribution's Linux? And have you looked in .subversion for the >> authorization file with your password stored in it? > > Thanks for referring me back to that area. I deleted the existing > authorization files and then found that the password is correctly stored. Not > sure what was going on but its fixed now. > >> Mind you, I provoundly loathe this feature. It puts your passwords in >> clear text in your $HOME/.subversion directory, and is one of the >> reasons I so strongly espouse svn+ssh access, which typically relies >> on SSH keys. > > Understood. We currently use https. Do you know how https speed compares with > svn+ssh?
In general, the protocol over which the connection is bundled is not a big performance limit. For moderately used repositories, complex post-commits doing silly things or doing checkouts to CIFS shares are such overwhelming performance hits that I've not noticed particular differences between svn+ssh and HTTPS. Heavily burdened servers running into resource limits for all that encryption, well, that's a different story.
