[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Why is my svn+ssh:// access asking 4 passwd twice?

From: Matthew England <mengland_at_mengland.net>
Date: 2005-04-12 05:35:05 CEST

>Are you using a single repository,


>or is there an external or other oddity going on?

I think the oddity that's happening is that I'm running every new checkout
test on a new, empty, client directory/folder. When I had presumed that
other repo's didn't have the (multi-password) problem...well I'm coming to
learn that that was probably due to the fact that I had already done a
checkout on that client dir in the past and therefore further access only
required on password.

>I can only say that when I started messing with svn+ssh, a checkout
>requested far more than four password authentications(I quit at 20+).

Ok, fair enough.

>That's when I just moved on to pageant.

What's pageant?

I think I'll be effectively renaming my inquiring to "_why_ does svn+ssh
have different access behavior for different kinds of repos?" rather then
"why is SVN breaking?"

I like to be intimately familiar with my development tools, particularly
when they are controlling the "family jewels"--my corporate source
code. When I don't know why something is occuring, and/or I have no
reasonable explanation for it, I dig deeper.

In this case, I'm willing to accept SVN "disconnecting and reconnecting"
approach, so long as I can find out why the system is behaving this way.

Maybe this was already stated and I'm simply not listening? (My apologies
if so.) I'll go back and re-read the emails (and maybe the FAQ again).

After a reread...

>Sounds to me like subversion has to make multiple connections to perform
>some operations, each of which is a separate repository access

...but _why_?

What is the ".svn" subdir with all of its stuff in there doing?
It's now pretty clear that that the first connection was to create the
".svn" stuff, and the other connection was to get all the "non-meta data"
files (ie, the data I really care about).

Why was .svn automatically created? It's not in the repo; it's just
autocreated in the client's local filesystem. The associated
.svn/README.txt says:

This is a Subversion working copy administrative directory.
Visit http://subversion.tigris.org/ for more information.

I suspect this is a bunch of control (and therefore very important)
meta-data (or something like meta-data) to manage and keep track of the
client-to-server relationship, changes, etc? ...and it's probably only
created (or at least initialized) upon the first checkout to a clean,
previously unused dir?

>Also, I wouldn't have even responded if you had specifically mentioned
>that you believed something was broken or that something had changed from
>an older scenario and the section I quoted didn't apply.

I'm not sure what is meant here. In any case, I hope I did not rub anyone
the wrong way.

Thanks for everyone's continued help and persistence here.

I see this as a detailed FAQ in the brewing. I'm happy to write a draft
for it (the FAQ) if this helps...so long as I can get all the
down-and-dirty info (or bet pointed towards it).


To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Tue Apr 12 05:37:00 2005

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.