sourceforge cvs

[Net Llama!]

On Tue, 28 Jan 2003, Chuck Mead wrote:
> On Tue, 28 Jan 2003, Net Llama! posted the following:
> NL>On Tue, 28 Jan 2003, Chuck Mead wrote:
> NL>> On Tue, 28 Jan 2003, Moritz Heiber posted the following:
> NL>> MH>> Sourceforge cvs is really, really, really a piece of shite!@#$%^&*
> NL>> MH>>
> NL>> MH>> Can't we please, please, please move it to where there are not 35,000
> NL>> MH>> other projects trying to use cvs at the same time?
> NL>> MH>>
> NL>> MH>> I'm begging here!
> NL>> MH>
> NL>> MH>I never had problems with SF.net CVS (except the one problem about 2 weeks
> NL>> MH>ago which has even been preannounced by SF.net) and therefore don't see the
> NL>> MH>point of moving it.
> NL>> MH>Could you specify which problems do you mean?
> NL>>
> NL>> sf cvs sends an end of file when it's busy (which it always is) on
> NL>> anonymous co and up's. It happens all the time to me when I am simply
> NL>> trying to do an anonymous co or up to build on a new machine. This means
> NL>> nestu's script has to be run repeatedly to automate the process (which
> NL>> is the goal of the script). The developer co is not as problematic (I
> NL>> use both depending on where I am) as it seems to work most of the time.
> NL>> It isn't that it doesn't work (it does) it's that we are sharing
> NL>> resources with a bazillion other people and the resulting congestion
> NL>> causes these problems. Since we have the capability to get around this
> NL>> issue (I *DO* have the resources to provide cvs) I do not see why we
> NL>> don't use it.
> NL>
> NL>Granted, i don't do anon checkouts as oftten as alot of the others here,
> NL>but i've never had problems.  Are you using z9 compression?  I've found
> NL>that makes a world of difference on high latency connections.
>
> I am sitting on a backbone with 622MB of broadband (RH headquarters)
> this is *NOT* a high latency connection... it is SF congestion with
> many, many people using cvs that is causing this problem so no amount of
> compression is going to help. (I have seen this before and recognize the
> behavior)

I'm not looking for an argument here, just trying to offer a helpful
suggestion.  I've seen the behavior that you're reporting before, and
increasing compression does resolve it.  You're not even willing to try
it?  BTW, i was referring to sf.net as the high latency connection, not
yours, no need to be so defensive.

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Lonni J Friedman				netllama at linux-sxs.org
Linux Step-by-step & TyGeMo		     http://netllama.ipfox.com