You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Branko Čibej <br...@xbc.nu> on 2003/07/15 22:16:18 UTC
Blast and issue 688 again
I finally figured out why _all_ python tests are failing on HP-UX. It's
issue 688 again, but this time in "svn co". Hey, I heard a funny one the
other day:
Q: Why is "svn co" like "svn diff URL URL"?
A: Because both open TWO sessions to the repository.
Yes, imagine that -- svn_client_checkout opens an RA session, then calls
svn_client_update, which opens an RA session...
Wish I could bash someone on the nose for this one... :-/
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Blast and issue 688 again
Posted by Branko Čibej <br...@xbc.nu>.
Glenn A. Thompson wrote:
> Hey,
>
> Branko Čibej wrote:
>
>> Greg Hudson wrote:
>>
>>
>>
>>> You might be able to get the HPUX tests going by arranging to close the
>>> first session before starting the second. But that would still be a
>>> rather inefficient use of server resources. Actually fixing the
>>> problem
>>> requires some thought, like maybe making svn_client__update_internal
>>> accept an optional ra baton.
>>>
>>>
>>>
>>
>> I'm looking at that solution, yes -- but there's a catch. The sessions
>> used by checkout and update aren't initialized in the same way, a few
>> parameters are different. I don't know yet if it's O.K. to just modify
>> the session params and keep on using it, or not.
>>
>
> Brane, I thought we semi agreed that this should be fixed in the FS.
> I'm pretty sure we can do it there with minimal impact.
>
> .... Yes, I'm still around.
> gat
>
> ** gat mumbles something about figuring a way to get money without
> customers.
Fixing issue 688 in the FS is one thing. But what I want to do is fix
"svn co" to not try to open two sessions in the first place.
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Blast and issue 688 again
Posted by "Glenn A. Thompson" <gt...@cdr.net>.
Hey,
Branko Čibej wrote:
>Greg Hudson wrote:
>
>
>
>>You might be able to get the HPUX tests going by arranging to close the
>>first session before starting the second. But that would still be a
>>rather inefficient use of server resources. Actually fixing the problem
>>requires some thought, like maybe making svn_client__update_internal
>>accept an optional ra baton.
>>
>>
>>
>>
>I'm looking at that solution, yes -- but there's a catch. The sessions
>used by checkout and update aren't initialized in the same way, a few
>parameters are different. I don't know yet if it's O.K. to just modify
>the session params and keep on using it, or not.
>
Brane, I thought we semi agreed that this should be fixed in the FS.
I'm pretty sure we can do it there with minimal impact.
.... Yes, I'm still around.
gat
** gat mumbles something about figuring a way to get money without
customers.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Blast and issue 688 again
Posted by Branko Čibej <br...@xbc.nu>.
Greg Hudson wrote:
>You might be able to get the HPUX tests going by arranging to close the
>first session before starting the second. But that would still be a
>rather inefficient use of server resources. Actually fixing the problem
>requires some thought, like maybe making svn_client__update_internal
>accept an optional ra baton.
>
>
I'm looking at that solution, yes -- but there's a catch. The sessions
used by checkout and update aren't initialized in the same way, a few
parameters are different. I don't know yet if it's O.K. to just modify
the session params and keep on using it, or not.
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Blast and issue 688 again
Posted by Greg Hudson <gh...@MIT.EDU>.
On Tue, 2003-07-15 at 18:16, Branko Čibej wrote:
> I finally figured out why _all_ python tests are failing on HP-UX. It's
> issue 688 again, but this time in "svn co". Hey, I heard a funny one the
> other day:
>
> Q: Why is "svn co" like "svn diff URL URL"?
> A: Because both open TWO sessions to the repository.
>
> Yes, imagine that -- svn_client_checkout opens an RA session, then calls
> svn_client_update, which opens an RA session...
>
> Wish I could bash someone on the nose for this one... :-/
Ah, you noticed it too. One session to get the current head revision in
order to initialize the wc, and another to perform the update operation.
You might be able to get the HPUX tests going by arranging to close the
first session before starting the second. But that would still be a
rather inefficient use of server resources. Actually fixing the problem
requires some thought, like maybe making svn_client__update_internal
accept an optional ra baton.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org