wrt complete checkout, some use of

svn checkout --depth=immediates
svn up dir --set-depth=immediates
svn up dir --set-depth=infinity

might help with network issues.  (it breaks the checkout into smaller
transactions)

Dennis E. Hamilton wrote on Thu, Aug 25, 2011 at 20:23:31 -0700:
> Thanks joe,
> 
> I already had a laborious additional SVN update stage running when I saw this 
> message.  So about 18 hours total into this, when it interrupted once again, 
> I started a new folder, this time on my local hard drive (I had been updating 
> onto a shared folder of a file server), and did a complete check-out in 30 
> minutes, 30 seconds.
> 
> I can now drag that baby over to the file server where I want to keep it.  
> Quickly.
> 
> Based on this, when the merge into the incubator/ooo/ SVN subtree happens, I 
> think I will nuke the tree I have and do a complete check-out the same way.
> 
>  - Dennis
> 
> -----Original Message-----
> From: Joe Schaefer [mailto:[email protected]] 
> Sent: Thursday, August 25, 2011 11:13
> To: [email protected]; [email protected]
> Subject: Re: Update on SVN dump load
> 
> Yes that is a painful way to proceed.  9 times out of 10
> it is way faster to nuke a partial checkout and retry than
> it is to use svn update to pick up where you left off.
> 
> I learned this while dealing with network issues during a
> FreeBSD checkout.  Wasted a full day waiting on svn up.
> 
> 
> 
>       
> ________________________________
> 
>       From: Dennis E. Hamilton <[email protected]>
>       To: [email protected]
>       Sent: Thursday, August 25, 2011 2:03 PM
>       Subject: RE: Update on SVN dump load
>       
>       I am clearly doing this wrong.  There must be a more-efficient way to 
> handle this than by an SVN check-out and, after the check-out is interrupted 
> for some reason, subsequent SVN updates to continue pulling down a working 
> copy of the repo, rinse-repeat whenever there are connection failures of some 
> kind.  
>       
>       I say that because I am around 12 hours into that process and I am 
> still pulling just the trunk (at about 1.5 GB including all of the .svn 
> stuff).
>       
>       Fortunately, it doesn't swamp my machine and I can do other work, such 
> as write emails [;<).  Don't think I'll try watching Netflix on-line though 
> [;<).
>       
>       - Dennis
>       
>       PS: I have, since June 1, had a lifetimes supply of ways to show myself 
> how stupid I am.  Walking onto a project of this magnitude without first 
> learning the toolcraft and customs on something smaller is not thrilling.  I 
> am going to find those smaller things to teeth on while I watch in horror how 
> complex this activity is.
>       
>       -----Original Message-----
>       From: Dennis E. Hamilton [mailto:[email protected]] 
>       Sent: Wednesday, August 24, 2011 16:22
>       To: [email protected]
>       Subject: RE: Update on SVN dump load
>       
>       Ah, the excitement builds ...
>       
>       One way to not do commits (and to avoid certificate warnings) is to use 
> the http:// address, not the https:// form.
>       
>       - Dennis
>       
>       -----Original Message-----
>       From: Rob Weir [mailto:[email protected]] 
>       Sent: Wednesday, August 24, 2011 15:17
>       To: [email protected]
>       Subject: Update on SVN dump load
>       
>       Our JIRA issue has been updated:
>       https://issues.apache.org/jira/browse/INFRA-3862
>       
>       Joe has done a test load onto:
>       https://svn-master.apache.org/repos/test/joes/ooo
>       
>       No commits to it, please, but yell out if you see anything wrong.  It
>       looks good so far.
>       
>       -Rob
>       
>       
>       
>       
> 
> 

Reply via email to