[cvsnt] Re: migration from cvs to cvsnt, binary delta, ku option

Tony Hoyle tmh at nodomain.org
Thu Jan 13 12:03:50 GMT 2005


Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.


Christian Kasper wrote:
> i want to migrate from cvs 1.11.13 to the recent cvsnt version. i am
> working under sunos 5.9 and i have ca. 30 repositories.
> in your installation guide you wrote
>  
> 	---There is no change in the repository structure and it should
> be problem free to move a complete repository to
> a new CVSNT server.---
>  
> but in a email from Thu Apr  1 20:55:03 2004 daniel wrote
>  
>      ---CVSNT newer versions store mergepoints and binary
>     deltas, so the repository format is not the same.---
>  
> *confused* 

These statements don't conflict... going from CVS->CVSNT is very easy - 
only your CVSROOT directory may need some changes (as the parameters to 
one or two of the xxxinfo files are different).

Mergepoints don't change the repository format (new keywords are allowed 
for in the RCS specs) - binary deltas technically do so going in the 
other direction (CVSNT->CVS) is trickier if you've used them.

> *???* is the binary delta stable and efficient ? 

It's quite good but could be better*, which is what I was writing about. 
  It's stable in
in that there are no known bugs at the moment, plus the repository 
format isn't going to change in the future even if I change the 
algorythm - your old files will still read correctly.

> i am working in a utf-8 environment, so i have to add all text and
> program files with the option –ku ???

No - -ku is for UCS-2/UTF-16 files.  UTF-8 are handled the same way as 
ordinary text files, except you get extra options (you can do on-the-fly 
translation to any supported codepage, should you need it).

Tony

* Using Xdelta as a benchmark, I've had test versions 15-20% more 
efficient than that, but the code to do so is really hairy.  The current 
code emphasises stability over efficiency, so it's a few% less efficient 
than Xdelta.  Overall though any binary diff is much more efficient than 
a standard text diff on binary data.



More information about the cvsnt mailing list
Download the latest CVSNT, TortosieCVS, WinCVS etc. for Windows 8 etc.
@CVSNT on Twitter   CVSNT on Facebook