[CLUE-Tech] On the subject of OpenOffice...

[greg at knaddison.com]

Quoting Ed Hill <ed at eh3.com>:

> On Fri, 2003-10-03 at 21:07, David Anselmi wrote:
> > greg at knaddison.com wrote:
> > [...]
> > > On the plus side, OOo is XML based which means that you can do things
> like store
> > > the source of documents in CVS and get useful output from the "diff"
> command.
> > 
> > The docs that I've seen are several XML files zipped together.  How do 
> > you store them in CVS?  I was thinking that I would have to figure this 
> > out, but since you mention it...
> 
> Hi folks,
> 
> I've been using CVS quite a bit recently, so heres what I'd do:
> 
>   1) create a binary data file: "data.bin"
> 
>   2) "cvs -ko add data.bin"
> 
>   3) "cvs ci data.bin"
> 
>   4) "cvs status data.bin" and notice that the "Sticky Options:" 
>      now includes "-ko" meaning that cvs will understand that its 
>      a binary file and shouldn't be diff-ed, etc.
> 

While that would work, it ignores some of the major benefits of CVS.  C
(Concurrent) VS can merge like text files, but not binary files and it can diff
various versions of text files, but again not binary files.  You can just unzip
the OOo files into a directory by the name of the document and then put that all
into CVS as normal text files, not binaries.  Another user could then check out
that folder, zip it, rename it, and open it with OOo.  I believe that the really
important file within the zip is the content.xml file and the rest are just
formatting.  

Ed's way works for binary files, but it bloats CVS (since it stores the entire
file rather than just the "diff" between the files) and it loses merge and diff
capabilities.

Greg