Ben Collins-Sussman writes on 20 Dec 2002 09:55:30 -0600
>Bob Gustafson <bobgus@rcnChicago.com> writes:
>
>> >Repeat after me: The working file and text base can be different even
>>if you
>> >didn't edit the working file.
>>
>> Why?
>
>Because the text-base is a pristine copy of the file as it exists in
>the repository.
>
>Repository files, by definition, always have keywords contracted, and
>if 'svn:eol-style == native', are stored with LF line-endings.
If one wanted to eliminate the 'extra disk storage requirement' [the point
of this thread, I think], then one could place some sort of 'wc <->
pristine_repo' delta into the .svn area, instead of the whole file.
The main thrust here is to figure out what is the essential information
required at each step of the process (and of course, decide on what is the
process too), and store only that much information.
There are a lot of tradeoffs, potential processing time, network traffic,
code development time, modularity of code, etc. It seems as though Subv is
doing a pretty good job as it is in many of these areas. However, as a
newby to the discussion, why not challenge some of the current
assumptions... At least that is where I am coming from.
BobG
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Dec 20 18:25:23 2002