[PD-dev] DesireData

[IOhannes m zmoelnig]

Mathieu Bouchard wrote:
> On Thu, 7 Feb 2008, IOhannes m zmoelnig wrote:
> 
>> "desiredata" is an equivalent of "pd", so in my nice little world, it 
>> would be /trunk/pd/src and /trunk/desiredata/src rather than 
>> /desiredata/pd/src (there is really no reason to have a "pd" subfolder 
>> in "desiredata") but this is just nitpicking.
> 
> Depends... I have another question... what if I want to fork/branch 
> (whatever one calls it) other folders outside of the /trunk/pd 
> hierarchy? CVS makes it easy to branch individual files, or specific 

why would you want to do so?
the only reason i see is, that you are already used to this practice.

svn has the concept of _larger_ atomic commits (very unlike cvs) - this 
is based on the assumption, that a change (or file,...) must always be 
seen in a larger context.
example: if i have a function defined in fileA and used in fileB; if i 
rename the function in fileA, i have to change fileB inoder to stay 
consistent; so the 2 changes are interdependent; in svn i can commit the 
changes to both files in a single atomic transaction to ensure 
consistency of the repository; CVS does not provide such a mechanism 
(e.g. if i managed to commit my change to fileA but _not_ to fileB 
because of a conflict, the entire repository will be in an inconsistent 
state)
why am i telling you this?
because single files usually do not make sense without their context 
(neighbouring files) - it therefore is useless to tag a single file.
if you are absolutely sure that you have to do so, you can - of course - 
just "svn copy" your file to a new position - and branched it is.


fgmasdr,
IOhannes