[darcs-users] version numbers vs. version names

zooko zooko at yumyum.zooko.com
Thu Mar 31 15:45:56 UTC 2005

> I mention it because, well... its the truth.  Its a lot easier to remember 
> and type in "3214" than "a TODO test which confirms that RT#266 is a 
> regression" or "Sat Mar 12 14:08:46 PST 2005".

I use tailor.py to mirror Shtoom's SVN repo.  Tailor makes names like this:

Mon Mar 28 11:10:54 AST 2005  zooko at zooko.com
  * [conflict resolver] merge the patch that adds shtoominto to scripts and the patch that adds shreadder to scripts

Mon Mar 28 10:40:49 AST 2005  anthony at interlink.com.au
  * [svn r1288] and more

Mon Mar 28 10:40:21 AST 2005  anthony at interlink.com.au
  * [svn r1287] more updates

Mon Mar 28 10:39:43 AST 2005  anthony at interlink.com.au
  * [svn r1286] more updates

Mon Mar 28 10:39:13 AST 2005  anthony at interlink.com.au
  * [svn r1285] updates to new schema methods

A few comments:

1.  The "[svn rXXX]" is automatically entered by copying the SVN rev #.

2.  They are very convenient.  Whenever I'm dealing with a patch that come from 
SVN, I usually type "darcs annotate -r1285" instead of 
"darcs annotate -r'updates to new schema'".

3.  Okay, so anthony is lazy about comments sometimes, since in SVN-world the 
comment isn't used for naming.  But I usually prefer to use the numbers even 
when the comments are distinct, useful, and memorable.  It's less to type.

4.  The one named "conflict resolver" I entered manually instead of via tailor.  
Too bad darcs didn't assign it an integer number for me.

5.  If I had to choose either numbers or patterns-of-names the way darcs 
currently does, I would choose patterns-of-names.  I *usually* use the numbers, 
but not always -- only when I am currently looking at the number so that I don't 
need to remember it.  If I had to have only one I would prefer the flexibility 
and meaningfulness of the patterns-of-names.  But as you can see it is very easy 
to get the best of both by automatically inserting a number into the name.



More information about the darcs-users mailing list