[darcs-users] darcs annotate format obscures the code

Ralph Corderoy ralph at inputplus.co.uk
Sun Mar 20 10:16:54 UTC 2005


Hi Michael,

> > > Where's the code? ?Code is distinguished by a single leading
> > > space, not much of a visual distinguisher at the best of times,
> > > but so are the (oft repeated) log messages embedded in the
> > > annotation making it almost impossible to tell code from
> > > annotation without careful study.
> > 
> > I usually output the annotation to a file and then open the file
> > with VI.  It makes all the stuff starting with '#' blue and the rest
> > is black.  Quite readable.  Although it usually is still enough to
> > make me use the search functions of my editor.
> 
> Which brings up an additional problem.  When I'm looking at a blame
> log I usually want to answer the question "why is line 452 in file
> Foo.pm there?" With the CVS/SVN/SVK annotation format line 452 is line
> 452 (well, 454 for SVK, they screw that up a bit).  With darcs... who
> knows?  The annotations add vertical space and thus throw off the line
> numbers entirely.  Now instead of a simple "jump to line 452" in my
> pager I must grep the code (and the log messages) to find my line.

In less do `g452/^ ', no?

I agree with the general problem.  I think the code up front with a
`change number', i.e. 1..n, followed by an `end of code' marker and the
blame details, in change number order.  Both have to be output at the
same time to avoid the repo changing inbetween generating the two.  I
suppose the command could send the code to stdout and store the blame
detail in a separate given file.

Cheers,


Ralph.





More information about the darcs-users mailing list