Re: Potential fix for issue 1504. (issue4237057)

[address@hidden]

On Mar 15, 2011, at 9:18 AM, Han-Wen Nienhuys wrote:

> On Tue, Mar 15, 2011 at 10:08 AM, address@hidden
> <address@hidden> wrote:
>> On Mar 15, 2011, at 9:03 AM, Han-Wen Nienhuys wrote:
>> 
>>> On Mon, Mar 14, 2011 at 9:16 AM,  <address@hidden> wrote:
>>> 
>>>> I've sketched this out using your suggestion above (calculating it once 
>>>> and returning the fraction for the called beam) - nevermind my previous 
>>>> question about redoing calculations. A new patch set is on-line.
>>> 
>>>> I still need to do the math for the longer slopes - I'll have time to do 
>>>> that later today or tomorrow.
>>>> 
>>>> In the spirit of the one-change-per-push idea, I'd like to push the fix to 
>>>> 1504 first before I push the change to feather-direction.  Does this seem 
>>>> like a good idea?
>>> 
>>> Do you mean: push an earlier version of this patch first?   I think
>>> it's not a good idea, because you would rewrite it directly after
>>> pushing, cluttering up the history of what is happening.  The idea of
>>> one-change-per-push is that all the individual changes are
>>> independent.
>> 
>> No, I mean that changing the feather-dir property from ly:dir to a pair 
>> seems like a different problem than fixing issue 1504.  It effectively adds 
>> a new feature to lilypond, and thus seems like it should be the object of 
>> its own patch/push.  However, if you think I
> 
> Ah right. My proposal was for feather-dir to be used to init
> feather-fractions (or whatever they're called.) - please do what you
> think is best, but if you are pushing 2 commits where the 2nd mostly
> rewrites the 1st, you might as well skip the 1st.

I'm going to push the first commit after people sign off on it and then work on 
the feather-dir bit (w/ the appropriate convert-ly rule).  It won't require 
many rewrites: it'll just require some more math in the calc_feather_fractions 
function.

Cheers,
MS