Re: Issue 2148 in lilypond: vertical skylines should use stencil integrals

[David Kastrup]

"address@hidden" <address@hidden> writes:

> On Mar 1, 2012, at 4:12 PM, address@hidden wrote:
>
>> 
>> Comment #90 on issue 2148 by address@hidden: vertical skylines should
>> use stencil integrals
>> http://code.google.com/p/lilypond/issues/detail?id=2148
>> 
>> If transparency is not supposed to have an impact anymore, why do we
>> still have different callbacks for "maybe-transparent" stencils even
>> in the latest iteration?  If this is dead code, it should be
>> removed.  There should only be one callback, and it should not look
>> at transparency.
>> 
>
> Because I haven't pushed my tabVoice code yet.  It doesn't have an
> impact in the regression tests you're talking about, but it would have
> an impact in all of the tab regtests.

Yes, it has an impact on the tab regtests.  That's how I noticed that
transparency _has_ an impact still.

> Once I push the tabVoice patch, I'll be able to get rid of this code.

Please don't let your code attempt to mask unrelated bugs.  It makes it
harder to understand and review and means that the unrelated bugs will
get harder to diagnose and fix _properly_, and that the interactions
with your code _after_ they get fixed get very hard to find and remove.

It makes it likely that the need for a proper fix for the original bug
will get overlooked, and it makes it likely that the necessity of
removing the workaround afterwards will get overlooked.  And it
complicated the review and makes it likely that it gets entangled in
code that should not be there anyway.

It is a triple recipe for trouble, and that only in order to make the
regtests appear better temporarily for no good reason.

-- 
David Kastrup