Re: Issue 5705: int->long in Stem::get_beaming and set_beaming (issue 561350044 by address@hidden)

[David Kastrup]

address@hidden writes:

> sorry to complain late about this change. I understand that this gets
> rid of a conversion warning, which is something that we want, but I am
> missing the big picture here. Is there a plan for the big picture?

As far as I can see the big picture is matching the used variables to
the used data types to that overflows become a non-issue.  LilyPond
already falls flat on its face for several "new complexity" kinds of
scores.  Baking human assumptions into the code and silencing conversion
warnings without addressing the underlying potential cause makes them a
lot harder to find and harder to exclude them as a cause of problems
than if one just keeps using the default sizes.

Just choosing the right types means that when a problem crops up likely
due to an exceeded size, this is not a place you need to check up on and
verify when going through all warning-silencing casts.

-- 
David Kastrup