Re: shortening a stem

[Janek Warchoł]

Hi,

(moving to -devel)

2013/4/8 David Nalesnik <address@hidden>:
> On Mon, Apr 8, 2013 at 3:37 AM, Werner LEMBERG <address@hidden> wrote:
>> Interesting.  Compiling your file with release/2.17.15-1-6-g0e77b32,
>> the stems disappear, regardless of the used offset value.  Something
>> apparently has changed internally.
>
> Yes, something has changed--I'm not sure what,  Unfortunately, I'm terribly
> short on time this week, so I can't look into this further.  I've attached a
> newer version of the function.  It works fine with versions up through
> 2.17.13, giving warnings where they are expected.  (There is provision for
> checking for whether offsetting a particular property fits the allowable
> cases.)  I've inserted your snippet.  With more recent files, you'll get a
> host of warnings concerning 'Y-offset of various objects including
> OttavaBracket.  *Arghh*

I've found the commit responsible for changing the behaviour of stem length:

commit 74e4d219b24ec6d6f28d663c0285418e6c8e122e
Author: Mike Solomon <address@hidden>
Date:   Tue Mar 5 21:03:55 2013 +0100

    Uses only unpure-pure containers to articulate unpure-pure
relationships (issue 3199)

It changes the default definition of Stem.length to be an unpure-pure
container.  I'm also pretty sure it's responsible for OttavaBracket
warnings, because it changes its Y-offset similarly.
Mike, can you comment on this?

best,
Janek