Re: Regtests for 2.15.38 - one significant difference

[David Kastrup]

"Phil Holmes" <address@hidden> writes:

> See the attached PNG - is this deliberate?

Yes.  The change in output is deliberate and now corresponds with the
input.  Whether it makes sense to change the regtest because one usually
would not want to see the string numbers, is a different question.

The previous behavior was that string numbers were only displayed when
placed on notes _inside_ of a chord.  That made no sense and was
inconsistent, but it did mean that simple monophonic examples needed no
special measures for letting the string numbers disappear in Voice
staffs.

The change is documented in changes.tely and
notation/fretted-strings.itely:

commit 15ca7b18a32b243c7bcacfba773b3da843a6dd36
Author: David Kastrup <address@hidden>
Date:   Sun Apr 22 12:14:48 2012 +0200

    Adapt documentation for fretted strings to rhythmic engraver changes.

 Documentation/changes.tely                   |   39 +++++++++---------
 Documentation/notation/fretted-strings.itely |   55 ++++++++++++++------------
 2 files changed, 48 insertions(+), 46 deletions(-)

commit f0ec420a50f51112d7b79c219530bba9695433c9
Author: David Kastrup <address@hidden>
Date:   Mon Apr 23 15:50:36 2012 +0200

    Stop Fingering_engraver from actively ignoring stroke-finger-events.

 lily/fingering-engraver.cc   |   12 ------------
 scm/define-event-classes.scm |    3 ++-
 2 files changed, 2 insertions(+), 13 deletions(-)

commit 84cad110c3bbaeca8e5144dba7f8756b89100396
Author: David Kastrup <address@hidden>
Date:   Sat Apr 21 22:30:55 2012 +0200

    Let rhythmic-engraver make its articulation-or-event decision based on curre
    
    This removes the dependency of the rhythmic engraver on a static list
    of unlistened event classes and thus is part of work on issue 2449.
    
    As one effect, string numbers on isolated notes in Voice contexts are
    now typeset by default since the string numbers have no listener in
    Voice contexts even though they would have one in TabVoice.

 lily/dispatcher.cc              |   21 +++++++++++++++++++++
 lily/include/dispatcher.hh      |    1 +
 lily/rhythmic-music-iterator.cc |   15 +++++----------
 3 files changed, 27 insertions(+), 10 deletions(-)

and you'll see in the code review
<URL:http://codereview.appspot.com/6098050> it is explicitly listed as
one effect in the summary, and in
<URL:http://code.google.com/p/lilypond/issues/detail?id=2449#c3> and
similar Regtest comments, the changes are listed and mentioned as
being expected.  In
<URL:http://code.google.com/p/lilypond/issues/detail?id=2449#c8> I asked
for opinions whether a change of the regtests was desirable or not, and
much to my utter and complete surprise, nobody bothered voicing an
opinion on that.  Would you believe that?

So in short: this is deliberate.  Feel free to change it if it worries
you.  But after all, one purpose of regtests is to show changes in
behavior, and this _is_ a change in behavior.

-- 
David Kastrup