\version "2.11.31"

\header { texidoc = " If the music is broken on an invisble bar (dividing in the middle of a normal bar), the bar number should be printed in parentheses.

Currently, no bar number is printed. " }

{ c'2 c' c' \bar "" \break c' c' c' \break c' c' }

\paper { ragged-right = ##t }

% EOF

Adds parenthesized measure numbers at mid-bar breaks.

http://codereview.appspot.com/5452057

Adds parenthesized measure numbers at mid-bar breaks.

http://codereview.appspot.com/5452057

Could it be possible (as an option) to print it in parentheses but also in italic?

I do not know if this is "standard practice" but IIRC I have seen it this way somewhere.

I'm not sure if it's standard practice either...

The easiest way for me to do this would be to figure out all of the different options and then roll them into one function.

Currently, I've heard: --) italic --) parenthesized --) not visible at partial measures

It's easy to create booleans for all of these things in a central function and then create wrapper functions (italic-parenthesized-at-partial-measure, parenthesized-at-partial-measure, italic-at-partial-measure, not-visible-at-partial-measures, etc.).

Once I get a solid list, I'll build it into the function.

Patchy the autobot says: patch does not apply

Adds parenthesized measure numbers at mid-bar breaks.

http://codereview.appspot.com/5452057

Patchy the autobot says: LGTM. in tuplet-broken.ly, I see a bar number on the third system but not the second. I assume that's correct?

I'm not sure if it's correct. The first measure number is normally never printed in a score, but I'm not sure about the first parenthesized measure number.

Cheers, MS

I suggest to interpret the rule as don't print a bar number at the beginning of the first system of a piece'. Consequently, I favour that we have(1)' in the second system for a broken first bar.

Adds parenthesized measure numbers at mid-bar breaks.

http://codereview.appspot.com/5452057

Patchy the autobot says: does not apply to master

Adds parenthesized measure numbers at mid-bar breaks.

http://codereview.appspot.com/5452057

Patchy the autobot says: LGTM. Patchy isnt going to reject this, but as a reviewer I want this behavior turned off by default. Seeing regtests with no bar numbers adding a bunch of (1) looks weird and wrong.

I don't disagree. But: How many regtests do we have with a single broken bar? Shouldn't we rather disable bar numbers in those tests?

I've deleted the regtest by now, but there were something like 5 regtests effected. The single broken bar didn't disturb me as much as

======

(1)======

======

(1)======

======

(1)======

that was present in Mike's favorite regtests of feathered beams. Come to think of it, shouldn't the parenthesized measures be disabled if you're not printing bar numbers?

Of course! Everything else is a bug. Mike?

Adds parenthesized measure numbers at mid-bar breaks.

http://codereview.appspot.com/5452057

Patchy's ability to communicate like a human, analyze musical information, and make musically-informed commentary is astounding.

New patch set up making this not a default.

Well, although I ended up dropping the artificial intelligence class, I've been having a blast with the machine learning class. http://www.ml-class.org/

BTW, it's running again in the spring, and I highly highly recommend it. Andrew Ng is an excellent professor (and I say this as somebody who spent 8 years full-time undergraduate, 1 year part-time undergraduate, and 2 years Masters -- I've seen a lot of professors). And it's free!

Patchy the autobot says: tuplet-broken.ly goes from nothing to (2). Look, this is supposed to catch unintentional problems. Are you testing these at all yourself?

Adds parenthesized measure numbers at mid-bar breaks.

http://codereview.appspot.com/5452057

Patchy the autobot says: LGTM. no changes is good changes

Comments by Reinhold on Rietveld

All the comments from Reinhold (and Neil) were addressed in the patch set from Dec 8. Is it ok to push the patch?

Cheers, MS

Yes, LGTM.

Just to make sure:

1. It will be possible to set the visibility of such "bar numbers for broken bars" independently of the visibility of bar numbers for normal (unbroken) bar?
2. It will be possible to print bar numbers for broken bars both in parentheses and in italic?
3. I do not see on Rietveld additions to the Doc; this new feature should be documented (as a snippet) in NR 1.2.5 Bars > Bar numbers IMHO. Please do not tag this issue as fixed unless it is documented or a new issue, type Documentation, has been added regarding this.

Thank you Mike!

Ignore Xavier's point 3. As per our usual patch-handling process, mark this fixed as soon as it is pushed (after another countdown). New features show up on a git diff of regtest directory.

Fixed as 123b51c50a462b14d3cbffe2f354eaa8ec772dc2.

1. It will be possible to set the visibility of such "bar numbers for broken bars" independently of the visibility of bar numbers for normal (unbroken) bar?

Yup

2. It will be possible to print bar numbers for broken bars both in parentheses and in italic?

Yup

The advantage of using a formatter function is that you can bend it however you want. Just copy and paste the function from translation-functions.scm (robust-bar-number-function) into your .ly file and tweak it however you like. If it gets to the point where users are using a specific tweak all the time, it'd be worth it to add that to LilyPond proper.

I've verified that the default behaviour is not to add parenthesized bar numbers, see image. I had a look at the patch and could not see how to turn on the new behaviour.

If one of the devs could advise how to turn on the enhanced behaviour, we can verify it.

hmm, there isn't even a regtest for this feature! Mike, could you add one?

If users cannot use the enhancement, then I suppose it is not fixed.

The only doc for barNumberVisibility is http://lsr.dsi.unimi.it/LSR/Item?id=592 so we could expand that. There were never any regression tests, so we could also expand ‘bar-number.ly’

There choices seem to be barNumberVisibility = #first-bar-number-invisible % #all-bar-numbers-visible % #(every-nth-bar-number-visible n) % #(modulo-bar-number-visible n m) % #first-bar-number-invisible-save-broken-bars % #first-bar-number-invisible-and-no-parenthesized-bar-numbers

User-oriented documentation is a completely distinct issue from new code features. If you don't trust a diff of regtests to spot such new features, then add a separate issue for that.

However, we do need a regtest, so I agree with marking this as not fixed.

Adds barNumberVisibility regtest

http://codereview.appspot.com/5501088

Patchy the autobot says: does not apply to current master

Should be "fixed", it was pushed months ago.

Mike could you do what Patchy wants in order to close this issue? Should I ask to open a new issue on the tracker in order to document this in NR 1.2.5 Bars > Bar numbers ? The regtest "bar-number-visibility.ly" is almost self-explanatory.

Thank you!

OK, marking as Fixed and Verified. Documentation should be added soon-ish; I'll open another issue if it isn't.

Fixed with 8d8982cc88ba1d0b20df5aff22b9555f76976a26.

Verified that a couple of commits were present in the repo:

For the new bar-number-visibility.ly regression test:

83c269ea982eea8b8b97c021844b3dfa91b56649

And for the code changes 8d8982cc88ba1d0b20df5aff22b9555f76976a26

Looks good.

Regarding comment #38: "Documentation should be added soon-ish; I'll open another issue if it isn't.". A cursory glance at the manuals does not reveal any documention. If documentation and/or an issue have been added, please add a reference to them to this issue so that people might find them when looking here.

Thanks