Dear developers,
The positions of the commas under ‘Holds and Pauses’ are not propagated.
It is intended if I correctly remember.
However, it would be more productive if the position is propagated. Please look at the following examples:
Score:
I had the same problem and I give a thumb up, that they should be propagated …
From a screenshot it’s difficult (impossible?) to know exactly how you’ve notated the end of the bar, with that gliss/port to a grace-sized note.
On the other hand, I can see from the attachment line that the comma is attached at a semantically unhelpful position.
What happens if you place the comma, using the caret, at (or close) to the end of the measure?
It’s very much possible that there is something funky going on, but, to be honest, it seems to me that Dorico is doing exactly what you told it to. The comma being close to the barline is perceptually salient to you, but not to a computer, as the offset is measured from the source, so it might just be possible that the same offset just doesn’t work for both layouts.
I echo Leo’s suggestions. Try first moving the comma to another rhythmic position by the end of the bar first. If that doesn’t yield any improvement, maybe consider renotating the passage with a hidden tuplet to place what I assume is a grace note shown before the barline in that very same bar and attaching the comma there.
Thank you for your answers.
I already tried the method suggested by pianoleo and LSalgueiro, and the followings happen:
I know the notation is not common. The long notes must slightly be raised by sliding. In other places, there are enough rests. There is only one weird situation where I used at the semantically unhelpful position… I could, of course, begin the sliding semiquaver earlier, but I do not want it. The following is the complete scene of these:
I would definitely pity the performers when the audience hears this and wonders why half of them are playing out of tune 
I suspect the issue is that, as far as Dorico’s concerned, the grace notes are attached to the following best. They don’t really have any grid time in their own right, so it’s not possible to put a grace note semantically between the grace note and the barline. I’m not sure quite what you can do to bypass this problem, short of finding a different way to notate the grace note (such as with a real note in a different voice, scaled.)
I was a bit elliptical in my last post, so I’ll flesh it out here: use a hidden tuplet to place the grace note. Tacking on an extra 16th note and it shouldn’t throw the note spacing off balance — so, something like 3:2 16ths in the last 8th.
(Also, I’m not sure how you expect the clarinets to do that in that register.)
Dear LSalgueiro,
Thank you for your idea! It could be a solution, but it needs more working steps.
I asked a clarinettist about it. He says that he can this. In some books I had read, sliding downward is strongly recommended if I still correctly remember. However, changing the lip position can make a small amount of upward sliding possible. The problem is that the clarinettist who will play the piece is not the same person whom I asked bout this.
