Vertical episemus on low notes ignored in vertical spacing

[rpspringuel]

When a score has a low note of a' or b' the presence of the episemus is ignored when determining the spacing between the notes and the text. This leads to the text overlapping with the episemus.

The problem appears to be in the first argument choices that Gregorio makes for \grefirstlinebottomspace. They should be:
0. note on first line or higher

1. note below first line (c)
2. note on 0th line (b) or note below first line with vertical episemus (c')
3. note below 0th line (a) or note on 0th line with vertical episemus (b')
4. note below 0th line with vertical episemus (a')

Cases appear to currently be:
0. note on first line or higher

1. note below first line (c)
2. note on 0th line (b) or note below first line with vertical episemus (c')
3. note below 0th line (a)

For consistency, case 1 should also be selected for note on first line with vertical episemus (d'), but my testing indicates that there is no spacing issue there right now.

[eschwab]

Is this still on the docket for 3.0?

[eroux]

Good point, as it might introduce an incompatible change, maybe it would be good to do it... I can do it (tomorrow), if someone wants to do it before, please do!

[henryso]

How incompatible a change? Does the user have to modify their gabc or tex files? If it's just a bug-fix, even if it changes the output (thus fixing the bug), then this can even be considered a patch release (i.e., 3.0.1).

Also, if I understand the bug description, it's a problem with the C code, right? That the C code chooses the incorrect value for the parameter?

Is this a long-standing bug? If it wasn't introduced recently (i.e., 2.4.2, 2.4.3, 3.0.0), then I'd vote to wait for the next release instead of rushing a fix.

[rpspringuel]

This should be a bugfix change. It will change scores visually, but there is no change in the underlying gabc or in a user's TeX. It is a long-standing bug. I discovered it while making my changes to the spacing code, but didn't fix it at the time.

Fixing it requires a change in the C code (to pick the correct value for the parameter) and making some modifications to \grefirstlinebottomspace in gregoriotex-spaces.tex. The latter needs a case 4 similar to what \gre@calculate@additionalspaces above it has. I don't know what changes need to be made for the former.

[eroux]

Oh ok, I thought incompatible changes would have to occur in the generated tex files, but if it's not the case, we can delay this feature.

[henryso]

As far as I can tell, the C code pas passing 4 for the first argument. I think the problem is that \grefirstlinebottomspace is simply not handling a case 4.

[rpspringuel]

Did anyone check to see that b' is getting mapped to case 3? Based on what I wrote when I found the bug, that wasn't happening (in addition to case 4 being missing).

[henryso]

I've confirmed that b' is mapping to case 3 in release-3.0 gregorio.