You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
feature request: turn off EUOUAE on for grenextscore #1758
In the screenshot you can see that I'm using a score where I have removed (or commented out, either way it's the same effect) the EUOUAE tags.
I would rather not clutter the directory. I have both bilingual sheets/booklets for a given Sunday or feast where I need the antiphon and a very much work-in-progress bilingual book that would involve page turns, and I've also copied the score from another directory that is Latin-only.
It'd be nicer to just search for one score file name without worrying about the appended no_euouae bit and then comment it out when the EUOUAE is not needed (for example, when there is no congregational chant at all or when the Magnificat is printed on the same page as its antiphon).
Something similar came up on the mailing list a long time ago and then I discovered a good use case for such a feature in building a booklet today.
Can you please clarify? For example, could you show what you would like the score/TeX to look like? Maybe \greseteuouae{keyword} where keyword is visible, invisible (takes up no space), or phantom (takes up space but does not print), as in #1683?
I think I would prefer invisible so the score prints at the last (::) and ignores everything after it; I don't personally see a use case for phantom since the will go to a new line if needed, which we probably don't want…
I second this request, but I think a different approach might be more widely applicable.
For instance, I experience the same inconvenience with the Euouae, but the intonation star (*) is also something that typically disappears the second time an antiphon is printed. So, like OP, I've been having to make two gabc files for each antiphon in a publication.
My proposal - and I have no idea if this is viable - is to make a tag in the gabc that could be toggled in tex. For instance, if we used <x>, then
could be invoked with something like \gresetfirstrepetitiontext{0} (I have no idea what to call it) to remove everything in the tags.
This, used in combination with \gresetinitiallines and firstsyllableinitial would improve QOL significantly for projects with repeated antiphons by eliminating the need for two gabc files.
I rarely print full antiphons twice in my own case, and I personally prefer doing something different than in real books (not using Ant. for an annotation, as I don't use a large initial for repeats). However, sometimes I do…and on those rare occasions I could be convinced to do something different if it meant only one file.
Personally I usually use * directly: taking that into account as well as the Gregorio-provided formats would work well, although those don't go in tags, and there's no need for that, as this would cause a lot of issues for existing scores.
I think the shorter the name with the fewer arguments the better: if that means multiple commands, so be it. Yes, one good thing about LaTeX is the number of arguments, but getting them right is such a pain…and David's proposal is short and clear.
I concede these points, and on a second reading, David's solution would work just as well; I would just need to put the asterisk in euouae tags.
That said, couldn't it be useful to have a second tag, so that the the visible / invisible / phantom functionality could be used independently of the <euouae> functionality?
Activity
davidweichiang commentedon May 21, 2026
Can you please clarify? For example, could you show what you would like the score/TeX to look like? Maybe
\greseteuouae{keyword}wherekeywordisvisible,invisible(takes up no space), orphantom(takes up space but does not print), as in #1683?MRoth1910 commentedon May 22, 2026
Sure.
Default should be visible.
I think I would prefer invisible so the score prints at the last (::) and ignores everything after it; I don't personally see a use case for phantom since the will go to a new line if needed, which we probably don't want…
fiat96 commentedon Aug 7, 2026
I second this request, but I think a different approach might be more widely applicable.
For instance, I experience the same inconvenience with the Euouae, but the intonation star (*) is also something that typically disappears the second time an antiphon is printed. So, like OP, I've been having to make two gabc files for each antiphon in a publication.
My proposal - and I have no idea if this is viable - is to make a tag in the gabc that could be toggled in tex. For instance, if we used
<x>, thencould be invoked with something like
\gresetfirstrepetitiontext{0}(I have no idea what to call it) to remove everything in the tags.This, used in combination with
\gresetinitiallinesandfirstsyllableinitialwould improve QOL significantly for projects with repeated antiphons by eliminating the need for two gabc files.MRoth1910 commentedon Aug 8, 2026
I rarely print full antiphons twice in my own case, and I personally prefer doing something different than in real books (not using Ant. for an annotation, as I don't use a large initial for repeats). However, sometimes I do…and on those rare occasions I could be convinced to do something different if it meant only one file.
Personally I usually use * directly: taking that into account as well as the Gregorio-provided formats would work well, although those don't go in tags, and there's no need for that, as this would cause a lot of issues for existing scores.
I think the shorter the name with the fewer arguments the better: if that means multiple commands, so be it. Yes, one good thing about LaTeX is the number of arguments, but getting them right is such a pain…and David's proposal is short and clear.
fiat96 commentedon Aug 10, 2026
I concede these points, and on a second reading, David's solution would work just as well; I would just need to put the asterisk in euouae tags.
That said, couldn't it be useful to have a second tag, so that the the
visible/invisible/phantomfunctionality could be used independently of the<euouae>functionality?