wise token
Please visit at https://github.com/pacaunt/typst-presentate/tree/main
It is still under development though, especially the syntax is very long, and there is no theme currently. However, I think that no states and counters is a way to escape the layout did not converge within 5 attempts warning, heavily happening on current package for creating slides (except touying, btw).
finite valley
Even touying have these converge issues?
wise token
Yeah, although it will be fixed on the next release, said by the author.
By the way, another key point of creating this project is to make the helper functions more flexible. So that I can modify more dynamic content, such as alternatives with change the current content by applying some function on it.
hasty hound
Yeah, although it will be fixed on the next release, said by the author.
Actually, that was a small bug that was introduced to implement a fake frozen state, and it was an easy problem to fix. So you can think that actually touying doesn't have this problem
To understand this, have you considered using numbering in your presentation? For example, here
https://github.com/touying-typ/touying/issues/38
so touying is also a "package that implements dynamic without states and counters"
However, I think it's still a great trial. Nice work.
wise token
so touying is also a "package that implements dynamic without states and counter...
In its maximum performance, yes. I feel like touying package really use the property of type "content" very very efficiently.
But the frozen is a very challenging issue, since it directly deals with typst internal states and counters, in my opinion. One day I will try that!
viral roost
Very nice! It's great to have an implementation of pause that is completely state-free.
I think most convergence issues you see in minideck are due to a change in Typst 0.12 (it will be fixed in the next release) but there are certainly still cases where there will be problems due to minideck's use of states.
I'm thinking now of offering something like your version of pause in the next minideck version. I have to think a bit more and play with it a bit...
wise token
To understand this, have you considered using numbering in your presentation? Fo...
I tried to mimic the frozen states by using the following approach: update the states and counters to the same value as it was in the beginning of the presentation. Does this count? https://github.com/pacaunt/typst-presentate?tab=readme-ov-file#numbering-and-frozen-states
viral roost
I tried to mimic the frozen states by using the following approach: update the ...
I think it's similar to something I tried, and also similar to what touying is doing in the current release (before the fix that will be on the next release). For me this method doesn't work well: as I increase the number of slides that use frozen states I always get convergence issues. Just 5 slides with equation + pause + equation (so 10 subslides in all) is enough to trigger the problem. So if it works for you that's great and I'll have to investigate why it didn't work in minideck
(I also tried with the same method as what touying does after the fix, and I still have these convergence issues in case of many slides)
wise token
Ah yes it was problematic when the number of slides increases too...
hasty hound
Yes, I mean touying will fix this in the next release.
https://github.com/touying-typ/touying/pull/124
viral roost
Yes, I mean touying will fix this in the next release.
https://github.com/touyin...
I tried that same approach but for some reason it's not helping much in my case
wise token
I find another way to implement the frozen counters that are updated via step() using calculations. Please check at https://github.com/pacaunt/typst-presentate/tree/dev with the example 3.
wise token
I tried that same approach but for some reason it's not helping much in my case
After some trial-and-errors, as long as the updates depend on themselves, typst will try resolving the .get() value even in the same location. That's why the warning still comes
viral roost
I find another way to implement the frozen counters that are updated via `step()...
thanks for sharing that, it works well in my testing, I'll try to do something similar in minideck
viral roost
After some trial-and-errors, as long as the updates depend on themselves, typst ...
you mean that s1.update(s1.at(loc)) is treated differently from s1.update(s2.at(loc)), because the first version will trigger a new iteration even if the s1 value at loc is not changed by the update?
wise token
yes, that's it
viral roost
interesting, thanks again
wise token
no problem!
viral roost
I wonder if that's something in typst that could be improved easily