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
Implements caching for videos which dramatically speeds up presentation
generation. The generation for the Google demo for example goes from 20
seconds to 2 seconds.
Unfortunately, `typst query` cannot extract the slides from the input
file. It can only get the speaker notes. Since the video cache key
requires the slide itself too, I therefore wrote some manual logic to
extract the slide content. That's didn't work because (as always)
parsing is hard. In this case, it also needs to handle slides such as
```typst
#slide[
#set page(margin: 2em)
#set text(size: 30pt)
#align(left)[
```typ
#import "@preview/polylux:0.4.0": *
#set page(paper: "presentation-16-9")
#slide[
Hello
#toolbox.pdfpc.speaker-note("
This page contains Hello
")
]
```
]
#toolbox.pdfpc.speaker-note("
To create the presentation, we use Typst.
Typst is a new typesetting system that is similar to LaTeX.
Here for example is a simple Typst document with one slide.
The slide contains the text Hello
and a speaker note with the text This page contains Hello
")
]
```
without parsing the nested `#slide` as another slide.
Instead, the tool now uses the slide images for the cache key. This
should work since
```sh
$ diff 1.png _out/image/1.png
$ diff 1.png _out/image/2.png
Binary files 1.png and _out/image/2.png differ
```
Also fixes#14.
0 commit comments