It makes me notice that one can copy almost any HTML in the editor, which is a bit weird, and full of bug (and maybe of security risks?). I think we should add a listener for onpaste and filter what is pasted (to only allow images or simple text)

from what I see, any html pasted is surrounded with <code>. There are no more risks than with the previous editor. Filtering/modifying what is pasted might be nice in some situations, sometime one just want to put some html elements not supported by plain markdown

Codecov Report

Merging #483 into master will increase coverage by 0.22%.
The diff coverage is 57.89%.

@@            Coverage Diff             @@
##           master     #483      +/-   ##
==========================================
+ Coverage   24.64%   24.87%   +0.22%     
==========================================
  Files          65       65              
  Lines        7120     7196      +76     
==========================================
+ Hits         1755     1790      +35     
- Misses       5365     5406      +41

tested with #486, the md tag is not interpreted as it is inside some html. Maybe I should put plain html instead of markdown when using cw

I think Markdown in HTML never worked. At least it doesn’t on GitHub:

So yes, going with HTML for everything seems to be a better solution. Ideally, we should have some kind of custom Markdown like ![132] where 132 is the idea of a media, and then Plume just replaces it with the full HTML code, but this is probably quite hard to achieve.

It might not be that hard actually, we “just” have to match with the Event Start(Image(_, _, _)) and map it to our custom html

User should input something like ![some title](123)
If the corresponding picture has a cw, it display this
And after clicking on it, it shows the picture (this is a toggle, so clicking again hide it back).

Hum, there is a small bug if the same picture with cw is present twice, clicking on the second will only activate the first one. I shall probably add some random to the generated id

it’s not from your experiments. It seems to came from the rule adding a 3em margin on main article img