So a big question : what is the interest to let accessible a plugin that is dead ? A least it would be cool to mention that it doesn't work since the version number xxx...
It's not a big question, it's a poorly considered question. How are the Joplin devs supposed to know that a plugin is no longer functioning without having to test every plugin against every release or pre-release? For the plugin to be marked as "not working from version xxx" it would need the author to mark it as such and if they are no longer maintaining the plugin that is unlikely to happen.
This is what this forum is for. For users of the plugins to flag up any that no longer work and/or are no longer supported. It seems that this works as the plugin no longer appears when searched for in Joplin.
Just out of interest I did a few tests before the plugin was pulled and found that the last Joplin version I could get it to work with was 3.0.15 which was published in August last year.
You are so true...
It's exactly what I have done (forum flag) and what happen (no more plugin).
I installed the plugin yesterday for the first time in my life.
This morning the plugin was there and now : no trace of it.
So, I don't know who did this but we can say it's "thanks to me" (sad and poor victory !) that the plugin is no longer present. In fact I wrote to the author of the plugin (no answer yet). Maybe it's him...
BUT...
Someone already flag about this plugin 6 months ago (maybe others before him):
I saw this topic (I gave the link in my previous post) after installing the plugin : it's only when I realized it's not working that I started a search on the forum.
So : I will not be the first and I will not be the last ( to try a plugin before searching information about it).
Some people were disappointed:
It means we have to find a way to avoid this kind of situation again. Forum in itself is not enough.
What do you propose ?
I think it's better to indicate a plugin is no longer working since a specific version of JOPLIN than to simply remove it (if there's no harm). First because some people will keep an old version of JOPLIN in order to have this plugin still working because they are really in need of it. So they need (in case) to have the plugin still accessible.
Second because other devs (than the original dev) will know the plugin exists and maybe will want to improve it and make it work again. It avoids to waste the work of the original dev and it avoids new ones to start from scratch...
I don't propose anything. You are the one that says the forum is not enough. What ideas do you have?
It's a nice idea if you are not the person implementing it. Who is going to to do that work and test every plugin against every release? I see no other way of finding out what Joplin version it is when a plugin stops working. Who is going to chase up the plugin developer to ensure that the project is actually "dead"? That is not a small amount of work.
I think that the easiest and least costly option is when a person finds a plugin that no longer works and appears abandoned they should just create a post with their findings and nominate that plugin for deletion. That way it is clear to the Joplin devs what is being suggested, rather than being a post on an existing thread that they may not see.
Whilst I do see your point about not wasting the work that has been done on these "dead" plugins I don't really think that holding back on upgrading Joplin is something that should be encouraged.
Unfortunately this is the problem with having a third-party plugin system. Plugin devs may abandon their project and unless someone else steps up or the Joplin devs feel that the plugin is useful enough to support themselves, that is it.
I have created a new emoji plugin for the markdown editor. Initially I created the plugin for myself but after reading this thread I decided to publish it the plugin. You can download the JPL file here or you you search for "Markdown Emoji Autocomplete" in the plugins section of the settings.
Normally we try very hard not to break any plugin, by ensuring that the API remains backward compatible and by deprecating rather than removing methods.
The one exception was the move from CodeMirror 5 to CodeMirror 6 for the Markdown editor. Even in that case a compatibility layer was developed and many plugins still worked fine, but some may have slip through the cracks and remain incompatible, including maybe this Emoji one.
In that case it's still possible to revert to the old Markdown editor if you prefer, although it means that newer plugins won't be compatible.
Another exception is the macOS theme plugin which is modifying the CSS and for this we don't currently provide a stable API, so it often breaks between versions.
