bug#50483: [R] Package r-shiny: Shiny server logic inactive

[zimoun]

Hi Charles,

On Thu, 16 Sep 2021 at 05:57, Charles <charles.b.jackson@protonmail.com> wrote:
> Hello, I am responsible to updating the minify build syste to use the
> updated uglifyjs (now node-uglify-js found in
> gnu/packages/uglifyjs.scm). When updating I noticed that some R
> packages had the input uglify-js (the old lisp one). I chose not mess
> with them at risk of breaking. What I failed to realize is that some R
> packages (like r-shiny) actually use the minify build-system to do the
> minifying, thus trying to call the updated uglifyer, while only
> importing the new one. Other R package like (r-shinytree) do not use
> the minify build system and instead directly call uglify-js using
> open-pipe* (these are the ones I didn't want to break). 
>
> I would highly suggest against using the old lisp uglifyer since it is
> outdated by 6 years, and if anyone decides to write some modern
> javascript, your package will break. Instead simply update the input
> to ("node-uglify-js" ,node-uglify-js) and don't forget to use module
> (gnu packages uglifyjs). For the packages like r-shinytree, I would
> suggest having them use the minify-build system like r-shiny (and also
> updating to the newer node-uglify-js). I have tested updaing the
> input, and it works for the ggplot2 example stated earlier in this
> thread.

Thanks for explaining.

It is “broken” since:

   CommitDate: Mon Jul 19 14:46:28 2021 +0300

so I propose to revert now or wait a couple of days for a fix.

Efraim, WDYT?

> Again I'm really sorry for not catching this. It is very late where I
> am, so if it isn't clear how to update, I can do it tomorrow.

Do not worry.  That happens. :-) The good is that the situation shows
some limitations and how to improve.

Cheers,
simon