search

darsain / sly

JavaScript library for one-directional scrolling with item based navigation support.
http://darsa.in/sly
2.87k stars 497 forks source link

added main to package.json #237

Closed capaj closed 8 years ago

capaj commented 8 years ago

so that package managers like JSPM or browserify or Webpack can be used with this package

capaj commented 8 years ago

also maybe publishing on npm would probably be a good idea since NPM is not just backend anymore-that is why I changed the name, because 'sly' is taken on npm...

darsain commented 8 years ago

Kinda same as #143, #188, and #216, even though they asked for bower.

The state of this library is: Sly needs to be rewritten, scaled down, more modular, better extensible API, ...

Tons of inlined crap needs to be offloaded to smaller libraries. Right now there is almost no difference between src/sly.js and dist/sly.js, but that will definitely change, so build step will be necessary to consume it.

I also want to never ever have to deal with dist folders in repositories again. They have no place there.

These two changes would make it no longer consumable by npm, unless I completely migrate to npm for all of the smaller libraries. Currently all of my stuff is either a fugly standalone abomination, or a component consumable by duo, which I consider to be the best package manager/builder available atm. Unfortunately it didn't win the popularity contest, so none is using it :sob:.

It seems that sooner or later I'll be forced to migrate and implement the (in duo absolutely unnecessary) step of publishing to npm, but until there is a good npm->font end build tool where assets like css, fonts, images, and other are a first class citizen and not only a hacky afterthought (like in browserify, jspm, and webpack), I will continue just staring into the wall, which is what I'm doing right now :neutral_face:. Hopefully https://github.com/makojs/core will be the thing.

Needless to say, I'm very irritated by the current front end package manager situation, and until it resolves I feel unmotivated to work on anything front end related. Component/Duo seemed like such a great solutions, but people decided to go with half crippled tools like bower, browserify, etc.

And when I finally migrate stuff to npm, I will definitely not let myself commit a sin of appending .js to the library name. It will definitely be namespaced as @darsain/sly :smile:.

So, until a foreseeable future I'm closing this. Sorry for any inconvenience this will cause. Sly in its current state isn't very good for what it tries to do anyway. Too big, too bloated, too complicated, with unresolved edge cases.

capaj commented 8 years ago

@darsain thanks for the wriiteup. I have few points I would like to make, even though I am not sure I can convince you in any of these:

  1. I think sly is the best library for making carousels ever created. Point me to a better one- I dare you. Yes there might be few edgecases unsolved, but I haven't really had them over the course of last 3 years using it, so they don't bother me as much.
  2. I agree with you on webpack and browserify being hacks, but why do you think JSPM is a hack?
  3. I totally feel your pain. I think duo was very interesting-I tried it on one or two small projects. I am currently using JSPM for my javascript, but before I settled on that I tried almost every solution. The reason why I think JSPM is on the best course to winning this pissing contest is that systemjs shadows https://github.com/whatwg/loader and therefore should be as close to what we will have in the browsers once the loader is in the browsers natively. That is also the reason why I believe it is necessary to spread it popularity everywhere I can. The sooner we uncover bugs and missing features in the spec, the better.
  4. I totally agree with you on the sly needing the refactor. I don't see how it relates to this PR, but ok.
  5. Sorry for appending '.js' in the name.