I agree that, if there's no animation, there should be no changes to styles of the element.

Also: The current fx_methods.js show/hide code changes the opacity in the case of a normal show() - I had a div with opacity set at .7, and it was re-set to 1.0 after the show().

I don't think my code is quite correct though.

Up for this. Zepto should not change anything else than display property. At the moment it breaks all css transitions applied to the element you want to show/hide.

Up for this. (2)

Same thing here! The 'transform' property realy cancels the 'fixed' positioned child elements, and for the 'absolute' positioned child elements, it acts like a 'relative' positioned element, preventing them to position themselves relative to their parent elements.

My suggested solution is this:

function hide(el, speed, scale, callback) {
  return anim(el, speed, 0, scale, function(){
    origHide.call($(this))
    callback && callback.call(this)
  })
}

function show(el, speed, scale, callback) {
  origShow.call(el)
  return anim(el, speed, 1, scale, callback)
}

$.fn.show = function(speed, callback) {
  return (speed === undefined) ? origShow.call(this) : show(this, speed, '1,1', callback)
}

$.fn.hide = function(speed, callback) {
  return (speed === undefined) ? origHide.call(this) : hide(this, speed, '0,0', callback)
}

Thus, by calling methods show() or hide() without the speed parameter setted, we exchanged only the 'display' property, and when the parameter is set, the event continues to be called the classic effect of element scale.

This breaks show with no speed given after a hide with an animation speed.

Perhaps the code could temporarily store the original opacity value before an animated hide, so that show can apply the correct value if no speed is given?

Any takers to implement this? If so, a unit test for fx_methods would be most welcome.

The show without speed should not set opacity 1.0 to element.