Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Timmy Willison
    @timmywil
    I skipped jquery/jquery#4466
    I'm fine with this.
    Time to say goodbye to my grunt plugin
    Michał Gołębiowski-Owczarek
    @mgol
    :D
    one less thing to remember, we had discrepancies between versions from time to time
    Timmy Willison
    @timmywil
    are the issues we had with package-lock.json still applicable? Also, what were they again? I recall a discrepancy between Windows and Mac.
    although, choosing npm over yarn isn't necessarily something I want
    Michał Gołębiowski-Owczarek
    @mgol
    yeah, that, plus from time to time the file was being changed when npm was updated
    so it wasn't stable between npm versions
    Timmy Willison
    @timmywil
    right
    Michał Gołębiowski-Owczarek
    @mgol
    and people use different Node & npm versions so it was all pretty unstable
    in my experience Yarn is way more stable with its lockfile format, for example
    Timmy Willison
    @timmywil
    that's pretty much why I use yarn
    Michał Gołębiowski-Owczarek
    @mgol
    me too :D
    Timmy Willison
    @timmywil
    that everything?
    Michał Gołębiowski-Owczarek
    @mgol
    plus, the thing between OSes is that e.g. fsevents is an optional package and only works on macOS so it's skipped on Windows. That means it doesn't get to package-lock when npm is run on Windows
    so Windows-based contributors would be unable to update dependencies, etc.
    @timmywil there's one PR where I need @gibson042's input
    Timmy Willison
    @timmywil
    ok, I trust he'll see the ping
    Michał Gołębiowski-Owczarek
    @mgol
    but I could use your opinion as well, Timmy
    jquery/sizzle#452
    the original PR: jquery/sizzle#439
    Timmy Willison
    @timmywil
    so it doesn't work with template selections
    Richard Gibson
    @gibson042
    I'm not sure what to think of it with the new information
    Michał Gołębiowski-Owczarek
    @mgol
    and the underlying issue is nodes inside of <template/> have the document fragment as their ownerDocument and that has a null documentElement
    Timmy Willison
    @timmywil
    wait, wouldn't the nodeType be 9?
    in that case
    I mean 11
    I wonder if we could check for 11
    Michał Gołębiowski-Owczarek
    @mgol
    nodeType is 9
    Timmy Willison
    @timmywil
    it's 9 on the document fragment and has no document element?
    Michał Gołębiowski-Owczarek
    @mgol
    yup
    Timmy Willison
    @timmywil
    that's stupid
    lol
    Richard Gibson
    @gibson042
    it seems like we generally want to catch the document fragment in the general case, but keep traversing up when it corresponds with a <template>
    Timmy Willison
    @timmywil
    I shame this fragment
    @gibson042 sounds about right
    I'm surprised the nodeType is 9
    Michał Gołębiowski-Owczarek
    @mgol
    right, it's weird...
    how to know we're inside of a template element, though?
    Timmy Willison
    @timmywil
    anyway, I guess there are other creative ways to accomplish what gibson said
    Michał Gołębiowski-Owczarek
    @mgol
    so that optimization got in but was accidentally reverted
    so what we currently have on Sizzle won't break jQuery tests
    therefore, this issue is not critical
    Timmy Willison
    @timmywil
    ok, I'm not sure the answer yet, but not breaking tests is good
    I think that's all for today. Thanks all!
    Michał Gołębiowski-Owczarek
    @mgol
    thank you!
    @gibson042 do you want to have a look at jquery/sizzle#453 before merging?
    Richard Gibson
    @gibson042
    yes please
    Michał Gołębiowski-Owczarek
    @mgol
    cool