Dupe loads of the jQuery JS crap; one set is broken links

Discussion in 'Feedback/Support/Suggestions' started by Wandering Aramean, Sep 15, 2011.  |  Print Topic

  1. Wandering Aramean
    Original Member

    Wandering Aramean Gold Member

    Messages:
    28,220
    Likes Received:
    61,766
    Status Points:
    20,020
    Given the apparent lack of interest from the site admins in creating a light-weight version of the interface that isn't burdened with all the client-side javascript junk I spent a bit of time this afternoon looking at the page source and the options in AdBlock Plus to see what I could do about getting it to load faster all on my own. I came across an interesting discovery. Each of the javascript loads is being called twice in the page source code.

    Even more interesting is that one of the calls is to a URL that I'm almost certain is not actually valid links. Here's what I'm seeing:
    http://milepoint.com/forums/js/xenforo/xenforo.js?_v=21d5099e
    http://milepoint.com/forums/forums/feedback-support-suggestions.46/js/xenforo/xenforo.js?_v=21d5099e

    The second entry returns an error when I try to load it.

    Getting rid of the bogus call will almost certainly improve server performance and client load times.

    And for anyone interested in a set of filters to make the pages perform MUCH faster without all the client-side junk drop me a line. It has made my MP experience significantly better.
     
    iolaire likes this.
  2. HaveMilesWillTravel
    Original Member

    HaveMilesWillTravel Gold Member

    Messages:
    12,504
    Likes Received:
    20,199
    Status Points:
    16,520
    Is the js not getting cached by the browser?
     
  3. Wandering Aramean
    Original Member

    Wandering Aramean Gold Member

    Messages:
    28,220
    Likes Received:
    61,766
    Status Points:
    20,020
    The performance problem is in the time it takes for all the client-side rendering to happen, not in the loading of the page. Watching the page fully load and then all the timestamps change, for example, is just one of many things that I'd rather not waste time on.
     
  4. Randy Petersen
    Original Member

    Randy Petersen Founder

    Messages:
    2,731
    Likes Received:
    15,136
    Status Points:
    10,520
    While I take exception to "apparent lack of interest from the site admins ..." there's actually plenty going on in that area and we are quite close to releasing the first release of both performance improvements and new features. Considering just five months into v1.0 there seems to be many things that are working well for our members.

    Among the performance changes which should contribute to even better performance is minified CSS and in an attempt to decrease the number of HTTP requests made (and thereby speed up pages), virtually all of our next release UI icon graphics come from just two sprites. And minified javascript is awaiting to be deployed as well.
     
    iolaire likes this.
  5. Tim
    Original Member

    Tim Admin

    Messages:
    445
    Likes Received:
    428
    Status Points:
    525
    Thanks for the report Wandering Aramean, I can't reproduce the slow page load you report, but in looking into your report of double loading of some resources I did find that for some reason, some resources are attempting to load more than once via different paths. We will be looking into this, and it does increase the number of requests per page which is not good, but doesn't result in the actual duplicate download of the resources. Thanks again for the report.
     
  6. Wandering Aramean
    Original Member

    Wandering Aramean Gold Member

    Messages:
    28,220
    Likes Received:
    61,766
    Status Points:
    20,020
    Sorry you take offense to the suggestion but there has been zero demonstrable progress that I've seen over the past 6 months to repeated requests for a "light" or mobile version of the site. Instead we got an iPhone and Android app which doesn't help the poor performance. And the discussion of those apps has taken over the thread titled "Mobile Site" (http://milepoint.com/forums/threads/mobile-site.16396/) so as a user there doesn't seem to me to be much action on this front. I'm not saying that other things aren't happening but this doesn't seem to be one of them.

    Back in March it was suggested that a mobile site was "on the list." http://milepoint.com/forums/threads/answered-slooow-not-just-today-always.5408/#post-210849
    In fact, it was apparently "high on the list." http://milepoint.com/forums/threads/answered-mobile-page.5657/
    In April the announcement was that there was an option but it wouldn't be implemented. http://milepoint.com/forums/threads/answered-down-arrow-wanted.8358/#post-320979
    In May a discussion was had amongst the founders and a test mobile skin was deemed a reasonable solution to the issue users are facing but that has not, to my knowledge, been implemented. http://milepoint.com/forums/threads...-mobile-apps-for-milepoint.10132/#post-418805

    I know you're an iOS guy and you love your devices but not everyone else is. If the mobile skin isn't coming soon then stop pretending it is. Telling us one thing and then not delivering it sucks.

    As for the speed issues, Tim, try a NetBook. You'll see it quite readily. It is all the client-side javascript crap, like changing the time stamps from absolute times to relative times, which churn processor cycles and cause redraws of the screen. This thread (also from March) best describes the issues: http://milepoint.com/forums/threads/answered-slooow-not-just-today-always.5408/#post-211593. Of course, it is marked as "answered" without considering that the problems still exist.

    Adding these three lines to ABP has changed the load time of pages significantly:
    ||milepoint.com/js/glossary-hover.js
    ||milepoint.com/forums/js/xenforo/xenforo.js?*
    ||milepoint.com/forums/cometchat/*

    I can easily reproduce an 8+ second load time with those three lines on and about 3 seconds with them excluded. It is a serious difference.
     
  7. HaveMilesWillTravel
    Original Member

    HaveMilesWillTravel Gold Member

    Messages:
    12,504
    Likes Received:
    20,199
    Status Points:
    16,520
    Wow. 8 seconds? I am using the previous generation of the MacBook Air (2.13 GHz Core 2 Duo chip) most of the time, and while I haven't measured the load time (which page anyway?) I don't think I have seen MP be that slow except maybe when I was on a really slow network or MP had problems. I am using Chrome 13 most of the time.

    A quick look at Firebug tells me that

    - there were 54 GET requests issued for http://milepoint.com/forums -- seems a bit high. 5.3 s load time

    - there were 42 GET requests issued for http://milepoint.com/forums/categories/airline-programs.115/ -- also high. 2.18s load time.

    - there were 52 GET requests issued for http://milepoint.com/forums/forums/united-mileage-plus-including-continental.99/ -- 2.57 seconds load time

    Don't have time right now to look further, but just to be clear, I would love to see performance improvements as well since I often use my iPad browser, which is obviously slower than the MB Air.
     
  8. Wandering Aramean
    Original Member

    Wandering Aramean Gold Member

    Messages:
    28,220
    Likes Received:
    61,766
    Status Points:
    20,020
    It isn't just the GET requests. The client-side stuff that the ABP filters I added kills is also pretty brutal. This is on my mid-level Toshiba, running W7 with 4GB RAM. Running on a NetBook with 1.5GB is an experience barely worth tolerating. :oops:
     
  9. HaveMilesWillTravel
    Original Member

    HaveMilesWillTravel Gold Member

    Messages:
    12,504
    Likes Received:
    20,199
    Status Points:
    16,520
    I'll try a slower machine later, though the MB Air with Core 2 Duo isn't exactly a CPU powerhouse either. Might also be worth looking at the site with YSlow, in addition to reviewing the JS re-munging of the page at load time that you are pointing out.

    Thanks for the filters.
     
  10. Tim
    Original Member

    Tim Admin

    Messages:
    445
    Likes Received:
    428
    Status Points:
    525
    Wandering Aramean, what browser are you using when you see the attempts to include the same files more than once?

    I was able to reproduce this yesterday and was using chrome. Strangely enough I'm not seeing this behavior today, and yesterday I could not identify where those second copies of files were included in the pages (I was actually seeing both JS files and some image files attempt to load from multiple places.)

    Thanks.
     
  11. Wandering Aramean
    Original Member

    Wandering Aramean Gold Member

    Messages:
    28,220
    Likes Received:
    61,766
    Status Points:
    20,020
    I am mostly using FF3.6 these days.
     

Share This Page