Announcement

Collapse
No announcement yet.

HTTPS By Default For Everyone

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Originally posted by cl333r View Post
    I was wondering (and I'm no expert), doesn't storing the articles by name in one big database table make them linearly more expensive to find as their count grows? I'm judging from the url params:
    "page=news_item&px=HTTPS-Default-Everyone"
    The whole query parameter concept in itself is already incredibly expensive compared to URL-based comparisons. Grabbing articles by ID by default would be much less taxing on the server.

    Comment


    • #12
      Michael the rss feed (https://www.phoronix.com/rss.php) still contains plain http links.

      Comment


      • #13
        Originally posted by roys View Post
        Michael the rss feed (https://www.phoronix.com/rss.php) still contains plain http links.
        When clicking them they become HTTPS. There was an issue at least in Thunderbird when making the RSS links HTTPS but still bookmarked the HTTP URL, it would result in double entries for each item with the HTTP and HTTPS link, something funky Thunderbird is doing, so I switched it back to just display HTTP
        Michael Larabel
        https://www.michaellarabel.com/

        Comment


        • #14
          Awesome! Good to hear.

          Comment


          • #15
            Score 15/100 | F | https://observatory.mozilla.org/anal...rd-party=false

            Some work to be done.

            Score A- | https://www.ssllabs.com/ssltest/anal...w.phoronix.com

            Weak ciphers included.

            Comment


            • #16
              excellent, thanks very much.

              Comment


              • #17
                Originally posted by cl333r View Post
                I was wondering (and I'm no expert), doesn't storing the articles by name in one big database table make them linearly more expensive to find as their count grows? I'm judging from the url params:
                "page=news_item&px=HTTPS-Default-Everyone"
                They are probably indexed and so the difference is almost non existent. I have had a page with 1.5 million active (2 million inactive/redirects) articles loaded in this manor on a single medium sized server and delivery was between 50-95 ms for the full article (dynamically rendered) on each request.

                Comment


                • #18
                  Originally posted by Michael View Post
                  Yeah unfortunately this new vBulletin theme is a mess and haven't been able to track down the URL source yet even when grepping all the files and searching the DB, but haven't spent too much time on it.
                  Many things inside and outside the forums link to http. For example the link at the top of each article discussion.
                  https://www.phoronix.com/vr.php will always redirect to http first, and only then to https...
                  But even things which link outside Phoronix are affected. Your Twitter account link at the end of each article still points to http.

                  Comment


                  • #19
                    Originally posted by cl333r View Post
                    I was wondering (and I'm no expert), doesn't storing the articles by name in one big database table make them linearly more expensive to find as their count grows? I'm judging from the url params:
                    "page=news_item&px=HTTPS-Default-Everyone"
                    Not an expert on his setup, but if he has a single column index on the article name.column in the DB, it shouldn't really be any worse than an integer based I'd.

                    Comment


                    • #20
                      Originally posted by mmstick View Post

                      The whole query parameter concept in itself is already incredibly expensive compared to URL-based comparisons. Grabbing articles by ID by default would be much less taxing on the server.
                      From a web server / web application viewpoint URL paths and query parameters are the same thing. Text is text.

                      Comment

                      Working...
                      X