6

It would be valuable to know what file a browser has found to read to render the home page of domain.

I know the browser goes to something like google.com and then starts to look for default file names/types such as index.html then perhaps index.htm and so on through a list of about a dozen other files.

I'm curious what file my browser has actually began to render (right clicking in the browser window and clicking "Save As" doesn't give the file name), and then secondly, I wonder if a site would begin to render faster if the file name present on the domain (e.g. index.php) was one of the initial file attempts the browser looked for vs. something more atypical (eg. placeholder.html).

  • 6
    On modern application driven websites, often there are no files, just a returned URI (resource) for the URL requested that is composed of output assembled by the app via templates from database content and other file resources such as images, CSS, etc. The above assumes a static website where content lives in individual files. – Fiasco Labs May 22 '16 at 3:28
14

Your browsers doesn't load any file, it requests a resource which the server then provides at his discretion (lengthy elaboration below).

If you type google.com into your browsers toolbar, the browser wil first append a protocol, either http:// or https. Then browser will look up the IP address belonging to google.com, which is 172.217.19.206. Your browser will then establish a socket with that server on the proper port (80 for http, 443 for https).

Your browser will then send the following request to the server:

GET / HTTP/1.1
Host: google.com

The web server will then decide what to do. This can involve a lot of steps.

A web server usually has something called a document root for any domain he serves. Files that the web server is allowed to serve the user usually reside inside this document root. For example, the document root for google.com might reside at /var/www/domains/google.com/htdocs/.

Now, when you request a resource the web server first inspects the resource, and then takes proper action. For example, if the resource ends wih .php, the web server might decide not to serve anything himself, but instead invokes the PHP interpreter, and lets the PHP interpreter execute the proper PHP-file for the requested resource, and then serves the user the output.

Take for example this request:

GET /article.php?id=123456 HTTP/1.1
Host: news.example.org

In this case, the web server on news.example.org is tasked to serve the resource /article.php?id=123456. What likely happens is that this web server will start the PHP interpreter. fetch the article.php file from the document root, feed it into the PHP interpreter and wait for the output. It will then send the output bck to the browser that requested it. In this case, this would likely be a site from a blog with certain content loaded from a database (the contents of the article stored with the id 12345).

But other things can happen, too.

Lets get back to the original example:

GET / HTTP/1.1
Host: google.com

What with any stanadard web server (Apache, Lighttpd, etc.) happens, is more or less the following:

  • Look for a file named index.html (in the document root) and serve it
  • If that doesn't exist, look for a file index.htm and serve it
  • If that doesn't exist, look for a file index.php and start the PHP interpreter, serve the output
  • If that doesn't exist, serve a 404 NOT FOUND error

The precedence of the extension is usually configurable on the side of the web server. The server might not serve any index.xxx file at all. For example, if you have a node.js server running, then the web server would task the node.js server to provide the resource /, which might just be anything the JS-App that is runnng on node decides it to be.

tl:dr; The browser doesn't look up a file. The browser requests a resource, the web server then handles the request and serves the content approrpiate for the requested resource, which might be a file, but might also be the output of a 3rd-party program.

As far as speed is concerned, this is dependent on the web server. But if you want your web server to always serve the asjkdjhfz9874jykdfndsk.html file when / is requested, you would usually configure the web server to look for such a file first, making it as fast as any other configuration.

Disclaimer This is not a full decsription how any web server works, nor tailored specific to one. Most web servers work similar, but especially sites like google.com will likely run some custom things that are tailored specifically to their needs.


Your browser will usually offer tools to inspect network activity. Using Chrome, you can open the "Dev tools" and inspect the headers. this is what my browser send to SE to let me edit this answer:

GET /posts/93567/edit HTTP/1.1
Host: webmasters.stackexchange.com

There are a few more that tell the server about caching, what language I am expecting, what browser I am using and where I'm coming from, but those aren#t interesting here. The point is, my browser requests the resource /posts/93567/edit. My browser will never have any idea about what file the web server serves. SE runs on ASP.NET MVC 5, which means that the web server (in SEs case, an IIS) will likely load some proper .asp file (that can be located anywhere), and lets the runtime evaluate it for the parameter postId=93567. The actual file or inner workings are never exposed to the browser, because the browser doesn't need to know (and because it is safer to hide that information for the one running the server).

The view will also show you any other resources (CSS files, JS files, images etc.) that your browser requests to correctly render the site. But with them, you will only learn about the resource, not wether or not those are actually files in the file system.

  • In your answer, you have some bullet points that show where "If X file doesn't exist, then look for the next file." While the answers here have said that what's served might be a "resource" vs a "file", the reality appears that there's some checking for files and their existence in many (not all?) cases. So, perhaps my question could be restated as "What did the web server land on (while it may vary between configs) to serve my browser. Was it index.php and then fired up the interpreter or something else?" – Tony DiNitto May 26 '16 at 13:52
  • 1
    The point is that the informaton what the web server actually did is never exposed to your browser. Your browser asks for the resource /. The server answers it. That is as much as your browser knows. You can do some educated guesses based on common configurations, but then again these educated guesses can easily be thrown off by deliberately configuring the web server to hide what he does (e.g. serving index.php requests by a node.js process instead of PHP). Note that in my example the resource / was requested, but the web server (in most common configs) answers by serving the index.html – Polygnome May 26 '16 at 14:03
14

The browser isn't looking for a file. It's just asking for a resource. The server then decides what that resource returns.

At it's most basic level that "file" is literally just a file. In the case of the default index page of a directory how the server is set up will determine which files is returned. Some servers are configured by default to return index.html if the file exists, then fall back to index.htm, etc. Others default to default.html, etc. They will keep trying until the list of default files is exhausted and will then return a 404 error.

In the case when server rewriting is turned on or dynamic pages are being constructed, the content being returned typically isn't a file at all. It resembles a file as the output is (typically) HTML just like a .html file would contain. But behind the scenes tens or hundreds of files create that content.

  • Thanks for answering - you're absolutely right. I'm wondering how I know what file is being rendered. I can eventually guess it accurately on a long enough timeline by simply explicitly calling on that filename in the URL. www.xyz.com/index.html fail to load anything? Then try www.xyz.com/index.htm and then so on until I get the site to render. I'm just looking for a shortcut to know what file my browser has loaded. – Tony DiNitto May 22 '16 at 0:20
  • I was not sure if IE still did it's rotation of trying different files if a request failed. Remember those days? It was a silly dance. – closetnoc May 22 '16 at 1:25
  • Sorry. I forgot to up-vote. Not sure what happened. Getting old maybe?? – closetnoc May 22 '16 at 4:25
  • 2
    @TonyDiNitto: It still doesn't work that way. There may not be a second URL that behaves exactly like http://www.example.com/ does. The HTTP specification certainly doesn't require it, at any rate. The only requirement is that you can ask for hypertext and the server either gives it to you or sends an appropriate error code. – Kevin May 22 '16 at 6:07
  • 7
    Even index.html might not be a plain file. A request for kitten.jpg might be read from a database rather than a filesystem. And so on... – Michael Hampton May 22 '16 at 9:04
4

I'm wondering how I know what file is being rendered. I can eventually guess it accurately on a long enough timeline by simply explicitly calling on that filename in the URL. www.xyz.com/index.html fail to load anything? Then try www.xyz.com/index.htm and then so on until I get the site to render. I'm just looking for a shortcut to know what file my browser has loaded.

I agree with John here that what you are requesting by specifying a URL is a resource (or for a better word, an object) from the server.

You'll never know 100% for sure what actual disk file is being read when a URL is requested. This is especially true if the server requires a third-party program to associate with it in order to produce output.

A typical third-party program is the PHP interpreter which is something Wordpress uses to deliver content. The interpreter can process code that may involve loading any number of files from the server's disk in order to construct the HTML data which is then delivered to the user's browser.

On top of that, special configuration can be applied to the server to assign special URLs to resources. This (in an apache environment) is known as URL rewriting, and this is very good since its the start to friendly URL creation.

The users won't know the exact filenames of the files loaded nor will they care (unless they are hackers) because all they care about is the actual content on the page.

It's also possible that some server admins decide not to use actual filenames in URLs for security reasons.

  • 1
    I take it one step further. Outside of images, robots.txt, style, and the sitemap, there are no files in my web space. The app works, however, if you start requesting files, you will get nothing because there is nothing. [insert cheese eating grin] It is part of my security strategy. Everything is out of the web space and environment languages such as PHP and Java are completely unavailable to Apache and not just disabled file types. – closetnoc May 22 '16 at 4:24
  • I'm guessing you invented your own server called the "closetnoche" lol. – Mike May 22 '16 at 17:19
  • Just because there aren't any files in the document root doesn't mean there aren't any resources (or files!) that can be delivered. If you are running a custom script (node.js, PHP, Java, whatever) that returns files, then this script might return files from arbitrary locations, even outside the document root, provided that the user the web server (or more precise, the script) is running as has access to it. Thats why you usually let the web server handle static files, because web servers have built-in protection against escaping the document root and e.g. printing ../../../../etc/passwd ;) – Polygnome May 23 '16 at 9:27
0

I'm limited to 2 links in my response... to talk about how links actually work, it's a challenge :-/ Perhaps I'll gain a few points and I could come back and improve this expiation soon.

As mention already, website behaviours and response is specific to server used and it's wide range of settings. Perhaps I could describe the 'typical' behaviour of a LAMP server (Linux Apache Mysql Php - Probably the most common web server used nowadays)

In the apache configuration of exemple.com you will have a DocumentRoot directive telling apache where to look for folder matching that site, lets say /www/

You can have a file called .htaccess that can hold some specific definitions for that directory and all it's subdirectory.

Placing .htaccess in /www/ will apply to the entire site (note that in unix convention, a filename starting by a dot is a hidden file)

Placing that file ins /www/test/ will apply to all calls (GET POST PUT etc...) starting with http://exemple.com/test/ A GET http :// exemple.com /test/ in reality is

GET /test/ HTTP/1.1
Host: exemple.com

But it's easier to write GET http : // exemple.com /test/

Apache will receive that call as it listens on port 80 When you see http://exemple.com:8080/ it means you force the port to be 8080

http : // exemple.com /test/ = http : // exemple.com :80 /test/

Apache will look for a .htaccess file in /www/ and in /www/test/

It will interpret first the one in /www/ then the one in /www/test/ (of course if the file exists)

So first things you want to have a look for those files as they are there to give specific directives that are not the default behaviour.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.