HTTP Cache: multiple URLs point to the same HTML file - http

There are times when multiple URLs point to the same HTML file.
For example, the classic use of the base tag
For example the main page can be in:
https://dom.com/schops/
but the following URLs also direct to that file:
https://dom.com/shops/mark1
https://dom.com/shops/mark1/products
https://dom.com/shops/mark3
My question is, is there a HTTP cache header or a way to mark that a URL actually has another source URL?
How do you usually solve this problem?
Anyway I am creating my implementation in Cache Storge API, but I would like its behavior to be similar to the standard using the necessary headers when appropriate.

Related

css image path with version for CDN

I'm working with the following paradigm for handling my CDN caching:
Each path contains "?version", for example: http://mycdn.com/some-javascript-file.js?123
The same paradigm is used for all of my resources (js, css, images), the problem I'm encountering is images paths in a css file.
For example, I have the following snippet in one of my css's:
"url (../../Images/example.png)"
The problem is that this image path doesn't use the version paradigm, I would like to add the version to the path somehow, is there a nice way to do this, except of the following methods:
1) For each image change - also change the css with some dummy version.
"url (../../Images/example.png?1)" - change 1
"url (../../Images/example.png?55)" - change 2
2) Transfer all of my css's files to be aspx files and to use the code-behind in order to define the version:
"<%= html.VersionUrl("../../Images/example.png")%>"
3) Use dotless lib: http://www.dotlesscss.org/
Any other simple/nice idea?
The best solution which I've found was to change the version tag to be at the beginning of the url and to use url rewrite in order to process the requests.
So if for example I used to had:
http://website/Content/Images/1.png?123456
this will become to:
http://website/123456/Content/Images/1.png
Notice that I use url rewrite in order the process the request so that http://website/123456/Content/Images/1.png will actually bring the data from http://website/Content/Images/1.png

Loading multiple CSS files with single http request

When I view the source code of yahoo mail, I see multiple css files in a link tag using an & symbol as shown below:
href="http://mail.yimg.com/zz/combo?kx/ucs/uh/css/271/yunivhead-min.css&kx/ucs/uh/css/221/logo-min.css&kx/ucs/avatar/css/17/avatar-min.css"
Does anyone know, how they separate each file and load them all using a single http request?
In this case, there seems to be a script that joins the css files into a single response.
The path to the script is http://mail.yimg.com/zz/combo. It accepts several parameters containing paths to CSS files, which will then be joined and possibly minified.
If you play around with the URL, you can see that you could remove the -min-Prefixes from the URL and you get the unminified CSS file in return: http://mail.yimg.com/zz/combo?kx/ucs/uh/css/271/yunivhead.css&kx/ucs/uh/css/221/logo.css&kx/ucs/avatar/css/17/avatar.css
There are several CSS minifiers around, for example CSSmin. But as this is a Yahoo page, they probably use their own CSS compressor, YUI. For details about how it works, see http://developer.yahoo.com/yui/compressor/#work.
Not familiar with the specifics, but the URL looks like a query string with the CSS files as unnamed parameters.
http://mail.yimg.com/zz/combo will be a service that loads the CSS, then concatenates and probably minifies the files before serving back to the client.
My guess is that http://mail.yimg.com/zz/combo is a small program / script which collects all params (like kx/ucs/uh/css/271/yunivhead-min.css, kx/ucs/uh/css/221/logo-min.css, kx/ucs/avatar/css/17/avatar-min.css), bundles them and minimizes them.
This is similar to the bundling feature for MVC, which you can read about at http://www.davidhayden.me/blog/asp.net-mvc-4-bundling-and-minification (or other sources).
If you take the URL apart what you see is that it's a request to something called "combo" passing in various querystring keys (note there's no values) that are the paths to some CSS files.
These keys will then be extracted in the standard way given the server side language being used and the CSS for that url parsed into a variable before being returned in its entirety to the response.
For their yui project, yahoo development have a project called yuiloader. While designed primarily for yui, the code seems like it can be set up to serve other files as well. This does more than COMBO. it also works out dependancies. with JS and CSS.
As Yahoo is the Y in YUI, this is probably their code base for mail.yimg.com.
The code can be found on https://github.com/yui/phploader.

Umbraco - Check if request is XmlHttpRequest

I'm doing an AJAX feature for my Umbraco website.
Basically what I'm trying to do is when someone requests for http://mysite.com/news with the XmlHttpRequest header then I want Umbraco to return JSON of the following page.
What I've been trying to do is finding the right place to check for the header in Umbraco. Basically I want to override or inherit some class where I could check the header and return the page in Json or HTML depending on the header.
I'm using Umbraco 4.7.0 (1.0.4090.21631)
Any feedback would be more than welcome! Thanks!
I'd probably use the URL Rewriting module for IIS7 to detect the content-type in the request, then redirect to the same URL with /jsonTemplate on the end of it. You would then create a masterpage called jsonTemplate, with a single macro inside it, which would return your JSON.
This is common for Umbraco - specifying a separate template with a separate macro to return a different view of the same content. Check out http://blog4umbraco.codeplex.com/SourceControl/changeset/view/68424#1138733 for an example of an XSLT macro that returns RSS using this approach.
You should check the Content-Type header from server side logic. Your Javascript should be specifying a Content-Type like application/json.

Shorter names for /ScriptResource.axd and/or /WebResource.axd in ASP.NET?

My page loads scripts and css from some custom assembly. For now, I use the ScriptManager to add scripts to the page (its CompositeScript) and the page's ClientScript.GetWebResourceUrl() for css files. So, on the page(s) my scripts are referenced by /ScriptResource.axd and css are referenced by /WebResource.axd. Works fine except the fact that those URLs look really ugly. Is there any manageable way to replace those urls with my own fake paths without moving/hosting script/css physical files somewhere? Any help or ideas would be greatly appreciated.
Use a Filter to find all the WebResource.axd and ScriptResource.axd src'es and replace them with something you like better. You then need a IHttpHandler, as Brian comments, that understands your new url-format and returns the same content as the original axd-files would have.
You should also look into combining all those script and css resources to save requests from the browser. How do I combine WebResource.axd and ScriptResource.axd files so as to result in less requests to my ASP.NET server?

Google Fonts CSS Include

Hey,
Since Google Fonts came out, I have had this question in mind. First see this below:
<link href='http://fonts.googleapis.com/css?family=Cantarell&subset=latin' rel='stylesheet' type='text/css'>
Here Google is linking to an external CSS file that doesn't have a file extension (.css)! Then Google also has another feature that if you want to inlude another font to this then just add the "|" sign and type the font name. How do you do this? Using Javascript, PHP or something?
Help is appreciated!
Thanks :)
The extension of a file does not have to mean anything at all about the contents of said file. It is merely a convention (one that Windows, for instance, uses to the point of making it seem like a requirement).
Any dynamic 'file' on a web site can return what ever kind of content it wants, any time it wants. The extension means nothing - aside from expected convention.
That URL could be a directory named css with a default 'document' that is a script, which handles the parameters to decide what content to give. Or, it could be a literal file named css which does the same thing. Or, it could not be a file or folder at all, instead merely part of a routing mechanism, which calls a controller based on the URL, and passes the parameters in.
Web servers return information in the response indicating what the MIME Type of the return value is, and the browser determines what to do with it based on that - not based on the extension of the file.
Yes, they have to be doing some sort of server-side processing when this URL is requested
http://fonts.googleapis.com/css
The querystring is parsed, and a text stream is returned with the CSS output. Allowing the user to add additional font families to the CSS is pretty trivial, as the server is just spitting back what you append to the query string.
You could do this in PHP or ASP.Net (and many others), but there is no indication of the underlying technology from Google's URL.
The easiest way to do this yourself would be to create a folder on your web server called "css", and then have a default script in there that does the processing. The URL could basically be almost identical to the Google url.

Resources