EPiServer 301-handling

Got some room to breath today, so decided to split a simple 301-handling I did some weeks ago into it’s own project and “put it in the git”. Also made a nuget for it if anyone wants to try it out.

https://github.com/studsboll/EPiServer301Handler

I like nugets, if done correctly they make it very easy and clean to try code if you want to, and I’m trying to make it a natural part of when i package stuff for safekeeping.

So what does this thing do then?

Basic 301-handling for EPiServer 7.5+. Support for Domain-redirects. Based on XML-files to be as simple as possible doing nothing more and nothing less.

If domain-redirects exists, that request will be routed first with persistent relative path.

Routes will catch what EPiServer does not want and send it to controller for checking if it’s a 301, if it is, a permanent redirect will trigger to new place. If not, a 404-exception will be thrown falling back on website configured error-handling.

Installation
Drop bin-file in your bin-directory. All is IInitializable modules and will hook onto EPiServer.

Optional configuration, add to appsettings

These are also the default paths if no appsettings exist.
XML should look as follows

I trim surrounding /, as well as ignore lower/uppercase on compare. So not that important.

So what about errorhandling?

To accompany this module I usually redirect to a page in EPiServer. Webconfig as

Then I set a page in EPiServer to answer on /error address. The page someone tried to access can be found in TempData[“originalUrl”].

SVG Fallback with DisplayTemplates in EPiServer Mvc 7.5

The requests about doing stuff with Svg starts to increase and it was time to think about how to handle it. I had probobly 5 different complicated ways to do it and since I’m effective (or lazy) i like it to be easy for me to handle.

So what I landed in was actually a pretty simple concept giving the DisplayTemplates some more responsibility. Think I’m pretty ok with this solution. There are some quirks for the EPiServer Editmode where I have some problems with it handling svg as images. But no time for that today. Doing simple failsafe for that place.

So what I do in my DisplayTemplate is to check if there is any more viewData that i care about in this state. Here i decided to use the cssClass on rendering aswell and now check for an svgModel property. If an svg property is present, I render that as main imagetag and I add the original as a fallback. (Here is also the Editmode workaround atm.)

Having this displaytemplate for my ContentReferences with UiHint.Image, i can now pass and extra ContentReference into it as so properties look like this.

and calling it looks like this

That is about as easy as it gets I think. Then I complete it with a js to take care of the fallback if we don’t have support for svg. Using Modernizr.

EPiServer 7.5 handling extra Routedata using Partial Routing

Querystrings are evil and does not make pretty links. We can agree to this yes? So what you normally do is put up routes that can handle all those quirky bastards and put em where they belong.

As an example I’m going to use a newslistpage with pagination. (I’m lazy so i’ll only write relative urls) /newslist. If i want to target a specific page in the pagination, i would normally set up a route to handle it so i could write /newslist/2 instead of the ugly /newslist/?page=2.

But if you are working with a Cms like EPiServer, the paths are not really known all the time and routing changes. This can be a tricky thing and I’ve seen a lot of people just ignoring it and carry on with querystrings since it’s the easy thing to do.

There is a very easy way to deal with this in EPiServer 7+ and it’s called Partial Routing. I was reading some posts about it on http://joelabrahamsson.com/ and thought, hey i can use that for simpler things aswell. So here is the implementation to use it for your own “extra” route data.

So what it does it add this extra routing to all pages of type NewsListPage. When resolving it, the routing checks if there are any parts of the routing left unhandled. In my case I’m expecting a possible paginationreference, so i try parse it to an int to see if its data i should bother with.

If it fits, I then add that extra data to the RouteData collection. This will make it easily accessible in the Controller.

So to recap, if there are parts left in the routing after hitting the NewsListPage, we handle it in the Partial Router, add it to the routedata collection and then we can fetch it like it was a querystring/normal route in the controller.

To initialize and attach this Partial Routing to EPiServer, we need to load it. I like to keep it pretty clean so I have a class handling it.

And that should be it. No excuse to use querystrings for simple tasks like this ^^

EPiServer 7.5 Mvc Parsing ContentAreas

So this other day I had to get all values from a Page in EPiServer (7.5+) to send to another service. I’ve done it before in earlier versions and thought there would be no problem. However the new ContentArea breaks the pattern a bit and left me confused. Looping through the pageproperties, you could do a switch on the PropertyDataType enum to see what we were dealing with, but not in case of ContentArea it seems.

The funny thing is that there is no ContentArea in that enum and the ContentArea itself says, by Type, that it’s a LongString? So I try do a check on OriginalType and compare to PropertyContentArea, successfully I can identify the ContentArea property. It’s never fun having to compare types though..

What i need to do now is parse everything that is in that contentarea and get the properties for those items aswell. Don’t know how often you do stuff like this but it’s a rather special case and something that is also pretty tight cached, so should be no worries.

The GetExternalUrl() that i use are extensions to get pretty links, which i want from everything that is a PageReference or a ContentReference. This of course varies what you are using em for. But should give you the basic idea.

LinqToTwitter Server OAuth Mvc c#4+

Recently I had to update some libs in a Project we are working on and one example was LinqToTwitter, which is a library I’m very found of, well documented and supported.

Still there are a few steps that documentation does not cover if you are a bit narrow in your knowledge of c# like me. One of the things I did not have the chance to work as much with before is Mvc and especially making stuff async.

LinqToTwitter library got full support for async right through, and also more or less forces you to implement it async aswell in the version i got from Nuget (3.0.4+).

It’s not a major issue since i do the Server OAuth by opening an extra view with Javascript, but still controller needs to be async for it to be happy. This was fairly simple to accomplish in the end (when i learned what to look for). So this is a basic async Mvc controller, note the inheritance on the base.

The thing with keyword “await” is that the following code will not be executed until response is given, but instead of locking the Thread waiting for the response it goes on and handle other stuff meanwhile.

I can see how this is a great thing when it comes to posting stuff, reading posts etc where you are not depending on what the response is. But I’m not so sure it is a good idea when it comes to Authentication. Fine if it’s about logging in and stuff, but when requesting OAuth permissions with tokens and stuff, i rather know the result before i move on. Like in this case we have settings depending on if the result was a success or not.

If you are trying to make a controller async but got trouble returning awaitable results from other libraries, for example if you have a big project and not all parts are awaitable, you can pretty easy put em into a queue anyway, making them awaitable.