Note: the issue referenced in this post has been fixed and will get released with ASP.NET Core 1.1.0.
While creating a small piece of middleware to modify the response body's content, I had trouble getting anything to appear in my browser. I would get a
200 response-code, but no content (i.e.
Content-Length: 0). Here is an example that creates a similar setup:
In a previous post, I demonstrated creating a Twitter configuration provider for ASP.NET Monster's #SummerOfConfig contest. This post is a follow-up demonstrating how to create a provider leveraging GPS coordinates to create a unique type of progress bar. Source code available on GitHub.
This is the first in an eventual series of posts related to ASP.NET Monster's #SummerOfConfig contest. The general idea is to implement zany configuration providers (i.e. not the kind you would expect to use in a production environment 😁 ). For my first entry, Twitter has been enlisted to provide "on-the-fly" configuration values. Imagine having the ability to change aspects of your site just by Tweeting! Read on to see how this jazz comes together.
Between the time I initially wrote about hybrid model binding, later about its updated behavior, and now, a series of changes were made to make the binding behavior safer and less verbose to implement.
Using Disqus is an easy way to get a comment-system on your site. Part of its initialization requires passing a unique Id per page so it knows what comments to retrieve. However, if you like to test your sites locally before deploying to production and use the same identifier for both environments, then you will find your production Disqus links reference your local-URI. When someone wants to share a page using Disqus, the link will point to your local site (which
probably won't should not work for anyone but you 😉 ). However, there are several ways to keep the links up-to-date.
In my previous introductory post on hybrid model binding, I demonstrated how a binder could be used with both body and URI content. After receiving feedback (thanks @buhakmeh), it became apparent being greedy by default was not the best option and would most-likely hinder adoption of the binder by the community.
Imagine you have a REST-ish API and your code follows the OMIOMO/Thunderdome principle.
GET-endpoints are easy to bind–just decorate the action-parameter with
[FromQuery]. But, what to do about
PUT? Now, there is content coming from the body of the request in addition to properies we want to bind from the URI. The oft-repeated answer is to have two or more parameters:
Here is the scenario for those of us who still use a pull-start mower:
you pull on the starter cord and instead of the usual feeling of the recoil spring being fully unwound, the cord keeps coming towards you. This is the end-result:
So, your new web app. is ready for QA and you have all your app. settings neatly separated using
appsettings files. But, when you launch the app. on your brand new server, you know something is wrong and quickly realize the app. just loaded production settings. How do you correct this problem? There are several ways and they all involve setting the hosting environment's environment-name.
Back in the RC1 days, you could just pass an argument on the command-line to set the host URI. It went like this:
> dnx web myapp --server.urls="http://localhost:10000/"
I recently read an article about integrating Azure Active Directory into an application. This allows the former to host login duties for the latter. The article contained several screenshots showing where to find information to complete the integration. The author obscured sensitive information by using Microsoft's Paint app. and its airbrush effect. One screenshot, in particular, contained a shot of the endpoints needed to handle logins of various technologies (i.e. WS-Federation, SAML-P, OAuth 2.0). However, this shot contained repeating information–a GUID that is part of a URI. Because the author manually obscured this data, the pattern of the airbrush effect varied between each row of URIs. Individually, the GUID could not be recovered. However, by stacking cropped frames of each row (seven in total), I was able to retrieve eighty-nine percent of the GUID (thirty-two out of thirty-six characters). What follows is a brief tutorial of how this was achieved. In this particular case, knowing the endpoint is not a crucial security leak, but the point that should be remembered is: repeating data intended to be obscured must be fully obscured. A better approach is to just not include the sensitive data in a screenshot–delete the particular section.
Note: this is only applicable for EF versions less than or equal to six. Version seven (part of .NET Core) has a completely different migration strategy.
Here is a classic scenario: Local would-be lumberjack decides to cut-down tall tree right beside house. Tree falls towards house. Disaster ensues. Others have a laugh at the homeowner's expense. This post will not help if you find yourself in a similar situation. Instead, this post will allow you to make an informed decision when that tree has a little bit of distance between itself and the house and its height makes the outcome a little blurry.
Update: this problem has been fixed and is confirmed working as-expected with the release of ASP.NET Core 1.0.
I love finding relationships in number-sets. A favorite memory from high school deals with a game we would play in Algebra class. The teacher would give us a set of numbers and one target number. The goal was to use the set of numbers and arithmetic operations to get to the target. Here is an example:
Took a short drive with the family to get lunch today. Among other items, we ordered small and medium-sized portions of fries. Everything tasted great, but I thought the amount of fries received between the two-sized boxes looked really similar. I began to think whether getting a medium was really worth the extra fourty-cents. So, I kept one of each and took some measurements to see whether ordering the larger size had any benefit. Ultimately, my interest is quantity per dollar. Keep in mind my method and results have some degree of error, but I will leave what percentage up to you.
First off, Happy Pi Day! I have been putting off owning and running a personal website for a long time. It seems weird, too, since I work predominantly with software (see About me). There are many options for hosting and storing content. But, I
have had a tendency to over-think solutions to problems (i.e. Analysis paralysis). Over the course of several years, I built up this collection of ideas and never held myself responsible for making deposits into my idea bank (see Idea Debt). While the ideas are varied, most have a writing-part where I want to share the knowledge gained from completing the idea. And in this case, one idea was to create the engine which drove the display of my writings.