I have a bug where special characters (danish 'ø' in this case) are shown correctly when running locally, but wrong when the code runs on a server (I get a pipe '|'). I was trying to solve this today, but I did not have access to the server, other than updating the code files. I will limited access monday, so I can at least write intermediate results to a log file, but to be prepared I thought I'd ask for suggestions here.
Program flow:
A C# test calls my own web service via a WebRequest.
This web service calls an external web service via another web request.
When running my own web service locally I get correct results returned from the external web service.
When deploying the code to the server the external web service returns an error; it didn't find the word I was searching for. Except that I entered 'Bøssemagergade', but the word it complains about is returned as 'B|ssemagergade'. What can cause this difference? AFAIK the server runs with the exact same regional settings as my local machine.
I'm sure this is common stuff, but I haven't been doing web services before (I do feel at home with C#/.NET though, and know about encodings etc.).
Thanks!
You need to check your data at every stage in the process. Make sure you understand which encoding you're using (try to avoid using Encoding.Default) and verify the binary data going across the wire with something like WireShark.
See my debugging unicode problems article for further information.
Related
It is not a specific problem because I have more than one experience that fiddler save my app.
A recent example:
I had an excel addin app write in C#. In the app I had a httpclient to connect to my server through ssl. And I ignored the certification in my code explicitly.
The app always worked well(I mean the connection part), but one day in a QA environment, it failed in no causes. The connection seemed to be blocked for long long time.
Coincidentally, we opened the Fiddler and tried to grab some http. Then everything worked fine.
(Fiddler may do something in the middle)
Then after closing Fiddler, we cannot recreate the problem because all things went back to fine again.
(Fiddler change something permanently!)
Here are some clues:
We use the normal C# http clients System.Net.Http .
Our server is ssl (https, but we must ignore the cert)
The app was good in the same environment at frist.
The lastest change is about the localization, and the test enviroment changed the windows local and language.
We are under a company network, using a proxy in the example above.
The failure is not the whole connection module, but a specific request.(In our case, a request that post two small xml files including excel itself)
Please give us some hint and guess, so we can feel better on this mystical problem.
How long is a "long long time" exactly?
Fiddler does not make any persistent changes to the system; if you start and stop it, the system is returned to the prior configuration. I suppose it's remotely possible that if some configuration setting were corrupt before Fiddler ran, that corruption wouldn't be replicated after Fiddler attempted to reapply it, but we don't know of any such "corrections."
This article describes some of the things folks find unexpected in terms of Fiddler "magically" fixing things while it runs.
I've inherited an ASP.NET web application written in C#. In many pages throughout the site the hostname is retrieved using:
BaseHost = Request.Url.Host;
Since I am using Visual Studio 2012 Express and it's installed local IIS Express server, I seem to be stuck with a port number appended to the hostname (localhost) when I debug/run locally. The above code does not result in the port number being included and as such breaks links that are generated by code (menu item links, redirects, etc).
I see that I can overcome the issue by changing the code to:
BaseHost = Request.Url.Authority;
This seems to fix it by including the port when I'm running locally (localhost:4652) and when published to my staging server (development.mysite.com).
My question: Is this bad form? Is there a time or a situation in which this is going to cause problems on my live site? It just seems a lot easier to do a quick replace of all these instances. I've considered writing a small routine to append : with Request.Url.Port, but it seems easier just to use Request.Url.Authority. Too easy maybe...
I've tried to research my question online and at MSDN, but I don't see an answer.
According to MSDN Authority includes the port number while Host does not. Another aspect is that Authority will escape reserved characters if need be.
Without knowing your application it is hard to say whether it is a good idea, but in general I would suspect that it won't break anything... so go ahead...
Another option is to run the application IIS instead of IIS Express...
My problem with this is that it ALWAYS adds the port, even when the port is not required. This can cause issues with multiple servers in some cases For example, in a production server environment behind a firewall on a pair of load-balanced web servers, it kept putting the firewall port in place, but that caused the URL to break because the port was tied to a specific web server in the server farm that wouldn't map correctly through the firewall. So I would be very careful with this method if you're using it across multiple servers. It caused a breaking issue with our application and had to be reverted back to using Url.Host. Plus, it made production web URL's look weird with the port number.
I'm trying to ensure that the server I'm currently running a piece of code on is a web front end server. I thought it might be as simple as this:
SPServer.Local.Role.ToString().Equals("WebFrontEnd")
However, if you are running your WFE in addition to app servers, etc on the same box, this will return "Application" and fail to correctly identify it as a web front end.
My idea is that by determining if the Microsoft SharePoint Foundation Workflow Web Application (service) is started and running on the server. This can be determined by going to Central Admin > System Settings > Manage Services on Server.
I need to do this programatically in C#. I'm fairly sure that these services and their statuses can be obtained via powershell which is a viable solution, but I'm not sure how to do it either way.
EDIT -- I'm aware of a way to loop through "services" using the following code:
SPServiceCollection services = SPFarm.Local.Services;
foreach (SPService service in services) {
}
However, this includes some items that look suspiciously similar to the list under "Services on Server" but are all listed with a status of "Online" and dont seem to include this.
I'm not on a machine to check, but I've a feeling you'll have more luck with SPServer.Local.ServiceInstances - that sounds like it should give the services on the server in particular rather than the farm in general.
Unfortunately not even that is reliable, as the Foundation Web Application service can be running on servers that are not actually front end servers (i.e. the load balancer never directs traffic to them). Ultimately it is the list of servers that the load balancing mechanism has that determines which servers are true front ends, and because there's various types of load balancers and no single interface to all of them I don't believe there is any one guaranteed method of determining the number of true web front ends in a farm.
I have just finished writing my c# console application, and I am contemplating embedding a web server into it (probably this one http://webserver.codeplex.com). I don't do much in the way over advanced web coding though, so I am not sure if I can do what I need to.
Basically, I would like to allow users to view the console output of my application in realtime just by visiting the site being served by my application. If I understand correctly, to do something like this it would require AJAX, which a simple C# Web Server wouldn't be able to handle.
Is this correct or is there an easy way to do this I am missing?
How to re-route console output
You will need to write your own TextWriter and make Console use it via Console.SetOut. This writer should notify connected web clients, as well as the original Console.Out.
How to host a COMET-like server
You can use HttpListener and some basic async programming to do this. If you wrap the HttpListenerContext.Response.OutputStream in a StreamWriter (with AutoFlush set to true) and set HttpListenerContext.Response.SendChunked to true clients will receive partial results - this means you can even do it in an IFRAME.
You will need to add rights to the URL for yourself if UAC is enabled:
netsh http add urlacl url=http://+:9090/ user=domain\username
Code?
I couldn't resist it; I have written a (poorly tested and mostly incomplete) sample.
That is incorrect.
AJAX is a client side technique relying on JavaScript. As long as the web server can respond to an HTTP request, it can server AJAX content (whatever that might mean:-).
I would use this UtilDev Cassini as me embedded web server of choice, it is based on the code from the embedded Visual Studio dev server and can run pretty much anything that runs in IIS. AJAX is a browser technology not a server technology, the server just sees http requests the same as any other. My last point would that it seems slightly odd to embed a web server in a console application. It would be more usual to do this with a windows service. Have you considered converting the console app into a windows service?
I have developped a web service using c# and Asp.net and this works fine on my local machine where I use vista and IIS7.0. I call it via Jquery using this url
http://localhost/myWebservice/vote
But when I try it on the preproduction site, I receive an 500 error from the server telling me the /vote is not good. I still can access it directly in the browser when typing ?op=vote, but I have then an error saying the formatting of my xml is not right.
Am I missing something in the configuration of the IIS 6.0?
To use routes in IIS 6 you probably need a wildcard mapping in IIS; along the same lines as this (for ASP.NET MVC, but the same concept) - although to be honest, I'd expect a 4xx for a file not found.
500 typically indicates a more serious problem - for example, the app-pool identity isn't configured correctly and can't execute as a service. Check the event log on the server for messages.
As an aside; always try to develop on something as close to the actual hosting platform as you can. For example, I run a Win2003 server in a VM for hosting IIS6. This saves a lot of unexpected surprises when you come to deploy.