I am migrating an existing .NET Framework application to .NET 6. Parts of the application were UI components in library projects, which I have updated to be Razor Class Libraries.
I have updated the relevant projects based on these microsoft docs:
consume-a-razor-component-from-an-rcl
razor-pages/ui-class
However, I have found a situation where an image, which exists in the RCL project(s), was being loaded in-memory, and modified using the Graphics/Bitmap APIs. I am trying to discover how to accomplish this in the .NET 6 world. I have such static assets properly loading in Views and such at runtime, but I suppose I am wondering how to leverage that runtime lookup in library code (as opposed to in Razor view code).
For example, given an image path to a content file from an RCL project like:
~/_content/My.RCL.Project/Images/foo.png
I can reference such a file in an image tag, like:
<img src="~/_content/My.RCL.Project/Images/foo.png" />
And this works just fine as long as the path is correct - the image is loaded as expected. This is being done successfully for many images in the site thus far.
But how can I load that file into memory in server-side/library code, e.g.
const string filePath = "~/_content/My.RCL.Project/Images/foo.png";
using(var bitmap = new Bitmap(filePath))
using(var graphics = Graphics.FromImage(bitmap))
{
//... omitted
}
I have tried to find the physical file via the IWebHostEnvironment APIs:
IWebHostEnvironment env; //obtained via DI
env.WebRootFileProvider.GetFileInfo(filePath);
env.ContentRootFileProvider.GetFileInfo(filePath);
However, both of these methods return a non-existent file.
Does such a mechanism/API exist for RCL content? Can this be done by calling the Razor engine directly in some form?
Related
I have a folder called Documentation inside the shared project, named App2 in this case. How do I access the files stored inside the Documentation folder? Attached image below shows the project structure.
Visual Studio Solution Page
I have tried out following commands but they aren't working :
System.Environment.GetFolderPath(System.Environment.SpecialFolder.Personal);
AppDomain.CurrentDomain.BaseDirectory
System.Reflection.Assembly.GetExecutingAssembly().CodeBase;
If it's troublesome to access the file in that folder, I'm open to hearing other alternatives.
This is how I have done it for JSON files in my shared project (using PCL). As Jason pointed out in the comments, if you are using .NET Standard, you can simply define the GetSharedFile method in your shared project instead of creating platform specific references.
Add the file to the shared project and set it as Embedded Resource
Create an IFileHelper interface in your shared project
public interface IFileHelper {
Stream GetSharedFile(string fileResourceName);
}
Create a new FileHelper class in each project (Android and iOS) with the following
public class FileHelper : IFileHelper {
public Stream GetSharedFile(string fileResourceName) {
Type type = typeof(IFileHelper); // We could use any type here, just needs to be something in your shared project so we get the correct Assembly below
return type.Assembly.GetManifestResourceStream(fileResourceName);
}
}
Add a documentation handler class in your shared project. Something like below (make sure to change the App namespace to match yours):
public class Documentation {
private const string ResourcePath = "App.Documentation.index.html"; // App would be your application's namespace, you may need to play with the Documentation path part to get it working
public string GetDocs() {
IFileHelper helper = DependencyService.Get<IFileHelper>(); // Your platform specific helper will be filled in here
Stream stream = helper.GetSharedFile(ResourcePath);
using (stream)
using (StreamReader reader = new StreamReader(stream)) {
return reader.ReadToEnd(); // This should be the file contents, you could serialize/process it further
}
}
}
I wrote this mostly by hand so let me know if something is not working. If you cannot load your file, I suggest trying to put it into the root of your shared project and then changing ResourcePath in the code above to the following (again using your app's namespace instead of App):
private const string ResourcePath = "App.index.html";
This question relates to an ASP.NET MVC WEP API, and Naudio
what I want
I was working prototype in WPF appliations when I use this code the wav file is converted to mp3
var retMs = new MemoryStream();
using (var ms = new MemoryStream(File.ReadAllBytes("sound.wav")))
using (var rdr = new WaveFileReader(ms))
using (var wtr = new LameMP3FileWriter(retMs, rdr.WaveFormat, 128))
{
rdr.CopyTo(wtr);
}
return retMs.ToArray();
But when using this code in api project I am getting error like this
Unable to load DLL 'libmp3lame.dll': The specified module could not be
found.
I know libmp3lame is unmanged dll, the WPF project I just copy the dll to bin folder and everything works fine, but how can I achieve this in Web API project, I mean asp.net 5 project
Note
the above code is working expected in wpf project
Also now I am only supported in my API is dotnetframwork means windows only
I removed other platform dependencies
Update:
Created Issue in ASP.Net MVC Repo
I'm a little late to the party but I'll share this for posterity. NAudio can convert a WAVE to an MP3 without the unmanaged DLL to avoid this issue all together. The only snag is it has to write out to a file (which can then be read back in). Assuming you have a MemoryStream where you read all the bytes in (like you did):
using (var wfr = new WaveFileReader(ms))
{
MediaFoundationApi.Startup();
MediaFoundationEncoder.EncodeToMp3(wfr, #"C:\Temp\test.mp3", 48000);
MediaFoundationApi.Shutdown();
}
Sounds like an issue with getting the mapping to the module in the bin library path, check out the following as may help -
How to get bin folder in ASP.NET Core 1.0
MVC4 App "Unable to load DLL 'libmp3lame.32.dll'
Is there a way to include regular javascript files in the Firefox SDK background script?
If I just have a script file that defines some variables, include it and access those in the background script.
Its my understanding so far that they MUST be CommonJS modules. I am porting a Chrome Ext that also uses alot of common code with a mobile app, etc which I'd rather not try converting to CommonJS modules.
Is this possible?
Its my understanding so far that they MUST be CommonJS modules.
That's correct.
If you want to include standard JS files that are already structured in some way, you'd either have to inject them into a page worker, which will
Create a permanent, invisible page and access its DOM,
then send in and out the few variables needed and resulting, respectively, using port, as I explained in your last question.
Or you could use some sort of file concatenation (if you minify your files, this should already happen), then save this new JS file in the lib folder, and require/export those same variables.
These approaches only require you to input/output the variables that are needed externally from the system of files you already have in place, so it's less of a pain than converting each file to commonJS.
NB: I use Angular for my webapp, and have used some modules for both like so
var syncHelper = function() {
this.filter = function(objects, prop) {
// do stuff
}
this.consolidate = function(local, server, id) {
//more stuff
}
}
// app is my angular webapp var
if (typeof app==='undefined') syncHelper.call(exports);
else app.service('syncHelper', syncHelper);
I'm using RGiesecke DLLExport library to produce a C# DLL that can be dynamically loaded from legacy application built on VC6. It exported methods and they were called from VC6 code. No problems. However, as long as I tried to declare a variable as of any one of my .net classes, it crashed.
//I tried CallingConvention = CallingConvention.StdCall too
[DllExport(CallingConvention = CallingConvention.Winapi)]
static void GetDwgReferences(string fileName)
{
//OK: inialize System classes of .net
DateTime dateTime = DateTime.Now;
//crashing here: declare a variable of my static class (.net assemebly)
//SafeString safeString;
//crashing here: declare a variable of my class (.net assemebly)
//Email email;
//crashing here: initialize an object of my class (.net assemebly)
//DwgXrefs dwgXrefs = new DwgXrefs();
//crashing here by declcare a variable of third-party library (.net assemebly)
//ExSystemServices _serv;
}
What's wrong? Please help.
I had a similar problem here trying to use unmanaged exports with Metatrader to load associated managed dlls.
After some digging I think I have found the problem. The app domain is probably not where you would expect it to be, the CLR is trying to resolve your assembly but failing with a nondescript error. In my case the app domain was actually executing in the directory of the host application, so I assume this is always the case.
What I would suggest you do is build a bare dll with no dependencies, and place in something such as the following:
static void Initialize()
{
SimpleLog.WriteLog("App -" + AppDomain.CurrentDomain.BaseDirectory);
}
[DllExport("Test", CallingConvention = CallingConvention.StdCall)]
public static void Test()
{
Initialize();
}
I am not sure but I think you possibly cannot use a static constructor here?
In the log you should see the executing directory for that domain. If you put your assemblies here it (hopefully) should work. It has fixed the problem for me here. I guess the next question is can we change the domain at runtime as I might not want to put these assemblies here.
Have a google if you need the source code for a simple logger - obviously do not use a third party logging framework with dll dependencies!
I think mine is an adaptation of this one:
http://www.codeproject.com/Articles/80175/Really-Simple-Log-Writer
As the other answer stated, it is difficult to know what error C# is throwing without explicitly catching the error in a try / catch block within each method of a C# dll.
You likely need to export as CallingConvention.StdCall and additionally marshal the incoming string as an unmanaged LPWStr type:
[DllExport(CallingConvention = CallingConvention.StdCall)]
static void GetDwgReferences([MarshalAs(UnmanagedType.LPWStr)] string fileName)
{
}
Please see Code to Export C# DLL to Metatrader Build 600+ for a working example using Robert Giesecke's C# Project Template for Unmanaged Exports to export a C# dll to a legacy application (MT4).
Additionally, you might find Native and .NET Interopability interesting though it is mostly geared toward accessing native code from within .NET.
In CRM 4.0 we could place dynamic content (aspx) in the ISV-folder in CRM, creating separate applications but with security and relative URLs to CRM, so for example a custom 360 view of account could be linked in an iframe using a relative URL along the lines of
/ISV/CrmMvcApp/Account.aspx/Overview?id=....
In CRM 2011 usage of the ISV folder is deprecated and Microsoft has some guidelines on how to transition into doing this in supported manner (MSDN gg309571: Upgrade Code in the ISV folder to Microsoft Dynamics CRM 2011). They say:
For scenarios that will not be satisfied by the Web resources feature, create your Web application in its own application pool with its own web.config.
The way I am reading this (coupled with the guidelines on supported/unsupported) is that we need a separate web site in IIS with its own binding as you are not allowed to add virtual directories etc. under the standard CRM app. This is unfortunate and does not allow relative paths/URLs in customizations and sitemap. This is troublesome especially when exporting and importing solutions from DEV, TEST and/or PROD.
Are my assumptions wrong?
Can we somehow have relative paths?
Have anyone else found a pragmatic and easy approach to having external content without doing the sitemap and customization changes for each environment?
EDIT: Confirmed with other sources that my understanding of the guidelines are correct, as this is also listed in the list of unsupported changes. Virtual folders and web apps are to be kept totally separated from the default CRM web site.
Creating an Internet Information Services (IIS) application inside the Microsoft Dynamics CRM website for any VDir and specifically within the ISV folder is not supported.
MSDN gg328350: Unsupported Customizations
If you primarily need to access CRM data/records, take a look at using a jScript web resources. You can do "most" CRUD operations using the REST OData services. If you use JQuery to parse the JSON it's very productive.
I have found a solution much like the javascript redirect, without the need for client execution and only configuring the environment details (servername, port) once. Additional logic can easily be added.
The solution creates a dependency into the customizations, but not an environment one like and can be used for unmanaged and managed solutions.
The solution was to place a file Redirect.aspx in the ISV folder. The code does not in any way interact with CRM and falls within the supported guidelines, however the solution is not future proof as the ISV folder is deprecated by Microsoft.
Redirect.aspxwill automatically pass along any parameter passed, so will work with or without the entity identifiers and so on.
Usage:
Place the file in the ISV folder on the CRM app server
Change the server name and port to match the current environment (must be done for each environment)
In customizations, for example for an iframe, use the following as a source:
/ISV/Redirect.aspx?redirect=http://SERVERREPLACE/CustomMvcApp/SomeControllerAction
Here is the content of Redirect.aspx
<%# Page Language="C#" %>
<html>
<script runat="server">
protected override void OnLoad(EventArgs e)
{
// must be customized for each environment
const string ServerBaseName = "appserver1:60001";
const string UrlParameterName = "redirect";
const string ReplacePattern = "SERVERREPLACE";
var parameterUrl = Request.Params[UrlParameterName].Replace(ReplacePattern, ServerBaseName);
var queryStringBuilder = new StringBuilder();
foreach (var key in Request.QueryString.AllKeys)
{
if (key == UrlParameterName)
{
continue;
}
queryStringBuilder.Append(!(queryStringBuilder.Length > 0) ? "?" : "&");
queryStringBuilder.Append(key + "=" + Request.QueryString[key]);
}
var completeRedirectString = parameterUrl + queryStringBuilder;
Response.Redirect(completeRedirectString);
}
</script>
<head>
<title>Redirecting</title>
</head>
</html>
Not quite "relative urls" as per your question, but a solution I use is to store "stub" or "root" urls in a config entity and read those records in JScript at runtime to determine the fully qualified destination for your custom links.