Bad CRC32 in GZIP stream - silverlight-4.0

I am using DevForce 2010 and Silverlight 4.
When saving entities that contain large amount of binary data, I get this error:
Unhandled Error in Silverlight Application The remote server returned an error: NotFound.
When debuging the application I see these errors:
Unhandled Error in Silverlight Application Insufficient memory to continue the execution of the program.
Bad CRC32 in GZIP stream.
I found this thread on Ideablades forum that discusses the problem: http://www.ideablade.com/forum/forum_posts.asp?TID=3361&PN=1&title=bad-crc32-in-gzip-stream
Is this a problem on the server or client?
Is this a problem that has been resolved in any new version of DevForce 2010?
My server has 4 GB memory. Would increasing the memory resolve the problem?
Or what would be the right solution?

Yes, the OnEndpointCreated overrides on both client and server are where you should add the customization. You can add the following to remove GZIP from the binding:
public override void OnEndpointCreated(System.ServiceModel.Description.ServiceEndpoint endpoint)
{
if (endpoint.Binding is CustomBinding)
{
var binding = endpoint.Binding as CustomBinding;
var elements = binding.CreateBindingElements();
// Swap out existing (GZIP) message encoding for binary
var encoding = elements.Find<MessageEncodingBindingElement>();
if (encoding != null)
{
elements.Remove(encoding);
encoding = new BinaryMessageEncodingBindingElement();
elements.Insert(0, encoding);
endpoint.Binding = new CustomBinding(elements);
}
}
}
DevForce will find your classes if they're in an assembly probed on the client/server.
This will turn off compression for everything from your DevForce client to the EntityServer, so may be a bit heavy-handed. You can turn on IIS compression to compress data sent to the client to help.

There haven't been any changes to GZIP processing since the 6.1.7 release of DevForce 2010. That thread still contains the best information of how to work around the problem: 1) modify the save logic or your entity definition to reduce the amount of data saved; 2) turn off use of GZIP; or 3) write a custom message encoder with another compression library.

Thank you Kim Johnson,
I have looked at the samples and I feel uncomfortable adding those config files and maybe breaking something that works fine today.
If I go the code-way, will I be ably to switch off GZIP and still retain the rest of the default settings for DevForce?
I guess the code below is what I should go for?
If I save these classes on the client and server, will DevForce automatically find these classes?
//Client
using System.ServiceModel.Channels;
using IdeaBlade.Core.Wcf.Extensions;
public class ProxyEvents : IdeaBlade.EntityModel.ServiceProxyEvents {
public override void OnEndpointCreated(System.ServiceModel.Description.ServiceEndpoint endpoint) {
base.OnEndpointCreated(endpoint);
// My client code turning GZIP off comes here?
}
public override void OnFactoryCreated(System.ServiceModel.ChannelFactory factory) {
base.OnFactoryCreated(factory);
}
}
//Server
using System.ServiceModel.Channels;
using IdeaBlade.Core.Wcf.Extensions;
public class ServiceEvents : IdeaBlade.EntityModel.Server.ServiceHostEvents {
public override void OnEndpointCreated(System.ServiceModel.Description.ServiceEndpoint endpoint) {
base.OnEndpointCreated(endpoint);
// My server code turning GZIP off comes here?
}
public override void OnServiceHostCreated(System.ServiceModel.ServiceHost host) {
base.OnServiceHostCreated(host);
}
}

Related

Is there a way to ensure the HTTP request body can be loaded into memory for possible multiple read requests?

Is there a way to ensure the HTTP request body can be loaded into memory? There are multiple middle-wares which will use same HTTP Request body to log at multiple levels.
I remember doing following in .Net Framework 4.6
await request.Content.LoadIntoBufferAsync();
/// *** Method snapshot from HttpContent - System.Net.Http library ***
/// <summary>Serialize the HTTP content to a memory buffer as an asynchronous operation.</summary>
/// <returns>The task object representing the asynchronous operation.</returns>
public Task LoadIntoBufferAsync()
{
return this.LoadIntoBufferAsync((long) int.MaxValue);
}
Can anyone help me find similar behavior in .Net Core?
EDIT
-- I think right answer here is to use EnableBuffering, but I am not to able to figure out which overloaded method should I use for EnableBuffering?
public static void EnableBuffering(this HttpRequest request)
{
BufferingHelper.EnableRewind(request);
}
public static void EnableBuffering(this HttpRequest request, int bufferThreshold)
{
BufferingHelper.EnableRewind(request, bufferThreshold);
}
public static void EnableBuffering(this HttpRequest request, long bufferLimit)
{
BufferingHelper.EnableRewind(request, bufferLimit: bufferLimit);
}
Size of HTTP Request in our application is varying from 50kb to 300mb.
For asp.net core 2.x you can use :
HttpContext.Request.EnableRewind();
For asp.net core 3.x you can use :
HttpContext.Request.EnableBuffering();
That methods ensure the request Body can be read multiple times. Normally buffers request bodies in memory :
https://learn.microsoft.com/en-us/dotnet/api/microsoft.aspnetcore.http.httprequestrewindextensions.enablebuffering?view=aspnetcore-3.1

VSTO-specific web request not working (System.Net.WebException)

I wrote a 2013/2016 VSTO app for Microsoft Word using C#. My app creates a new toolbar with buttons. One such button runs my app, which launches a basic Windows Form.
Before the user can work with my app, they need to enter information like their license code and email address. My code in turns sends a basic request to my licensing server and awaits a response.
All my code has been running just fine and now it no longer is. Now, when I run the code, I receive the following two error messages:
System.Net.WebException: 'The underlying connection was closed: An
unexpected error occurred on a send.' Inner Exception: IOException:
Unable to read data from the transport connection: An existing
connection was forcibly closed by the remote host.
and
System.Net.WebException: 'The underlying connection was closed: An
unexpected error occurred on a send.' Inner Exception:
SocketException: An existing connection was forcibly closed by the
remote host
I decided to run the code using a standard console app to see if I received the same error message, and sure enough, it worked great! Now I am wondering if Word or the Microsoft VSTO technology is blocking my app from accessing my server.
Here is the code in VSTO that does not work
Note 1: Created a basic 2013/2016 C# VSTO add-in, added a toolbar, and added
Note 2: Added a reference to System.Web.
Note 3: Modified the website link and the query strings as I did not want to publish them on this public forum.
using System;
using Microsoft.Office.Tools.Ribbon;
using System.Web;
using System.Net;
namespace WordAddIn3
{
public partial class Ribbon1
{
private void Ribbon1_Load(object sender, RibbonUIEventArgs e)
{
}
private void button1_Click(object sender, RibbonControlEventArgs e)
{
// Attempt to activate the product using the licensing server on the website.
Console.WriteLine("** ActivateLicense");
//build the url to call the website's software licensing component.
var builder = new UriBuilder("https://validwebsite.com");
builder.Port = -1;
//build the query string.
var query = HttpUtility.ParseQueryString(builder.Query);
query["license_key"] = "validactivationcdode";
query["product_id"] = "validproductid";
query["email"] = "validemailaddress";
builder.Query = query.ToString();
string url = builder.ToString();
Console.WriteLine("activation request:");
Console.WriteLine(url); //display the REST endpoint.
//make the synchronous call to the web service.
var syncClient = new WebClient();
var responseStream = syncClient.DownloadString(url);
Console.WriteLine("Response stream:");
Console.WriteLine(responseStream); //display the server json response.
}
}
}
Here is what is pretty much the same exact code in a console app that does work
Note 1: Created a basic C# console app.
Note 2: Added a reference to System.Web.
Note 3: Modified the website link and the query strings as I did not want to publish them on this public forum. You will receive an error, but that is due to the sample website not having a licensing server.
using System;
using System.Net;
using System.Web;
namespace ConsoleApp1
{
class Program
{
static void Main(string[] args)
{
// Attempt to activate the product using the licensing server on the website.
Console.WriteLine("** ActivateLicense");
//build the url to call the website's software licensing component.
var builder = new UriBuilder("https://validwebsite.com");
builder.Port = -1;
//build the query string.
var query = HttpUtility.ParseQueryString(builder.Query);
query["license_key"] = "validactivationcdode";
query["product_id"] = "validproductid";
query["email"] = "validemailaddress";
builder.Query = query.ToString();
string url = builder.ToString();
Console.WriteLine("activation request:");
Console.WriteLine(url); //display the REST endpoint.
//make the synchronous call to the web service.
var syncClient = new WebClient();
var responseStream = syncClient.DownloadString(url);
Console.WriteLine("Response stream:");
Console.WriteLine(responseStream); //display the server json response.
Console.ReadKey();
}
}
}
Can you help me determine why the code is no longer working in the add-in where it did before (with no code changes)?
I read a lot online and there seem to be too many reasons why this might happen. As an FYI, the website with the licensing server is running. It is (and always has been) a little slow, but when running the code with VSTO, the response is immediate (suggesting no timeout). The Console code runs and there is never a timeout.. I always get a response from the licensing server.
On another thread for a similar problem, someone recommended running WireShark. I am not really familiar with the product, but during my working console run, I received no error messages and instead I got messages like these:
Standard query 0x626a AAAA mywebsite.com
and
Standard query response 0x626a AAAA mywebsite.com
However, if I run the same code in VSTO, I get additional messages that are errors (this one shows up twice):
TCP 60 443 → 50308 [RST, ACK] Seq=1 Ack=125 Win=32768 Len=0

how can webflux handle global error, like 404 page not found

i use #restcontrolleradvice and #ExceptionHandler , but i can handle controller exception. server error like 404, 500 can't handle.
#RestControllerAdvice
public class HttpExceptionHandler {
private static final Logger logger = LoggerFactory.getLogger(HttpExceptionHandler.class);
#ExceptionHandler(value = Exception.class)
public String exceptions(Exception e) {
String code = Global.ERR_UNKNOWN;
if (e instanceof MethodNotAllowedException) {
code = Global.ERR_HTTP_METHOD;
}
return code;
}
}
If you're using Spring Boot, this is already done for you and you can customize this support as well quite easily (see Spring Boot reference docs).
If you're using plain Spring Framework, then you need to register a custom WebExceptionHandler bean to handle that (see Spring Framework reference docs). Because those errors can happen at any point during request handling (i.e. not only during the controller handling phase, but also during response encoding, within a WebFilter...), the API there is quite low level and you need to deal with raw DataBuffer instances. If you're looking for inspiration on how to achieve higher level error handling support, you can also take a look at what's done in Spring Boot.

.Net Web API 2 POST consuming gzip compressed content

I'm writing a .Net Web Api (2) that have this one POST method. This method is currently deserializing it's only parameter by using the standard JSON formatter. We are also writing the Client that will consume this Api a C# Client using System.Net.Http.HttpClient to communicate.
There is the potential to be moving a large volume of data. This made us look into reducing the footprint of the request.
After searching this site, I came across some alternatives using gzip compression. I already have a working proof of concept:
Client side something down the lines of this
Server side something down the lines of this
So, my question...
Do I really need to write all this custom code for this? Is there a built in way to accomplish lowering the footprint of the request?
Some articles that came across mention about enabling gzip (or deflate) in IIS (see Enable IIS7 gzip). This was not working for me (I enabled it, I'm still doing the compression on the Client side, removed the DelegatingHandler from the Server...but nothing, I end up with a null parameter in the controller method)
I ended up implementing a DelegatingHandler to look for a header with ContentEncoding "gzip" and decompress accordingly.
using System;
using System.Linq;
using System.Net.Http;
using System.Threading;
using System.Threading.Tasks;
namespace WebApi.MessageHandlers
{
/// <summary>
/// GZip message handler.
/// </summary>
public class GZipMessageHandler : DelegatingHandler
{
protected override Task<HttpResponseMessage> SendAsync(HttpRequestMessage request, CancellationToken cancellationToken)
{
if (IsRequestCompressed(request))
{
request.Content = Descompress(request.Content);
}
return base.SendAsync(request, cancellationToken);
}
private bool IsRequestCompressed(HttpRequestMessage request)
{
return request.Content.Headers.ContentEncoding.Contains("gzip", StringComparer.OrdinalIgnoreCase);
}
private HttpContent Descompress(HttpContent content)
{
// Handle compression...
throw new NotImplementedException();
}
}
}

Issue with Spring Security, Spring Webflow, file uploads and UTF-8

I have a problem very similar to the one described here: File Upload using Spring WebFlow 2.4.0, parameter not binded, but that one didn't mention anything about UTF-8 issues. I'm using Spring Framework 4.1.6, Spring Security 4.0.2 and Spring Webflow 2.4.2.
It revolves around StandardServletMultipartResolver vs. CommonsMultipartResolver as far as I can tell, but I'm not sure. If I use CommonsMultipartResolver I can upload files on any page except for Webflow pages fine and UTF-8 encoding works as well on all pages. However on the Webflow pages an exception is thrown trying to access the file . If I use StandardServletMultipartResolver then all of the file uploads work, including Webflow, but on any page that has a UTF-8 character, e.g., caractère, I get garbage.
The wierd thing is I can see in FireBug that the file is being posted when I use the commons resolver. Also, if I debug the RequestContext coming from Webflow I can also see the file buried 4 levels deep in requests. The code for the common resolver (see end of post for the standard resolver code):
public FileResult uploadFile(Recipe recipe, RequestContext requestContext) {
ServletExternalContext context = (ServletExternalContext) requestContext.getExternalContext();
MultipartHttpServletRequest multipartRequest = new DefaultMultipartHttpServletRequest((HttpServletRequest)context.getNativeRequest());
MultipartFile file = multipartRequest.getFile("file");
So, is this a Spring Security issue or a Spring Webflow problem? I suspect the commons resolver would work if I could cast the RequestContext above correctly, but I've tried numerous combinations with no luck. Any guidance on this would be greatly appreciated.
Here are some relevant configurations and code:
WebMvcConfig
#Bean
public CommonsMultipartResolver filterMultipartResolver() {
CommonsMultipartResolver resolver = new CommonsMultipartResolver();
resolver.setDefaultEncoding("UTF-8");
return resolver;
}
SecurityConfig
#Override
protected void configure(HttpSecurity http) throws Exception {
CharacterEncodingFilter characterEncodingFilter = new CharacterEncodingFilter();
characterEncodingFilter.setEncoding("UTF-8");
characterEncodingFilter.setForceEncoding(true);
http
//.csrf().disable()
.addFilterBefore(characterEncodingFilter, CsrfFilter.class)
...more settings...
SecurityInitializer
#Override
protected void beforeSpringSecurityFilterChain(ServletContext servletContext) {
insertFilters(servletContext, new MultipartFilter());
}
Webflow Action
<action-state id="uploadFile">
<evaluate expression="fileActions.uploadFile(recipe, flowRequestContext)"/>
<transition to="review"/>
</action-state>
Upload file method
public FileResult uploadFile(Recipe recipe, RequestContext requestContext) {
ServletExternalContext context = (ServletExternalContext) requestContext.getExternalContext();
MultipartHttpServletRequest multipartRequest = new StandardMultipartHttpServletRequest((HttpServletRequest)context.getNativeRequest());
MultipartFile file = multipartRequest.getFile("file");
...rest of code to save the file...
Turns out you can cast the RequestContext to get at the underlying MultipartHttpServletRequest but it's not pretty. Here's what I ended up with:
Upload file method
public FileResult uploadFile(Recipe recipe, RequestContext requestContext) {
logger.debug("uploadFile");
ServletExternalContext context = (ServletExternalContext) requestContext.getExternalContext();
SecurityContextHolderAwareRequestWrapper wrapper1 = (SecurityContextHolderAwareRequestWrapper)context.getNativeRequest();
HttpServletRequestWrapper wrapper2 = (HttpServletRequestWrapper)wrapper1.getRequest();
FirewalledRequest firewall = (FirewalledRequest)wrapper2.getRequest();
MultipartHttpServletRequest multipartRequest = (DefaultMultipartHttpServletRequest)firewall.getRequest();
MultipartFile file = multipartRequest.getFile("file");
...rest of code to save the file...
Using this I get to keep the CommonsMultipartResolver, all file uploads in the app work whether Webflow or not, and I have no issues with UTF-8 and character mangling.
I'm not particularly happy with this solution (even though it works) since it's dependent upon a specific nesting of requests that could change in the future(?). I'm be interested if anyone else has run into the same UTF-8 issue and how they solved it, but for now I'm going to test the heck out of this and move on.