I have a nice question about the loop-expression of a for-loop in C#.
I always like to learn and extend my knowledge so I often use my reflector to look and learn from other. Lately I was looking around in an assembly which I assume is a C# assembly.
I came across this:
public virtual void DoRows(Parse rows)
{
for (; rows != null; {
Parse more;
rows = more;
}
)
{
more = rows.More;
this.DoRow(rows);
}
}
I never knew I could use a code-block in my loop-expression. So I fired up MSDN and went looking in the language reference of C# but found nothing.
The C++ reference shows some interesting things tho (http://msdn.microsoft.com/en-us/library/b80153d8.aspx), but I am not a C++ developper and I have a strong feeling the assmbly was not written in C++.
dotPeek says the assembly is a v4.0.30319 / msil assembly.
Now here are the questions:
- Is this a C++ or a C# code-construct?!
- Can we do things like this in C#?
I am not able to reproduce code that looks like that and compiles in VS2010.
#edit: changed the word assembly in code-construct.
#edit II: Thanks for all your answers. I think this is either a bug in the dotPeek reflector or we found some for-loop-expression easter eggs. This is what Red Gate says about the same method:
public virtual void DoRows(Parse rows)
{
while (rows != null)
{
Parse more = rows.More;
this.DoRow(rows);
rows = more;
}
}
There is no such thing as a C++ or C# assembly. Once the code is compiled, it is ILCode. The reflector takes this ILCode and decompiles it to the language you selected.
This can result in code that is valid in ILCode, but not valid in the target language.
Compilation to IL is a many-to-one operation and decompilers have rules regarding which of the many choices they choose when decompiling. Often, different languages produce IL that is not possible to reproduce directly in another language.
In this instance either the decompiler has chosen poorly and the produced C# is invalid (a bug in the decompiler), or it cannot decompile the code into equivalent C#.
Related
Main context
We're actually trying to get a multi-threading version of Ghostscript x64 DLL, to make use of it through Ghostscript .NET. This component is supposed to "allow runing multiple Ghostscript instances simultaneously within a single process", but, as we have checked in our project, works fine until concurrent requests are made to the application. Same behavior can be replicated lauching same method using Tasks. The error description that raises in both cases, just when a call to the process is made until the last is being executed, is:
An error occured when call to 'gsapi_new_instance' is made: -100
Even it does no seem to be related with .NET directly, I will post a sample of our C# method code, just for contextualize.
// Define switches...
string[] switchesArray = switches.ToArray();
using (GhostscriptProcessor procesador = new GhostscriptProcessor())
{
try
{
procesador.StartProcessing(switchesArray, null);
byte[] destinationFile = System.IO.File.ReadAllBytes(destinationPath);
return destinationFile;
}
catch (Exception ex)
{
throw ex;
}
finally
{
System.IO.File.Delete(sourceFile);
}
}
THREADSAFE solution
Starting our investigation, we found this KenS's answer on this post, indicating that Ghostscript DLL must be generated with GS_THREADSAFE compiler definition.
To clarify, as we make use of Ghostscript 9.52 x64 to generate our PDFs, we need this x64 DLL compiled for Release configuration. After trying to compile Ghostscript sources on Windows 10 x64 machine, using Visual Studio Community 2017 and Visual Studio Community 2019, we finally managed to build and generate all items (only with VS Community 2019) without GS_THREADSAFE parameter, just to confirm that compilation is fine, and we check that the DLLs and executables are working. For this process we took in mind all we found in Ghostscript official documentation.
As we have no other guide to include this GS_THREADSAFE parameter, we followed the instructions given in this solution, including XCFLAGS="-DGS_THREADSAFE=1" on nmake build commands, usign this sentence for Rebuild all option:
cd .. && nmake -f psi\msvc32.mak WIN64= SBR=1 DEVSTUDIO= XCFLAGS=-DGS_THREADSAFE=1 && nmake -f psi\msvc32.mak WIN64= DEVSTUDIO= XCFLAGS=-DGS_THREADSAFE=1 bsc
This approach, rises an error during build:
Error LNK2019 unresolved external symbol errprintf_nomem referenced in
function gs_log_error File \mkromfs.obj 1
As it seems, the file mkromfs.c has a method called errprintf_nomem, which can't be found when GS_THREADSAFE is set.
Questions
1 - Is there any public release of Ghostscript that include x64 DLLs compiled to be THREADSAFE?
And, if not (that's what I'm guessing...)
2 - Is it possible to get this DLL to be THREADSAFE without changing the source code?
3- Could anyone provide, please, a step by step guide or walkthrough to build a x64 Ghostscript DLL using GS_THREADSAFE using Visual Studio (or even any other possible alternative) over Windows 10 x64?
4 - A few posts talk about people achive to manage multithreading using Ghostscript .NET. I assume this examples are all using a GS_THREADSAFE DLL... Is any other workaround we have passed?
Thank a lot in advance.
To summarize all this questions, and as a guide for future developers having this same trouble, these are the answers we've found until now:
AS #KenS mentions in his reply: "No, the Ghostscript developers don't actually build thread-safe versions of the binaries."
At this very moment, clearly not, as it has been reported on this opened bug.
As it seems to be a matter of commercial licensing support, we avoid comment on this point anymore.
Thanks again to #HABJAN. I absolutely take back what I've stated on my question, as it is possible to have Ghostscript .NET working on multi-threading scenarios. Below comes the solution we applied, in case it could be useful for someone.
Based on HABJAN example, what we have done to achieve this was to create a custom class to capture Ghostscript logging:
protected class ConsoleStdIO : Ghostscript.NET.GhostscriptStdIO
{
public ConsoleStdIO(bool handleStdIn, bool handleStdOut, bool handleStdErr) : base(handleStdIn, handleStdOut, handleStdErr)
{
}
public override void StdIn(out string input, int count)
{
char[] userInput = new char[count];
Console.In.ReadBlock(userInput, 0, count);
input = new string(userInput);
}
public override void StdOut(string output)
{
//log
}
public override void StdError(string error)
{
//log
}
}
For our previous method, we simple include a call to this class and this avoids errors when multiple tasks are executed at the same time:
// Define switches...
string[] switchesArray = switches.ToArray();
using (GhostscriptProcessor procesador = new GhostscriptProcessor())
{
try
{
procesador.StartProcessing(switchesArray, new ConsoleStdIO(true, true, true));
byte[] destinationFile = System.IO.File.ReadAllBytes(destinationPath);
return destinationFile;
}
catch (Exception ex)
{
throw ex;
}
finally
{
System.IO.File.Delete(sourceFile);
}
}
Well, it seems to me that you are asking here for technical support.
You clearly want to use Ghostscript in a commercial undertaking, indeed one might reasonably say you want an enterprise version of Ghostscript. Presumably you don't want to alter the source in order to permit you to use an open source license, because you don't want to pay for a commercial license.
With that in mind the answers to your questions are:
No, the Ghostscript developers don't actually build thread-safe versions of the binaries.
Currently, no. That's probably an oversight.
That would be a technical support question, there's no guarantee of technical support to free users, it's the one of the few areas of leverage for dual license vendors to persuade people to take up a commercial license. So I hope you will understand that I'm not going to provide that.
as far as I can see, no.
I am working on code written in Visual C++17. But a MS specific feature is used namely properties.
Please note that I am not talking about .NET, but the behavior is similar. Most of my collegues including myself have a lot of experience in C# and hence at first glance these properties seem a handy feature.
This is not managed C++, to emphasize and avoid confusion!
The definition in Visual C++ is as follows :
virtual long GetDatumLening() = 0;
virtual void SetDatumLening(long l) = 0;
__declspec(property(get = GetDatumLening, put = SetDatumLening)) long DatumLening;
And using it in the code is like this :
if (hoofdLening->NietHypothecaireLening && hoofdLening->DatumLening < lControleDatum)
Both NietHypothecaireLening and DatumLening are properties defined in the explained manner.
Now my problem is that when I am debugging the application if I hover over those properties the values are not shown whereas properties in C# are shown even if there is some code behind it which is quite nice.
Is there a setting of some kind that I am missing in my Visual Studio environment or is this something I will have to live with ?
The fact that both the getter and setter are virtual should not make a difference, there is of course a non abstract class that implements these methods.
Thanks in advance.
I just filed this bug on Microsoft Connect regarding the inability to compile the following toy snippet of code:
template <typename... P> struct S {
template <void(*F)(P...)> static void T() { }
};
void A(int, float) { }
int main() { S<int, float>::T<&A>(); }
The error is:
test.cpp(2): error C3520: 'P' : parameter pack must be expanded in this context
Essentially, I cannot unpack a variadic type inside of a function signature when used as a template parameter. This code is (I think) legal; at least, GCC 4.7, Clang 3.0, and ICC 13 all support it.
I've seen this SO question but no workarounds are requested or given, which is what I'm looking for.
While not super critical (obviously I've been getting by without variadic templates for many years now) this pattern is of particular importance to some of the work I'd like to do and some articles I'd like to do on C++11 reflection techniques for serialization, script binding, etc. which is something I'd like to be useful to Visual Studio users (as it is by far the dominant compiler toolset in my industry). I'd like to hope that Microsoft's engineers will be able to fix this by 2013 RTM, but I'm not holding my breath.
A toy (from memory this time) sample of how this is being used is something like (minus the macros that make it slightly easier to use):
Reflect<MyType>("MyType")
.bind("GetMatrix", mat44, &MyType::GetMatrix>()
.bind("Display", void, &MyType::Display>();
Of course this can all be done without variadic templates. It just takes large masses of code and accepting limitations to the maximum arity of bound member functions, of course. And yes, passing the functions as a template parameter is of import, due to the nature of how member function pointers work in Visual Studio (variable size) and a desire to achieve efficiency on par with Impossibly Fast C++ Delegates (in my niche of the C++ community, this level of optimization can sometimes actually matter) which negates the option of using std::function or similar designs.
Is there a work-around for this VS bug? Or any other way to use variadic templates to use a (compile-time) function pointer parameter in VC++12?
Not sure why but simplifying with a typedef seems to work:
template <typename... P>
struct S
{
typedef void (*MyFunc)(P...);
template <MyFunc myFunc>
static void foo() {}
};
void foo2(int, float) {}
int main()
{
S<int, float>::foo<&foo2>();
}
At least on the Visual Studio 2013 Ultimate Preview.
Using the MFCApplication wizard in Visual C++ 2012, if "Generate attributed database class" is checked, a header with some special syntax (attributed C++ classes) are generated, which look like this:
// MFCApplication2Set.h: interface of the CMFCApplication2Set class
//
#pragma once
// code generated on March-05-13, 9:26 AM
[
db_source(L"Provider=SQLNCLI11.1;..."),
db_table(L"dbo.tblEmployee")
]
class CMFCApplication2Set
{
public:
... big list of stuff that corresponds to the fields in your db table omitted ...
}
The above header corresponds to a mostly empty implementation file:
// MFCApplication2Set.cpp : implementation of the CMFCApplication2Set class
//
#include "stdafx.h"
#include "MFCApplication2.h"
#include "MFCApplication2Set.h"
#ifdef _DEBUG
#define new DEBUG_NEW
#endif
// CMFCApplication2Set implementation
That was the WHOLE implementation class. Notice that:
A. No parent class name is specified anywhere.
B. There's some "Visual C++ magic" going on here, that is, as a novice to modern C++, I'm lost.
db_source is documented here but the documentation is pretty thin or at least opaque to me.
Questions:
I was mystified by this syntax when I first saw it, but I figured out it's probably a variant of this Attributes feature in Visual C++. It is that, right?
How am I meant to I use these generated "attributed database class" objects? I found this documentation but if you look carefully at that documentation, that code sample is showing people the Old Way and the New Way, and is not telling me what I want to know which is how to use this new object that the IDE can not give me any documentation or code completion features for. Also since the generated code for the CMFCApplication2Set class generated by the wizard does not reference any types or class names, I'm lost. If I could even use some IDE feature to know what methods and stuff have been Magically Injected into this Magical mystery Object, I'd be better off. The only think I can think to do is to learn the old way and learn all the things you can call from the old two-separate-ATL-types world, and then somehow learn to combine them.
In a nutshell, I'm looking for the minimum syntax I need to know to actually use one of these Attributed Database Class instances, variables, as they are generated in a new MFC app by the wizard. The instance shown below is a member of an MFC document class and CMFCApplication2Set m_MFCApplication2Set is declared as a field inside the MFC document class.
What I have tried is to use this "untyped object". This object appears to have lots of data fields (m_X) and has only one method that shows up in IDE code completion, called GetRowSetProperties. Thanks to whatever magic or injection is going on, this generated Attributed Database Class (which does not visibly inherit anything) is a complete mystery to me at edit time and compile time.
Here's me just trying to inspect this thing to see if it even initialized itself when its constructor ran:
BOOL CMFCApplication2Doc::OnNewDocument() { if (!CDocument::OnNewDocument()) return FALSE;
TRACE( m_MFCApplication2Set.m_AStringFieldName); // outputs NOISE.
return TRUE; }
At compile time and edit time, the IDE gives me NO help about the types involved in this "anonymous" class that inherits from nothing, but which gets lots of secret powers due to some kind of Injection via those attributes I'm guessing. At debug time, I can see that there is more than just a bunch of data fields in this C++ class, but this still doesn't help me know how to use it. A minimal code sample of using one of these to go get a recordset from the database, would be great.
Update: Calling OpenDataSource is fun, because it compiles but the IDE doesn't think it should be valid. Nevertheless, it runs, and returns 0 as the result, but that doesn't actually initialize this CThingyThatVisualStudioGaveYouThatYouDontKnowWhatItIs:
This is a deprecated feature of attributed C++ code. Pre-processor expands code and replaces attributes with actual base classes. If you enable generation of these intermediate files, things are going to be more clear to you:
You will have XXX.mrg.cpp and XXX.mrg.h files generated, which you can review and see the real C++ code forwarded to compiler.
The attributes will be replaced with substituted bases classes, maps like BEGIN_COLUMN_MAP etc. The attributed source code is compact, but the feature is deprecated and looking into expanded code it should be easy (if necessary) to strip the attributes and copy expanded code right into source. It's easy with DB attributes, and more difficult with COM attributes since the internal dependencies are most sophisticated.
Since I don't know exactly what part of it alone that triggers the error, I'm not entirely sure how to better label it.
This question is a by-product of the SO question c# code seems to get optimized in an invalid way such that an object value becomes null, which I attempted to help Gary with yesterday evening. He was the one that found out that there was a problem, I've just reduced the problem to a simpler project, and want verification before I go further with it, hence this question here.
I'll post a note on Microsoft Connect if others can verify that they too get this problem, and of course I hope that either Jon, Mads or Eric will take a look at it as well :)
It involves:
3 projects, 2 of which are class libraries, one of which is a console program (this last one isn't needed to reproduce the problem, but just executing this shows the problem, whereas you need to use reflector and look at the compiled code if you don't add it)
Incomplete references and type inference
Generics
The code is available here: code repository.
I'll post a description below of how to make the projects if you rather want to get your hands dirty.
The problem exhibits itself by producing an invalid cast in a method call, before returning a simple generic list, casting it to something strange before returning it. The original code ended up with a cast to a boolean, yes, a boolean. The compiler added a cast from a List<SomeEntityObject> to a boolean, before returning the result, and the method signature said that it would return a List<SomeEntityObject>. This in turn leads to odd problems at runtime, everything from the result of the method call being considered "optimized away" (the original question), or a crash with either BadImageFormatException or InvalidProgramException or one of the similar exceptions.
During my work to reproduce this, I've seen a cast to void[], and the current version of my code now casts to a TypedReference. In one case, Reflector crashes so most likely the code was beyond hope in that case. Your mileage might vary.
Here's what to do to reproduce it:
Note: There is likely that there are more minimal forms that will reproduce the problem, but moving all the code to just one project made it go away. Removing the generics from the classes also makes the problem go away. The code below reproduces the problem each time for me, so I'm leaving it as is.
I apologize for the escaped html characters in the code below, this is Markdown playing a trick on me, if anyone knows how I can rectify it, please let me know, or just edit the question
Create a new Visual Studio 2010 solution containing a console application, for .NET 4.0
Add two new projects, both class libraries, also .NET 4.0 (I'm going to assume they're named ClassLibrary1 and ClassLibrary2)
Adjust all the projects to use the full .NET 4.0 runtime, not just the client profile
Add a reference in the console project to ClassLibrary2
Add a reference in ClassLibrary2 to ClassLibrary 1
Remove the two Class1.cs files that was added by default to the class libraries
In ClassLibrary1, add a reference to System.Runtime.Caching
Add a new file to ClassLibrary1, call it DummyCache.cs, and paste in the following code:
using System;
using System.Collections.Generic;
using System.Runtime.Caching;
namespace ClassLibrary1
{
public class DummyCache<TModel> where TModel : new()
{
public void TriggerMethod<T>()
{
}
// Try commenting this out, note that it is never called!
public void TriggerMethod<T>(T value, CacheItemPolicy policy)
{
}
public CacheItemPolicy GetDefaultCacheItemPolicy()
{
return null;
}
public CacheItemPolicy GetDefaultCacheItemPolicy(IEnumerable<string> dependentKeys, bool createInsertDependency = false)
{
return null;
}
}
}
Add a new file to ClassLibrary2, call it Dummy.cs and paste in the following code:
using System;
using System.Collections.Generic;
using ClassLibrary1;
namespace ClassLibrary2
{
public class Dummy
{
private DummyCache<Dummy> Cache { get; set; }
public void TryCommentingMeOut()
{
Cache.TriggerMethod<Dummy>();
}
public List<Dummy> GetDummies()
{
var policy = Cache.GetDefaultCacheItemPolicy();
return new List<Dummy>();
}
}
}
Paste in the following code in Program.cs in the console project:
using System;
using System.Collections.Generic;
using ClassLibrary2;
namespace ConsoleApplication23
{
class Program
{
static void Main(string[] args)
{
Dummy dummy = new Dummy();
// This will crash with InvalidProgramException
// or BadImageFormatException, or a similar exception
List<Dummy> dummies = dummy.GetDummies();
}
}
}
Build, and ensure there are no compiler errors
Now try running the program. This should crash with one of the more horrible exceptions. I've seen both InvalidProgramException and BadImageFormatException, depending on what the cast ended up as
Look at the generated code of Dummy.GetDummies in Reflector. The source code looks like this:
public List<Dummy> GetDummies()
{
var policy = Cache.GetDefaultCacheItemPolicy();
return new List<Dummy>();
}
however reflector says (for me, it might differ in which cast it chose for you, and in one case Reflector even crashed):
public List<Dummy> GetDummies()
{
List<Dummy> policy = (List<Dummy>)this.Cache.GetDefaultCacheItemPolicy();
TypedReference CS$1$0000 = (TypedReference) new List<Dummy>();
return (List<Dummy>) CS$1$0000;
}
Now, here's a couple of odd things, the above crash/invalid code aside:
Library2, which has Dummy.GetDummies, performs a call to get the default cache policy on the class from Library1. It uses type inference var policy = ..., and the result is an CacheItemPolicy object (null in the code, but type is important).
However, ClassLibrary2 does not have a reference to System.Runtime.Caching, so it should not compile.
And indeed, if you comment out the method in Dummy that is named TryCommentingMeOut, you get:
The type 'System.Runtime.Caching.CacheItemPolicy' is defined in an assembly that is not referenced. You must add a reference to assembly 'System.Runtime.Caching, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a'.
Why having this method present makes the compiler happy I don't know, and I don't even know if this is linked to the current problem or not. Perhaps it is a second bug.
There is a similar method in DummyCache, if you restore the method in Dummy, so that the code again compiles, and then comment out the method in DummyCache that has the "Try commenting this out" comment above it, you get the same compiler error
OK, I downloaded your code and can confirm the problem as described.
I have not done any extensive tinkering with this, but when I run & reflector a Release build all seems OK (= null ref exception and clean disassembly).
Reflector (6.10.11) crashed on the Debug builds.
One more experiment: I wondered about the use of CacheItemPolicies so I replaced it with my own MyCacheItemPolicy (in a 3rd classlib) and the same BadImageFormat exception pops up.
The exception mentions : {"Bad binary signature. (Exception from HRESULT: 0x80131192)"}