Edit:
For personn interested in a cleaner way to implemenent that, have a look to that answer.
In my job I often need to use third-made API to access remote system.
For instance to create a request and send it to the remote system:
#include "external_lib.h"
void SendRequest(UserRequest user_request)
{
try
{
external_lib::Request my_request;
my_request.SetPrice(user_request.price);
my_request.SetVolume(user_request.quantity);
my_request.SetVisibleVolume(user_request.quantity);
my_request.SetReference(user_request.instrument);
my_request.SetUserID(user_request.user_name);
my_request.SetUserPassword(user_request.user_name);
// Meny other member affectations ...
}
catch(external_lib::out_of_range_error& e)
{
// Price , volume ????
}
catch(external_lib::error_t& e)
{
// Here I need to tell the user what was going wrong
}
}
Each lib's setter do checks the values that the end user has provided, and may thow an exception when the user does not comply with remote system needs. For instance a specific user may be disallowed to send a too big volume. That's an example, and actually many times users tries does not comply: no long valid instrument, the prices is out of the limit, etc, etc.
Conseqently, our end user need an explicit error message to tell him what to modify in its request to get a second chance to compose a valid request. I have to provide hiim such hints
Whatever , external lib's exceptions (mostly) never specifies which field is the source
of aborting the request.
What is the best way, according to you, to handle those exceptions?
My first try at handling those exceptions was to "wrap" the Request class with mine. Each setters are then wrapped in a method which does only one thing : a try/catch block. The catch block then throws a new exceptions of mine : my_out_of_range_volume, or my_out_of_range_price depending on the setter. For instance SetVolume() will be wrapped this way:
My_Request::SetVolume(const int volume)
{
try
{
m_Request.SetVolume(volume);
}
catch(external_lib::out_range_error& e)
{
throw my_out_of_range_volume(volume, e);
}
}
What do you think of it? What do you think about the exception handling overhead it implies? ... :/
Well the question is open, I need new idea to get rid of that lib constraints!
If there really are a lot of methods you need to call, you could cut down on the code using a reflection library, by creating just one method to do the calling and exception handling, and passing in the name of the method/property to call/set as an argument. You'd still have the same amount of try/catch calls, but the code would be simpler and you'd already know the name of the method that failed.
Alternatively, depending on the type of exception object that they throw back, it may contain stack information or you could use another library to walk the stack trace to get the name of the last method that it failed on. This depends on the platform you're using.
I always prefer a wrapper whenever I'm using third party library.
It allows me to define my own exception handling mechanism avoiding users of my class to know about external library.
Also, if later the third party changes the exception handling to return codes then my users need not be affected.
But rather than throwing the exception back to my users I would implement the error codes. Something like this:
class MyRequest
{
enum RequestErrorCode
{
PRICE_OUT_OF_LIMIT,
VOLUME_OUT_OF_LIMIT,
...
...
...
};
bool SetPrice(const int price , RequestErrorCode& ErrorCode_out);
...
private:
external_lib::Request mRequest;
};
bool MyRequest::SetPrice(const int price , RequestErrorCode& ErrorCode_out)
{
bool bReturn = true;
try
{
bReturn = mRequest.SetPrice(price);
}
catch(external_lib::out_of_range_error& e)
{
ErrorCode_out = PRICE_OUT_OF_LIMIT;
bReturn = false;
}
return bReturn;
}
bool SendRequest(UserRequest user_request)
{
MyRequest my_request;
MyRequest::RequestErrorCode anErrorCode;
bool bReturn = my_request.SetPrice(user_request.price, anErrorCode);
if( false == bReturn)
{
//Get the error code and process
//ex:PRICE_OUT_OF_LIMIT
}
}
I think in this case I might dare a macro. Something like (not tested, backslashes omitted):
#define SET( ins, setfun, value, msg )
try {
ins.setfun( value );
}
catch( external::error & ) {
throw my_explanation( msg, value );
}
and in use:
Instrument i;
SET( i, SetExpiry, "01-01-2010", "Invalid expiry date" );
SET( i, SetPeriod, 6, "Period out of range" );
You get the idea.
Although this is not really the answer you are looking for, but i think that your external lib, or you usage of it, somehow abuses exceptions. An exception should not be used to alter the general process flow. If it is the general case, that the input does not match the specification, than it is up to your app to valid the parameter before passing it to the external lib. Exceptions should only be thrown if an "exceptional" case occurrs, and i think whenever it comes to doing something with user input, you usually have to deal with everything and not rely on 'the user has to provide the correct data, otherwise we handle it with exceptions'.
nevertheless, an alternative to Neil's suggestions could be using boost::lambda, if you want to avoid macros.
In your first version, you could report the number of operations that succeeded provided the SetXXX functions return some value. You could also keep a counter (which increases after every SetXXX call in that try block) to note what all calls succeeded and based on that counter value, return an appropriate error message.
The major problem with validating each and every step is, in a real-time system -- you are probably introducing too much latency.
Otherwise, your second option looks like the only way. Now, if you have to write a wrapper for every library function and why not add the validation logic, if you can, instead of making the actual call to the said library? This IMO, is more efficient.
Related
Suppose I have this in a custom Stack implementation:
void Pop (Stack & s) {
assert (!isEmpty(s));
// implementation details of popping omitted
}
Suppose I want to catch bad client uses like popping an empty stack. "Assert" is great until you compile the production version and disable it via the NDEBUG flag. What's the most professional way of dying gracefully if you detect an unrecoverable error, assuming that you are going to turn off assertions eventually?
Yes, I know there are a bunch of options: use "exit"; change the function to return a C-like error code; use C++ exception handling; just ignore the offending operation (make bad calls into a no-op); implement a personal version assert called something else that won't get turned off, etc.
What the "most pro" thing to do here? I just want to die quickly, with a helpful message.
For fun, I created a vector instance from the standard library and popped it empty. This caused a seg fault, which might be acceptable for the standard library, but I want to catch and report such a problem before dying.
It is often used self defined assert macro with logging in production code. For example,
#define MYCOMPANY_ASSERT_FATAL(expression, msg, ret_val) if (!expression){ logger.fatal(msg); return ret_val;}
#define MYCOMPANY_ASSERT_WARN(...) .....
#define MYCOMPANY_ASSERT_ERROR(...) .....
Basically, the severity level of the error which should be regarded as exiting or messaging or so is dependent on your application program context layer. In short, some_lib_func() couldn't decide to exit an own process in most cases. That's why some_lib_func() should propagate the error information to caller. For example,
// library or middle module func couldn't decide to handle own application process, so after only logging return to caller
bool read_csv(double& val) {
double val = 0;
bool ret = parse_csv_from_file(val);
MYCOMPANY_ASSERT_ERROR(ret, "failed to parse_csv_from_file", false);
// return to caller with logging and error info(true/false here)
....
....
}
// application layer and need to handle for error which should be continue or message for user..
bool show_weather_in_foreign_country() {
bool ret = read_csv();
if (!ret) {
show_error_message();
}
// in this case read_csv error is trivial and you want to continue to process with proper message
...
}
In this case, showing weather is considered as trivial process in your application and you don't want to exit a whole application if error, so after showing a appropriate message and continue to process. In the other hand, another case below is critical, despite using the same function of read_csv.
// application layer and need to handle for error which should be continue or message for user..
bool send_your_account_balance_in_securiities_to_your_wife() {
double val = 0;
bool ret = read_csv(val);
MYCOMPANY_ASSERT_FATAL(ret, "critical in my account balance", false);
// in this application context, the failure of read_csv is critical and never continue to process
// and return to caller. Caller should probably exit an own application
send_email_your_wife(val);
// if send 0$ with some kind of mistake to you wife, she might die or kill you ...
...
}
Therefore, preparing some of defined macro for propagating error and logging is so useful and make your code simple and safe, and then you need to use them properly depending on your application contexts.
I am building a .Net Profiler for some custom requirement.
I want to hook at Enter and Leave for some specific methods. To achieve this, I have tried below two approaches.
IL Rewrite - I am able to inject the custom code at both the places. It is successfully getting injected and calling the custom code. I am also able to get the input arguments, 'this' and return value. Injecting at the Enter is not much difficult. However, it's complex to inject at Leave as there could be return at multiple places in a method. I have to inject the code at every place where return statement is written.
It's bit complex but somewhat doable. But, if there any exception, the execution is not reaching to return statement and hence my injected code is not getting invoked.
Subscribe to Enter/Leave through SetEnterLeaveFunctionHooks2 in ICorProfilerInfo2 as per the sample code given here.
In both the cases, hook at the Leave is not getting invoked in case of exception in the method.
How to handle this? I want a return value in all the scenarios. In case of an exception, I should know there is an exception; I will consider as 'No return value'. Probably, I may need exception details as well.
Below is a sample method. I want to hook at Enter and Leave for GetString method. It has multiple returns. I am able to capture the return value in a normal scenario. But in case of exception, execution stops immediately and due to that the hook at return is not getting invoked.
public int GetInt()
{
//int retVal = 10;
int retVal = 1010;
//throw new Exception("test");
return retVal;
}
public string GetString()
{
var retunValue = "Return string ";
if (GetInt() > 100)
{
retunValue += " inside IF > 100";
return retunValue;
}
return retunValue + " at last return";
}
To get the exception notification when using IL re-writing, you need to inject a try-finally or try-catch-throw. Since the ret instruction is not valid within a try block, you will need to replace them with a leave instruction that branches to an instruction after the inserted exception handler and return from there.
Another option is to include COR_PRF_MONITOR_EXCEPTIONS in your call to SetEventMask and listen on the ExceptionUnwindFunctionEnter and ExceptionUnwindFunctionLeave callbacks. These callbacks don't include the thrown exception however. You could track the exception from ExceptionThrown, but this may be misleading when an exception leaves a filter block as the runtime will treat it as returning false and continue with the previous exception, but IIRC, there is no callback to indicate when this happens.
I made a ProtocolBuffer object from the proto class I usually use and I need to Serialize it. Now, I take the object and call SerializeToArray() on it like this:
int size = messageObject.ByteSize();
void* buffer = malloc(size);
messageObject.SerializeToArray(buffer, size);
As far as I know there is no problem with this since the object has data in it (I checked it by breaking right before the Serialize line).
When the method calls however it triggers an abort() which I don't know anything about.
I have no idea what it could be. The only data that is included in this object is a "type" enumerator (which I can set to the type of data that is being used in this object since it can include different sorts of messages) and it holds one message object of the repeatable type.
message MessageID
{
enum Type { LOGINDATA = 1; PLAYERDATA = 2; WORLDDATA = 3; }
// Identifies which field is filled in.
required Type type = 1;
// One of the following will be filled in.
repeated PlayerData playerData = 2;
optional WorldData worldData = 3;
optional LoginData loginData = 10;
}
This is the base message. So, Type is 2 in this case which stands for PLAYERDATA. Also, playerData is being set with a single object of the type PlayerData.
An help is appreciated.
Any time that the protobuf library aborts (which, again, should only be in debug mode or in sever circumstances), it will print information about the problem to the console. If your app doesn't have a console, you can use google::protobuf::SetLogHandler to direct the information somewhere else:
https://developers.google.com/protocol-buffers/docs/reference/cpp/google.protobuf.common#SetLogHandler.details
typedef void LogHandler(LogLevel level, const char* filename,
int line, const std::string& message);
LogHandler* SetLogHandler(LogHandler* new_func);
The protobuf library sometimes writes warning and error messages to stderr.
These messages are primarily useful for developers, but may also help end users figure out a problem. If you would prefer that these messages be sent somewhere other than stderr, call SetLogHandler() to set your own handler. This returns the old handler. Set the handler to NULL to ignore log messages (but see also LogSilencer, below).
Obviously, SetLogHandler is not thread-safe. You should only call it at initialization time, and probably not from library code. If you simply want to suppress log messages temporarily (e.g. because you have some code that tends to trigger them frequently and you know the warnings are not important to you), use the LogSilencer class below.
The only reason for an abort that I know of (which only applies in debug builds) is if some required field isn't set. You say that the type field is set, so there must be a required field in PlayerData which is not set.
I am trying to determine if a given XML file is valid (has proper syntax and structure), and I am using Xerces. I have been able to succesfully read proper files but when I give it files with incorrect syntax, no errors are thrown.
I have been fishing around and found out that I might have to use an Error handler and user setErrorHandler to catch the errors instead of the traditional try-throw-catch exception handling.
The problem that I am having though is that I am very confused how to declare the proper handler, set it to my parser and then read the errors if there are any that show up.
Is there any chance somebody could shed some light on my situation?
// #input_parameter from function: const string & xmlConfigArg
xercesc::DOMDocument* doc = NULL;
string xmlConfig(xmlConfigArg);
Handler handler; // I'm not sure what type of handler to use
_parser->setErrorHandler(&handler);
try{
_parser->parse(xmlConfigArg.c_str());
doc = _parser-> getDocument();
}catch(...){
//Nothing is ever caught here
}
You need to derive a class from ErrorHandler (< xercesc/sax/ErrorHandler.hpp >)
then overwrite all the virtual methods there.
After doing so, You can get the error code from the class you created. No exceptions will be thrown in the parsing, so you can wave the try/cache block (or keep it for a different use).
I'm designing an interface that can be used to report errors in C++. (I'm working with a legacy system where exceptions are out of question.) In my youthful naivety, I started along these lines while designing my API:
bool DoStuff(int amount, string* error);
Return value signals success/failure, while error is used to report a human readable explanation. So far so good. Subroutine calls passed along the error pointer and everything was hunky-dory.
I ran into the following problems with this design (so far):
Cannot report warnings.
Not thread-safe.
Next, I decided to go with the following interface, instead of plain string:
class Issues {
public:
void Error(const string& message);
void Warning(const string& message);
void Merge(const Issues& issues);
}
So that I can change my API like this:
bool DoStuff(int amount, Issues* issues);
I'm wondering, is there a more generic/standard API out there that deals with this problem? If yes, I'd like to take a look.
UPDATE: I'm not looking for a logging library. For those who are curious, imagine you're writing a query engine that includes a compiler. The compiler issues warnings and errors, and those need to be returned to the user, as part of the response. Logging has its place in the design, but this is not it.
I usually use things like boost::signals or .NET delegates to report errors/warning/logging/whatever. You report errors with no changes to the interface, and the library user plugs whatever she wants to the signal to get the error reports (writing to a file, updating a console window, aborting the program, throwing an exception, ignoring warnings, etc).
Something like this, at eg. global scope:
boost::signal<void(std::string const&)> logError;
boost::signal<void(std::string const&)> logWarning;
and then
void routineWhichMayFail()
{
...
if (answer != 42)
{
logError("Universal error");
return;
}
}
and you connect something to logError and logWarning at initialization:
void robustErrorHandler(std::string const& msg)
{
std::cerr << "Error: " << msg << "\n";
std::exit(EXIT_FAILURE);
}
void initializeMyProgram()
{
logError.connect(&robustErrorHandler);
}
You can even throw exceptions in the error handler instead of exiting, and use fancier things than bare functions (logging classes, "delegates" -- pointers to methods with a this object bundled, RPC to a distant server). This way, you decouple the error handling from error reporting, which is good. You can also report to multiple destinations, you can even have your handlers return a boolean telling whether the action should be eg. retried.
From your explanation it sounds like you are trying to implement a logging library for your project. You can look at log4cpp or Boost.Log.