I am using this Julia blog Capture Output in Julia to redirect the errors. However, due to asynchronous nature atexit function is called before STDERR is populated and the program hang when trying to read from empty stream. Related problems have been discussed in some Julia forums but I couldn't make a complete sense out of it.
I have partially solved the problem by catching the exception and then writing the exception to STDERR but the trace gets lost in the process and I only get to know the exception and not it's origin.
Can someone please help me? Is there a better solution to it? - Thanks
I am using Julia v0.5 on Mac Os X El Capitan v10.11.4
Below is a sample code that I am trying to run
errorRead, errorWrite = redirect_stderr()
atexit(function()
close(errorWrite)
redirect_stderr(STDOUT)
for err in eachline(errorRead)
println(err)
end
close(errorRead)
end)
try erroneousfunc()
catch err
println(STDERR, err)
exit(0)
end
Related
I'm having trouble with my Deno program. I'm getting messages like this:
error: Uncaught (in promise) Error: No such host is known. (os error 11001)
at deno:core/01_core.js:106:46
at unwrapOpResult (deno:core/01_core.js:126:13)
at async Object.connect (deno:extensions/net/01_net.js:219:13)
Then deno exits.
I don't know how to debug this. This stack trace only points to code that comes with Deno, not to my code.
I've searched my code and I've put a .catch() or a try/catch everywhere I can think of, but that did not help.
Is there anything I can do to help me find the problem? I'd love it if I could get a complete stack dump. Or if I could have the debugger pause at the problem. Or if you have any other suggestions.
Thanks!
Edit 8/29/2021
I found two bugs in my code. Here are the actual bugs. It was a serious pain to track these down. I'm still looking for a tool or process to help the next time I make a silly mistake like this.
Bug #1:
I was using try/catch (shown in red) when I should have been using .catch() (shown in green). My try/catch did nothing. If there was an error sending the data, that would cause my program to crash.
Bug #2:
const promise = Deno.connect(options);
promise.catch(reportError);
promise.then(longRunningTask);
await someOtherPromise;
promise.then(connection => {
// We never get to here.
try {
connection.close();
} catch {
console.log("🙁");
}
});
// And we never get to here.
The code I've shown here was spread throughout a much longer program. I did not understand the rules regarding promises. The second .then() requires a second .catch().
Here was one of my attempts to solve this problem. I told VS code to break on all exceptions. It seems to ignore my request. I never got to a breakpoint, but the debug console shows that the program crashed because of an exception.
In Lablgtk, whenever there is an exception in a callback, the exception is automatically caught and an error message is printed in the console, such as:
(prog:12345) LablGTK-CRITICAL **: gtk_tree_model_foreach_func:
callback raised an exception
This gives no stack trace and no details about the exception, and because it is caught I cannot retrieve this information myself.
Can I enable more detailed logging information for this case? Or prevent the exception from being caught automatically?
I guess the best way to do so is to catch your exception manually and handle it yourself.
let callback_print_exn f () =
try f () with
e -> my_exn_printer e
Assuming val my_exn_printer : exn -> unit is your custom exception printer, you can simply print your callbacks exceptions by replacing ~callback:f by ~callback:(callback_print_exn f) in your code.
Of course, you can also with that method send that exception to another
thread, register a "callback id" that would be passed along with your exception...
About the stack trace, I'm not sure you can retrieve it easily. As it's launched as a callback, you probably want to know the signal used and that can be stored in your callback handler.
I had another similar issue, but this time it was harder to find where to put the calls to intercept the exception.
Fortunately, this time there was a very specific error message coming from the Glib C code:
GLib-CRITICAL **: Source ID ... was not found when attempting to remove it`
Stack Overflow + grep led me to the actual C function, but I could not find which of the several Lablgtk functions bound to this code was the culprit.
So I downloaded the Glib source, added an explicit segmentation fault to the code, compiled it and used LD_LIBRARY_PATH to force my modified Glib version to be loaded.
Then I ran the OCaml binary with gdb, and I got my stack trace, with the precise line number where the Lablgtk function was being called. And from there it was a quick 3-line patch.
Hacks like this one (which was still faster than trying to find where to intercept the call) could be avoided by having a "strict mode" preventing exceptions from being automatically caught. I still believe such a switch should be available for Lablgtk users, and hope it will eventually be available.
I'm working on some flash app. Now, to test customer side of it I can use Flash Player debugger version that will save logs and show error messages. When it's deployed on the customer side - they will have a regular Flash Player version which means I will have no access to error messages if errors will happen. So I would like to equip it with some tool that would capture all of my trace messages in code and errors text. As for trace messages that's fairly simple, I just override the function in my code so it sends a POST request with trace message to a logger server, but how can I get a hold of the error message? Is there a known approach to this or some trick that somebody can suggest?
You can install the debug version of flash as your browser's default (in Chrome, you must disable the built-in player), so if you wanted to test user experience and debug, this would be the ideal solution.
However, to answer your question: there's no method for universally catching all errors, and redirecting them (that I know of). You'd have to encapsulate problem code ahead of time with try...catch statements, and send the property back on catch. For example:
try {
this["foo"]();
} catch (e:Error) {
trace(e);
}
In the debug version, the traced value would be TypeError: Error #1006: value is not a function. And while the standard version will only output TypeError: Error #1006, (a notably less descriptive error), what we're missing is any reference to where the error occured. To get this, we need to use Error.getStackTrace() to see the call stack and the line where the error occurred. In debug, this outputs the following:
TypeError: Error #1006: value is not a function.
at Shell_fla::MainTimeline/init()[C:\Projects\shell.as:91
In the standard client, we get a dissapointing null. In short, you cannot get any valuable info from the client versions.
The best advice I can give is to write around your problem code with your own custom error reports. For example, catch IO errors and trace the file it failed to load, or if you're expecting an object.foo, first try if (object.hasOwnProperty("foo")) { // do something } else { trace("foo not found in " + object.name) }. Code defensively.
Cheers,
I've discovered this post on StackOverflow:
How to catch all exceptions in Flex?
It answers my question, strange that I haven't ran into it while I was googling prior to asking.
I've just been scratching my hairs for a long time with an error I couldn't find. It turned out to be an arity exception but apparently because it happened on the EDT I couldn't "see" it. It didn't show in the "lein run" terminal when run from a terminal and it didn't show in any Emacs buffer when run from Emacs.
After a very long time I ended up doing this:
(try (function-call-with-arity-error ...) (catch Exception e (println e)
and, at last, thanks to this, I got to see this printed:
#<ArityException clojure.lang.ArityException: wrong number of args passed to...
and hence I've been able to find my error.
And if I do this:
(do
(println "trying...")
(arity-error-here-on-purpose) ; this ones throws the arity error
(println "done")
)
Then the terminal prints "trying..." but never gets to "done...".
I tried setting a default uncaught exception handler: the exception isn't getting caught. It's as if the program or the EDT was "stuck" after the arity exception (without printing anything anywhere).
How am I suppose to deal with this the next time? Because I couldn't see any message anywhere it took me a really long time to find. Once again: nothing in the 'lein run' terminal and nothing in any Emacs buffer.
Should I create a function that wraps calls that should happen on the EDT inside try / catch manually and that then logs/println the exception?
Also note that this is in a relatively "long" Clojure app: 1000 lines of code, so I can't paste it here and I couldn't reproduce that behavior in a short example (but it happens consistently in my app).
You can register an uncaught exception handler with a Thread so you don't have to wrap all the calls that can occur on the EDT.
Thread.UncaughtExceptionHandler
Try registering one with the EDT and seeing if this gives you the desired behavior.
I have a DLL with some COM objects. Sometimes, this objects crashes and register an error event in the Windows Event Log with lots of hexadecimal informations. I have no clue why this crashes happens.
So, How can I trace those COM objects exceptions?
The first step is to lookup the Fail code's hex value (E.G. E_FAIL 0x80004005). I've had really good luck with posting that value in Google to get a sense of what the error code means.
Then, I just use trial and error to try to isolate the location in code that's failing, and the root cause of the failure.
If you just want a really quick way to find out what the error code means, you could use the "Error Lookup" tool packaged with Visual Studio (details here). Enter the hex value, and it will give you the string describing that error code.
Of course, once you know that, you've still got to figure out why it's happening.
A good way to look up error (hresult) codes is HResult Plus or welt.exe (Windows Error Lookup Tool).
I use logging internally in the COM-classes to see what is going on. Also, once the COM-class is loaded by the executable, you can attach the VS debugger to it and debug the COM code with breakpoints, watches, and all that fun stuff.
COM objects don't throw exceptions. They return HRESULTs, most of which indicate a failure. So if you're looking for the equivalent of an exception stack trace, you're out of luck. You're going to have to walk through the code by hand and figure out what's going on.