<div dir="ltr">Sorry to confuse this thread with a second Michael.<div><br></div><div>I spoke with Michael Sloan about this proposal before he made it. I'm usually very hesitant to introduce breaking changes to core APIs, but in this case (1) there's a very good use case that's currently excluded by the API, and (2) Michael Sloan figured out some great ways to minimize the breakage. You could even argue that this proposal has *no* API breakage.</div><div><br></div><div>I'm +1 on including it. I'm also hopeful that we can address the `error = errorWithStackTrace`, but that would really be a step 2 after the changes to SomeException are made.<br><br><div class="gmail_quote">On Wed, Apr 15, 2015 at 5:56 AM Carter Schonwald <<a href="mailto:carter.schonwald@gmail.com">carter.schonwald@gmail.com</a>> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Hey Michael,<div>I actually proposed something along these lines that got OK'd by libraries early this past fall, the main challenge we hit was actually doing the enginering to add the stack traces to exceptions! theres some nasty module cycles in base that happen when you try to weave things around so that the standard error "message here" call includes some stack trace info. Have you tried to do that simple starter patch to base? </div><div><br></div><div>Chris Allen and I spent like 2 days trying to get it to work and just gave up because of the cycles. We (and others) would probably love some headway on that front.</div><div><br></div><div>Theres also some in progress work to use the dwarf debugging info data in >7.10 to provide useful stack traces in the default builds for GHC afaik, 'cause the stack trace functionality you're pointing at currenlty only work on profiled builds</div><div><br></div><div>cheers</div></div><div dir="ltr"><div>-Carter</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Apr 14, 2015 at 2:38 PM, Michael Sloan <span dir="ltr"><<a href="mailto:mgsloan@gmail.com" target="_blank">mgsloan@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Control.Exception currently lacks a good way to supply extra<br>information along with exceptions. For example, exceptions could be<br>thrown along with their callstack[1] or implicit stack[2], but we have<br>no generic way to include this information with exceptions.<br><font face="arial, helvetica, sans-serif"><br>Proposed Solution<br>=================<br><br></font>The proposed solution is to add a list of `SomeExceptionInfo` to the<br>`SomeException` datatype. This list stores additional information<br>about the exception. These `ExceptionInfo` instances use a mechanism<br>which is pretty much identical to the dynamic way the `Exception` type<br>works:<br><br><font face="arial, helvetica, sans-serif"> data SomeException = forall e . Exception e =><br> SomeExceptionWithInfo e [SomeExceptionInfo]<br><br> data SomeExceptionInfo = forall a . ExceptionInfo a =><br> SomeExceptionInfo a<br><br> class Typeable a => ExceptionInfo a where<br> displayExceptionInfo :: a -> String<br><br> addExceptionInfo<br> :: (ExceptionInfo a, Exception e)<br> => a -> e -> SomeException<br> addExceptionInfo x (toException -> SomeExceptionWithInfo e xs) =<br> SomeExceptionWithInfo e (SomeExceptionInfo x : xs)<br></font><br>`ExceptionInfo` lacks the to / from functions that `Exception` has,<br>because I don't see much point in supporting a hierarchy for exception<br>info. The `Typeable` superclass constraint supplies the necessary<br>casting.<br><br>`SomeExceptionInfo` could validly instead just use the constraint<br>`(Typeable a, Show a)`. However, I believe it's good to have a new<br>class for this so that:<br><br> * The user can specify a custom `displayExceptionInfo`<br> implementation, for when this extra info is presented to the user.<br> This function would be invoked by the `show` implementation for<br> `SomeException`.<br><br> * Types need to opt-in to be usable with `SomeExceptionInfo`.<br> Similarly to exceptions, I imagine that a type with a<br> `ExceptionInfo` instance won't be used for anything but acting as<br> such an annotation. Having a class for this allows you to ask GHCI<br> about all in-scope exception info types via `:info ExceptionInfo`.<br><br>Backwards Compatibility<br>=======================<br><br>GHC 7.10 adds support for bidirectional pattern synonyms. This means<br>that this change could be made without breaking code:<br><br> pattern SomeException x <- SomeExceptionWithInfo x _ where<br> SomeException x = SomeExceptionWithInfo x []<br><br>Note that consumers of this do not need to enable `-XPatternSynonyms`.<br><br>Applications<br>============<br><br>Callstacks<br>----------<br><br>As mentioned at the beginning, this can be used to add callstacks to<br>exceptions:<br><br> newtype ExceptionCallStack =<br> ExceptionCallStack { unExceptionCallStack :: [String] }<br> deriving Typeable<br><br> instance ExceptionInfo ExceptionCallStack where<br> displayExceptionInfo = unlines . unExceptionCallStack<br><br> throwIOWithStack :: Exception e => e -> IO a<br> throwIOWithStack e = do<br> stack <- currentCallStack<br> if null stack<br> then throwIO e<br> else throwIO (addExceptionInfo (ExceptionCallStack stack) e)<br><br>I see little downside for making something like this the default<br>implementation `throwIO`. Each rethrowing of the `SomeException`<br>would add an additional stacktrace to its annotation, much like the<br>output of `+RTS -xc`. Unlike this debug output, though, the<br>stacktraces would be associated with the exception, rather than just<br>listing locations that exceptions were thrown. This makes it<br>tractable to debug exceptions that occur in concurrent programs, or in<br>programs which frequently throw exceptions during normal functioning.<br><br>Throwing Exceptions in Handlers<br>-------------------------------<br><br>Example:<br><br> main =<br> throwIO InformativeErrorMessage `finally`<br> throwIO ObscureCleanupIssue<br><br>While `InformativeErrorMessage` got thrown, the user doesn't see it,<br>since `ObscureCleanupIssue` is thrown instead. This causes a few<br>issues:<br><br>1. If the exception is handled by the default handler and yielded to<br> the user, then the more informative error is lost.<br><br>2. Callers who expect to catch the "Informative error message" won't<br> run their handlers for this exception type.<br><br>Problem 1 can now easily be resolved by adding some info to the<br>exception:<br><br> data ExceptionCause = ExceptionCause<br> { unExceptionCause :: SomeException }<br> deriving Typeable<br><br> instance ExceptionInfo ExceptionCause where<br> displayExceptionInfo fe =<br> "thrown while handling " ++<br> displayException (unExceptionCause fe)<br><br> catch :: Exception e => IO a -> (e -> IO a) -> IO a<br> catch f g = f `oldCatch` handler<br> where<br> handler ex = g ex `oldCatch` \(ex' :: SomeException) -><br> throwIO (addExceptionInfo info ex')<br> where<br> info = ExceptionCause (toException ex)<br><br>This implementation of `catch` is written in a backwards-compatible<br>way, such that the exception thrown during finalization is still the<br>one that gets rethrown. The "original" exception is recorded in the<br>added info. This is the same approach used by Python 3's<br>`__context__` attribute[3]. This was brought to my attention in a<br>post by Mike Meyer[4], in a thread about having bracket not suppress<br>the original exception[5].<br><br>This doesn't directly resolve issue #2, due to this backwards<br>compatibility. With the earlier example, a `catch` handler for<br>`InformativeErrorMessage` won't be invoked, because it isn't the<br>exception being rethrown. This can be resolved by having a variant of<br>catch which instead throws the original exception. This might be a<br>good default for finalization handlers like `bracket` and `finally`.<br><br>Asynchronous Exceptions<br>-----------------------<br><br>Currently, the only reliable way to catch exceptions, ignoring async<br>exceptions, is to fork a new thread. This is the approach used by the<br>enclosed-exceptions[6] package. I think it's quite ugly that we need<br>to go to such lengths due to the lack of one bit of information about<br>the exception! This would resolve ghc trac #5902[7].<br><br>base-4.7 added the `SomeAsyncException` type, but this doesn't enforce<br>anything. Any exception can be thrown as a sync or async exception.<br>Instead, we ought to have a reliable way to know if an exception is<br>synchronous or asynchronous. Here's what this would look like:<br><br> data IsAsync = IsAsync<br> deriving (Typeable, Show)<br><br> instance ExceptionInfo IsAsync where<br> displayExceptionInfo IsAsync = "thrown asynchronously"<br><br> throwTo :: Exception e => ThreadId -> e -> IO ()<br> throwTo tid = oldThrowTo tid . addExceptionInfo IsAsync<br><br>The details of this get a bit tricky: What happens if `throwIO` is<br>used to rethrow a `SomeException` which has this `IsAsync` flag set?<br>I'm going to leave out my thoughts on this for now as the interactions<br>between unsafePerformIO and the concept of "rethrowing" async<br>exceptions. Such details are explained in a post by Edsko de Vries[8]<br>and ghc trac #2558[9].<br><br>Issue: fromException loses info<br>===============================<br><br>I can think of one main non-ideal aspect of this proposal:<br><br>Currently, the `toException` and `fromException` methods usually form<br>a prism. In other words, when `fromException` yields a `Just`, you<br>should get the same `SomeException` when using `toException` on that<br>value.<br><br>For example,<br><br> fail "testing 1 2 3" `catch` \(ex :: SomeException) -> throwIO ex<br><br>is equivalent to<br><br> fail "testing 3 4 5" `catch` \(ex :: IOError) -> throwIO ex<br><br>However, with exception info added to just `SomeException`, and no<br>changes to existing `Exception` instances, this<br>doesn't hold. Exceptions caught as a specific exception type get<br>rethrown with less information.<br><br>One resolution to this is be to add `[SomeExceptionInfo]` as a field<br>to existing `Exception` instances. This would require the use of<br>non-default implementations of the `toException` and `fromException`<br>instances.<br><br>Another approach is to have variants of `catch` and `throw` which also<br>pass around the `[SomeExceptionInfo]`.<br><br>[1] <a href="https://hackage.haskell.org/package/base-4.8.0.0/docs/GHC-Stack.html#currentCallStack" target="_blank">https://hackage.haskell.org/package/base-4.8.0.0/docs/GHC-Stack.html#currentCallStack</a><br>[2] <a href="https://ghc.haskell.org/trac/ghc/wiki/ExplicitCallStack/ImplicitLocations" target="_blank">https://ghc.haskell.org/trac/ghc/wiki/ExplicitCallStack/ImplicitLocations</a><br>[3] <a href="https://www.python.org/dev/peps/pep-3134/" target="_blank">https://www.python.org/dev/peps/pep-3134/</a><br>[4] <a href="https://mail.haskell.org/pipermail/haskell-cafe/2014-July/114987.html" target="_blank">https://mail.haskell.org/pipermail/haskell-cafe/2014-July/114987.html</a><br>[5] <a href="https://mail.haskell.org/pipermail/haskell-cafe/2014-July/114986.html" target="_blank">https://mail.haskell.org/pipermail/haskell-cafe/2014-July/114986.html</a><br>[6] <a href="https://hackage.haskell.org/package/enclosed-exceptions" target="_blank">https://hackage.haskell.org/package/enclosed-exceptions</a><br>[7] <a href="https://ghc.haskell.org/trac/ghc/ticket/5902" target="_blank">https://ghc.haskell.org/trac/ghc/ticket/5902</a><br>[8] <a href="http://www.edsko.net/2013/06/11/throwTo/" target="_blank">http://www.edsko.net/2013/06/11/throwTo/</a><br>[9] <a href="https://ghc.haskell.org/trac/ghc/ticket/2558" target="_blank">https://ghc.haskell.org/trac/ghc/ticket/2558</a><br></div>
<br>_______________________________________________<br>
Libraries mailing list<br>
<a href="mailto:Libraries@haskell.org" target="_blank">Libraries@haskell.org</a><br>
<a href="http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries" target="_blank">http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries</a><br>
<br></blockquote></div><br></div>
_______________________________________________<br>
Libraries mailing list<br>
<a href="mailto:Libraries@haskell.org" target="_blank">Libraries@haskell.org</a><br>
<a href="http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries" target="_blank">http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries</a><br>
</blockquote></div></div></div>