更新:最新版本的 Intellij IDEA implements 正是我正在寻找的。问题是如何在 IDE 之外实现它(这样我就可以将异步堆栈跟踪转储到日志文件),理想情况下不使用检测代理。
自从我将我的应用程序从同步模型转换为异步模型后,我在调试失败时遇到了问题。
当我使用同步 API 时,我总是在异常堆栈跟踪中找到我的类,因此我知道如果出现问题从哪里开始查找。使用异步 API,我得到的堆栈跟踪既不引用我的类,也不指示是什么请求触发了失败。
我会给你一个具体的例子,但我对这类问题的通用解决方案很感兴趣。
我使用 Jersey 发出 HTTP 请求:
new Client().target("http://test.com/").request().rx().get(JsonNode.class);
其中 rx() 指示请求应该异步发生,返回 CompletionStage<JsonNode> 而不是直接返回 JsonNode。如果这个调用失败,我得到这个堆栈跟踪:
javax.ws.rs.ForbiddenException: HTTP 403 Authentication Failed
at org.glassfish.jersey.client.JerseyInvocation.convertToException(JerseyInvocation.java:1083)
at org.glassfish.jersey.client.JerseyInvocation.translate(JerseyInvocation.java:883)
at org.glassfish.jersey.client.JerseyInvocation.lambda$invoke$1(JerseyInvocation.java:767)
at org.glassfish.jersey.internal.Errors.process(Errors.java:316)
at org.glassfish.jersey.internal.Errors.process(Errors.java:298)
at org.glassfish.jersey.internal.Errors.process(Errors.java:229)
at org.glassfish.jersey.process.internal.RequestScope.runInScope(RequestScope.java:414)
at org.glassfish.jersey.client.JerseyInvocation.invoke(JerseyInvocation.java:765)
at org.glassfish.jersey.client.JerseyInvocation$Builder.method(JerseyInvocation.java:456)
at org.glassfish.jersey.client.JerseyCompletionStageRxInvoker.lambda$method$1(JerseyCompletionStageRxInvoker.java:70)
at java.util.concurrent.CompletableFuture$AsyncSupply.run(CompletableFuture.java:1590)
注意事项:
因此,我无法追踪异常的源头。
如果深入挖掘,您会发现 Jersey is invoking :
CompletableFuture.supplyAsync(() -> getSyncInvoker().method(name, entity, responseType))
用于 rx() 调用。因为供应商是由 Jersey 构建的,所以没有对用户代码的引用。
I tried filing a bug report 针对 Jetty 提出了一个不相关的异步示例,随后因安全原因被拒绝。
相反,我一直在添加如下上下文信息:
makeHttpRequest().exceptionally(e ->
{
throw new RuntimeException(e);
});
意思是,我在代码中的每个 HTTP 请求之后手动添加 exceptionally()。 Jersey 抛出的任何异常都包含在引用我的代码的辅助异常中。生成的堆栈跟踪如下所示:
java.lang.RuntimeException: javax.ws.rs.ForbiddenException: HTTP 403 Authentication Failed
at my.user.code.Testcase.lambda$null$1(Testcase.java:25)
at java.util.concurrent.CompletableFuture.uniExceptionally(CompletableFuture.java:870)
... 6 common frames omitted
Caused by: javax.ws.rs.ForbiddenException: HTTP 403 Authentication Failed
at org.glassfish.jersey.client.JerseyInvocation.convertToException(JerseyInvocation.java:1083)
at org.glassfish.jersey.client.JerseyInvocation.translate(JerseyInvocation.java:883)
at org.glassfish.jersey.client.JerseyInvocation.lambda$invoke$1(JerseyInvocation.java:767)
at org.glassfish.jersey.internal.Errors.process(Errors.java:316)
at org.glassfish.jersey.internal.Errors.process(Errors.java:298)
at org.glassfish.jersey.internal.Errors.process(Errors.java:229)
at org.glassfish.jersey.process.internal.RequestScope.runInScope(RequestScope.java:414)
at org.glassfish.jersey.client.JerseyInvocation.invoke(JerseyInvocation.java:765)
at org.glassfish.jersey.client.JerseyInvocation$Builder.method(JerseyInvocation.java:456)
at org.glassfish.jersey.client.JerseyCompletionStageRxInvoker.lambda$method$1(JerseyCompletionStageRxInvoker.java:70)
at java.util.concurrent.CompletableFuture$AsyncSupply.run(CompletableFuture.java:1590)
... 3 common frames omitted
我不喜欢这种方法,因为它容易出错并且会降低代码的可读性。如果我在某些 HTTP 请求中错误地忽略了它,我最终会得到一个模糊的堆栈跟踪并花费大量时间来追踪它。
此外,如果我想将这个技巧隐藏在实用程序类之后,那么我必须在 CompletionStage 之外实例化一个异常;否则,实用程序类将出现在堆栈跟踪中,而不是实际的调用站点。在 CompletionStage 之外实例化异常非常昂贵,因为即使异步调用从未抛出任何异常,此代码也会运行。
是否有一种健壮、易于维护的方法来向异步调用添加上下文信息?
或者,是否有一种有效的方法可以在没有这些上下文信息的情况下将堆栈跟踪跟踪回源?
最佳答案
鉴于这个问题已经将近一个月没有收到任何答案,我将发布迄今为止我找到的最佳解决方案:
DebugCompletableFuture.java:
/**
* A {@link CompletableFuture} that eases debugging.
*
* @param <T> the type of value returned by the future
*/
public final class DebugCompletableFuture<T> extends CompletableFuture<T>
{
private static RunMode RUN_MODE = RunMode.DEBUG;
private static final Set<String> CLASS_PREFIXES_TO_REMOVE = ImmutableSet.of(DebugCompletableFuture.class.getName(),
CompletableFuture.class.getName(), ThreadPoolExecutor.class.getName());
private static final Set<Class<? extends Throwable>> EXCEPTIONS_TO_UNWRAP = ImmutableSet.of(AsynchronousException.class,
CompletionException.class, ExecutionException.class);
private final CompletableFuture<T> delegate;
private final AsynchronousException asyncStacktrace;
/**
* @param delegate the stage to delegate to
* @throws NullPointerException if any of the arguments are null
*/
private DebugCompletableFuture(CompletableFuture<T> delegate)
{
requireThat("delegate", delegate).isNotNull();
this.delegate = delegate;
this.asyncStacktrace = new AsynchronousException();
delegate.whenComplete((value, exception) ->
{
if (exception == null)
{
super.complete(value);
return;
}
exception = Exceptions.unwrap(exception, EXCEPTIONS_TO_UNWRAP);
asyncStacktrace.initCause(exception);
filterStacktrace(asyncStacktrace, element ->
{
String className = element.getClassName();
for (String prefix : CLASS_PREFIXES_TO_REMOVE)
if (className.startsWith(prefix))
return true;
return false;
});
Set<String> newMethods = getMethodsInStacktrace(asyncStacktrace);
if (!newMethods.isEmpty())
{
Set<String> oldMethods = getMethodsInStacktrace(exception);
newMethods.removeAll(oldMethods);
if (!newMethods.isEmpty())
{
// The async stacktrace introduces something new
super.completeExceptionally(asyncStacktrace);
return;
}
}
super.completeExceptionally(exception);
});
}
/**
* @param exception an exception
* @return the methods referenced by the stacktrace
* @throws NullPointerException if {@code exception} is null
*/
private Set<String> getMethodsInStacktrace(Throwable exception)
{
requireThat("exception", exception).isNotNull();
Set<String> result = new HashSet<>();
for (StackTraceElement element : exception.getStackTrace())
result.add(element.getClassName() + "." + element.getMethodName());
for (Throwable suppressed : exception.getSuppressed())
result.addAll(getMethodsInStacktrace(suppressed));
return result;
}
/**
* @param <T2> the type returned by the delegate
* @param delegate the stage to delegate to
* @return if {@code RUN_MODE == DEBUG} returns an instance that wraps {@code delegate}; otherwise, returns {@code delegate}
* unchanged
* @throws NullPointerException if any of the arguments are null
*/
public static <T2> CompletableFuture<T2> wrap(CompletableFuture<T2> delegate)
{
if (RUN_MODE != RunMode.DEBUG)
return delegate;
return new DebugCompletableFuture<>(delegate);
}
/**
* Removes stack trace elements that match a filter. The exception and its descendants are processed recursively.
* <p>
* This method can be used to remove lines that hold little value for the end user (such as the implementation of utility functions).
*
* @param exception the exception to process
* @param elementFilter returns true if the current stack trace element should be removed
*/
private void filterStacktrace(Throwable exception, Predicate<StackTraceElement> elementFilter)
{
Throwable cause = exception.getCause();
if (cause != null)
filterStacktrace(cause, elementFilter);
for (Throwable suppressed : exception.getSuppressed())
filterStacktrace(suppressed, elementFilter);
StackTraceElement[] elements = exception.getStackTrace();
List<StackTraceElement> keep = new ArrayList<>(elements.length);
for (StackTraceElement element : elements)
{
if (!elementFilter.test(element))
keep.add(element);
}
exception.setStackTrace(keep.toArray(new StackTraceElement[0]));
}
@Override
public <U> CompletableFuture<U> thenApply(Function<? super T, ? extends U> fn)
{
return wrap(super.thenApply(fn));
}
@Override
public <U> CompletableFuture<U> thenApplyAsync(Function<? super T, ? extends U> fn)
{
return wrap(super.thenApplyAsync(fn));
}
@Override
public <U> CompletableFuture<U> thenApplyAsync(Function<? super T, ? extends U> fn, Executor executor)
{
return wrap(super.thenApplyAsync(fn, executor));
}
@Override
public CompletableFuture<Void> thenAccept(Consumer<? super T> action)
{
return wrap(super.thenAccept(action));
}
@Override
public CompletableFuture<Void> thenAcceptAsync(Consumer<? super T> action)
{
return wrap(super.thenAcceptAsync(action));
}
@Override
public CompletableFuture<Void> thenAcceptAsync(Consumer<? super T> action, Executor executor)
{
return wrap(super.thenAcceptAsync(action, executor));
}
@Override
public CompletableFuture<Void> thenRun(Runnable action)
{
return wrap(super.thenRun(action));
}
@Override
public CompletableFuture<Void> thenRunAsync(Runnable action)
{
return wrap(super.thenRunAsync(action));
}
@Override
public CompletableFuture<Void> thenRunAsync(Runnable action, Executor executor)
{
return wrap(super.thenRunAsync(action, executor));
}
@Override
public <U, V> CompletableFuture<V> thenCombine(CompletionStage<? extends U> other,
BiFunction<? super T, ? super U, ? extends V> fn)
{
return wrap(super.thenCombine(other, fn));
}
@Override
public <U, V> CompletableFuture<V> thenCombineAsync(CompletionStage<? extends U> other,
BiFunction<? super T, ? super U, ? extends V> fn)
{
return wrap(super.thenCombineAsync(other, fn));
}
@Override
public <U, V> CompletableFuture<V> thenCombineAsync(CompletionStage<? extends U> other,
BiFunction<? super T, ? super U, ? extends V> fn,
Executor executor)
{
return wrap(super.thenCombineAsync(other, fn, executor));
}
@Override
public <U> CompletableFuture<Void> thenAcceptBoth(CompletionStage<? extends U> other,
BiConsumer<? super T, ? super U> action)
{
return wrap(super.thenAcceptBoth(other, action));
}
@Override
public <U> CompletableFuture<Void> thenAcceptBothAsync(CompletionStage<? extends U> other,
BiConsumer<? super T, ? super U> action)
{
return wrap(super.thenAcceptBothAsync(other, action));
}
@Override
public <U> CompletableFuture<Void> thenAcceptBothAsync(CompletionStage<? extends U> other,
BiConsumer<? super T, ? super U> action,
Executor executor)
{
return wrap(super.thenAcceptBothAsync(other, action, executor));
}
@Override
public CompletableFuture<Void> runAfterBoth(CompletionStage<?> other, Runnable action)
{
return wrap(super.runAfterBoth(other, action));
}
@Override
public CompletableFuture<Void> runAfterBothAsync(CompletionStage<?> other, Runnable action)
{
return wrap(super.runAfterBothAsync(other, action));
}
@Override
public CompletableFuture<Void> runAfterBothAsync(CompletionStage<?> other, Runnable action, Executor executor)
{
return wrap(super.runAfterBothAsync(other, action, executor));
}
@Override
public <U> CompletableFuture<U> applyToEither(CompletionStage<? extends T> other, Function<? super T, U> fn)
{
return wrap(super.applyToEither(other, fn));
}
@Override
public <U> CompletableFuture<U> applyToEitherAsync(CompletionStage<? extends T> other, Function<? super T, U> fn)
{
return wrap(super.applyToEitherAsync(other, fn));
}
@Override
public <U> CompletableFuture<U> applyToEitherAsync(CompletionStage<? extends T> other, Function<? super T, U> fn,
Executor executor)
{
return wrap(super.applyToEitherAsync(other, fn, executor));
}
@Override
public CompletableFuture<Void> acceptEither(CompletionStage<? extends T> other, Consumer<? super T> action)
{
return wrap(super.acceptEither(other, action));
}
@Override
public CompletableFuture<Void> acceptEitherAsync(CompletionStage<? extends T> other, Consumer<? super T> action)
{
return wrap(super.acceptEitherAsync(other, action));
}
@Override
public CompletableFuture<Void> acceptEitherAsync(CompletionStage<? extends T> other, Consumer<? super T> action,
Executor executor)
{
return wrap(super.acceptEitherAsync(other, action, executor));
}
@Override
public CompletableFuture<Void> runAfterEither(CompletionStage<?> other, Runnable action)
{
return wrap(super.runAfterEither(other, action));
}
@Override
public CompletableFuture<Void> runAfterEitherAsync(CompletionStage<?> other, Runnable action)
{
return wrap(super.runAfterEitherAsync(other, action));
}
@Override
public CompletableFuture<Void> runAfterEitherAsync(CompletionStage<?> other, Runnable action, Executor executor)
{
return wrap(super.runAfterEitherAsync(other, action, executor));
}
@Override
public <U> CompletableFuture<U> thenCompose(Function<? super T, ? extends CompletionStage<U>> fn)
{
return wrap(super.thenCompose(fn));
}
@Override
public <U> CompletableFuture<U> thenComposeAsync(Function<? super T, ? extends CompletionStage<U>> fn)
{
return wrap(super.thenComposeAsync(fn));
}
@Override
public <U> CompletableFuture<U> thenComposeAsync(Function<? super T, ? extends CompletionStage<U>> fn,
Executor executor)
{
return wrap(super.thenComposeAsync(fn, executor));
}
@Override
public CompletableFuture<T> exceptionally(Function<Throwable, ? extends T> fn)
{
return wrap(super.exceptionally(fn));
}
@Override
public CompletableFuture<T> whenComplete(BiConsumer<? super T, ? super Throwable> action)
{
return wrap(super.whenComplete(action));
}
@Override
public CompletableFuture<T> whenCompleteAsync(BiConsumer<? super T, ? super Throwable> action)
{
return wrap(super.whenCompleteAsync(action));
}
@Override
public CompletableFuture<T> whenCompleteAsync(BiConsumer<? super T, ? super Throwable> action,
Executor executor)
{
return wrap(super.whenCompleteAsync(action, executor));
}
@Override
public <U> CompletableFuture<U> handle(BiFunction<? super T, Throwable, ? extends U> fn)
{
return wrap(super.handle(fn));
}
@Override
public <U> CompletableFuture<U> handleAsync(BiFunction<? super T, Throwable, ? extends U> fn)
{
return wrap(super.handleAsync(fn));
}
@Override
public <U> CompletableFuture<U> handleAsync(BiFunction<? super T, Throwable, ? extends U> fn,
Executor executor)
{
return wrap(super.handleAsync(fn, executor));
}
@Override
public boolean complete(T value)
{
return delegate.complete(value);
}
@Override
public boolean completeExceptionally(Throwable ex)
{
return delegate.completeExceptionally(ex);
}
}
RunMode.java:
/**
* Operational modes.
*/
public enum RunMode
{
/**
* Optimized for debugging problems (extra runtime checks, logging of the program state).
*/
DEBUG,
/**
* Optimized for maximum performance.
*/
RELEASE
}
AsynchronousException.java
/**
* Thrown when an asynchronous operation fails. The stacktrace indicates who triggered the operation.
*/
public final class AsynchronousException extends RuntimeException
{
private static final long serialVersionUID = 0L;
public AsynchronousException()
{
}
}
用法:
DebugCompletableFuture.wrap(CompletableFuture.supplyAsync(this::expensiveOperation));
好处:您将获得相对干净的异步堆栈跟踪。
缺点:每次创建 future 时都构造一个新的 AsynchronousException 非常昂贵。具体来说,如果您要生成大量 future,这会在堆上生成大量垃圾,并且 GC 开销会变得很明显。
我仍然希望有人能提出性能更好的方法。
关于java - 如何创建异步堆栈跟踪?,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/49269620/
我正在学习如何使用Nokogiri,根据这段代码我遇到了一些问题:require'rubygems'require'mechanize'post_agent=WWW::Mechanize.newpost_page=post_agent.get('http://www.vbulletin.org/forum/showthread.php?t=230708')puts"\nabsolutepathwithtbodygivesnil"putspost_page.parser.xpath('/html/body/div/div/div/div/div/table/tbody/tr/td/div
总的来说,我对ruby还比较陌生,我正在为我正在创建的对象编写一些rspec测试用例。许多测试用例都非常基础,我只是想确保正确填充和返回值。我想知道是否有办法使用循环结构来执行此操作。不必为我要测试的每个方法都设置一个assertEquals。例如:describeitem,"TestingtheItem"doit"willhaveanullvaluetostart"doitem=Item.new#HereIcoulddotheitem.name.shouldbe_nil#thenIcoulddoitem.category.shouldbe_nilendend但我想要一些方法来使用
出于纯粹的兴趣,我很好奇如何按顺序创建PI,而不是在过程结果之后生成数字,而是让数字在过程本身生成时显示。如果是这种情况,那么数字可以自行产生,我可以对以前看到的数字实现垃圾收集,从而创建一个无限系列。结果只是在Pi系列之后每秒生成一个数字。这是我通过互联网筛选的结果:这是流行的计算机友好算法,类机器算法:defarccot(x,unity)xpow=unity/xn=1sign=1sum=0loopdoterm=xpow/nbreakifterm==0sum+=sign*(xpow/n)xpow/=x*xn+=2sign=-signendsumenddefcalc_pi(digits
关闭。这个问题是opinion-based.它目前不接受答案。想要改进这个问题?更新问题,以便editingthispost可以用事实和引用来回答它.关闭4年前。Improvethisquestion我想在固定时间创建一系列低音和高音调的哔哔声。例如:在150毫秒时发出高音调的蜂鸣声在151毫秒时发出低音调的蜂鸣声200毫秒时发出低音调的蜂鸣声250毫秒的高音调蜂鸣声有没有办法在Ruby或Python中做到这一点?我真的不在乎输出编码是什么(.wav、.mp3、.ogg等等),但我确实想创建一个输出文件。
给定这段代码defcreate@upgrades=User.update_all(["role=?","upgraded"],:id=>params[:upgrade])redirect_toadmin_upgrades_path,:notice=>"Successfullyupgradeduser."end我如何在该操作中实际验证它们是否已保存或未重定向到适当的页面和消息? 最佳答案 在Rails3中,update_all不返回任何有意义的信息,除了已更新的记录数(这可能取决于您的DBMS是否返回该信息)。http://ar.ru
我在我的项目目录中完成了compasscreate.和compassinitrails。几个问题:我已将我的.sass文件放在public/stylesheets中。这是放置它们的正确位置吗?当我运行compasswatch时,它不会自动编译这些.sass文件。我必须手动指定文件:compasswatchpublic/stylesheets/myfile.sass等。如何让它自动运行?文件ie.css、print.css和screen.css已放在stylesheets/compiled。如何在编译后不让它们重新出现的情况下删除它们?我自己编译的.sass文件编译成compiled/t
我正在寻找执行以下操作的正确语法(在Perl、Shell或Ruby中):#variabletoaccessthedatalinesappendedasafileEND_OF_SCRIPT_MARKERrawdatastartshereanditcontinues. 最佳答案 Perl用__DATA__做这个:#!/usr/bin/perlusestrict;usewarnings;while(){print;}__DATA__Texttoprintgoeshere 关于ruby-如何将脚
Rackup通过Rack的默认处理程序成功运行任何Rack应用程序。例如:classRackAppdefcall(environment)['200',{'Content-Type'=>'text/html'},["Helloworld"]]endendrunRackApp.new但是当最后一行更改为使用Rack的内置CGI处理程序时,rackup给出“NoMethodErrorat/undefinedmethod`call'fornil:NilClass”:Rack::Handler::CGI.runRackApp.newRack的其他内置处理程序也提出了同样的反对意见。例如Rack
使用带有Rails插件的vim,您可以创建一个迁移文件,然后一次性打开该文件吗?textmate也可以这样吗? 最佳答案 你可以使用rails.vim然后做类似的事情::Rgeneratemigratonadd_foo_to_bar插件将打开迁移生成的文件,这正是您想要的。我不能代表textmate。 关于ruby-使用VimRails,您可以创建一个新的迁移文件并一次性打开它吗?,我们在StackOverflow上找到一个类似的问题: https://sta
在选择我想要运行操作的频率时,唯一的选项是“每天”、“每小时”和“每10分钟”。谢谢!我想为我的Rails3.1应用程序运行调度程序。 最佳答案 这不是一个优雅的解决方案,但您可以安排它每天运行,并在实际开始工作之前检查日期是否为当月的第一天。 关于ruby-如何每月在Heroku运行一次Scheduler插件?,我们在StackOverflow上找到一个类似的问题: https://stackoverflow.com/questions/8692687/