#fillinstacktrace search results

When profiling allocations with async-profiler I see, that Throwable#fillInStackTrace is very expensive, at least it is shown as a major all allocator. What would happen if I created a custom Exception and over wrote the #fillInStackTrace method as @heinzkabutz wrote in #129? 🤔


17% of all measured TLAB allocs are causes by #fillInStackTrace. Do I have a problem? 🙀


Except in situations when you use exceptions on #fastpath to (for example) simulate non-structured #goto. @graalvm compiler can compile throw & catch, but #fillInStackTrace has to run in #Interpreter mode: graalvm.org/truffle/javado…

Do not try to be smart: Do not disable exception stacktraces. You are doing something wrong if stacktrace population is your performance bottleneck. Especially when the exceptions are in your public API.



17% of all measured TLAB allocs are causes by #fillInStackTrace. Do I have a problem? 🙀


When profiling allocations with async-profiler I see, that Throwable#fillInStackTrace is very expensive, at least it is shown as a major all allocator. What would happen if I created a custom Exception and over wrote the #fillInStackTrace method as @heinzkabutz wrote in #129? 🤔


Except in situations when you use exceptions on #fastpath to (for example) simulate non-structured #goto. @graalvm compiler can compile throw & catch, but #fillInStackTrace has to run in #Interpreter mode: graalvm.org/truffle/javado…

Do not try to be smart: Do not disable exception stacktraces. You are doing something wrong if stacktrace population is your performance bottleneck. Especially when the exceptions are in your public API.



No results for "#fillinstacktrace"
No results for "#fillinstacktrace"
Loading...

Something went wrong.


Something went wrong.