adrian_hawryluk
Contributor
Joined 11 years ago
User Profile
User Widgets
Contributions
Duplicate Routines entry/exit
I don't understand. Why can I not Start profiling when entering a function and Stop profiling when exiting the same function? I would think that this would be a very common thing to do. For that matter, why is there a limit of only one anyway? I can think of other use cases, such as where I'd want to stop and generate a report at the exiting of a function. What's up with this? Why the limitation?Re: Generating a UML sequence diagram
Say you have 3 functions, x, y and z and you needed to know what order they were called in from function w. How would you be able to determine that from the function call graph results? It is that functionality that I'm looking for. From what I've seen, the call graph results only indicated how long it took to execute not when it was executed. (As a further complication , it only shows an average, and not how long each call took, but that's not relevant to the functionality in currently looking for, just an observation.) Do you see what I'm getting at Alex? ARe: Generating a UML sequence diagram
I don't mean just who called who and how long, but also when and in what order. The reason for the functionality is to do a post mortem analysis of the code path between objects and threads. This includes crashes or just for records so that reexecution of the code isn't necessary. This is especially useful in old complex legacy code where documentation is scarce, but I'm sure would still be useful in non legacy applications too where interactions of an unexpected nature occur frequently or (even better) rarely and are difficult to track statically. Thanks, A
