You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It would be useful to have some kind of optional instrumentation on the generated messages so that we could more easily track things like construction, destruction, and coping of the messages. Other kinds of data might be useful as well, but I can't think of any off the top of my head.
AC:
Messages generate optional (compile time enabled) mechanism for tracking calls to the message's constructor, destructor, and any copy/move constructors.
Ideally this could be disabled/enabled by the consuming package (like with a #define before including the message's header), and that if you do not use the instrumentation then it does not affect performance at all. But importantly, the enabling/disabling of the instrumentation cannot affect the behavior of the message class.
The text was updated successfully, but these errors were encountered:
This is handy as extra proof of correct zero-copy behavior, but for now we'll rely on people understanding how the intraprocess system works (and possibly instrument it somewhere, with manual verification). We can come back to this extra feature later.
It would be useful to have some kind of optional instrumentation on the generated messages so that we could more easily track things like construction, destruction, and coping of the messages. Other kinds of data might be useful as well, but I can't think of any off the top of my head.
AC:
Ideally this could be disabled/enabled by the consuming package (like with a
#define
before including the message's header), and that if you do not use the instrumentation then it does not affect performance at all. But importantly, the enabling/disabling of the instrumentation cannot affect the behavior of the message class.The text was updated successfully, but these errors were encountered: