Hi, I have noticed that in

getChange

(osquery/core/watcher.cpp),

change

is uninitialized if

tryTo

throws an exception. And

change.footprint

is assigned to

state.initial_footprint

. Does it matter?

are you seeing this manifesting as an error, or in reading the code? I would generally say that matters… I’m not a c++, but it looks like

change

is declared and it’s various members worked with. Are you referring to

change.footprint

specifically?

Oh, I'm sorry. It's not

tryTo

. It may be

at

or

takeOr

. Anyway, I mean if something is thrown, do we need to initialize

change

in the catch block?

I'm not enough of a c++ person to be sure here. There is a try/catch block. And

change

is declared outside it. So if the members of that stuct are intialized as 0 when declared it seems fine. If they're undef it'd be bad. But I can't tell which is which.

Thanks for the reply. I'll open a blueprint issue about this.

It’s not blueprint. You’re asking a c++ question.

From first principles. I'd like to initialize

change

when it is declared (or in the catch block). Should I ask it here or open an issue?

Well, I’m the only one that answered here. So I suspect you’re not going to get a better response 😞

That is what I want (to initialize when declared). Thanks anyway. I'll try opening an issue to get more replies.

If it’s easy, you could also just PR it. I think it’s like a 4 line diff.

I've submitted a PR, just a 2 line diff.