https://bugs.kde.org/show_bug.cgi?id=428689
--- Comment #5 from Andrew Somerville ---
I'm going to experiment a little with `libdw` from `elfutils` just to inform
myself a bit. I'm a newbie compared to you, but the idea has become interesting
enough that I'm curious to learn more.
>From what
https://bugs.kde.org/show_bug.cgi?id=428689
--- Comment #4 from Milian Wolff ---
I think it's more reliable to query for the size of the type size using GDB
directly, as accessing the function parameters of inlined frames would not be
possible in a release build - and the allocator will most
https://bugs.kde.org/show_bug.cgi?id=428689
--- Comment #3 from Andrew Somerville ---
It would seem that it might be possible to search the call stack some small
number of frames for function names matching a regex and find allocators for
type T and then lookup the sizeof info for T in the DWARF
https://bugs.kde.org/show_bug.cgi?id=428689
--- Comment #2 from Andrew Somerville ---
> Would you think it should look like this:
>
> [1x 100b]
> [a ] [1x 200b ]
> [b ] [1x 200b ] [1x 100b]
> [c]
Yeah, that's
https://bugs.kde.org/show_bug.cgi?id=428689
--- Comment #1 from Milian Wolff ---
Can you make a rough mockup of how you envision this to look like? I have a
hard time envisioning it, especially in flame graphs for code like this:
void a() { malloc(100); }
void b() { a(); malloc(200); }
void