Arpad Borsos

``` export SDKROOT=$(xcrun --sdk macosx --show-sdk-path) ``` That is mentioned in the readme. Otherwise you need to correctly activate the xcode build tools, that is always a bit confusing.

The breakpad handler (and in-proc) will write the event directly to disk in the case of a hard crash. On restart, the event will be loaded from disk and passed...

> I am assuming this second step will still use the custom transport. Yes thats true. You can take a look at the directory that you configured via: https://github.com/getsentry/sentry-native/blob/0e94e49f442b75da4cf4048e25d577cd581a7b09/include/sentry.h#L960-L990

That is a good idea! In which form would you prefer that?

I’m not sure making this dynamic (as in: changable at runtime) is a good idea in general. The other thing we have to consider is that sentry-native has its own...

This looks like a problem with upstream crashpad. We should update our fork soon anyway, though not sure that will magically fix this issue. Is the `-Wmissing-field-initializers` flag specific to...

I think this should be supported, I don’t see a reason why not. Though none of us engineers have such a machine to actually test this.

Looking into this a bit, and seeing that the code does not seem to belong to any loaded module/library. I *think* what happens is that PACE is decrypting/decompressing the code...

I have looked through the crashpad API again, but I don’t see a way to manually register a "module". I wonder what happens when you attach a debugger to that...

its possible that some of the codesize comes from us vendoring a libunwindstack fork, which is heavy on c++, which is not known for small code size. Anyhow, it is...