Arpad Borsos

So my ideas about "separation of concerns" comes directly from working on symbolic/ator, and a thought experiment I did about how I would design a clean-room minidump processor (or stack...

> That's definitely better than passing a trait outparameter, which is what we do at the moment. Yes, that is indeed one of the annoying parts of current API. The...

#404 was rather about validating invalid input. This here is about valid minidumps with *tons* of data in them. As the linked symbolic issue shows, we do occasionally get real...

You mean writing of thread names?

The Android event has source context, and makes it obvious why the mapping fails:  I can take a closer look at the debug info for this particular function, but...

Can you by any chance look at the generated code for the other examples? I wonder why that code is not being uploaded. Probably we are running into the file...

In that case il2cpp does not seem to generate a line mapping, so there is nothing we can do :-(

I get the following error when I have more than one assert.response test: uncaught: Error: ECONNREFUSED, Connection refused at Client._onConnect (net.js:548:18) at IOWatcher.onWritable [as callback](net.js:165:12) But I only run into...

We internally have a split between stack unwinding, and symbolication. Both as phases in code, as well as in intermediate cache formats we use. And those cache formats are being...

I think that the original motivation was rather that it depended on some Windows system libraries to read pdbs, on Windows only. I’m not entirely sure about unwind info, but...