Build a Crash Handler

Crash Handler/Reporter Proposal

Purpose and Problem

Currently our C++ Tools dont have specific Crash hanlders.
this means that the only Crash data we get are errors of parent tools.

this makes it hard for us to get the data we actually want.

We want to build a custom crash Hanlder that wraps around existing tools to give us the specific data in a specific format for us to Process.

We also want a Database to collect all the reports together to later learn from them.

Implementation

We want a Singleton Style implementation to ensure only one Crash handler exists in a single Application Unit.

Multi instantiation should still add data to the Crash Handler.

What do we Want to catch

• stack trace file (backward-cpp) (minidump's)
• machine info (network, cpu, ram)
  • this helps to find out if we have a specific hardware set that causes problems
• installed libs
  • this helps us identify if a user has incorrect library installed
  • or if we conflict with libs
• Core Dumps

Inspiration / Research

Research

Links

--------------- Repos

• https://github.com/google/breakpad
• https://github.com/bombela/backward-cpp

----------- Tool Docs

• https://www.robopenguins.com/less-crashy-crashes/
• https://chromium.googlesource.com/crashpad/crashpad/+/master/README.md
• https://chromium.googlesource.com/breakpad/breakpad/

---------------- Infos Write Ups

• https://raygun.com/documentation/language-guides/cpp/crash-reporting/installation/
• https://stackoverflow.com/questions/3899870/how-to-print-a-stack-trace-whenever-a-certain-function-is-called/54365144#54365144
• https://www.robopenguins.com/core-dumps/
• https://www.boost.org/doc/libs/master/doc/html/stacktrace/getting_started.html#stacktrace.getting_started.handle_terminates

Inspiration

we can take inspiration by what side fx Houdini is doing with there info section.

Start by reviewing the Breakpad, backward-cpp, and Crashpad references linked in the issue, then define the crash handler's scope and report format. Done should include singleton-style behavior, support for multiple instances contributing data, and collection of stack traces, machine information, installed libraries, and core dumps for later database processing.