Bug 1839697 Comment 3 Edit History

Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.

(In reply to Chris H-C :chutten from comment #2)
> We provide the tools, you decide how to use them : ) But I do have opinions.
> 
> I would definitely discourage implicit distinctions. Make it explicit regardless of how so that even if in the future we consolidate on `stack_traces` for language-independent stacks, we can still tell Java stacks from not-Java stacks.
> 
> As a data consumer myself, I'd not expect to find this information in `processType` or `reason`. I'd be looking for something like a string `crash_language: java|native` or a bool `is_exception: true|false` or some other standard metric. Are those acceptable options?

Okay, that's what I figured you'd say. I'm thinking something like `cause: java_exception|native_error`. My reasoning is that `cause` more accurately describes the domain of the value than `crash_language`. I.e., `cause` implies that the `java_exception`/`native_error` relates to the initial frame/crash reason, whereas `crash_language` might be taken as relating to all of the information in the crash, but I think in the future we will try to have cross-language exception info. I don't like `is_exception` because it doesn't distinguish the language (though it could).
(In reply to Chris H-C :chutten from comment #2)
> We provide the tools, you decide how to use them : ) But I do have opinions.
> 
> I would definitely discourage implicit distinctions. Make it explicit regardless of how so that even if in the future we consolidate on `stack_traces` for language-independent stacks, we can still tell Java stacks from not-Java stacks.
> 
> As a data consumer myself, I'd not expect to find this information in `processType` or `reason`. I'd be looking for something like a string `crash_language: java|native` or a bool `is_exception: true|false` or some other standard metric. Are those acceptable options?

Okay, that's what I figured you'd say. I'm thinking something like `cause: java_exception|native_error`. My reasoning is that `cause` more accurately describes the domain of the value than `crash_language`. I.e., `cause` implies that the `java_exception`/`os_fault` relates to the initial frame/crash reason, whereas `crash_language` might be taken as relating to all of the information in the crash, but I think in the future we will try to have cross-language exception info. I don't like `is_exception` because it doesn't distinguish the language (though it could).

Back to Bug 1839697 Comment 3