You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
By "redundant" do you mean the client could produce the name itself in some way?
I just meant that the line and module are already present in the StackFrame, so I didn't understand if this flag means the client is required to also put the information into name.
I think it was a mistake to include parameters. I believe the author of this part of the spec was borrowing too heavily from Visual Studio's debug interface (FIF_FUNCNAME_ARGS) but here is how I believe this was intended to work:
parameters == false: Don't indicate anything about the parameters to the method. Debugger doesn't need to even look at parameterTypes, parameterNames or parameterValues
parameters == true: Include the arguments in a way that makes sense for the language being debugged, using parameterTypes, parameterNames, and parameterValues as makes sense
Sorry to necro this thread, but I was curious in regards to the parameters surfacing, what about the actual values/arguments specified to those parameters?
For example if I have this call stack:
Not sure where I am supposed to surface those Argument values, since it's not a property on the StackTrace itself or the Source
Activity
connor4312 commentedon Jul 27, 2023
I'm not sure actually. Gregg Miskelly (@gregg-miskelly) do you have historical context on this one?
gregg-miskelly commentedon Jul 27, 2023
Correct
By "redundant" do you mean the client could produce the name itself in some way? If so, that is correct, but it is done this way so that:
-and-
scopesrequesttromey commentedon Jul 28, 2023
I just meant that the line and module are already present in the
StackFrame, so I didn't understand if this flag means the client is required to also put the information intoname.connor4312 commentedon Jul 28, 2023
Thanks for the context. I suggest replacing "for the stack frame" to "in the
StackFrame.name" for the relevant properties.tromey commentedon Oct 6, 2023
A related question here is what does
parametersmean -- how does it differ fromparameterTypes,parameterNames, andparameterValues?gregg-miskelly commentedon Oct 6, 2023
I think it was a mistake to include
parameters. I believe the author of this part of the spec was borrowing too heavily from Visual Studio's debug interface (FIF_FUNCNAME_ARGS) but here is how I believe this was intended to work:parameters== false: Don't indicate anything about the parameters to the method. Debugger doesn't need to even look atparameterTypes,parameterNamesorparameterValuesparameters== true: Include the arguments in a way that makes sense for the language being debugged, usingparameterTypes,parameterNames, andparameterValuesas makes senseJustinGrote commentedon Oct 12, 2024
Sorry to necro this thread, but I was curious in regards to the parameters surfacing, what about the actual values/arguments specified to those parameters?
For example if I have this call stack:

Not sure where I am supposed to surface those Argument values, since it's not a property on the StackTrace itself or the Source
connor4312 commentedon Oct 25, 2024
The guidance would be to put them in the
StackFrame.namein a 'way that makes sense' for users of the DA.If I had my druthers, it would be a bit more structured, but we are where we are.
I will make a PR to clarify
StackFrameFormata bit.fix: clarify how StackTraceArguments.format applies
fix: clarify how StackTraceArguments.format applies (#510)
JustinGrote commentedon Oct 30, 2024
Thank you!