DAP Server stack trace should use startFrame and levels parameters

[dpeacock]

Your environment

• ruby -v: ruby 3.1.2p20 (2022-04-12 revision 4491bb740a) [arm64-darwin21]
• rdbg -v: rdbg 1.5.0

Describe the bug
The DAP server does not use the parameters on the StackTrace request. In some instances this can result in a very large stack trace being returned, which can take extra time. VS code for example requests a depth of 20 at a time and has a 'more' button to retrieve additional data. Currently these are ignored, and the more button just resends the same data every time.

To Reproduce
Set a breakpoint in VS Code where the stack depth is > 20. VS Code will request 20 levels, but the dap server will ignore this and send the entire stack trace.

Expected behavior
Only send the requested levels.

[ko1]

Thank you, I don't know this behavior.

[ko1]

I understand levels is information how many levels are needed (20 frames for example).
What happens on "more" button is pushed?
StackTrace event with 40 levels?
StackTrace event with more 20 levels? (but I can't find the start level).

[dpeacock]

So the first request would be:
startFrame: 0, levels: 20
If less than 20 frames are returned, the more button will disappear. If 20 frames are returned, clicking the more button will request:
startFrame: 20, levels: 20 which will return the next 20 frames. This will continue until a request occurs where there are less than 20 frames available indicating all the frames have been retrieved.

[ko1]

memo;

def rec0
  a = 0
end

(1..1000).each{|i|
  eval("def rec#{i} = (a = #{i}; rec#{i-1})\n")
}

rec1000```