service/dap: Support local attach #2260
Conversation
|
I think there is a race condition that I need to debug. |
|
I think it would be better if the test for this was more like TestAttachDetach in proc_test and used testnextnethttp, which will wait forever, instead of changing loopprog. |
I changed loopprog to print status messages a bit less often, so I could use it for manual testing. The change is not necessary for the unittest. I can remove it. Using testnextnethttp has more moving parts. I would need to wait for the server to start listening, then have a client and send a request to trigger the code that has the breakpoint. What's the advantage of using that? I am not quite following. Also, my current test harness doesn't have a good spot to run the client code while guaranteeing that it would trigger the breakpoint after the debugger is fully attached and ready to process the breakpoint. |
There was a problem hiding this comment.
What's the advantage of using that?
You could test that detach without kill works, but on second thought this is fine.
It seems that dap tests are getting stuck on freebsd, I don't know if it's your fault or a preexisting problem in the freebsd backend, if it's the latter you can disable this new test on freebsd.
How can you tell? I can't seem to make sense of this new output format :( Also, the output is not verbose and doesn't have RUN + PASS/FAIL messages, so I can't even tell which test functions worked. |
You can force the verbose output by temporarily changing _scripts/make.go. |
* Support local attach * Undo loopprog change * Remove test check for system-specific error message * Skip attach tests on freebasd Co-authored-by: Polina Sokolova <polinasok@users.noreply.github.com>
Updates #1515