Fix race issue

[daveads]

fix: Resolve race condition in HTTP server handling

• Implement thread-safe safeServer to wrap http.Server
• Modify Serve function to use safeServer for improved concurrency
• Ensure atomic operations for server start and shutdown

This change addresses a race condition in the HTTP server handling,particularly affecting CLI tests. By introducing a safeServer struct with atomic operations and proper synchronization, and prevent potential data races during server startup and shutdown.

Fix for terrastruct/d2#2087

[daveads]

@alixander reverted....

is it okay now ?

[alixander]

So what I was asking originally is whether this PR can be reverted: #19

If so, I'd like to see the revert in this PR.

[alixander]

you seem to be reverting it in one commit, and then your next commit just undoes it. I'm confused what you're doing. Please stop closing and opening new PRs too

[daveads]

So what I was asking originally is whether this PR can be reverted: #19

If so, I'd like to see the revert in this PR.

oh... i tried undoing the changes on that PR buh it wasn't working the way i expected it to.

buh can we test this now ? to see if it fixes the race issues...

[daveads]

@alixander reverting back undo the changes i recently made... and revert back to what was already there...

is this what you wanted ?

it doesn't keep my current changes that fix the race issue.

[alixander]

i've reverted the previous fix that didn't work in master, please rebase

[daveads]

@alixander it's ready for merge.

[alixander]

There's some miscommunication here.

Your previous PR (#19) did not work. So it looks like your second attempt is trying a new, completely different strategy. So I wanted the previous one reverted. And now you've just included both in this PR. I have said the same thing in an earlier comment #23 (comment)

Please also stop pinging me repeatedly, just request review if needed. Your behavior in general is strange, with the closing and re-opening issues, attempting to revert and undo the revert of your last PR. This pattern is frustrating and if you continue down this line, I will no longer be accepting your contributions.

[daveads]

There's some miscommunication here.

Your previous PR (#19) did not work. So it looks like your second attempt is trying a new, completely different strategy. So I wanted the previous one reverted. And now you've just included both in this PR. I have said the same thing in an earlier comment #23 (comment)

Please also stop pinging me repeatedly, just request review if needed. Your behavior in general is strange, with the closing and re-opening issues, attempting to revert and undo the revert of your last PR. This pattern is frustrating and if you continue down this line, I will no longer be accepting your contributions.

Okay, closing and reopening the PR was an attempt to resolve an issue.
And sure, I'll stop pinging—I didn’t realize it was considered unusual. I tend to do that often.

[daveads]

Git rebase is something I don't particularly like messing with, but I hope this does what you needed.

[alixander]

Thank you for your attempt @daveads but you're not understanding me and I've tried many times now, I'm no longer accepting your contributions.