Rework forge command invocation (broken EPIPE issue)

[antico5]

This PR changes the handling of the child process spawned when running forge, currently only used for formatting. The broken EPIPE error happens when a process tries to write to the stdin of a dead process.

[codecov]

Codecov Report

Attention: Patch coverage is 18.75000% with 39 lines in your changes missing coverage. Please review.

Project coverage is 54.32%. Comparing base (e5372f3) to head (09a4470).

Files with missing lines	Patch %	Lines
server/src/utils/operatingSystem.ts	8.69%	21 Missing ⚠️
server/src/services/formatting/forgeFormat.ts	27.27%	8 Missing ⚠️
...er/src/services/formatting/onDocumentFormatting.ts	14.28%	6 Missing ⚠️
...rver/src/frameworks/Foundry/resolveForgeCommand.ts	33.33%	2 Missing ⚠️
server/src/utils/errors.ts	50.00%	2 Missing ⚠️

Additional details and impacted files

@@               Coverage Diff               @@
##           development     #642      +/-   ##
===============================================
- Coverage        54.52%   54.32%   -0.21%     
===============================================
  Files              226      227       +1     
  Lines             5199     5230      +31     
  Branches           800      809       +9     
===============================================
+ Hits              2835     2841       +6     
- Misses            2127     2152      +25     
  Partials           237      237              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[kanej]

I think we need to clear the timeout on process exit.
I am tempted to say we should rework the timeout Error, so that when it is caught by the try/catch in onDocumentFormatting we report a more specific sentry status for timeout: { status: "internal_error", result: null }. This would let us separate out timeouts from anything else.

[kanej]

As a sidepoint - the codecov here is really painful to review. If it is not pulling its weight we should consider turning it off.
What do you think would be required to bring execWithInput under test?

[antico5]

@kanej addressed all comments. About codecov, I'd be happy to remove it. We can tell on PR reviews what needs and what doesn't need coverage. For execWithInput, I think it's not viable to unit test. It would involve a lot of stubbing and mocking, and the added value would be almost nonexistent. For a non-stub testing case, we'd have to see what's possible in terms of process manipulation from our test case, and it can be tricky to have it be multi platform.

[kanej]

My previous suggestion caused a build error. So this type guard, while cumbersome does the job. We should port the error helpers from Hardhat!