add --elide-lines override flag for workspace filtering

[martinamps]

What does this PR do?

This adds a new flag --elide-lines to bun test to allow overriding the default value of 10. Lots of other folks ran into this in #12431

How did you verify your code works?

Manually and via unit tests

before:

✗ bun run  --filter gateway test
gateway test $ bun test
│ [28 lines elided]
│
│ tests/unit/openai/structuredOutput.test.ts:
│ (pass) OpenAI Structured Output > should handle structured output for research paper extraction [0.80ms]
│ (pass) OpenAI Structured Output > should handle structured output for math reasoning [0.28ms]
│ (pass) OpenAI Structured Output > should handle invalid structured output [0.29ms]
│
│  22 pass
│  0 fail
│  155 expect() calls
│ Ran 22 tests across 4 files. [66.00ms]
└─ Done in 77 ms

after:

✗ /Users/martinamps/code/bun/build/release/bun run --elide-lines 0 --filter gateway test
gateway test $ bun test
│ bun test v1.1.39 (1d485617)
│
│ tests/unit/model-definitions/getDefinition.test.ts:
│ (pass) Model Definitions > should return correct OpenAI chat model definition
│ (pass) Model Definitions > should return correct Anthropic chat model definition [0.14ms]
│ (pass) Model Definitions > should handle unknown model gracefully
│ (pass) Model Definitions > should type check different model types correctly [0.03ms]
│ (pass) Model Definitions > provider fields should be present and correct type [0.04ms]
│
│ tests/unit/openai/functionCalling.test.ts:
│ Cache Middleware is not enabled because caches is not defined.
│ (pass) OpenAI Function Calling > should handle function calling with a single function [3.81ms]
│ (pass) OpenAI Function Calling > should handle function calling with multiple functions [0.67ms]
│ (pass) OpenAI Function Calling > should handle function calling with a specified function [0.53ms]
│ (pass) OpenAI Function Calling > should handle function execution and follow-up [0.53ms]
│ (pass) OpenAI Function Calling > should handle errors in function calling [0.33ms]
│ (pass) OpenAI Function Calling > should handle function calling with invalid arguments [0.46ms]
│ (pass) OpenAI Function Calling > should handle streaming responses [1.75ms]
│
│ tests/unit/openai/chatCompletions.test.ts:
│ (pass) OpenAI Chat Completions > should forward a basic chat completion request [0.49ms]
│ (pass) OpenAI Chat Completions > should handle rate limiting [0.25ms]
│ mamps here3
│ (pass) OpenAI Chat Completions > should handle network errors [0.33ms]
│ (pass) OpenAI Chat Completions > should reject requests with invalid content type [0.21ms]
│ (pass) OpenAI Chat Completions > should use the correct API key [0.28ms]
│ (pass) OpenAI Chat Completions > should handle very large requests [0.32ms]
│ (pass) OpenAI Chat Completions > should handle streaming responses [0.36ms]
│
│ tests/unit/openai/structuredOutput.test.ts:
│ (pass) OpenAI Structured Output > should handle structured output for research paper extraction [0.87ms]
│ (pass) OpenAI Structured Output > should handle structured output for math reasoning [0.28ms]
│ (pass) OpenAI Structured Output > should handle invalid structured output [0.31ms]
│
│  22 pass
│  0 fail
│  155 expect() calls
│ Ran 22 tests across 4 files. [68.00ms]
└─ Done in 77 ms

The coupling of elide to pretty output was a little tight so my matching for the output is a little relaxed as a result and I had to make minor tweaks to the test runner to pass in the extra env flags

[jakeboone02]

Might be nice to have this as a bunfig option as well, so it doesn't have to be included in every command.

[Jarred-Sumner]

Nice. Looks like there's some test failures, but this is a good idea and I would like to merge it once the tests pass.

[martinamps]

@jakeboone02 sounds reasonable! I attempted to add it here https://github.com/oven-sh/bun/blob/main/src/bunfig.zig#L592 and it didn't do anything - happen to know why off the top of your head before I debug?

@Jarred-Sumner awesome 🙏 which tests? looks like they're all passing in CICD afaict

[jakeboone02]

@jakeboone02 sounds reasonable! I attempted to add it here https://github.com/oven-sh/bun/blob/main/src/bunfig.zig#L592 and it didn't do anything - happen to know why off the top of your head before I debug?

I don't know but be aware of #12216 (although that seems like the opposite problem...🤷‍♂️).

[RiskyMH]

which tests? looks like they're all passing in CICD afaict

I'm not sure why its green status, but if you press "details" it will send you to the page: https://buildkite.com/bun/bun/builds/8169

[martinamps]

aha ty @RiskyMH should've clicked through. Fixed cc @Jarred-Sumner (2 fail on canary because my code isn't there yet I guess)

@jakeboone02 I agree with you it'd be nice to have it in bunfig but unclear to me immediately why that isn't working like --silent.. Ok if we make that a follow-up PR when I have more cycles?

[jakeboone02]

@jakeboone02 I agree with you it'd be nice to have it in bunfig but unclear to me immediately why that isn't working like --silent.. Ok if we make that a follow-up PR when I have more cycles?

Sure! The sooner the better for the CLI flag IMO. A bunfig option can wait.

[Jarred-Sumner]

I woud like to merge this but there are Windows test failures

[martinamps]

ah-- I think it's because windowsIsTerminal() does not respect FORCE_COLOR which is how I trigger the code path in the other environments https://github.com/oven-sh/bun/blob/main/src/cli/filter_run.zig#L534 -- now if you try use it in a non TTY environment it'll error, and I explicitly test that on windows. Hopefully passes this time 😅 sorry!

[Veskel01]

My goodness - I'm waiting for the merge. This is the most annoying thing about Bun-based monorepo

[Jarred-Sumner]

Thank you. This will be part of Bun v1.1.43