meta(tracing): tracking issue for tracing improvements

[mds1]

Component

Forge, Cast

Describe the feature you would like

forge test, forge script, and cast run all produce traces. Depending on your use case, there are various display modes that may be the most helpful:

1. Hiding addresses if you have a label vs. showing label + addr vs. always showing addr (ref feat(tracing): add flag to hide address when label is available #3381)
2. Truncated address like 0x123...def vs full addresses (this has come up in an issue before, can't find it now though)
3. Overall tx state diffs vs. no state diffs (ref feat(cheatcodes): ability to capture and store state diffs #2846)
4. Call-level state diffs within the trace (arguably would get cluttered, but still is useful)
5. Showing only logs instead of a full trace
6. Showing vs. hiding function parameter names
7. Showing vs. hiding decoded SSTOREs and SLOADs in a call, similar to https://tx.eth.samczsun.com/#txhash=0x3443cd8f0c722fa9d83e12033f29eb590b177804ae625f2072625dfb79b329c2
8. Hide/modify certain calls in the traces, for example file/json reading/writing large files can quickly clutter logs. Perhaps a flag that takes an array of VM function names/signatures and replaces them with some other value

I think we should support these various options in two ways:

• A new [trace] section in the config file (more useful for forge)
• CLI flags (more useful for cast)

Additional context

No response

[odyslam]

Linking also this as relevant: #3599

I will probably PR a quick fix for #3599, as it's really painful. But, generally, really like this "proper" solution to this family of use cases.