feat(forge): truncate large data in execution traces

[grandizzy]

Component

Forge

Describe the feature you would like

from @hexonaut - add an arg to truncate function call input/output in the execution trace - we have some functions that pass large byte arrays and it makes the execution trace pretty useless.

Additional context

• should be handled as a config in revm-insepctor TraceWriter in https://github.com/paradigmxyz/revm-inspectors/blob/bf724817ce484cf241076d2915d162e47a46f326/src/tracing/writer.rs#L102-L106

Edit: Started on revm-inspectors side paradigmxyz/revm-inspectors#283

[yash-atreya]

We're blocked by the revm migration here #10183 if we want to include this config as part of revm-inspectors since inspectors have already been bumped to new revm

cc @grandizzy @zerosnacks

[drortirosh]

Suggestion: instead of just truncation, maybe reduce repeated dumps:
That is, the parser "remembers" structs, and if a struct is dumped again , either as part of input or output, just "name" it instead of full dump.

This would also help dumps of erc4337, where a "UserOperation" struct is passed several times, and at times it can be few Ks.

[hexonaut]

Suggestion: instead of just truncation, maybe reduce repeated dumps: That is, the parser "remembers" structs, and if a struct is dumped again , either as part of input or output, just "name" it instead of full dump.

This would also help dumps of erc4337, where a "UserOperation" struct is passed several times, and at times it can be few Ks.

In the case I mentioned the byte array changes every time, so not sure if this will help.

[jenpaff]

blocked by revm bump