LGTM! Mind fixing the lint?

Offenses:

lib/macho/view.rb:34:28: C: [Correctable] Style/RedundantSelf: Redundant self detected.
      "#<#{self.class}:0x#{self.__id__.to_s(16)} @endianness=#{@endianness.inspect}, @offset=#{@offset.inspect}, length=#{@raw_data.length}>"
                           ^^^^

11 files inspected, 1 offense detected, 1 offense auto-correctable
Error: Process completed with exit code 1.

Doing so now - also adding a MachOFile reference to MachOView so that you can navigate backwards from a command to the file (some commands make use of other commands in a file such as MH_FILESET_ENTRY to the LC_LOAD at the same address)

also adding a MachOFile reference to MachOView so that you can navigate backwards from a command to the file (some commands make use of other commands in a file such as MH_FILESET_ENTRY to the LC_LOAD at the same address)

Cool, that should be fine as long as we don't break any APIs that allow for dynamic construction/creation of LCs 🙂

Looking to exercise new code paths now

also adding a MachOFile reference to MachOView so that you can navigate backwards from a command to the file (some commands make use of other commands in a file such as MH_FILESET_ENTRY to the LC_LOAD at the same address)

Cool, that should be fine as long as we don't break any APIs that allow for dynamic construction/creation of LCs 🙂

Simple solution was to attach the MachOFile to the MachOView so navigation would be command -> view -> macho_file -> command which could be shortened by delegation but hey

Ok I'm a bad person... This PR expanded scope to cover the use case.

The E2E capability is load BootKernelCollection in pry (fast due to the @raw_data fix)

Iterate over FH_FILESET_ENTRY and map each to .segment and extract

Added some tests to make CI happy

LGTM structurally, although I'm a little bit confused by what's going on with FilesetEntryCommand#segment here -- it looks like you're using it to find a corresponding segment load command for the embedded Mach-O, but I was under the impression that fileoff could be within a segment and not just at the start of one. Is that a misunderstanding on my part?

LGTM structurally, although I'm a little bit confused by what's going on with FilesetEntryCommand#segment here -- it looks like you're using it to find a corresponding segment load command for the embedded Mach-O, but I was under the impression that fileoff could be within a segment and not just at the start of one. Is that a misunderstanding on my part?

Let me do some more testing - it seems to me that the .kc file has __REGION{digit} that have a file offset - the FILESET_INFO has a matching fileoff and annotates

I'm thinking of creating a Git LFS repo for text mach-os (large ones like kc files and prelink kernels that can be brought into testing if you want the other larger types.

Makes sense! I didn't realize each had a __REGION. If that's the observed behavior (knowing Apple, I'm sure it isn't documented), then I'm good with this approach.

We have a unified CI configuration set up for Homebrew, so I think we're good without commit hooks for the time being.

Oh, I misunderstood. Sure, as long as they're opt-in, I would accept a PR enabling them for rubocop.