Is there a way to print these in rust-lldb?

[vjpr]

I am using CLion with the Rust plugin.

Trying to debug a variable in the swc-project/swc project.

$ p src

(string_cache::atom::Atom<swc_atoms::JsWordStaticSet>) src = {
  unsafe_data = 140464742296960 {
    0 = 140464742296960
  }
  phantom = {}
}

[vjpr]

Work is ongoing here intellij-rust/intellij-rust#3000

[vjpr]

I actually will need help with this.

Is there a way to get the string from the unsafe_data ptr?

[vjpr]

This seems to work for me, but would be interested in input. I am new to Rust.

unsafe_data points to dynamic_set::Entry, which I then just read as a CString - which seems to work although I thought Rust's strs are not null-terminated.

This is using IntelliJ with the Rust plugin. Then you run a debug session.

---

~/.lldbinit

command script import ./summaries.py
command script import ./lldb_init.py

lldb_init.py

def __lldb_init_module(debugger, internal_dict):
  summary_cmd = (
    "type summary add "
    "Atom "
    "--expand "
    "--hide-empty "
    '--regex "string_cache::atom::Atom" '
    "--python-function summaries.AtomSummary "
    "--category Rust "
  )

  synthetic_cmd = (
      "type synthetic add "
      "Atom "
      "--python-class summaries.AtomProvider "
      '--regex "string_cache::atom::Atom" '
      "--category Rust "
    )


  # NOTE: If this command is incorrect `unknown option` is printed with no other debugging info.
  debugger.HandleCommand(summary_cmd)
  # debugger.HandleCommand(synthetic_cmd)

summaries.py

import lldb
import sys

PY3 = sys.version_info[0] == 3
if PY3:
  from typing import Optional

def AtomSummary(valobj, internal_dict):
  # type: (SBValue, dict) -> str

  process = valobj.GetProcess()

  # length = valobj.GetChildMemberWithName("length").GetValueAsUnsigned()
  length = 200
  if length == 0:
    return '""'

  unsafe_data = valobj.GetChildMemberWithName("unsafe_data")

  # From: https://github.com/servo/string-cache/blob/master/src/atom.rs
  TAG_MASK = 0b_11  # u8
  DYNAMIC_TAG = 0b_00  # u8
  INLINE_TAG = 0b_01  # b8 - len in upper nybble
  STATIC_TAG = 0b_10  # u8
  TAG_MASK = 0b_11
  LEN_OFFSET = 4
  LEN_MASK = 0xF0
  # ---

  # *const dynamic_set::Entry
  # https://docs.rs/string_cache/0.8.1/src/string_cache/dynamic_set.rs.html
  unsafe_data = unsafe_data.GetChildAtIndex(0).GetValueAsUnsigned()

  tag = unsafe_data & TAG_MASK  # u8

  type = 'unknown'
  if (tag == DYNAMIC_TAG): type = 'dynamic'
  if (tag == INLINE_TAG):  type = 'inline_tag'
  if (tag == STATIC_TAG): type = 'static_tag'

  if (unsafe_data == 0):
    return '""'

  # TODO(vjpr): Check tag to decide what to do.

  error = lldb.SBError()
  ptr_to_entry = process.ReadPointerFromMemory(unsafe_data, error)
  if error.Success():
    print('pointer: 0x%x' % ptr_to_entry)
  else:
    print('error: ', error)
    return '""'

  error = lldb.SBError()
  str = process.ReadCStringFromMemory(ptr_to_entry, length, error)

  print("ptr_to_entry: 0x{:x}, entry_ptr: 0x{:x}, tag: {}".format(unsafe_data, ptr_to_entry, tag))
  print("string: {}".format(str))

  return '"%s"' % str

[SimonSapin]

Can lldb printing be configured to call native code? If not, I’m afraid there’s no way around duplicating in python many implementation details of the library that can change in any future version