================
@@ -147,96 +148,111 @@ class ProcessInstanceInfo : public ProcessInfo {
ProcessInstanceInfo() = default;
ProcessInstanceInfo(const char *name, const ArchSpec &arch, lldb::pid_t pid)
- : ProcessInfo(name, arch, pid), m_euid(UINT32_MAX), m_egid(UINT32_MAX),
- m_parent_pid(LLDB_INVALID_PROCESS_ID) {}
+ : ProcessInfo(name, arch, pid) {}
void Clear() {
ProcessInfo::Clear();
- m_euid = UINT32_MAX;
- m_egid = UINT32_MAX;
- m_parent_pid = LLDB_INVALID_PROCESS_ID;
+ m_euid = std::nullopt;
+ m_egid = std::nullopt;
+ m_parent_pid = std::nullopt;
}
- uint32_t GetEffectiveUserID() const { return m_euid; }
+ uint32_t GetEffectiveUserID() const { return m_euid.value(); }
----------------
feg208 wrote:
I don't have particularly strong feelings here but we should keep it consistent
between this class and the parent class which has GetUID and GetGID neither of
which returns a std::optional. Given the risk of std::bad_optional_access
getting thrown thoughout I am coming to the notion that maybe we shouldn't use
optionals here? If we have optionals I guess we are mandating that the paired
IsValid must exist and must always be called. That seems like it'll create
challenges. Can we go back to using invalid values like UINT32_MAX and maybe
alter the interface to use optionals in a subsequent pr? This way we could
debate the merits of this approach independent of adding a niceness field.
https://github.com/llvm/llvm-project/pull/91544
_______________________________________________
lldb-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
