labath added inline comments.
================
Comment at: lldb/include/lldb/Utility/SharedCluster.h:52
ClusterManager() : m_objects() {}
-
- llvm::SmallVector<T *, 16> m_objects;
+ llvm::SmallPtrSet<T *, 16> m_objects;
std::mutex m_mutex;
----------------
JDevlieghere wrote:
> labath wrote:
> > JDevlieghere wrote:
> > > I assume pointers cannot be modified once they're in the set. Can this be
> > > `const T *`?
> > Did you mean pointers temselves, or the objects they point to? The pointers
> > are automatically const (you can't modify the map copy in-place), and if
> > someone wanted to store const objects, he could always instantiate this
> > class with `const T` instead.
> I meant the latter. You could indeed have an instance of this class with a
> `const T`, but that has a different meaning. My point was that the
> `m_objects` could always store const objects (even though it's somewhat
> counter intuitive that you can delete a const pointer). Anyway, it's no big
> deal, and I guess not making this `const T*` allows you to have an actual
> `const T` template argument while otherwise we'd need `const
> std::remove_const<T>` to make that case work, which obviously isn't worth it.
Ok, I see what you mean. That sounds reasonable, and I don't think the
remove_const thingy would be necessary, as it should be possible add multiple
`const` qualifiers in an idempotent manner.
Also agree, its not a big deal.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D131996/new/
https://reviews.llvm.org/D131996
_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
