[Lldb-commits] [PATCH] D50912: Don't cancel the current IOHandler when we push a handler for an utility function run.

Raphael Isemann via Phabricator via lldb-commits lldb-commits at lists.llvm.org
Thu Aug 23 11:31:14 PDT 2018

teemperor added inline comments.

Comment at: source/Target/Process.cpp:4696-4697
+    // the IOHandler for Editline).
+    bool cancel_top_handler = m_mod_id.IsRunningUtilityFunction();
+    GetTarget().GetDebugger().PushIOHandler(io_handler_sp, cancel_top_handler);
     return true;
clayborg wrote:
> Do we still need this extra bool to PushIOHandler? Can't we do the code snippet I suggested above?
Do we know for sure that omitting this IOHandler push doesn't break something? From the comments above it seems that this is actually changing LLDB's behavior. And we can't easily revert this commit once I stack the expression completion + follow-ups on top of it.

I'm fine with doing this change as a direct follow-up commit that won't have any dependencies attached.


More information about the lldb-commits mailing list