[Lldb-commits] [PATCH] D142341: [LLDB][NFC] Fix valobj_sp null pointer checks in lldb/source/Plugins/Language/

[Dave Lee via Phabricator via lldb-commits]

kastiglione added inline comments.


================
Comment at: lldb/source/Plugins/Language/CPlusPlus/LibCxx.cpp:225
   if (valobj_sp)
+    SyntheticChildrenFrontEnd(*valobj_sp);
     Update();
----------------
Michael137 wrote:
> Michael137 wrote:
> > Michael137 wrote:
> > > this won't initialise the parent constructor though will it? Just creates a temporary and immediately destructs it? You might need to put it back into the initialiser list and use a ternary
> > Oh but `SyntheticChildrenFrontEnd` seems to always take a reference. The interfaces seem to be a little mismatched
> Perhaps the null-check in the constructor is just redundant and should actually be an assert.
> 
> @jingham might know more about the assumptions here
@xgupta reiterating Michael's point, I think this change results in mis-construction. Have you run the test suite, I would expect some failures with this change.

I agree that this should be an assert. The other option would be to change the constructor's signature, using a `lldb::ValueObject &` instead of `lldb::ValueObjectSP`, which puts the onus on callers to handle any null-ness. Are we aware of any callers passing null?




Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142341/new/

https://reviews.llvm.org/D142341