Fix typos

Fixed typo. Sorry!

Add (Initializer && and assert to else branch per Tom's comments.

Add assert at the beginning of blocks pointed out by Coverity

In D139148#3969383, @tahonermann wrote:

When committing changes to address Coverity reported issues, I think it would be useful to include the Coverity analysis in the commit message. A textual representation can be obtained using cov-format-errors --emacs-style from the command line with access to the intermediate directory. I'm not sure if a textual representation can be obtained through the Coverity UI though.

Perhaps the` assert` should be added after line 5926 above.

Done.

I think the checks for Initializer being non-null following the addition of an assert should be removed.

I don't see any after this point.

Move the assert to after the branches that handle the cases where Initializer may be null

Add assert instead of multiple checks in if conditions.

Add more nullptr checks per Shafik's comments

In D139148#3965404, @shafik wrote:

This looks like the correct thing to do but I see further down there are also unconditional uses of Initializer that should also be guarded with a check. It would be good to fix those as well since we are here.

Yes, it does make sense.

I am guessing the answer is no since this was found using a static analysis tool but do you have a test case that triggers this failure?

Indeed, the answer is no.

In D115320#3195090, @DavidSpickett wrote:

I've gone ahead and added the REQUIRES myself.

Change CHECK-LABEL to STORE-LINE-LABEL

Rebasing patch to see if the test failure will go away.

https://reviews.llvm.org/harbormaster/unit/view/1536685/
Test failure seems unrelated to this patch. Unable to reproduce locally.

Fix test format

Changing test's CHECK lines

1. Removed returnNullTBAA() method
2. Add NO-TBAA check-prefix to the test

In D115320#3182581, @jeroen.dobbelaere wrote:

When I try out the example on llvm-13, I get a 'omnipotent char' tbaa description. That should be ok in general. When I replace the 'float _Complex' with 'double', I do get the 'double' tbaa. That might be a better example for the testcase ?

Updated test per Jeroen's suggestion

In D115320#3178682, @jeroen.dobbelaere wrote:

Do you have a testcase ?

Add test case
Fix formatting

In D108712#2965433, @aaron.ballman wrote:

LGTM! There is no way for this to be null because emitRethrow() is what should be called in that case.

I'll land this on your behalf sometime shortly (perhaps as late as tomorrow).

In D108481#2960875, @aaron.ballman wrote:

LGTM, thanks for the cleanup! Do you need someone to commit on your behalf? If so, what name and email address would you like me to use for patch attribution?

1. Removed the useless assert on TC.
2. Retained the change from cast_or_null to cast