In D65148#1629103, @lebedev.ri wrote:

In D65148#1629010, @dmgreen wrote:

That is a valid transformation: https://rise4fun.com/Alive/bxq

I see, because of the truncs back into an i16. That makes sense.

Ok, and that one is done too:

$ ./bin/clang -target arm-none-eabi -mcpu=cortex-m0plus -O3 -mllvm -phi-node-folding-threshold=2 -S -o old.s /tmp/test.c 
$ ./bin/clang -target arm-none-eabi -mcpu=cortex-m0plus -O3 -mllvm -phi-node-folding-threshold=3 -S -o new.s /tmp/test.c 
$ diff old.s new.s 
$ <no diff>
$ sha512sum *.s
9332ab2169151fa21ab03e1fb218178aa16bed2c0dfdd0a40b178a75dbd4a865ba7a7b1b4b37f252928ea794e38776260f84f6c1ae3684d5a445118f518351e2  new.s
9332ab2169151fa21ab03e1fb218178aa16bed2c0dfdd0a40b178a75dbd4a865ba7a7b1b4b37f252928ea794e38776260f84f6c1ae3684d5a445118f518351e2  old.s

Minor update.

I would think this should go into DAGCombiner under hasBitTest() hook.

I'm testing this patch on our codebase and it is working pretty well. We use the Allman style and the lambda problem has been an issue for many years. One thing to note in this patch is that some of the files have CRLF line endings but should be LF endings, which is why they're showing so many edits. I'm also seeing a clang tidy test failing with this patch. The readability-braces-around-statements tests seem to fail because the indent width is appending double of what it should.
void test() {

   if (cond("if0") /*comment*/) {
      do_something("same-line");
}

Add a missing null check.

• Add missing renames