On 7/27/20 5:53 PM, Martin Sebor wrote:
The tests I committed with the change didn't exercise any of this so that's my bad. I'm still not sure I understand how the problem with the incomplete type comes up (I haven't had a chance to look into the recent updates on the bug yet) but to retain the optimization (and keep the comments in sync with the code) I think a better solution than restricting the function to integers is to limit it to complete types. Beyond that, extending the function to also constant arrays or nonzero aggregates will also enable the optimization for those.
Hello. I must admit that I'm not super-familiar with that code I modified. Can you please assign the PR and propose a proper fix? I can then test it on chromium and I'm also deferring backport of the patch I installed to master. Thanks, Martin
