Don't use std::conj if its argument could be real

[mhoemmen]

This is a follow-on to PR #87.

[lums658]

What was the decision on this — metaprogram to dispatch to identity (no-op) if real? When we were writing out iterative template library, we needed to write the algorithms (solvers mostly) to handle the complex case as well as the real, so we needed to write them with dot_conj (and conj) — and they needed to do the right thing with real arguments. Regards, Andrew Lumsdaine On Oct 18, 2021, at 1:21 PM, Mark Hoemmen ***@***.******@***.***>> wrote: This is a follow-on to PR #87<#87>. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#115>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AGKPFDMDOR2CHNGBEU52NF3UHR6UJANCNFSM5GHPX43A>. Triage notifications on the go with GitHub Mobile for iOS<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[mhoemmen]

@lums658 wrote:

What was the decision on this — metaprogram to dispatch to identity (no-op) if real?

It's not really a "decision" -- P1673 already requires that hermitian_* etc. work for noncomplex element types, so we just need to write some overloaded conj_if_complex thing that is the identity for everything other than std::complex.

[mhoemmen]

This issue is subsumed by ORNL/cpp-proposals-pub#236. Thanks all!