Jim King

Results 105 comments of Jim King

Does this have any relation to #13?

There are also tests that explicitly set `BOOST_NO_CXX11_VARIADIC_TEMPLATES` to test C++03 compatibility instead of just forcing that test to run if cxxstd=03. I found that both chrono and optional now...

I've been able to reproduce this reliably and grabbed information from the debugger. Here's what I'm seeing: 1. PMIx fence is stuck: fenced = 0 after 1+ billion iterations 2....

I was able to use https://github.com/tonistiigi/binfmt/releases/tag/deploy%2Fv9.2.2-52 in place of this repo to work around this issue.

You can also handle the error result when it is already initialized, and ignore it.

I added this to my model base class, it's quite effective: ```python @contextmanager def no_history(self): """ Disable history generation for this instance. """ assert not hasattr( # nosec self, "skip_history_when_saving"...

Looks reasonable; submit a pull request?

You'll want to align this to https://github.com/apache/thrift/pull/3049 which I just merged.

This should be locked down. See issue 305 comments.

The return value optimization to use a reference is probably a wash, since the value is going to sent through the protocol anyway. Maybe it provides a tiny memory savings....