Winston Chang
Winston Chang
@jimhester, @kevinushey Do you guys know if this change will cause problems on platforms that don't support C++17?
Thanks, guys. There already is a configure script, so that at least reduces the amount of work needed.
@vtamara I don't think the test is correct -- on the Ubuntu CI, you can see that it says that C++17 is required, but that hasn't been true in the...
@vtamara I've given some feedback, but don't currently have time to test and fix on various platforms for you. I don't want to introduce breakage on platforms that are widely...
One more thought: it may be simpler to do something like this: - If C++17 is available, use `CXX_STD=CXX17` - else use `CXX_STD=CXX11`
I still think the logic isn't quite right. I think it should be something like this: - If a C++17 compiler is available, we know that `` must provide `timespec_get`...
I understand your point of view, in trying to be conservative about the changes. If that's the idea, then this logic makes sense to me: - First try to compile...
@danieldjewell Can you add to src/README.md explaining this change? We've made other modifications to tinycthread and have kept a record of the changes in that file. Otherwise, looks good to...
I think it could make sense to do this when at least one callback is run when the console is idle (that is, not from an explicit call to `run_now()`....
I just realized a problem with doing this. Most callbacks have no output, and if R is running in the terminal or in RStudio, then it will print a new...