Here’s my notes for day 2!
Zombie pointers are ABA problem pointers.
Hazard pointers don’t solve the problem.
Paper - http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2369.pdf
I asked a question! It was “what are some things the standards commmittee can do to help with poor error messages?” I don’t think the committee liked that question. Some people laughed at me after hearing it. The first answer I got was along the lines of “complain to your vendor.” But after I pressed it, they talked a bit about what you can do to teach students about error messages. Things like “show them the error messages in a controlled environment” and “translate the error messages for them the first time.”
I told my friend Anushree that I hadn’t done any internships yet, and I think she felt bad for me, because we spent the next hour talking to recruiters who’d set up shop at the conference.
Constructions separated from assignment, output parameters, and other bad practices in C++. Solution is to put const everywhere.
These were kinda disappointing. I was hoping to see something more like Chandler Carruth’s lightning talk from 2017, so I guess I was bound to be disappointed, but in general the lightning talk format felt pretty constraining for the presenters, and the content felt a lot less cool than the regular talks. I kinda wish I’d just gone to bed.