Serious Chrome zero-day

Re:Serious Chrome zero-day by chii on Wednesday March 06, 2019 @07:41PM (#19324576)

> I can write a safe C++ app on my own >> But only if you don't use any external libraries

using an external library is not "on your own"

[ Parent | on HN ]

Re:Serious Chrome zero-day by dymk on Wednesday March 06, 2019 @08:10PM (#19324729)

Can you use the STL? What boundary is considered trusted?

[ Parent | on HN ]

Re:Serious Chrome zero-day by chii on Thursday March 07, 2019 @09:34PM (#19334643)

The point i was making is that such boundary doesn't exist - no one can write everything "on their own".

Therefore, language features to prevent a class of exploits should be a high priority when considering a project.

[ Parent | on HN ]

Re:Serious Chrome zero-day by Agathos on Wednesday March 06, 2019 @08:21PM (#19324776)

I can trust the compiler... I hope?

[ Parent | on HN ]

Re:Serious Chrome zero-day by stcredzero on Wednesday March 06, 2019 @08:24PM (#19324799)

Hope...

[ Parent | on HN ]

Re:Serious Chrome zero-day by pjmlp on Thursday March 07, 2019 @02:43AM (#19326258)

Not at all, check the Linux 5.0 bug introduced by a gcc "optimization" regarding UB.

[ Parent | on HN ]

Re:Serious Chrome zero-day by Miltnoid on Wednesday March 06, 2019 @08:56PM (#19324937)

Use CompCert! (And C instead of C++)

[ Parent | on HN ]

Re:Serious Chrome zero-day by Johnny555 on Thursday March 07, 2019 @03:21PM (#19331929)

Don't most people link at least with basic libraries to help with I/O and other standard operating system interfaces?

Do many people really use the kernel syscall interface directly?

[ Parent | on HN ]