Thanks for your help.
I am a postgraduate student, and want to do some things about Sanitizers. Since Thread Sanitizer works good but still have a lot space (the cost and time of program) to work on, I think that doing some research on that is not a bad idea.
But the imformation online valuable were posted almost three years ago, and no new paper works on that. Obvious I have a question about why researchers don’t take it important or just I got the wrong pages on that. Is what researchers did is good enough or not other space left?
By the way I have read the web on llvm.org and GitHub - google/sanitizers: AddressSanitizer, ThreadSanitizer, MemorySanitizer.
Could anyone tell me something they don’t like about the ThreadSanitizer ?
That’ good enough for me.
Thank you again.
I do not like Thread Sanitizer because it is not that user friendly.
Thanks for your reply, could you tell me which part make you fell ‘‘not user friendly’’. For example, maybe the out message is not readable well? Or it just cost too much resource of the computer, which is not workable in real production environment?
One commonly cited issue is that it doesn’t supports fences, see e.g. Eliminating Data Races in Firefox - A Technical Report - Mozilla Hacks - the Web developer blog, although there are workarounds. There was an attempt to add support in ⚙ D47107 Vector clock algorithm improvement in case of blocking release sequence (the maintainer pointed me to that in TSan:data race using memory fence · Issue #1144 · google/sanitizers · GitHub).
Thanks for your reply, I will do some research on that. Actually, I am reading your words on the github. Thank you agian.