Groups | Search | Server Info | Login | Register
Groups > comp.compilers > #3470
| From | Cameron McInally <cameron.mcinally@nyu.edu> |
|---|---|
| Newsgroups | comp.compilers |
| Subject | Re: binary search debugging of compilers |
| Date | 2023-05-17 13:40 -0400 |
| Organization | Compilers Central |
| Message-ID | <23-05-014@comp.compilers> (permalink) |
| References | <CADSkJJVN7RGqhgVDZaz_K5be6uEDaMofMu5OF3RR4Y5-fDu00Q@mail.gmail.com> <23-05-013@comp.compilers> |
On Wed, May 17, 2023 at 13:13 Russ Cox <rsc@swtch.com> wrote: > Thanks to everyone for the replies and pointers. > > As far as 'git bisect' is concerned, we have found 'git bisect' > incredibly useful too, of course. It is clearly another instance of > binary search to find bugs, but it is binary search over time, while > the system I described is binary search over the program being > compiled. "What change broke the compiler?" vs "What code does that > change miscompile?". The two are independent and complementary. As for a binary search over the program being compiled, LLVM has a mechanism to bisect optimization passes. Not exactly the same, but worth noting. https://llvm.org/docs/OptBisect.html Also, Viet is still in compilers at Coherent Logix.
Back to comp.compilers | Previous | Next — Previous in thread | Next in thread | Find similar
Re: binary search debugging of compilers Russ Cox <rsc@swtch.com> - 2023-05-17 10:55 -0400
Re: binary search debugging of compilers Cameron McInally <cameron.mcinally@nyu.edu> - 2023-05-17 13:40 -0400
Re: binary search debugging of compilers Kaz Kylheku <864-117-4973@kylheku.com> - 2023-05-17 18:47 +0000
Re: binary search debugging of compilers Spiros Bousbouras <spibou@gmail.com> - 2023-05-18 10:50 +0000
Re: binary search debugging of compilers Kaz Kylheku <864-117-4973@kylheku.com> - 2023-05-19 03:44 +0000
binary search debugging of compilers Cliff Click <cclick0@gmail.com> - 2023-05-19 10:47 -0700
Re: binary search debugging of compilers mrs@kithrup.com (Mike Stump) - 2023-05-20 18:04 +0000
csiph-web