[Bug bootstrap/125598] [16/17 regression] Comparing stages 2 and 3 with -with-build-config=bootstrap-O3
"yizeyi18 at mail dot nankai.edu.cn via Gcc-bugs" <[email protected]>
| Newsgroups | gmane.comp.gcc.bugs |
|---|---|
| Message-ID | <[email protected]/bugzilla/> |
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125598 --- Comment #9 from 伊泽平 <yizeyi18 at mail dot nankai.edu.cn> --- (In reply to Sam James from comment #8) > With empty BUILD_CONFIG, it's likely PR126526. I don't know if OP's issue is > the same or not though. Seems the same; my agent also says a lot about var tracking caused difference here, which I am not familiar with. To be helpful, it also hints me that the mechanism ENABLE_GC_CHECKING take effect is that ENABLE_GC_CHECKING affects init_ggc_heuristics()(from gcc/ggc-common.cc:1127) then somewhat heap size changed then some template initialize-associated variable address changed for different GC behavior; finally this addr diff yields different debug section.