Hi Richard, On 12/02/26 2:19 pm, Richard Biener wrote: > On Tue, Dec 16, 2025 at 8:22 AM Eric Botcazou <[email protected]> > wrote: >> >>> Since you are the maintainer for the "cprop_hardreg" pass, can I request you >>> to allow the backport? >> >> The thing is, the gcc-13 branch is old now and only one remaining release >> will >> be made from it, so we must make sure not to introduce regressions at the >> last >> minute. The patch is not particularly risky of course, but I'm not sure if >> it >> is really appropriate to backport it at this point. If other maintainers or >> the release managers are nevertheless OK with it, then I'll join them. > > I'll note that fixing on trunk and the 13 branch only is definitely a no-go, > the 15 and 14 branches would need to receive the same fix. Fixing the 15 > branch looks reasonable given this is wrong-code. I understand the > testcase only fails on the 13 branch? The bugzilla isn't very well > maintained, > no reported against version, no version that works/fails recorded.
The issue exists on trunk too. Yes, I understand that backporting to the 13 branch will involve backporting to 14 and 15 branches too. I asked for the 13 branch as I wanted permission to backport till gcc-13. I have updated the bugzilla. -Surya > > Richard. > >> >> -- >> Eric Botcazou >> >>
