https://sourceware.org/bugzilla/show_bug.cgi?id=33127
--- Comment #20 from Indu Bhagat <indu.bhagat at oracle dot com> --- (Apologies for the delay.) I was trying to reason for SFrame if the generated R_*_NONE entries will always be towards the end. To which the answer is : yes. SFrame merging code does rely on one-to-one order between FDEs and the relas. So this patch should be OK. Also, looks like the presence of : 000000000030 000000000000 R_X86_64_NONE 0 000000000030 000000000000 R_X86_64_NONE 0 in relas in general should not trip any existing consumer of these relocatable binaries ... (including ld after this patch). -- You are receiving this mail because: You are on the CC list for the bug.
