We've observed this in two RELs so far - zeroed data not going in a .bss section but rather a .data section, with a locally applied #pragma explicit_zero_data on as a workaround. Is there a different way this can happen?
Enabling #pragma explicit_zero_data on globally would probably break a ton of files but removing the = nullptr or = 0.0f could restore them. It's probably not the right solution though.
We've observed this in two RELs so far - zeroed data not going in a .bss section but rather a .data section, with a locally applied
#pragma explicit_zero_data onas a workaround. Is there a different way this can happen?Enabling
#pragma explicit_zero_data onglobally would probably break a ton of files but removing the= nullptror= 0.0fcould restore them. It's probably not the right solution though.