Hi Jerry!
Am 06.06.26 um 10:40 PM schrieb Jerry D:
On 6/6/26 1:20 PM, Harald Anlauf wrote:
Dear all,
this patch fixes a 15/16/17 regression when using the LOC() intrinsic
on SELECT TYPE temporaries, where gfc_is_simply_contiguous could
return false and gfc_conv_array_parameter would create unnecessarily
messy code that lead to an ICE.
I looked at how C_LOC solved this, and adjusted gfc_conv_intrinsic_loc
accordingly.
Regtested on x86_64-pc-linux-gnu. OK for mainline / 16 / 15?
Thanks,
Harald
Yes, looks good.
Thanks for the swift review!
Pushed as r17-1396-g2dcdb2463f8993 to mainline so far.
Harald
Thanks Harald.
Jerry
