[PATCH] vect: Add assert to expand_vector_conversion [PR118616]

2025-04-05 Thread Andrew Pinski
In some cases (after inliing due to LTO and -O3), GCC cannot figure out that the length of the converts vect is not empty when supportable_indirect_convert_operation returns true. So we get an extra warning because we loop through all but the last entry and GCC decided that `converts.length () - 1`

Re: [PATCH] vect: Add assert to expand_vector_conversion [PR118616]

2025-03-24 Thread Andrew Pinski
On Mon, Mar 24, 2025 at 2:00 AM Richard Biener wrote: > > On Sun, Mar 23, 2025 at 9:46 PM Andrew Pinski > wrote: > > > > In some cases (after inliing due to LTO and -O3), GCC cannot > > figure out that the length of the converts vect is not empty > > when supportable_indirect_convert_operation r

Re: [PATCH] vect: Add assert to expand_vector_conversion [PR118616]

2025-03-24 Thread Richard Biener
On Sun, Mar 23, 2025 at 9:46 PM Andrew Pinski wrote: > > In some cases (after inliing due to LTO and -O3), GCC cannot > figure out that the length of the converts vect is not empty > when supportable_indirect_convert_operation returns true. So > we get an extra warning because we loop through all