https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69177
--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Because the new number of lines before columns are lost are way too small for lots of larger sources. Having half as many lines or so before we trigger the column loss is IMHO acceptable, hitting it 32 times earlier is just too much. I agree with Richi on this.
