https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119792
--- Comment #17 from rguenther at suse dot de <rguenther at suse dot de> --- On Mon, 21 Apr 2025, ebotcazou at gcc dot gnu.org wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119792 > > --- Comment #16 from Eric Botcazou <ebotcazou at gcc dot gnu.org> --- > > Eric, can you try to read through > > https://gcc.gnu.org/wiki/document-middle-end-type-system and amend > > it with comments about Ada? > > I added a minimal comment about TYPE_CANONICAL a couple of days ago. > > > I'm especially curious what exactly the properties are that are not useless > > in the VIEW_CONVERT_EXPRs that get wrongly stripped here? > > The VIEW_CONVERT_EXPR gives the (fixed) size of the RHS of an assignment, the > RHS being the dereference of a pointer to a self-referential type > (type_contains_placeholder_p). So WITH_SIZE_EXPR isn't applicable here? OK, that's a thing only introduced by gimplification at the moment, IIRC. So in case the VIEW_CONVERT_EXPR is "just" to make the gimplifier emit the appropriate WITH_SIZE_EXPR I wonder whether we should change GENERIC to make that the prefered communication vehicle? Note TREE_TYPE of the WITH_SIZE_EXPR should agree with the TREE_TYPE of its operand 0. The other option would be to make sure to not apply the GIMPLE useless_type_conversion_p to GENERIC before creating the WITH_SIZE_EXPR (which would be a fix anyway, IMO). In case it's really just about WITH_SIZE_EXPR ...
