That's one possibility. Another one is to do as I suggested a couple of days ago. Set a combination of flags in ob_flags to tell XaAES to treat ob_spec as a pointer to a gradient-structure where the first long of thestruct contains the original ob_spec (so perhaps the indirect-flag shouldbe set too). Then it would be possible to define gradient for every type of objects without having to add a system call.How would you inform the AES that a gradient has changed, so it will recalculate it?
Hmm... Are the gradients precalculated, or made on the fly?
Your suggestion is more elegant, but I don't think application developersshould be encouraged to use this feature as it will break the concept of themes/skins.This is a basic discussion: Should apps have their own themes (like e.g. opera on windows),or should everything forced to look the same?
The answer is simple: Everything must look the same. This is a major factor in good UI design, and it's also necessary to allow successfull skinning. Unfortunately very few GEM-applications will allow proper skinning. Most applications were made in the 90's and are supposed to work on plain TOS. So there's userdefs everywhere due to the lack of useful stuff like popups, frames and keyboard shortcuts. These apps will never look good with a skin/theme that differs much from the standard AES look.
Jo Even