Fixes 2024 12 take 1#5
Conversation
The bug in SBCL has been reported and fixed in thread https://sourceforge.net/p/sbcl/mailman/sbcl-bugs/thread/tencent_9C9549ADDEE1D71B17C21F6A48A451531206@qq.com.
d4d5ef5 to
99bd221
Compare
|
Hello Madhu! Could you please add test cases for the modifications you've made to |
…-existent slots * macros.lisp: (define-package-cobject-classes): Unwrap the loop which pushes forward declarations for slots and correctly handle slots which are foreign pointer typedefs: Do not try to find slots when "class" doesn't have slots. i.e. when "class" is a cffi::foreign-pointer-type, a typedef of a pointer to a struct. Push the fwd decls of "pointed-to" class instead.
99bd221 to
5fb2b77
Compare
|
Sorry the earlier commits were bogus. They did not fix the actual problem I was facing with cobj::define-package-cobject-classes. I think I've addressed the problem just now. I'm still working on the tests to go with it.. To demonsrate the problem I'll need a patch which lets us specify the order in which definitions are processed by cobj::define-package-cobject-classes. Before this patch is applied on executing the above form, cffi-object bombs out on sbcl Any comments appreciated, thankx |
5fb2b77 to
1d96ab8
Compare
|
I pushed the wrong commit, hopefully I've pushed the correct branch now.. Can you please delete the earlier files pushed on the branch. |
1d96ab8 to
21af316
Compare
|
It seems you are tackling forward references of structs, which was already supported in the previous version of |
|
The proposed patch only fixes a type mismatch in the implementation. |
Hello, here are some more miscellaneous patches for your review., do let
me know if these are correct or should be done differently.