Matthew Flatt <mflatt@...> writes: > In version 5.1.3 and earlier, you could write > > (define-cstruct _git_index_entry_unmerged > ([mode0 _uint] > [mode1 _uint] > [mode2 _uint] > [oid0 _git_oid] ; assuming _git_oid is defined > [oid1 _git_oid] > [oid2 _git_oid] > [path _string])) > > With the latest development version (or v5.1.3.3 and later), you can > write > > (define-cstruct _git_index_entry_unmerged > ([mode (_array _uint 3)] > [oid (_array _git_oid 3)] ; assuming _git_oid is defined > [path _string])) > >
Thanks a lot! I think I will work on other parts of the library and then I'll definitely go for the development version option, which seems to be just what I was looking for. In fact, _git_oid was already a struct with a static byte array of length 40 as its only member. Since every function I had come up with so far used a (maybe double) pointer to _git_oid, I could just declare it as (_bytes <mode> 40)... until I found the structs in my first email. The first solution is probably too ugly when each oid has a length of 40. Cheers and thanks again! _________________________________________________ For list-related administrative tasks: http://lists.racket-lang.org/listinfo/users
