From ae819deb38c3a962e3badf020705c3d0c6c84e67 Mon Sep 17 00:00:00 2001
From: Guillaume Melquiond
Date: Thu, 25 Mar 2021 12:03:40 +0100
Subject: Make it more obvious when the calling convention of APPLY changes.

Despite their names, APPLY1 to APPLY4 are completely different from
APPLY(n) with n = 1 to 4. Indeed, the latter assumes that the return
address was already pushed on the stack, before the arguments were. On the
other hand, APPLY1 to APPLY4 insert the return address in the middle of
the already pushed arguments.
---
 kernel/vmbytecodes.mli | 1 +
 1 file changed, 1 insertion(+)

(limited to 'kernel/vmbytecodes.mli')

diff --git a/kernel/vmbytecodes.mli b/kernel/vmbytecodes.mli
index eeca0d2ad1..d9e2d91177 100644
--- a/kernel/vmbytecodes.mli
+++ b/kernel/vmbytecodes.mli
@@ -30,6 +30,7 @@ type instruction =
   | Kpush                               (** sp = accu :: sp *)
   | Kpop of int                         (** sp = skipn n sp *)
   | Kpush_retaddr of Label.t            (** sp = pc :: coq_env :: coq_extra_args :: sp ; coq_extra_args = 0 *)
+  | Kshort_apply of int                 (** number of arguments (arguments on top of stack) *)
   | Kapply of int                       (** number of arguments (arguments on top of stack) *)
   | Kappterm of int * int               (** number of arguments, slot size *)
   | Kreturn of int                      (** slot size *)
-- 
cgit v1.2.3