Taken from https://gitlab.com/embeddable-common-lisp/ecl/-/commit/f3d4cf4b66ab6c3cd8629ab6d0c7f7c50d7fd8a4
with "src/cmp/cmppass2-loc.lsp" changed to its old name "src/cmp/cmploc.lsp"

diff --git a/src/cmp/cmpc-wt.lsp b/src/cmp/cmpc-wt.lsp
index 2f5f4063ca7050975d8468a322165dcd098f713e..1a681455c5e3888fffdc59637900a0123e339758 100644
--- a/src/cmp/cmpc-wt.lsp
+++ b/src/cmp/cmpc-wt.lsp
@@ -19,18 +19,7 @@
 (defun wt1 (form)
   (cond ((not (floatp form))
          (typecase form
-           (INTEGER
-            (princ form *compiler-output1*)
-            (princ
-             (cond ((typep form (rep-type->lisp-type :int)) "")
-                   ((typep form (rep-type->lisp-type :unsigned-int)) "U")
-                   ((typep form (rep-type->lisp-type :long)) "L")
-                   ((typep form (rep-type->lisp-type :unsigned-long)) "UL")
-                   ((typep form (rep-type->lisp-type :long-long)) "LL")
-                   ((typep form (rep-type->lisp-type :unsigned-long-long)) "ULL")
-                   (t (baboon :format-control "wt1: The number ~A doesn't fit any integer type." form)))
-             *compiler-output1*))
-           ((or STRING CHARACTER)
+           ((or INTEGER STRING CHARACTER)
             (princ form *compiler-output1*))
            (VAR (wt-var form))
            (t (wt-loc form))))
diff --git a/src/cmp/cmppass2-loc.lsp b/src/cmp/cmppass2-loc.lsp
index c6ec0a6637e399268cfaf8fd1021ca0ef08f7432..a1fa9fd8704e1eaba0561e1df23bba254451c30b 100644
--- a/src/cmp/cmploc.lsp
+++ b/src/cmp/cmploc.lsp
@@ -181,10 +181,30 @@
 (defun wt-temp (temp)
   (wt "T" temp))
 
+(defun wt-fixnum (value &optional vv)
+  (declare (ignore vv))
+  (princ value *compiler-output1*)
+  ;; Specify explicit type suffix as a workaround for MSVC. C99
+  ;; standard compliant compilers don't need type suffixes and choose
+  ;; the correct type themselves. Note that we cannot savely use
+  ;; anything smaller than a long long here, because we might perform
+  ;; some other computation on the integer constant which could
+  ;; overflow if we use a smaller integer type (overflows in long long
+  ;; computations are taken care of by the compiler before we get to
+  ;; this point).
+  #+msvc (princ (cond ((typep value (rep-type->lisp-type :long-long)) "LL")
+                      ((typep value (rep-type->lisp-type :unsigned-long-long)) "ULL")
+                      (t (baboon :format-control
+                                 "wt-fixnum: The number ~A doesn't fit any integer type."
+                                 value)))
+                *compiler-output1*))
+
 (defun wt-number (value &optional vv)
+  (declare (ignore vv))
   (wt value))
 
 (defun wt-character (value &optional vv)
+  (declare (ignore vv))
   ;; We do not use the '...' format because this creates objects of type
   ;; 'char' which have sign problems
   (wt value))
diff --git a/src/cmp/cmptables.lsp b/src/cmp/cmptables.lsp
index 814f5f8410ce50b3a9b00d626bb9c641ecd526a9..e649bf8834fb355b188de4123c38d3b0a4b7ae5b 100644
--- a/src/cmp/cmptables.lsp
+++ b/src/cmp/cmptables.lsp
@@ -182,7 +182,7 @@
 
     (temp . wt-temp)
     (lcl . wt-lcl-loc)
-    (fixnum-value . wt-number)
+    (fixnum-value . wt-fixnum)
     (long-float-value . wt-number)
     (double-float-value . wt-number)
     (single-float-value . wt-number)