Get fork() working on Windows

This is done without using Microsoft's internal APIs. MAP_PRIVATE
mappings are copied to the subprocess via a pipe, since Microsoft
doesn't want us to have proper COW pages. MAP_SHARED mappings are
remapped without needing to do any copying. Global variables need
copying along with the stack and the whole heap of anonymous mem.
This actually improves the reliability of the redbean http server
although one shouldn't expect 10k+ connections on a home computer
that isn't running software built to serve like Linux or FreeBSD.

ape/ape.S

@@ -225,9 +225,14 @@ pc:	cld
 3:	call	pcread
 	dec	%di
 	jnz	3b
-6:	mov	%ax,XLM(LOADSTATE)+0
-	mov	%cx,XLM(LOADSTATE)+2
-	mov	%dx,XLM(LOADSTATE)+4
+6:	mov	$XLM(LOADSTATE),%di		# ax,cx,dx,es
+	stosw
+	xchg	%cx,%ax
+	stosw
+	xchg	%dx,%ax
+	stosw
+	mov	%es,%ax
+	stosw
 	ljmp	$0,$REAL(realmodeloader)
 	.endfn	pc,globl,hidden
 

ape/config.h

@@ -126,7 +126,7 @@
 #define XLM_BADIDT                          0x2230
 #define XLM_BADIDT_SIZE                     6
 #define XLM_LOADSTATE                       0x2240
-#define XLM_LOADSTATE_SIZE                  6
+#define XLM_LOADSTATE_SIZE                  8
 #define XLM_SIZE                            ROUNDUP(XLM_LOADSTATE + XLM_LOADSTATE_SIZE, 0x1000)
 #define IMAGE_BASE_REAL                     (XLM_BASE_REAL + XLM_SIZE)