Optimize chkPS

Use EXX instead of the stack for HL protection and remove all
spurious uses of chkPS,

I wanted to inline chkPS in next because of its "tight loop" status,
but for reasons I don't understand, doing so breaks Collapse OS.
Later...

blk/003

@@ -10,8 +10,7 @@ the dictionary (B30) for a word reference.
 
 Contents
 
- 4 Number literals              6 Compilation vs meta-comp.
+ 5 Number literals              6 Compilation vs meta-comp.
  8 Interpreter I/O             11 Signed-ness
 14 Addressed devices           17 DOES>
-18 Disk blocks                 21 How blocks are organized
-22 Addressed devices
+18 Disk blocks                                          (cont.)

blk/004

@@ -1,11 +1 @@
-Number literals
-
-Traditional Forth often use HEX/DEC switches to go from decimal
-to hexadecimal parsing. Collapse OS parses literals in a way
-that is closer to C.
-
-Straight numbers are decimals, numbers starting with "0x"
-are hexadecimals (example "0x12ef"), "0b" prefixes indicate
-binary (example "0b1010"), char literals are single characters
-surrounded by ' (example 'X'). Char literals can't be used for
-whitespaces.
+21 How blocks are organized    22 Addressed devices

blk/005

@@ -0,0 +1,12 @@
+Number literals
+
+Traditional Forth often use HEX/DEC switches to go from decimal
+to hexadecimal parsing. Collapse OS parses literals in a way
+that is closer to C.
+
+Straight numbers are decimals, numbers starting with "0x"
+are hexadecimals (example "0x12ef"), "0b" prefixes indicate
+binary (example "0b1010"), char literals are single characters
+surrounded by ' (example 'X'). Char literals can't be used for
+whitespaces.
+

blk/216

@@ -1,8 +1,10 @@
 : OP1 CREATE C, DOES> C@ A, ;
 0xf3 OP1 DI,                   0xfb OP1 EI,
 0xeb OP1 EXDEHL,               0xd9 OP1 EXX,
+0x08 OP1 EXAFAF',
 0x76 OP1 HALT,                 0xe9 OP1 JP(HL),
 0x12 OP1 LD(DE)A,              0x1a OP1 LDA(DE),
+0x02 OP1 LD(BC)A,              0x0a OP1 LDA(BC),
 0x00 OP1 NOP,                  0xc9 OP1 RET,
 0x17 OP1 RLA,                  0x07 OP1 RLCA,
 0x1f OP1 RRA,                  0x0f OP1 RRCA,

blk/285

@@ -1,11 +1,11 @@
 CODE (?br)           ( 0x67 )
-    HL POPqq, chkPS,
+    HL POPqq,
     HLZ,
     JRZ, L2 BWR ( BR + 2. False, branch )
     L1 BSET ( loop will jump here )
     ( True, skip next 2 bytes and don't branch )
     IY INCss, IY INCss,
-    JPNEXT,
+    JPNEXT, NOP, NOP, NOP,
 CODE (loop) ( 0x77 )
     0 IX+ INC(IXY+), IFZ, 1 IX+ INC(IXY+), THEN, ( I++ )
     ( Jump if I <> I' )

blk/286

@@ -1,11 +1,11 @@
 CODE >R ( 0xa8 )
-    HL POPqq, chkPS,
+    HL POPqq,
     17 BCALL, ( 17 == pushRS )
-;CODE
+;CODE NOP, NOP, NOP,
 CODE 2>R ( 0xb9 )
-    DE POPqq, HL POPqq, chkPS,
+    DE POPqq, HL POPqq,
     17 BCALL, ( 17 == pushRS ) EXDEHL, 17 BCALL,
-;CODE
+;CODE NOP, NOP, NOP,
 CODE R> ( 0xce )
     20 BCALL, ( 20 == popRS )
     HL PUSHqq,

blk/299

@@ -1,10 +1,13 @@
 PC ORG @ 0x1e + ! ( chkPS )
-    HL PUSHqq,
+( Note that you only need to call this in words that push
+  back to PSP. If they don't, calling chkPS is redundant with
+  check in next )
+    EXX,
 ( We have the return address for this very call on the stack
-  and protected registers. 4 - is to compensate that. )
-    HL PS_ADDR 4 - LDddnn,
+  and protected registers. 2 - is to compensate that. )
+    HL PS_ADDR 2 - LDddnn,
     SP SUBHLss,
-    HL POPqq,
+    EXX,
     CNC RETcc,      ( PS_ADDR >= SP? good )
     JR, L2 BWR      ( abortUnderflow-B298 )
 

blk/310

@@ -1,6 +1,5 @@
 CODE PICK
     HL POPqq,
-    chkPS,
     ( x2 )
     L SLAr, H RLr,
     SP ADDHLss,

blk/312

@@ -2,7 +2,6 @@
 CODE 2DROP
     HL POPqq,
     HL POPqq,
-    chkPS,
 ;CODE
 
 ( a b -- a b a b )

blk/322

@@ -1,7 +1,6 @@
 CODE !
     HL POPqq,
     DE POPqq,
-    chkPS,
     (HL) E LDrr,
     HL INCss,
     (HL) D LDrr,

blk/323

@@ -1,7 +1,6 @@
 CODE C!
     HL POPqq,
     DE POPqq,
-    chkPS,
     (HL) E LDrr,
 ;CODE
 

blk/324

@@ -1,7 +1,6 @@
 CODE PC!
     BC POPqq,
     HL POPqq,
-    chkPS,
     L OUT(C)r,
 ;CODE
 

recipes/rc2014/eeprom/README.md

@@ -34,19 +34,13 @@ I don't think you need a schematic. It's really simple.
 
 ### Building the binary
 
-You build the binary by modifying the base recipe's `xcomp` unit. This binary
-is missing 2 things: Addressed devices and the AT28 Driver.
+The binary from the base recipe has almost all it needs to write to EEPROM. The
+only thing it needs is the AT28 driver from B590. You could add it to the
+`xcomp` unit and rebuild, but the driver is so tiny, you're probably better off
+loading it at runtime.
 
-Addressed devices are at B140. If you read that block, you'll see that it tells
-you to load block 142. Open the `xcomp` unit and locate the ACIA driver loading
-line. Insert your new load line after that one.
-
-Do the same thing with the AT28 driver (B590)
-
-You also have to modify the initialization sequence at the end of the `xcomp`
-unit to include `ADEV$`.
-
-Build again, write `os.com` to EEPROM.
+If your system has mass storage, it's as easy as a LOAD. If it doesn't, you
+can use `/tools/exec` to send `blk/591` to the RC2014.
 
 ## Writing contents to the AT28
 
@@ -58,7 +52,7 @@ run this from your modern computer:
     ./upload <tty device> a000 <filename>
 
 Then, activate `AT28!` with `' AT28! A!* !` and then run
-`0xa000 0x2000 <size-of-bin> AMOVE`. `AT28!` checks every myte for integrity,
+`0xa000 0x2000 <size-of-bin> AMOVE`. `AT28!` checks every byte for integrity,
 so it there's no error, you should be fine. Your content is now on the EEPROM!
 
 Why not upload content directly to `0x2000` after having activated `AT28!`?