put a redundant null byte after the string

this will cause issues if the word has a null byte in the name.
so don't do that. obviously

2 files changed, 7 insertions, 0 deletions

diff --git a/jefs.s b/jefs.s
index c0b1b82..3234b44 100644
--- a/jefs.s
+++ b/jefs.s
@@ -21,6 +21,7 @@
 		%define mac_latest lfa_%2
 	ffa_%2:	db %3 ; FFA
 	nfa_%2:	dw slen ; NFA
+		db 0
 		db %1
 %endmacro
 
diff --git a/readme.md b/readme.md
index 97ed56a..51a8f1c 100644
--- a/readme.md
+++ b/readme.md
@@ -18,11 +18,17 @@ the dictionary follows a fairly standard format.
 | flag | 1 byte | FFA (flag field address) |
 | string length | 2 bytes | NFA (name field address) |
 | string | variable length | still NFA |
+| null byte | 1 byte |
 | code | variable length | CFA (code field address) |
 
 probably, some bitmask antics could be done to store the string length
 and flags together. but alternatively: no.
 
+the stored string is a counted string but also has a
+null byte at the end. bit of a waste of space but i can use
+x86's `repz` for it. maybe there is a counted string version
+though, i don't know
+
 ## forth registers
 there are a set of 'virtual registers' (see moving forth part 1)