From 44ee13246e6326eafe378a9b061f990fa392d901 Mon Sep 17 00:00:00 2001
From: janis <janis@nirgendwo.xyz>
Date: Tue, 28 Oct 2025 17:07:05 +0100
Subject: [PATCH] is_number

---
 lang/src/tokeniser.asm | 48 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)

diff --git a/lang/src/tokeniser.asm b/lang/src/tokeniser.asm
index 4341603..46fb908 100644
--- a/lang/src/tokeniser.asm
+++ b/lang/src/tokeniser.asm
@@ -365,8 +365,56 @@ is_ident:
     pop rbp
     ret
 
+;; Numbers are sequences of numeric characters, interspersed with underscores
+;; The leading character must be numeric
+;; In the future, numbers may be prefixed with '0x' for hexadecimal or '0b' for binary.
 is_number:
+    push rbp
+    mov rbp, rsp
+    push r12
+    push r13
+    push r14
+
+    mov rax, [rel cursor]
+    mov r12, [rel buffer]
+    mov r13, [rel buffer_len]
+    add r12, rax
+    sub r13, rax
+
+    mov dil, [r12]
+    call is_numeric
+    test rax, rax
+    je .not_number
+
+    mov r14, 1
+.loop:
+    cmp r14, r13
+    jge .number
+    mov dil, [r12 + r14]
+    call is_whitespace
+    test rax, rax
+    je .number
+    cmp dil, '_'
+    je .loop_next
+    call is_numeric
+    test rax, rax
+    je .not_number
+.loop_next:
+    inc r14
+    jmp .loop
+.number:
+    mov rax, [rel cursor]
+    add rax, r14
+    mov [rel cursor], rax
+    mov rax, r14
+    jmp .epilogue
+.not_number:
     xor rax, rax
+.epilogue:
+    pop r14
+    pop r13
+    pop r12
+    pop rbp
     ret
 
 skip_whitespaces: