boot2

commit ce22557fc8cb164acd92288a244669cc4651ec03
parent dca693c3073799b8d46cc9f6479839ea4c8d913d
Author: Ryan Sepassi <rsepassi@gmail.com>
Date:   Thu, 23 Apr 2026 17:04:40 -0700

m1pp: add strlen expression op

Diffstat:
M
m1pp/m1pp.c
 | 
28
+++++++++++++++++++++++++++-
M
tests/m1pp/04-expr-ops.M1pp
 | 
7
+++++++
M
tests/m1pp/04-expr-ops.expected
 | 
7
+++++++
A
tests/m1pp/_04-strlen-badarg.M1pp
 | 
7
+++++++

4 files changed, 48 insertions(+), 1 deletion(-)
diff --git a/m1pp/m1pp.c b/m1pp/m1pp.c
@@ -102,6 +102,7 @@ enum ExprOp {
     EXPR_LE,
     EXPR_GT,
     EXPR_GE,
+    EXPR_STRLEN,
     EXPR_INVALID
 };
 
@@ -804,6 +805,9 @@ static enum ExprOp expr_op_code(const struct Token *tok)
     if (token_text_eq(tok, ">=")) {
         return EXPR_GE;
     }
+    if (token_text_eq(tok, "strlen")) {
+        return EXPR_STRLEN;
+    }
     return EXPR_INVALID;
 }
 
@@ -932,6 +936,7 @@ static int apply_expr_op(enum ExprOp op, const long long *args, int argc, long l
         }
         *out = (args[0] >= args[1]);
         return 1;
+    case EXPR_STRLEN:
     case EXPR_INVALID:
         break;
     }
@@ -1021,13 +1026,34 @@ static int eval_expr_range(struct TokenSpan span, long long *out)
             if (op == EXPR_INVALID) {
                 return fail("bad expression");
             }
+            pos++;
+            if (op == EXPR_STRLEN) {
+                /* strlen is degenerate: argument is a TOK_STRING atom,
+                 * not a recursive expression. Handle inline and yield
+                 * the string's raw byte count (span.len - 2). */
+                skip_expr_newlines(&pos, span.end);
+                if (pos >= span.end || pos->kind != TOK_STRING) {
+                    return fail("bad expression");
+                }
+                if (pos->text.len < 2 || pos->text.ptr[0] != '"') {
+                    return fail("bad expression");
+                }
+                value = (long long)(pos->text.len - 2);
+                pos++;
+                skip_expr_newlines(&pos, span.end);
+                if (pos >= span.end || pos->kind != TOK_RPAREN) {
+                    return fail("bad expression");
+                }
+                pos++;
+                have_value = 1;
+                continue;
+            }
             if (frame_top >= MAX_EXPR_FRAMES) {
                 return fail("expression overflow");
             }
             frames[frame_top].op = op;
             frames[frame_top].argc = 0;
             frame_top++;
-            pos++;
             continue;
         }
 
diff --git a/tests/m1pp/04-expr-ops.M1pp b/tests/m1pp/04-expr-ops.M1pp
@@ -33,4 +33,11 @@ $((>= 5 6))
 
 # nested expressions
 $((+ (* 2 3) (- 7 4) (/ 12 3)))
+
+# strlen: raw byte count between the quotes (matches what M1's "..." emits
+# before appending NUL). Composes with arithmetic like any other op.
+%((strlen "hello"))
+%((+ (strlen "hello") 1))
+!((strlen "x"))
+%((strlen ""))
 END
diff --git a/tests/m1pp/04-expr-ops.expected b/tests/m1pp/04-expr-ops.expected
@@ -33,4 +33,11 @@
 
 
 '0D00000000000000'
+
+
+
+'05000000'
+'06000000'
+'01'
+'00000000'
 END
diff --git a/tests/m1pp/_04-strlen-badarg.M1pp b/tests/m1pp/_04-strlen-badarg.M1pp
@@ -0,0 +1,7 @@
+# Malformed: strlen requires a double-quoted TOK_STRING argument.
+# Single-quoted '...' hex literals are meaningless for strlen and must
+# be rejected. Expected behavior: non-zero exit from the expander.
+# (Underscore-prefixed filename so test.sh skips this fixture.)
+
+%((strlen 'deadbeef'))
+END