>From c43f1ee93ccfe84ddf4504e7d232f645416dc354 Mon Sep 17 00:00:00 2001
From: Szabolcs Nagy <nsz@port70.net>
Date: Sun, 20 Sep 2026 11:10:41 +0000
Subject: [PATCH 17/25] regex: struct changes for memory saving

the type changes should be safe: backref is in -1..9, assertions is
in 0..511, node type is 0..3, nullable is -1..1. Note: the node type
was an enum which could warn about missed switch cases.

tre_pos_and_tags_t and tre_tnfa_transition_t may be used O(n^2) times
and dominate memory use. the size saving is 20-30% on 64bit targets:
before -> now (-> with char tre_ctype_t -> packed assert/class)

  tre_ast_node_t: 48 -> 40  (-> 40 -> 40)
  tre_pos_and_tags_t: 56 -> 40  (-> 32 -> 24)
  tre_tnfa_transition_t: 56 -> 48  (-> 40 -> 32)

tre_ctype_t can be char instead of wctype_t (but not portable outside
musl). assertions+backref+class+neg_classes can be packed into 32bits.
however this patch avoids source/logic changes.
---
 src/regex/regcomp.c | 12 ++++++------
 src/regex/tre.h     |  8 ++++----
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/src/regex/regcomp.c b/src/regex/regcomp.c
index c3400c3d..c541e2fd 100644
--- a/src/regex/regcomp.c
+++ b/src/regex/regcomp.c
@@ -48,11 +48,11 @@ typedef struct {
   int position;
   int code_min;
   int code_max;
-  int *tags;
-  int assertions;
+  short assertions;
+  signed char backref;
   tre_ctype_t class;
   tre_ctype_t *neg_classes;
-  int backref;
+  int *tags;
 } tre_pos_and_tags_t;
 
 
@@ -84,10 +84,10 @@ typedef enum {
 /* A generic AST node.  All AST nodes consist of this node on the top
    level with `obj' pointing to the actual content. */
 typedef struct {
-  tre_ast_type_t type;   /* Type of the node. */
-  void *obj;             /* Pointer to actual node. */
-  int nullable;
+  char type;             /* Type of the node. */
+  signed char nullable;
   int submatch_id;
+  void *obj;             /* Pointer to actual node. */
   int num_submatches;
   int num_tags;
   tre_pos_and_tags_t *firstpos;
diff --git a/src/regex/tre.h b/src/regex/tre.h
index 55b17257..0ad77f1d 100644
--- a/src/regex/tre.h
+++ b/src/regex/tre.h
@@ -98,19 +98,19 @@ struct tnfa_transition {
   tre_tnfa_transition_t *state;
   /* ID number of the destination state. */
   int state_id;
-  /* -1 terminated array of tags (or NULL). */
-  int *tags;
   /* Assertion bitmap. */
-  int assertions;
+  short assertions;
   /* Assertion parameters. */
   union {
     /* Character class assertion. */
     tre_ctype_t class;
     /* Back reference assertion. */
-    int backref;
+    signed char backref;
   } u;
   /* Negative character class assertions. */
   tre_ctype_t *neg_classes;
+  /* -1 terminated array of tags (or NULL). */
+  int *tags;
 };
 
 
-- 
2.52.0