From 4e206589d7223dbffc6ae95e1eef50ada4caa2e2 Mon Sep 17 00:00:00 2001
From: Evan Miller <emmiller@gmail.com>
Date: Wed, 24 Dec 2025 12:38:58 -0500
Subject: [PATCH] Convert DTA notes to UTF-8

---
 src/stata/readstat_dta_read.c | 19 ++++++++++++++++++-
 src/test/test_list.h          | 15 +++++++++++++++
 2 files changed, 33 insertions(+), 1 deletion(-)

diff --git a/src/stata/readstat_dta_read.c b/src/stata/readstat_dta_read.c
index ce07c333..27d86d14 100644
--- a/src/stata/readstat_dta_read.c
+++ b/src/stata/readstat_dta_read.c
@@ -184,6 +184,7 @@ static readstat_error_t dta_read_expansion_fields(dta_ctx_t *ctx) {
     readstat_error_t retval = READSTAT_OK;
     readstat_io_t *io = ctx->io;
     char *buffer = NULL;
+    char *utf8_note = NULL;
 
     if (ctx->expansion_len_len == 0)
         return READSTAT_OK;
@@ -271,7 +272,21 @@ static readstat_error_t dta_read_expansion_fields(dta_ctx_t *ctx) {
             int index = 0;
             if (strncmp(&buffer[0], "_dta", 4) == 0 &&
                     sscanf(&buffer[ctx->ch_metadata_len], "note%d", &index) == 1) {
-                if (ctx->handle.note(index, &buffer[2*ctx->ch_metadata_len], ctx->user_ctx) != READSTAT_HANDLER_OK) {
+                const char *note_src = &buffer[2*ctx->ch_metadata_len];
+                size_t note_src_len = len - 2*ctx->ch_metadata_len;
+                size_t utf8_note_len = 4*note_src_len + 1;
+
+                if ((utf8_note = readstat_realloc(utf8_note, utf8_note_len)) == NULL) {
+                    retval = READSTAT_ERROR_MALLOC;
+                    goto cleanup;
+                }
+
+                retval = readstat_convert(utf8_note, utf8_note_len, note_src,
+                        strnlen(note_src, note_src_len), ctx->converter);
+                if (retval != READSTAT_OK)
+                    goto cleanup;
+
+                if (ctx->handle.note(index, utf8_note, ctx->user_ctx) != READSTAT_HANDLER_OK) {
                     retval = READSTAT_ERROR_USER_ABORT;
                     goto cleanup;
                 }
@@ -291,6 +306,8 @@ static readstat_error_t dta_read_expansion_fields(dta_ctx_t *ctx) {
 cleanup:
     if (buffer)
         free(buffer);
+    if (utf8_note)
+        free(utf8_note);
 
     return retval;
 }
diff --git a/src/test/test_list.h b/src/test/test_list.h
index 90f4f6d1..bd4cc169 100644
--- a/src/test/test_list.h
+++ b/src/test/test_list.h
@@ -795,6 +795,21 @@ static rt_test_group_t _test_groups[] = {
                         .label_set = "somelbl"
                     }
                 }
+            },
+            {
+                .label = "UTF-8 note",
+                .test_formats = RT_FORMAT_DTA_118_AND_NEWER,
+                .rows = 0,
+                .notes_count = 1,
+                .notes = {
+                    "Stra" "\xc3\x9f" "e"
+                },
+                .columns = {
+                    {
+                        .name = "var1",
+                        .type = READSTAT_TYPE_DOUBLE
+                    }
+                }
             }
         }
     },