From a6ec3e0c3ee0dd36cc281e5c2e793f374f051716 Mon Sep 17 00:00:00 2001
From: Marc Bernard <59966492+mbtools@users.noreply.github.com>
Date: Tue, 22 Nov 2022 13:14:07 -0500
Subject: [PATCH] Refactor: BOM handling (#5878)

Co-authored-by: Lars Hvam <larshp@hotmail.com>
---
 src/repo/zcl_abapgit_dot_abapgit.clas.abap    | 24 ++++--------
 src/utils/zcl_abapgit_convert.clas.abap       | 24 +++++-------
 .../zcl_abapgit_convert.clas.testclasses.abap | 37 +++++++++++++++++++
 src/xml/zcl_abapgit_xml.clas.abap             | 13 +------
 4 files changed, 55 insertions(+), 43 deletions(-)

diff --git a/src/repo/zcl_abapgit_dot_abapgit.clas.abap b/src/repo/zcl_abapgit_dot_abapgit.clas.abap
index d87836ecb..490a495eb 100644
--- a/src/repo/zcl_abapgit_dot_abapgit.clas.abap
+++ b/src/repo/zcl_abapgit_dot_abapgit.clas.abap
@@ -221,6 +221,11 @@ CLASS zcl_abapgit_dot_abapgit IMPLEMENTATION.
   ENDMETHOD.
 
 
+  METHOD get_version_constant.
+    rv_version_constant = ms_data-version_constant.
+  ENDMETHOD.
+
+
   METHOD is_ignored.
 
     DATA: lv_name     TYPE string,
@@ -277,22 +282,10 @@ CLASS zcl_abapgit_dot_abapgit IMPLEMENTATION.
 
   METHOD serialize.
 
-    DATA: lv_xml  TYPE string,
-          lv_mark TYPE string.
+    DATA lv_xml TYPE string.
 
     lv_xml = to_xml( ms_data ).
 
-    "unicode systems always add the byte order mark to the xml, while non-unicode does not
-    "this code will always add the byte order mark if it is not in the xml
-    TRY.
-        lv_mark = zcl_abapgit_convert=>xstring_to_string_utf8( cl_abap_char_utilities=>byte_order_mark_utf8 ).
-      CATCH zcx_abapgit_exception ##NO_HANDLER.
-* In non-unicode systems, the byte order mark throws an error
-    ENDTRY.
-    IF lv_xml(1) <> lv_mark.
-      CONCATENATE lv_mark lv_xml INTO lv_xml.
-    ENDIF.
-
     rv_xstr = zcl_abapgit_convert=>string_to_xstring_utf8_bom( lv_xml ).
 
   ENDMETHOD.
@@ -317,14 +310,12 @@ CLASS zcl_abapgit_dot_abapgit IMPLEMENTATION.
     ms_data-starting_folder = iv_path.
   ENDMETHOD.
 
-  METHOD get_version_constant.
-    rv_version_constant = ms_data-version_constant.
-  ENDMETHOD.
 
   METHOD set_version_constant.
     ms_data-version_constant = iv_version_constant.
   ENDMETHOD.
 
+
   METHOD to_file.
     rs_file-path     = zif_abapgit_definitions=>c_root_dir.
     rs_file-filename = zif_abapgit_definitions=>c_dot_abapgit.
@@ -349,5 +340,4 @@ CLASS zcl_abapgit_dot_abapgit IMPLEMENTATION.
     ASSERT sy-subrc = 0.
 
   ENDMETHOD.
-
 ENDCLASS.
diff --git a/src/utils/zcl_abapgit_convert.clas.abap b/src/utils/zcl_abapgit_convert.clas.abap
index 8e0892f6c..1acdee17d 100644
--- a/src/utils/zcl_abapgit_convert.clas.abap
+++ b/src/utils/zcl_abapgit_convert.clas.abap
@@ -211,8 +211,10 @@ CLASS zcl_abapgit_convert IMPLEMENTATION.
         ENDIF.
 
         go_convert_out->convert(
-          EXPORTING data = iv_string
-          IMPORTING buffer = rv_xstring ).
+          EXPORTING
+            data   = iv_string
+          IMPORTING
+            buffer = rv_xstring ).
 
       CATCH cx_parameter_invalid_range
             cx_sy_codepage_converter_init
@@ -226,21 +228,15 @@ CLASS zcl_abapgit_convert IMPLEMENTATION.
 
   METHOD string_to_xstring_utf8_bom.
 
-    DATA: lv_hex     TYPE x LENGTH 1 VALUE '23',
-          lv_hex_bom TYPE x LENGTH 3 VALUE 'EFBBBF'.
+    IF iv_string IS INITIAL.
+      RETURN.
+    ENDIF.
 
     rv_xstring = string_to_xstring_utf8( iv_string ).
 
-    "unicode systems always add the byte order mark to the xml, while non-unicode does not
-    "in class ZCL_ABAPGIT_XML~TO_XML byte order mark was added to XML as #
-    "In non-unicode systems zcl_abapgit_convert=>xstring_to_string_utf8( cl_abap_char_utilities=>byte_order_mark_utf8 )
-    "has result # as HEX 23 and not HEX EFBBBF.
-    "So we have to remove 23 first and add EFBBBF after to serialized string
-    IF rv_xstring(3) <> cl_abap_char_utilities=>byte_order_mark_utf8
-    AND rv_xstring(1) = lv_hex.
-      REPLACE FIRST OCCURRENCE
-        OF lv_hex IN rv_xstring WITH lv_hex_bom IN BYTE MODE.
-      ASSERT sy-subrc = 0.
+    " Add UTF-8 BOM
+    IF xstrlen( rv_xstring ) < 3 OR rv_xstring(3) <> cl_abap_char_utilities=>byte_order_mark_utf8.
+      rv_xstring = cl_abap_char_utilities=>byte_order_mark_utf8 && rv_xstring.
     ENDIF.
 
   ENDMETHOD.
diff --git a/src/utils/zcl_abapgit_convert.clas.testclasses.abap b/src/utils/zcl_abapgit_convert.clas.testclasses.abap
index 2acebabd2..4d90af4b6 100644
--- a/src/utils/zcl_abapgit_convert.clas.testclasses.abap
+++ b/src/utils/zcl_abapgit_convert.clas.testclasses.abap
@@ -10,6 +10,7 @@ CLASS ltcl_convert DEFINITION FOR TESTING RISK LEVEL HARMLESS DURATION SHORT FIN
     METHODS split_string FOR TESTING.
     METHODS convert_bitbyte FOR TESTING RAISING zcx_abapgit_exception.
     METHODS string_to_xstring_utf8 FOR TESTING RAISING zcx_abapgit_exception.
+    METHODS string_to_xstring_utf8_bom FOR TESTING RAISING zcx_abapgit_exception.
     METHODS xstring_to_string_utf8 FOR TESTING RAISING zcx_abapgit_exception.
     METHODS xstring_to_string_not_utf8 FOR TESTING RAISING zcx_abapgit_exception.
     METHODS base64_to_xstring FOR TESTING.
@@ -113,6 +114,12 @@ CLASS ltcl_convert IMPLEMENTATION.
 
     DATA lv_result TYPE xstring.
 
+    lv_result = zcl_abapgit_convert=>string_to_xstring_utf8( `` ).
+
+    cl_abap_unit_assert=>assert_equals(
+      act = lv_result
+      exp = `` ).
+
     lv_result = zcl_abapgit_convert=>string_to_xstring_utf8( 'abc' ).
 
     cl_abap_unit_assert=>assert_equals(
@@ -121,10 +128,40 @@ CLASS ltcl_convert IMPLEMENTATION.
 
   ENDMETHOD.
 
+  METHOD string_to_xstring_utf8_bom.
+
+    DATA lv_result TYPE xstring.
+
+    lv_result = zcl_abapgit_convert=>string_to_xstring_utf8_bom( `` ).
+
+    cl_abap_unit_assert=>assert_equals(
+      act = lv_result
+      exp = `` ).
+
+    lv_result = zcl_abapgit_convert=>string_to_xstring_utf8_bom( 'a' ).
+
+    cl_abap_unit_assert=>assert_equals(
+      act = lv_result
+      exp = 'EFBBBF61' ).
+
+    lv_result = zcl_abapgit_convert=>string_to_xstring_utf8_bom( 'abcd' ).
+
+    cl_abap_unit_assert=>assert_equals(
+      act = lv_result
+      exp = 'EFBBBF61626364' ).
+
+  ENDMETHOD.
+
   METHOD xstring_to_string_utf8.
 
     DATA lv_result TYPE string.
 
+    lv_result = zcl_abapgit_convert=>xstring_to_string_utf8( `` ).
+
+    cl_abap_unit_assert=>assert_equals(
+      act = lv_result
+      exp = `` ).
+
     lv_result = zcl_abapgit_convert=>xstring_to_string_utf8( '616263' ).
 
     cl_abap_unit_assert=>assert_equals(
diff --git a/src/xml/zcl_abapgit_xml.clas.abap b/src/xml/zcl_abapgit_xml.clas.abap
index aa70e5fec..30111cef6 100644
--- a/src/xml/zcl_abapgit_xml.clas.abap
+++ b/src/xml/zcl_abapgit_xml.clas.abap
@@ -148,10 +148,8 @@ CLASS zcl_abapgit_xml IMPLEMENTATION.
 
     DATA: li_ostream       TYPE REF TO if_ixml_ostream,
           li_renderer      TYPE REF TO if_ixml_renderer,
-          lv_mark          TYPE string,
           li_streamfactory TYPE REF TO if_ixml_stream_factory.
 
-
     li_streamfactory = mi_ixml->create_stream_factory( ).
 
     li_ostream = li_streamfactory->create_ostream_cstring( rv_xml ).
@@ -162,16 +160,7 @@ CLASS zcl_abapgit_xml IMPLEMENTATION.
 
     li_renderer->render( ).
 
-    "unicode systems always add the byte order mark to the xml, while non-unicode does not
-    "this code will always add the byte order mark if it is not in the xml
-    TRY.
-        lv_mark = zcl_abapgit_convert=>xstring_to_string_utf8( cl_abap_char_utilities=>byte_order_mark_utf8 ).
-      CATCH zcx_abapgit_exception ##NO_HANDLER.
-        ASSERT 0 = 1.
-    ENDTRY.
-    IF rv_xml(1) <> lv_mark.
-      CONCATENATE lv_mark rv_xml INTO rv_xml.
-    ENDIF.
+    " handling of BOM moved to zcl_abapgit_convert=>string_to_xstring_utf8_bom
 
   ENDMETHOD.
 ENDCLASS.