XML PARSE

DEPENDENCIES

@@ -116,7 +116,7 @@ The following libraries ARE required WHEN :
 
     BOTH runtime AND development components required.
 
-    libxml2 >= 2.5.10 - https://gitlab.gnome.org/GNOME/libxml2/-/wikis/home
+    libxml2 >= 2.6.1 - https://gitlab.gnome.org/GNOME/libxml2/-/wikis/home
 
     libxml2 is distributed under MIT License (Expat variant).
 

DEPENDENCIES.md

@@ -102,7 +102,7 @@ XML Support
 
 Support for GENERATE XML is provided by:
 
-*   [libxml2](https://gitlab.gnome.org/GNOME/libxml2/-/wikis/home) 2.5.10 or later
+*   [libxml2](https://gitlab.gnome.org/GNOME/libxml2/-/wikis/home) 2.6.1 or later
 
     libxml2 is distributed under MIT License (Expat variant).
 

NEWS

@@ -20,11 +20,29 @@ NEWS - user visible changes				-*- outline -*-
    customization can be done using COB_PROF_FILE, COB_PROF_MAX_DEPTH and
    COB_PROF_FORMAT
 
-** initial support for XML PARSE
-   while this feature is not implemented fully and some events are not
-   implemented yet, we'd like you to test this feature already;
-   note that for the XML-EVENT EXCEPTION, the stored exception-values in
-   XML-CODE and XML-TEXT registers differ from other implementations
+** support for XML PARSE
+   Events sent to COBOL code try to follow as closely as possible what IBM
+   does with a few differences linked to internal XML engine differences:
+   - we cannot guarantee that incomplete CONTENT-CHARACTERS and 
+     ATTRIBUTE-CHARACTERS XML-EVENTS will be sent at the same time, 
+     and cut in the same place as with IBM's parser.
+     GnuCOBOL can merge short CONTENT-CHARACTERS events together across
+     END-OF-INPUT chunks boundary.
+   - For the EXCEPTION XML-EVENT, the reported XML-CODE will often only 
+     correspond to generic error or warning instead of the full error code.
+     In COMPAT mode, generic warnings use code 99 and generic errors use 201.
+     In XMLNSS mode, the low order bits for generic errors and warnings are 
+     set to zero.
+     Moreover, for those generic exceptions, XML-TEXT will contain the error
+     text as reported by libxml2.
+     We may map more of libxml2's errors on IBM's error codes in the future.
+   - The order between EXCEPTION events and other events is not guaranteed 
+     to be the same as with an IBM parser.
+   For most codebases, we do not expect those differences to break existing
+   COBOL code.
+   Support for XML PARSE RETURNING NATIONAL and XML PARSE VALIDATING 
+   is experimental, remain untested and will show a warning at compile time 
+   when used.
 
 ** new runtime configuration COB_HIDE_CURSOR, allows to hide the cursor during
    extended ScreenIO operations

cobc/ChangeLog

@@ -1,4 +1,13 @@
 
+2026-04-14  Guillaume Bertholon <guillaume.bertholon@ocamlpro.com>
+
+	* parser.y: remove the CB_PENDING warning on XML PARSE but still warn for 
+	  untested XML PARSE RETURNING NATIONAL and XML PARSE VALIDATING.
+	* typeck.c: remove invalid call to cob_check_based for XML-* builtin variable 
+	  length registers (like XML-TEXT)
+	* codegen.c: remove the uninitialized and unused b_* field for XML-* builtin 
+	  variable length registers
+
 2026-06-08  Nicolas Berthier <nicolas.berthier@ocamlpro.com>
 
 	* tree.h, parser.y: change type of cobc_cs_check flags to permit

cobc/codegen.c

@@ -1091,6 +1091,8 @@ output_base (struct cb_field *f, const cob_u32_t no_output)
 		} else {
 			output ("cob_local_ptr");
 		}
+	} else if (f01->storage == CB_STORAGE_LINKAGE && f01->flag_internal_register) {
+		output ("%s%d.data", CB_PREFIX_FIELD, f01->id);	
 	} else {
 		output ("%s%d", CB_PREFIX_BASE, f01->id);
 	}
@@ -7274,7 +7276,7 @@ output_xml_parse (struct cb_xml_parse *p)
 {
 	int flags = 0;
 	if (cb_xml_parse_xmlss) {
-		flags |= COB_XML_PARSE_XMLNSS;
+		flags |= COB_XML_PARSE_XMLSS;
 	}
 	if (p->returning_national && current_prog->xml_ntext) {
 		flags |= COB_XML_PARSE_NATIONAL;
@@ -11905,7 +11907,9 @@ output_internal_function (struct cb_program *prog, cb_tree parameter_list)
 	/* Dangling linkage section items */
 	seen = 0;
 	for (f = prog->linkage_storage; f; f = f->sister) {
-		if (f->redefines) {
+		if (f->redefines || f->flag_internal_register) {
+			/* XML-TEXT and other XML-* registers are in linkage_storage but do not use the 
+			   corresponding b_* field. */
 			continue;
 		}
 		for (l = parameter_list; l; l = CB_CHAIN (l)) {

cobc/parser.y

@@ -18130,7 +18130,6 @@ xml_parse_statement:
   xml PARSE
   {
 	begin_statement (STMT_XML_PARSE, TERM_XML);
-	CB_PENDING ("XML PARSE");
 	__CS_ENTER (CB_CS_XML_PARSE);
 	cb_set_register_receiving (current_program->xml_code, 1);
 	cb_set_register_receiving (current_program->xml_event, 1);
@@ -18141,8 +18140,6 @@ xml_parse_statement:
 		cb_set_register_receiving (current_program->xml_namespace_prefix, 1);
 		cb_set_register_receiving (current_program->xml_nnamespace, 1);
 		cb_set_register_receiving (current_program->xml_nnamespace_prefix, 1);
-	}
-	if (cb_xml_parse_xmlss) {
 		cb_set_register_receiving (current_program->xml_information, 0);
 	}
   }
@@ -18182,14 +18179,28 @@ _with_encoding:
 ;
 
 _returning_national:
-/* empty */         		{ $$ = NULL; }
-| RETURNING NATIONAL		{ $$ = cb_true; }
+/* empty */
+  { 
+	$$ = NULL; 
+  }
+| RETURNING NATIONAL
+  { 
+	CB_PENDING ("XML PARSE RETURNING NATIONAL");
+	$$ = cb_true;
+  }
 ;
 
 _validating_with:
-/* empty */                 	{ $$ = NULL; }
+/* empty */                 	
+  { 
+	$$ = NULL; 
+  }
 | VALIDATING _with
-  schema_file_or_record_name	{ $$ = $3; }
+  schema_file_or_record_name	
+  { 
+	CB_PENDING ("XML PARSE VALIDATING");
+	$$ = $3;
+  }
 ;
 
 schema_file_or_record_name:

cobc/typeck.c

@@ -2544,7 +2544,7 @@ cb_build_identifier (cb_tree x, const int subchk)
 		 && !current_statement->flag_no_based) {
 			if (p->flag_item_based
 			 || (p->storage == CB_STORAGE_LINKAGE &&
-				!(p->flag_is_pdiv_parm || p->flag_is_returning))) {
+				!(p->flag_is_pdiv_parm || p->flag_is_returning || p->flag_internal_register))) {
 				current_statement->null_check = CB_BUILD_FUNCALL_2 (
 					"cob_check_based",
 					cb_build_address (cb_build_field_reference (p, NULL)),

config/mf.words

@@ -678,15 +678,7 @@ reserved:	WRITE
 reserved:	WRITING
 reserved:	XML
 reserved:	XML-DECLARATION*
-reserved:	XML-EVENT             # note: this is a register, move as soon as supported
-#reserved:	XML-INFORMATION       # note: this is a register, move as soon as supported
-#reserved:	XML-NAMESPACE         # note: this is a register, move as soon as supported
-#reserved:	XML-NAMESPACE-PREFIX  # note: this is a register, move as soon as supported
-#reserved:	XML-NNAMESPACE        # note: this is a register, move as soon as supported
-#reserved:	XML-NNAMESPACE-PREFIX # note: this is a register, move as soon as supported
-reserved:	XML-NTEXT             # note: this is a register, move as soon as supported
 reserved:	XML-SCHEMA
-reserved:	XML-TEXT              # note: this is a register, move as soon as supported
 reserved:	YYYYDDD*
 reserved:	YYYYMMDD*
 reserved:	ZERO
@@ -721,14 +713,14 @@ register:	SORT-RETURN
 # register:	TIME-OF-DAY		# only available in OS/VS mode
 # register:	WHEN-COMPILED		# only available in OS/VS or VSC2 modes (note the format of the date differs with the mode!)
 register:	XML-CODE
-# register:	XML-EVENT
+register:	XML-EVENT
 # register:	XML-INFORMATION
 # register:	XML-NAMESPACE
 # register:	XML-NAMESPACE-PREFIX
 # register:	XML-NNAMESPACE
 # register:	XML-NNAMESPACE-PREFIX
-# register:	XML-NTEXT
-# register:	XML-TEXT
+register:	XML-NTEXT
+register:	XML-TEXT
 
 # disable all functions
 not-intrinsic-function:	DIALECT-ALL

configure.ac

@@ -1132,7 +1132,7 @@ AS_IF([test "$with_xml2" = yes -o "$with_xml2" = check], [
 	LIBS="$LIBS $LIBCOB_LIBS_extern $XML2_LIBS"
 	# note: PKG_CONFIG and xml2-config set -I/path/to/libxml2 which contains a "libxml" folder where
 	#       all the files we look for are included
-	for header in xmlwriter xmlversion uri parser tree; do
+	for header in xmlwriter xmlversion uri parser tree SAX2; do
 		AC_CHECK_HEADER([libxml/$header.h], [],
 			[if test "$with_xml2" = yes; then
 			   AC_MSG_ERROR([headers for libxml2 are required for --with-xml2, you may adjust XML2_CFLAGS])

libcob/ChangeLog

@@ -1,4 +1,24 @@
 
+2026-04-14  Guillaume Bertholon <guillaume.bertholon@ocamlpro.com>
+
+	* common.h: rename COB_XML_PARSE_XMLNSS into COB_XML_PARSE_XMLSS to match 
+	  the IBM option name
+	* mlio.c: Fix issues in XML PARSE handling most notably a use 
+	  after free error if the internal buffer needs to grow during the parsing. 
+	  This fix changes the definition of xml_event to store offsets in the 
+	  buffer instead of pointers
+	* mlio.c: Respect the high order half-word for exception XML-CODE, 
+	  but do not expose internal libxml2 error codes 
+	* mlio.c: Reduce the number of parsing states by removing useless ones, 
+	  and encode eof in these states
+	* mlio.c: Handle XML chunks with more than one recoverable error
+	* mlio.c: Trigger ON EXCEPTION code after EXCEPTION XML events
+	* mlio.c: Remove spurious events when there is no <?xml?> declaration in the file
+	* mlio.c: Handle incomplete CONTENT-CHARACTERS events (with XML-INFORMATION = 2).
+	  Note that, compared to IBM, we may merge short contiguous CONTENT-CHARACTERS 
+	  events across END-OF-INPUT boundaries. 
+	  This is due to libxml2 internal details.
+
 2026-03-02  Fabrice Le Fessant <fabrice.le_fessant@ocamlpro.com>
 
 	* coblocal.h, common.c, profiling.c: rename is_test to cob_is_test
@@ -13,6 +33,11 @@
 	* fileio.c (cob_file_close): close file depending on internal state, not
 	  depending on file organization
 
+2025-11-19  Chuck Haatvedt <chuck.haatvedt+cobol@gmail.com>
+
+	* mlio.c (xml_startDocument, myStructuredErrorHandler, xml_parse):
+	  compat for LIBXML_VERSION < 21400
+
 2025-11-19  Oğuzcan Kırmemiş <oguzcan.kirmemis@gmail.com>
 
 	* common.c (cob_set_signal): add the enum COB_SIGNAL_REGIME to toggle
@@ -87,6 +112,22 @@
 	* intrinsic.c (cob_intr_char): raise COB_EC_ARGUMENT_FUNCTION
 	  when calling CHAR with an argument outside the collation range
 
+2025-08-15  Chuck Haatvedt <chuck.haatvedt+cobol@gmail.com>
+
+	* mlio.c: modified to support xml parse with xmlss.
+	  eliminated the xml_event_data structure and moved that data
+	  into the xml_event structure. Created a new enum cob_xml_registers
+	  and added it to the add_xml_event_data function. This function was
+	  modified to update the xml_event structure. All of the context parser
+	  callback functions were modified to use the add_xml_event_data function.
+	  the cob_xml_parse and xml_parse functions were modified to support
+	  the new end_of_input event required by xmlss. a new eof variable
+	  was added to the xml_state structure so that the endDocument callback
+	  function could be triggered by the parser in the xml_parse funtction.
+
+	  TODO ==> logic needs to be added to support returning NATIONAL data
+	  this needs to support the RETURNING NATIONAL phrase.
+
 2025-07-28  Simon Sobisch <simonsobisch@gnu.org>
 
 	* common.h, fileio.c: new externalized typedef EXTFH_FUNC used in

libcob/common.h

@@ -1116,7 +1116,7 @@ enum cob_statement {
 #define COB_JSON_CJSON			1
 #define COB_JSON_JSON_C			2
 
-#define COB_XML_PARSE_XMLNSS	(1U << 0)
+#define COB_XML_PARSE_XMLSS		(1U << 0)
 #define COB_XML_PARSE_NATIONAL	(1U << 1)
 #define COB_XML_PARSE_VALIDATE_FILE	(1U << 2)
 
@@ -1327,11 +1327,11 @@ typedef struct __cob_module {
 	const char	*gc_version;	/* module version, until 3.1.2: set by cob_check_version */
 
 	unsigned char		xml_mode;		/* Mode to handle XML PARSE (may be extended) */
-		/* similar to XMLPARSE(XMLNSS) Micro Focus,
+		/* similar to XMLPARSE(XMLSS) Micro Focus,
 		   IBM may be different (_very_ likely for error codes);
 		   but the main difference is to "COMPAT" */
 		#define COB_XML_COMPAT 		0
-		#define COB_XML_XMLNSS		1
+		#define COB_XML_XMLSS		1
 	struct cob_frame_ext *frame_ptr;	/* current frame ptr, note: if set then cob_frame in this
 										   module is of type "struct cob_frame_ext",
 										   otherwise "struct cob_frame" */