COBOL: Import new upstream candidate parser

This fixes support for COBOL symbols after the recent breakage of regex
parsers, as well as introducing additional features and bug fixes.

Also import some of the tests.

https://github.com/universal-ctags/ctags/pull/2076

Part of #2119.

ctags/parsers/cobol.c

@@ -8,42 +8,516 @@
 *   files.
 */
 
+/* Some references:
+ * - https://www.cs.vu.nl/grammarware/browsable/cobol/
+ * - https://www.cs.vu.nl/grammarware/browsable/vs-cobol-ii/
+ * - https://open-cobol.sourceforge.io/guides/grammar.pdf
+ * - http://mapage.noos.fr/~bpinon/a_cobol_parser.htm
+ * - https://en.wikipedia.org/wiki/COBOL
+ */
+
 /*
 *   INCLUDE FILES
 */
 #include "general.h"	/* must always come first */
+#include "debug.h"
+#include "entry.h"
+#include "keyword.h"
+#include "nestlevel.h"
 #include "parse.h"
+#include "read.h"
 #include "routines.h"
 
-static tagRegexTable cobolTagRegexTable[] = {
-	{"^[ \t]*[0-9]+[ \t]+([A-Z0-9][A-Z0-9-]*)[ \t]+("
-	 "BLANK|OCCURS|IS|JUST|PIC|REDEFINES|RENAMES|SIGN|SYNC|USAGE|VALUE"
-	 ")", "\\1",
-	 "d,data,data items", "i", NULL},
-	{"^[ \t]*[FSR]D[ \t]+([A-Z0-9][A-Z0-9-]*)\\.", "\\1",
-	 "f,file,file descriptions (FD, SD, RD)", "i", NULL},
-	{"^[ \t]*[0-9]+[ \t]+([A-Z0-9][A-Z0-9-]*)\\.", "\\1",
-	 "g,group,group items", "i", NULL},
-	{"^[ \t]*([A-Z0-9][A-Z0-9-]*)\\.", "\\1",
-	 "p,paragraph,paragraphs", "i", NULL},
-	{"^[ \t]*PROGRAM-ID\\.[ \t]+([A-Z0-9][A-Z0-9-]*)\\.", "\\1",
-	 "P,program,program ids", "i", NULL},
-	{"^[ \t]*([A-Z0-9][A-Z0-9-]*)[ \t]+SECTION\\.", "\\1",
-	 "s,section,sections", "i", NULL},
+typedef enum {
+	K_FILE,
+	K_GROUP,
+	K_PROGRAM,
+	K_SECTION,
+	K_DIVISION,
+	K_PARAGRAPH,
+	K_DATA,
+	K_SOURCEFILE,
+} cobolKind;
+
+typedef enum {
+	COBOL_SOURCEFILE_COPIED,
+} cobolSourcefileRole;
+
+static roleDefinition CobolSourcefileRoles [] = {
+	{ true, "copied", "copied in source file" },
 };
 
-/*
-*   FUNCTION DEFINITIONS
-*/
+static kindDefinition CobolKinds[] = {
+	{ true, 'f', "fd", "file descriptions (FD, SD, RD)" },
+	{ true, 'g', "group", "group items" },
+	{ true, 'P', "program", "program ids" },
+	{ true, 's', "section", "sections" },
+	{ true, 'D', "division", "divisions" },
+	{ true, 'p', "paragraph", "paragraphs" },
+	{ true, 'd', "data", "data items"      },
+	{ true, 'S', "sourcefile", "source code file",
+	  .referenceOnly = true, ATTACH_ROLES(CobolSourcefileRoles)},
+};
+
+static langType Lang_cobol;
+
+enum {
+	KEYWORD_FD,
+	KEYWORD_SD,
+	KEYWORD_RD,
+	KEYWORD_SECTION,
+	KEYWORD_DIVISION,
+	KEYWORD_CONTINUE,
+	KEYWORD_END_EXEC,
+	KEYWORD_FILLER,
+	KEYWORD_BLANK,
+	KEYWORD_OCCURS,
+	KEYWORD_IS,
+	KEYWORD_JUST,
+	KEYWORD_PIC,
+	KEYWORD_REDEFINES,
+	KEYWORD_RENAMES,
+	KEYWORD_SIGN,
+	KEYWORD_SYNC,
+	KEYWORD_USAGE,
+	KEYWORD_VALUE,
+	KEYWORD_PROGRAM_ID,
+	KEYWORD_EXIT,
+	KEYWORD_COPY,
+};
+
+static const keywordTable cobolKeywordTable[] = {
+#define DEFINE_KEYWORD(n) { #n, KEYWORD_##n }
+	DEFINE_KEYWORD (FD),
+	DEFINE_KEYWORD (SD),
+	DEFINE_KEYWORD (RD),
+	DEFINE_KEYWORD (SECTION),
+	DEFINE_KEYWORD (DIVISION),
+	DEFINE_KEYWORD (CONTINUE),
+	{ "END-EXEC", KEYWORD_END_EXEC },
+	DEFINE_KEYWORD (EXIT),
+	DEFINE_KEYWORD (FILLER),
+	DEFINE_KEYWORD (BLANK),
+	DEFINE_KEYWORD (OCCURS),
+	DEFINE_KEYWORD (IS),
+	DEFINE_KEYWORD (JUST),
+	DEFINE_KEYWORD (PIC),
+	{ "PICTURE", KEYWORD_PIC },
+	DEFINE_KEYWORD (REDEFINES),
+	DEFINE_KEYWORD (RENAMES),
+	DEFINE_KEYWORD (SIGN),
+	DEFINE_KEYWORD (SYNC),
+	DEFINE_KEYWORD (USAGE),
+	DEFINE_KEYWORD (VALUE),
+	{ "VALUES", KEYWORD_VALUE },
+	{ "PROGRAM-ID", KEYWORD_PROGRAM_ID },
+	DEFINE_KEYWORD (COPY),
+};
+
+#define INDICATOR_COLUMN 7
+#define PROGRAM_NAME_AREA_COLUMN 73
+
+#define isIdentifierChar(c) (isalnum(c) || (c) == '-')
+#define isQuote(c) ((c) == '\'' || (c) == '"')
+
+typedef enum {
+	/* Fixed: program starts at column 8, ends at column 72 */
+	FORMAT_FIXED	= 0x1,
+	/* Free: program starts at column 1, no specific end */
+	FORMAT_FREE		= 0x2,
+	/* Variable: program starts at column 8, no specific end */
+	FORMAT_VARIABLE	= FORMAT_FIXED | FORMAT_FREE
+} CobolFormat;
+
+static struct {
+	vString *line;
+	unsigned long int lineNumber;
+	MIOPos filePosition;
+	const char *nextLine;
+	CobolFormat format;
+} CblInputState;
+
+static void cblppInit (const CobolFormat format)
+{
+	CblInputState.line = vStringNew ();
+	CblInputState.lineNumber = 0;
+	CblInputState.nextLine = NULL;
+	CblInputState.format = format;
+}
+
+static void cblppDeinit (void)
+{
+	vStringDelete (CblInputState.line);
+}
+
+static const char *cblppGetColumn (const char *line,
+								   const unsigned int column)
+{
+	unsigned int col = 0;
+
+	for (; *line; line++)
+	{
+		col += (*line == '\t') ? 8 : 1;
+		if (col >= column)
+			return line;
+	}
+
+	return NULL;
+}
+
+static void cblppAppendLine (vString *buffer,
+							 const char *line)
+{
+	if (CblInputState.format & FORMAT_FIXED)
+	{
+		const char *indicator = cblppGetColumn (line, INDICATOR_COLUMN);
+
+		if (indicator && *indicator && *indicator != '*' && *indicator != '/')
+		{
+			const char *lineStart = indicator + 1;
+			const char *lineEnd = cblppGetColumn (line, PROGRAM_NAME_AREA_COLUMN);
+
+			if (*indicator == '-')
+			{
+				vStringStripTrailing (buffer);
+				while (isspace (*lineStart))
+					lineStart++;
+			}
+
+			if (CblInputState.format == FORMAT_FIXED)
+				vStringNCatS (buffer, lineStart, lineEnd - lineStart);
+			else
+				vStringCatS (buffer, lineStart);
+		}
+	}
+	else if (line[0] != '*' && line[0] != '/')
+		vStringCatS (buffer, line);
+}
+
+/* TODO: skip *> comments */
+static const char *cblppGetLine (void)
+{
+	const char *line;
+
+	if (CblInputState.nextLine)
+	{
+		line = CblInputState.nextLine;
+		CblInputState.nextLine = NULL;
+	}
+	else
+		line = (const char *) readLineFromInputFile ();
+
+	CblInputState.lineNumber = getInputLineNumber ();
+	CblInputState.filePosition = getInputFilePosition ();
+
+	if (!line)
+		return NULL;
+
+	vStringClear (CblInputState.line);
+	cblppAppendLine (CblInputState.line, line);
+
+	/* check for continuation lines */
+	if (CblInputState.format & FORMAT_FIXED)
+	{
+		while (true)
+		{
+			const char *indicator;
+			line = (const char *) readLineFromInputFile ();
+			if (! line)
+				break;
+			indicator = cblppGetColumn (line, INDICATOR_COLUMN);
+			if (indicator && *indicator == '-')
+				cblppAppendLine (CblInputState.line, line);
+			else
+				break;
+		}
+
+		CblInputState.nextLine = line;
+	}
+
+	return vStringValue (CblInputState.line);
+}
+
+static void initCOBOLRefTagEntry (tagEntryInfo *e, const char *name,
+								  const cobolKind kind, const int role)
+{
+	initRefTagEntry (e, name, kind, role);
+	e->lineNumber = CblInputState.lineNumber;
+	e->filePosition = CblInputState.filePosition;
+}
+
+static void initCOBOLTagEntry (tagEntryInfo *e, const char *name, const cobolKind kind)
+{
+	initCOBOLRefTagEntry (e, name, kind, ROLE_INDEX_DEFINITION);
+}
+
+static int makeCOBOLRefTag (const char *name, const cobolKind kind, const int role)
+{
+	if (CobolKinds[kind].enabled)
+	{
+		tagEntryInfo e;
+
+		initCOBOLRefTagEntry (&e, name, kind, role);
+
+		return makeTagEntry (&e);
+	}
+
+	return CORK_NIL;
+}
+
+static int makeCOBOLTag (const char *name, const cobolKind kind)
+{
+	return makeCOBOLRefTag (name, kind, ROLE_INDEX_DEFINITION);
+}
+
+#define CBL_NL(nl) (*((unsigned int *) (nestingLevelGetUserData (nl))))
+
+static NestingLevel *popNestingLevelsToLevelNumber (NestingLevels *levels, const unsigned int levelNumber)
+{
+	NestingLevel *nl;
+
+	while (true)
+	{
+		nl = nestingLevelsGetCurrent (levels);
+		if (! nl || CBL_NL (nl) < levelNumber)
+			break;
+		nestingLevelsPop (levels);
+	}
+
+	return nl;
+}
+
+static bool isNumeric (const char *nptr, unsigned long int *num)
+{
+	char *endptr;
+	unsigned long int v;
+
+	v = strtoul (nptr, &endptr, 10);
+	if (nptr != endptr && *endptr == 0)
+	{
+		if (num)
+			*num = v;
+		return true;
+	}
+	return false;
+}
+
+static void findCOBOLTags (const CobolFormat format)
+{
+	NestingLevels *levels;
+	const char *line;
+
+	cblppInit (format);
+
+	levels = nestingLevelsNew (sizeof (unsigned int));
+
+	while ((line = cblppGetLine ()) != NULL)
+	{
+		char word[64];
+		int keyword;
+		unsigned long int levelNumber;
+
+#define READ_WHILE(word, cond) \
+	do { \
+		unsigned int i; \
+		for (i = 0; i < (ARRAY_SIZE (word) - 1) && *line && (cond); line++) \
+			word[i++] = *line; \
+		word[i] = 0; \
+	} while (0)
+#define READ_LITERAL(word) \
+	do { \
+		const char READ_LITERAL__q = isQuote (*line) ? *line++ : 0; \
+		READ_WHILE (word, (READ_LITERAL__q && READ_LITERAL__q != *line) || \
+		                   isIdentifierChar (*line)); \
+		if (READ_LITERAL__q && READ_LITERAL__q == *line) \
+			line++; \
+		keyword = lookupCaseKeyword (word, Lang_cobol); \
+	} while (0)
+#define READ_WORD(word, keyword) \
+	do { \
+		READ_WHILE (word, isIdentifierChar (*line)); \
+		keyword = lookupCaseKeyword (word, Lang_cobol); \
+	} while (0)
+#define READ_KEYWORD(keyword) \
+	do { \
+		char READ_KEYWORD__word[64]; \
+		READ_WORD (READ_KEYWORD__word, keyword); \
+	} while (0)
+#define SKIP_SPACES() do { while (isspace (*line)) line++; } while (0)
+
+		SKIP_SPACES ();
+		READ_WORD (word, keyword);
+		SKIP_SPACES ();
+
+		switch (keyword)
+		{
+		case KEYWORD_FD:
+		case KEYWORD_SD:
+		case KEYWORD_RD:
+			READ_WORD (word, keyword);
+			SKIP_SPACES ();
+			if (*word && *line == '.')
+				makeCOBOLTag (word, K_FILE);
+			break;
+
+		case KEYWORD_PROGRAM_ID:
+			if (*line == '.')
+			{
+				line++;
+				SKIP_SPACES ();
+			}
+			READ_LITERAL (word);
+			if (*word)
+				makeCOBOLTag (word, K_PROGRAM);
+			break;
+
+		case KEYWORD_COPY:
+			READ_WORD (word, keyword); // FIXME: also allow LITERAL
+			if (*word)
+				makeCOBOLRefTag (word, K_SOURCEFILE, COBOL_SOURCEFILE_COPIED);
+			break;
+
+		case KEYWORD_CONTINUE:
+		case KEYWORD_END_EXEC:
+		case KEYWORD_EXIT:
+		case KEYWORD_FILLER:
+			/* nothing, just ignore those in following cases */;
+			break;
+
+		default:
+			if (isNumeric (word, &levelNumber))
+			{
+				READ_WORD (word, keyword);
+				SKIP_SPACES ();
+
+				if (*word && keyword != KEYWORD_FILLER)
+				{
+					int kind = KIND_GHOST_INDEX;
+
+					if (*line == '.')
+						kind = K_GROUP;
+					else
+					{
+						int keyword2;
+
+						READ_KEYWORD (keyword2);
+						switch (keyword2)
+						{
+						case KEYWORD_BLANK:
+						case KEYWORD_OCCURS:
+						case KEYWORD_IS:
+						case KEYWORD_JUST:
+						case KEYWORD_PIC:
+						case KEYWORD_REDEFINES:
+						case KEYWORD_RENAMES:
+						case KEYWORD_SIGN:
+						case KEYWORD_SYNC:
+						case KEYWORD_USAGE:
+						case KEYWORD_VALUE:
+							kind = K_DATA;
+						}
+					}
+
+					if (kind != KIND_GHOST_INDEX)
+					{
+						NestingLevel *nl;
+						tagEntryInfo entry;
+						int r;
+						unsigned int nestingLevelNumber;
+
+						/* for nesting purposes, level 77 is identical to 1,
+						 * and 66 to 2 */
+						switch (levelNumber)
+						{
+						default: nestingLevelNumber = levelNumber; break;
+						case 77: nestingLevelNumber = 1; break;
+						case 66: nestingLevelNumber = 2; break;
+						}
+
+						nl = popNestingLevelsToLevelNumber (levels, nestingLevelNumber);
+						initCOBOLTagEntry (&entry, word, kind);
+						if (nl && CBL_NL (nl) < nestingLevelNumber)
+							entry.extensionFields.scopeIndex = nl->corkIndex;
+						r = makeTagEntry (&entry);
+						if (levelNumber < 50 /* exclude special levels */)
+						{
+							nl = nestingLevelsPush (levels, r);
+							CBL_NL (nl) = levelNumber;
+						}
+					}
+				}
+			}
+			else if (*word && *line == '.')
+				makeCOBOLTag (word, K_PARAGRAPH);
+			else
+			{
+				int keyword2;
+
+				READ_KEYWORD (keyword2);
+				SKIP_SPACES ();
+
+				if (keyword2 == KEYWORD_DIVISION && *line == '.')
+					makeCOBOLTag (word, K_DIVISION);
+				else if (keyword2 == KEYWORD_SECTION && *line == '.')
+					makeCOBOLTag (word, K_SECTION);
+			}
+		}
+	}
+
+	nestingLevelsFree (levels);
+	cblppDeinit ();
+}
+
+static void findCOBOLFixedTags (void)
+{
+	findCOBOLTags (FORMAT_FIXED);
+}
+
+static void findCOBOLFreeTags (void)
+{
+	findCOBOLTags (FORMAT_FREE);
+}
+
+static void findCOBOLVariableTags (void)
+{
+	findCOBOLTags (FORMAT_VARIABLE);
+}
+
+static void initializeCobolParser (langType language)
+{
+	Lang_cobol = language;
+}
+
+static parserDefinition* commonCobolParserDefinition (const char *name,
+													  simpleParser parser)
+{
+	parserDefinition* def = parserNew (name);
+	def->initialize = initializeCobolParser;
+	def->parser = parser;
+	def->kindTable = CobolKinds;
+	def->kindCount = ARRAY_SIZE(CobolKinds);
+	def->keywordTable = cobolKeywordTable;
+	def->keywordCount = ARRAY_SIZE(cobolKeywordTable);
+	def->useCork = true;
+	return def;
+}
 
 extern parserDefinition* CobolParser (void)
 {
 	static const char *const extensions [] = {
-			"cbl", "cob", "cpy", "CBL", "COB", NULL };
-	parserDefinition* def = parserNew ("Cobol");
+			"cbl", "cob", "CBL", "COB", NULL };
+	parserDefinition* def = commonCobolParserDefinition ("Cobol",
+														 findCOBOLFixedTags);
 	def->extensions = extensions;
-	def->tagRegexTable = cobolTagRegexTable;
-	def->tagRegexCount = ARRAY_SIZE (cobolTagRegexTable);
-	def->method     = METHOD_NOT_CRAFTED|METHOD_REGEX;
 	return def;
 }
+
+extern parserDefinition* CobolFreeParser (void)
+{
+	return commonCobolParserDefinition ("CobolFree", findCOBOLFreeTags);
+}
+
+extern parserDefinition* CobolVariableParser (void)
+{
+	return commonCobolParserDefinition ("CobolVariable", findCOBOLVariableTags);
+}

src/symbols.c

@@ -490,10 +490,12 @@ static void add_top_level_items(GeanyDocument *doc)
 			tag_list_add_groups(tag_store,
 				&tv_iters.tag_class, _("Program"), ICON_CLASS,
 				&tv_iters.tag_function, _("File"), ICON_METHOD,
+				&tv_iters.tag_interface, _("Divisions"), ICON_NAMESPACE,
 				&tv_iters.tag_namespace, _("Sections"), ICON_NAMESPACE,
 				&tv_iters.tag_macro, _("Paragraph"), ICON_OTHER,
 				&tv_iters.tag_struct, _("Group"), ICON_STRUCT,
 				&tv_iters.tag_variable, _("Data"), ICON_VAR,
+				&tv_iters.tag_externvar, _("Copies"), ICON_NAMESPACE,
 				NULL);
 			break;
 		case GEANY_FILETYPES_CONF:

src/tagmanager/tm_parser.c

@@ -423,11 +423,13 @@ static TMParserMapEntry map_R[] = {
 
 static TMParserMapEntry map_COBOL[] = {
 	{'d', tm_tag_variable_t},
+	{'D', tm_tag_interface_t},
 	{'f', tm_tag_function_t},
 	{'g', tm_tag_struct_t},
 	{'p', tm_tag_macro_t},
 	{'P', tm_tag_class_t},
 	{'s', tm_tag_namespace_t},
+	{'S', tm_tag_externvar_t},
 };
 
 static TMParserMapEntry map_OBJC[] = {
@@ -768,6 +770,7 @@ gboolean tm_parser_has_full_context(TMParserType lang)
 		case TM_PARSER_C:
 		case TM_PARSER_CPP:
 		case TM_PARSER_CSHARP:
+		case TM_PARSER_COBOL:
 		case TM_PARSER_D:
 		case TM_PARSER_FERITE:
 		case TM_PARSER_GLSL:

tests/ctags/Makefile.am

@@ -1,9 +1,6 @@
 
 dist_check_SCRIPTS = runner.sh
 
-# Disabled tests:
-#	simple.cbl
-
 NULL =
 test_sources = \
 	1795612.js						\
@@ -117,6 +114,11 @@ test_sources = \
 	case_sensitivity.php			\
 	char-selector.f90				\
 	classes.php						\
+	cobol/helloworld.cbl			\
+	cobol/helloworld2.cbl			\
+	cobol/levels.cbl				\
+	cobol/quoted-program-id.cbl		\
+	cobol/simple.cbl				\
 	common.f						\
 	complex-return.js				\
 	continuation.f90				\

tests/ctags/cobol/helloworld.cbl

@@ -0,0 +1,14 @@
+ 00000* VALIDATION OF BASE COBOL INSTALL                                00050000
+ 01000 IDENTIFICATION DIVISION.                                         00060000
+ 01100 PROGRAM-ID. 'HELLO'.                                             00070000
+ 02000 ENVIRONMENT DIVISION.                                            00080000
+ 02100 CONFIGURATION SECTION.                                           00090000
+ 02110 SOURCE-COMPUTER.  GNULINUX.                                      00100000
+ 02120 OBJECT-COMPUTER.  HERCULES.                                      00110000
+ 02200 SPECIAL-NAMES.                                                   00120000
+ 02210     CONSOLE IS CONSL.                                            00130000
+ 03000 DATA DIVISION.                                                   00140000
+ 04000 PROCEDURE DIVISION.                                              00150000
+ 04100 00-MAIN.                                                         00160000
+ 04110     DISPLAY 'HELLO, WORLD' UPON CONSL.                           00170000
+ 04900     STOP RUN.                                                    00180000

tests/ctags/cobol/helloworld.cbl.tags

@@ -0,0 +1,11 @@
+# format=tagmanager
+00-MAINÌ65536Ö0
+CONFIGURATIONÌ256Ö0
+DATAÌ32Ö0
+ENVIRONMENTÌ32Ö0
+HELLOÌ1Ö0
+IDENTIFICATIONÌ32Ö0
+OBJECT-COMPUTERÌ65536Ö0
+PROCEDUREÌ32Ö0
+SOURCE-COMPUTERÌ65536Ö0
+SPECIAL-NAMESÌ65536Ö0

tests/ctags/cobol/helloworld2.cbl

@@ -0,0 +1,19 @@
+ 00000* VALIDATION OF BASE COBOL INSTALL
+	IDENTIFICATION DIVISION.
+ 01100 PROGRAM-ID. 'HELLO WORLD'.
+ 02000 ENVIRONMENT DIV
+      -ISION.
+ 02100 CONFIGURATION SECTION.
+ 02110 SOURCE-COMPUTER.  GNULINUX.
+ 02120 OBJECT-COMPUTER.  HERCULES.
+ 02200 SPECIAL
+      --NAMES.
+ 02210     CONSOLE IS CONSL.
+ 03000 DATA DIVISION.
+ 04000          PROCEDUR
+      -   E DIVISION.
+ 04100 00-MAIN.
+ 04110     DISPLAY
+           'HELLO, WORLD'
+           UPON CONSL.
+ 04900     STOP RUN.

tests/ctags/cobol/helloworld2.cbl.tags

@@ -0,0 +1,11 @@
+# format=tagmanager
+00-MAINÌ65536Ö0
+CONFIGURATIONÌ256Ö0
+DATAÌ32Ö0
+ENVIRONMENTÌ32Ö0
+HELLO WORLDÌ1Ö0
+IDENTIFICATIONÌ32Ö0
+OBJECT-COMPUTERÌ65536Ö0
+PROCEDUREÌ32Ö0
+SOURCE-COMPUTERÌ65536Ö0
+SPECIAL-NAMESÌ65536Ö0

tests/ctags/cobol/levels.cbl

@@ -0,0 +1,27 @@
+       IDENTIFICATION DIVISION.
+       PROGRAM-ID. Test-Items.
+
+       DATA DIVISION.
+       WORKING-STORAGE SECTION.
+       01 RECORD1.
+              05 ITEM1 PIC X(1).
+              05 ITEM2 PIC A(1).
+                     88 ODD VALUES 1, 3, 5, 7, 9.
+                     88 EVEN VALUES 2, 4, 6, 8.
+              05 ITEM3 PIC X(6).
+              66 RDITEM4 RENAMES ITEM1 THRU ITEM2.
+              66 RDITEM5 RENAMES ITEM1 THROUGH ITEM3.
+       77 STDLN1 PIC A(4).
+       01 REC2.
+              02 G1.
+                     05 ITEM1 PICTURE X(10).
+                     05 ITEM2 PIC X(10).
+              66 OTHERNAME1 RENAMES ITEM1 IN REC2.
+              66 OTHERNAME2 RENAMES G1.
+
+       PROCEDURE DIVISION.
+       MAIN-PROCEDURE.
+            DISPLAY "Hello world".
+            DISPLAY OTHERNAME2.
+            STOP RUN.
+       END PROGRAM Test-Items.

tests/ctags/cobol/levels.cbl.tags

@@ -0,0 +1,22 @@
+# format=tagmanager
+DATAÌ32Ö0
+EVENÌ16384ÎRECORD1.ITEM2Ö0
+G1Ì2048ÎREC2Ö0
+IDENTIFICATIONÌ32Ö0
+ITEM1Ì16384ÎREC2.G1Ö0
+ITEM1Ì16384ÎRECORD1Ö0
+ITEM2Ì16384ÎREC2.G1Ö0
+ITEM2Ì16384ÎRECORD1Ö0
+ITEM3Ì16384ÎRECORD1Ö0
+MAIN-PROCEDUREÌ65536Ö0
+ODDÌ16384ÎRECORD1.ITEM2Ö0
+OTHERNAME1Ì16384ÎREC2Ö0
+OTHERNAME2Ì16384ÎREC2Ö0
+PROCEDUREÌ32Ö0
+RDITEM4Ì16384ÎRECORD1Ö0
+RDITEM5Ì16384ÎRECORD1Ö0
+REC2Ì2048Ö0
+RECORD1Ì2048Ö0
+STDLN1Ì16384Ö0
+Test-ItemsÌ1Ö0
+WORKING-STORAGEÌ256Ö0

tests/ctags/cobol/quoted-program-id.cbl

@@ -0,0 +1,7 @@
+       IDENTIFICATION DIVISION.
+       PROGRAM-ID. 'Test program name'.
+       PROCEDURE DIVISION.
+       MAIN.
+           DISPLAY 'This is a test program'.
+       STOP RUN.
+       END PROGRAM 'Test program name'.

tests/ctags/cobol/quoted-program-id.cbl.tags

@@ -0,0 +1,5 @@
+# format=tagmanager
+IDENTIFICATIONÌ32Ö0
+MAINÌ65536Ö0
+PROCEDUREÌ32Ö0
+Test program nameÌ1Ö0

tests/ctags/cobol/simple.cbl.tags

@@ -0,0 +1,22 @@
+# format=tagmanager
+AUTHORÌ65536Ö0
+BeginÌ65536Ö0
+DATAÌ32Ö0
+DW-DAYS-IN-MONTHÌ16384ÎGroup-Name.Data-Item1.DW-DAYS-IN-MONTHSÖ0
+DW-DAYS-IN-MONTHSÌ16384ÎGroup-Name.Data-Item1Ö0
+Data-Item1Ì16384ÎGroup-NameÖ0
+ENVIRONMENTÌ32Ö0
+FILEÌ256Ö0
+FILE-CONTROLÌ65536Ö0
+File-Data-ItemÌ16384ÎFile-Group-NameÖ0
+File-Group-NameÌ2048Ö0
+File-NameÌ16Ö0
+Group-NameÌ2048Ö0
+IDENTIFICATIONÌ32Ö0
+INPUT-OUPUTÌ256Ö0
+PROCEDUREÌ32Ö0
+Program-NameÌ1Ö0
+SH-WORK-MMDDYYYYÌ16384ÎGroup-Name.Data-Item1Ö0
+SH-WORK-MMDDYYYY-2Ì16384ÎGroup-Name.Data-Item1Ö0
+Subprogram-NameÌ65536Ö0
+WORKING-STORAGEÌ256Ö0

tests/ctags/simple.cbl.tags

@@ -1,19 +0,0 @@
-# format=tagmanager
-AUTHORÌ65536Ö0
-BeginÌ65536Ö0
-DW-DAYS-IN-MONTHÌ16384Ö0
-DW-DAYS-IN-MONTHSÌ16384Ö0
-Data-Item1Ì16384Ö0
-FILEÌ256Ö0
-FILE-CONTROLÌ65536Ö0
-File-Data-ItemÌ16384Ö0
-File-Group-NameÌ2048Ö0
-File-NameÌ16Ö0
-Group-NameÌ2048Ö0
-INPUT-OUPUTÌ256Ö0
-PROGRAM-IDÌ65536Ö0
-Program-NameÌ1Ö0
-SH-WORK-MMDDYYYYÌ16384Ö0
-SH-WORK-MMDDYYYY-2Ì16384Ö0
-Subprogram-NameÌ65536Ö0
-WORKING-STORAGEÌ256Ö0