Adding upstream version 2:9.0.1378.upstream/2%9.0.1378upstream

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

1 files changed, 4698 insertions, 0 deletions

diff --git a/src/tag.c b/src/tag.c
new file mode 100644
index 0000000..86f8ac4
--- /dev/null
+++ b/src/tag.c
@@ -0,0 +1,4698 @@
+/* vi:set ts=8 sts=4 sw=4 noet:
+ *
+ * VIM - Vi IMproved	by Bram Moolenaar
+ *
+ * Do ":help uganda"  in Vim to read copying and usage conditions.
+ * Do ":help credits" in Vim to see a list of people who contributed.
+ * See README.txt for an overview of the Vim source code.
+ */
+
+/*
+ * Code to handle tags and the tag stack
+ */
+
+#include "vim.h"
+
+/*
+ * Structure to hold pointers to various items in a tag line.
+ */
+typedef struct tag_pointers
+{
+    // filled in by parse_tag_line():
+    char_u	*tagname;	// start of tag name (skip "file:")
+    char_u	*tagname_end;	// char after tag name
+    char_u	*fname;		// first char of file name
+    char_u	*fname_end;	// char after file name
+    char_u	*command;	// first char of command
+    // filled in by parse_match():
+    char_u	*command_end;	// first char after command
+    char_u	*tag_fname;	// file name of the tags file. This is used
+				// when 'tr' is set.
+#ifdef FEAT_EMACS_TAGS
+    int		is_etag;	// TRUE for emacs tag
+#endif
+    char_u	*tagkind;	// "kind:" value
+    char_u	*tagkind_end;	// end of tagkind
+    char_u	*user_data;	// user_data string
+    char_u	*user_data_end;	// end of user_data
+    linenr_T	tagline;	// "line:" value
+} tagptrs_T;
+
+/*
+ * Return values used when reading lines from a tags file.
+ */
+typedef enum
+{
+    TAGS_READ_SUCCESS = 1,
+    TAGS_READ_EOF,
+    TAGS_READ_IGNORE,
+} tags_read_status_T;
+
+/*
+ * States used during a tags search
+ */
+typedef enum
+{
+    TS_START,		// at start of file
+    TS_LINEAR,		// linear searching forward, till EOF
+    TS_BINARY,		// binary searching
+    TS_SKIP_BACK,	// skipping backwards
+    TS_STEP_FORWARD	// stepping forwards
+} tagsearch_state_T;	// Current search state
+
+/*
+ * Binary search file offsets in a tags file
+ */
+typedef struct
+{
+    off_T	low_offset;	// offset for first char of first line that
+				// could match
+    off_T	high_offset;	// offset of char after last line that could
+				// match
+    off_T	curr_offset;	// Current file offset in search range
+    off_T	curr_offset_used; // curr_offset used when skipping back
+    off_T	match_offset;	// Where the binary search found a tag
+    int	low_char;		// first char at low_offset
+    int	high_char;		// first char at high_offset
+} tagsearch_info_T;
+
+/*
+ * Return values used when matching tags against a pattern.
+ */
+typedef enum
+{
+    TAG_MATCH_SUCCESS = 1,
+    TAG_MATCH_FAIL,
+    TAG_MATCH_STOP,
+    TAG_MATCH_NEXT
+} tagmatch_status_T;
+
+/*
+ * Arguments used for matching tags read from a tags file against a pattern.
+ */
+typedef struct
+{
+    int	matchoff;		// tag match offset
+    int	match_re;		// TRUE if the tag matches a regexp
+    int	match_no_ic;		// TRUE if the tag matches with case
+    int	has_re;			// regular expression used
+    int	sortic;			// tags file sorted ignoring case (foldcase)
+    int	sort_error;		// tags file not sorted
+} findtags_match_args_T;
+
+/*
+ * The matching tags are first stored in one of the hash tables.  In
+ * which one depends on the priority of the match.
+ * ht_match[] is used to find duplicates, ga_match[] to keep them in sequence.
+ * At the end, all the matches from ga_match[] are concatenated, to make a list
+ * sorted on priority.
+ */
+#define MT_ST_CUR	0		// static match in current file
+#define MT_GL_CUR	1		// global match in current file
+#define MT_GL_OTH	2		// global match in other file
+#define MT_ST_OTH	3		// static match in other file
+#define MT_IC_OFF	4		// add for icase match
+#define MT_RE_OFF	8		// add for regexp match
+#define MT_MASK		7		// mask for printing priority
+#define MT_COUNT	16
+
+static char	*mt_names[MT_COUNT/2] =
+		{"FSC", "F C", "F  ", "FS ", " SC", "  C", "   ", " S "};
+
+#define NOTAGFILE	99		// return value for jumpto_tag
+static char_u	*nofile_fname = NULL;	// fname for NOTAGFILE error
+
+static void taglen_advance(int l);
+
+static int jumpto_tag(char_u *lbuf, int forceit, int keep_help);
+#ifdef FEAT_EMACS_TAGS
+static int parse_tag_line(char_u *lbuf, int is_etag, tagptrs_T *tagp);
+#else
+static int parse_tag_line(char_u *lbuf, tagptrs_T *tagp);
+#endif
+static int test_for_static(tagptrs_T *);
+static int parse_match(char_u *lbuf, tagptrs_T *tagp);
+static char_u *tag_full_fname(tagptrs_T *tagp);
+static char_u *expand_tag_fname(char_u *fname, char_u *tag_fname, int expand);
+#ifdef FEAT_EMACS_TAGS
+static int test_for_current(int, char_u *, char_u *, char_u *, char_u *);
+#else
+static int test_for_current(char_u *, char_u *, char_u *, char_u *);
+#endif
+static int find_extra(char_u **pp);
+static void print_tag_list(int new_tag, int use_tagstack, int num_matches, char_u **matches);
+#if defined(FEAT_QUICKFIX) && defined(FEAT_EVAL)
+static int add_llist_tags(char_u *tag, int num_matches, char_u **matches);
+#endif
+static void tagstack_clear_entry(taggy_T *item);
+
+static char_u	*tagmatchname = NULL;	// name of last used tag
+
+#if defined(FEAT_QUICKFIX)
+/*
+ * Tag for preview window is remembered separately, to avoid messing up the
+ * normal tagstack.
+ */
+static taggy_T ptag_entry = {NULL, {{0, 0, 0}, 0}, 0, 0, NULL};
+#endif
+
+#ifdef FEAT_EVAL
+static int  tfu_in_use = FALSE;	    // disallow recursive call of tagfunc
+static callback_T tfu_cb;	    // 'tagfunc' callback function
+#endif
+
+// Used instead of NUL to separate tag fields in the growarrays.
+#define TAG_SEP 0x02
+
+#if defined(FEAT_EVAL) || defined(PROTO)
+/*
+ * Reads the 'tagfunc' option value and convert that to a callback value.
+ * Invoked when the 'tagfunc' option is set. The option value can be a name of
+ * a function (string), or function(<name>) or funcref(<name>) or a lambda.
+ */
+    char *
+did_set_tagfunc(optset_T *args UNUSED)
+{
+#ifdef FEAT_EVAL
+    free_callback(&tfu_cb);
+    free_callback(&curbuf->b_tfu_cb);
+
+    if (*curbuf->b_p_tfu == NUL)
+	return NULL;
+
+    if (option_set_callback_func(curbuf->b_p_tfu, &tfu_cb) == FAIL)
+	return e_invalid_argument;
+
+    copy_callback(&curbuf->b_tfu_cb, &tfu_cb);
+#endif
+
+    return NULL;
+}
+#endif
+
+# if defined(EXITFREE) || defined(PROTO)
+    void
+free_tagfunc_option(void)
+{
+# ifdef FEAT_EVAL
+    free_callback(&tfu_cb);
+# endif
+}
+# endif
+
+#if defined(FEAT_EVAL) || defined(PROTO)
+/*
+ * Mark the global 'tagfunc' callback with "copyID" so that it is not garbage
+ * collected.
+ */
+    int
+set_ref_in_tagfunc(int copyID UNUSED)
+{
+    int	abort = FALSE;
+
+    abort = set_ref_in_callback(&tfu_cb, copyID);
+
+    return abort;
+}
+
+/*
+ * Copy the global 'tagfunc' callback function to the buffer-local 'tagfunc'
+ * callback for 'buf'.
+ */
+    void
+set_buflocal_tfu_callback(buf_T *buf UNUSED)
+{
+    free_callback(&buf->b_tfu_cb);
+    if (tfu_cb.cb_name != NULL && *tfu_cb.cb_name != NUL)
+	copy_callback(&buf->b_tfu_cb, &tfu_cb);
+}
+#endif
+
+/*
+ * Jump to tag; handling of tag commands and tag stack
+ *
+ * *tag != NUL: ":tag {tag}", jump to new tag, add to tag stack
+ *
+ * type == DT_TAG:	":tag [tag]", jump to newer position or same tag again
+ * type == DT_HELP:	like DT_TAG, but don't use regexp.
+ * type == DT_POP:	":pop" or CTRL-T, jump to old position
+ * type == DT_NEXT:	jump to next match of same tag
+ * type == DT_PREV:	jump to previous match of same tag
+ * type == DT_FIRST:	jump to first match of same tag
+ * type == DT_LAST:	jump to last match of same tag
+ * type == DT_SELECT:	":tselect [tag]", select tag from a list of all matches
+ * type == DT_JUMP:	":tjump [tag]", jump to tag or select tag from a list
+ * type == DT_CSCOPE:	use cscope to find the tag
+ * type == DT_LTAG:	use location list for displaying tag matches
+ * type == DT_FREE:	free cached matches
+ *
+ * for cscope, returns TRUE if we jumped to tag or aborted, FALSE otherwise
+ */
+    int
+do_tag(
+    char_u	*tag,		// tag (pattern) to jump to
+    int		type,
+    int		count,
+    int		forceit,	// :ta with !
+    int		verbose)	// print "tag not found" message
+{
+    taggy_T	*tagstack = curwin->w_tagstack;
+    int		tagstackidx = curwin->w_tagstackidx;
+    int		tagstacklen = curwin->w_tagstacklen;
+    int		cur_match = 0;
+    int		cur_fnum = curbuf->b_fnum;
+    int		oldtagstackidx = tagstackidx;
+    int		prevtagstackidx = tagstackidx;
+    int		prev_num_matches;
+    int		new_tag = FALSE;
+    int		i;
+    int		ic;
+    int		no_regexp = FALSE;
+    int		error_cur_match = 0;
+    int		save_pos = FALSE;
+    fmark_T	saved_fmark;
+#ifdef FEAT_CSCOPE
+    int		jumped_to_tag = FALSE;
+#endif
+    int		new_num_matches;
+    char_u	**new_matches;
+    int		use_tagstack;
+    int		skip_msg = FALSE;
+    char_u	*buf_ffname = curbuf->b_ffname;	    // name to use for
+						    // priority computation
+    int		use_tfu = 1;
+    char_u	*tofree = NULL;
+
+    // remember the matches for the last used tag
+    static int		num_matches = 0;
+    static int		max_num_matches = 0;  // limit used for match search
+    static char_u	**matches = NULL;
+    static int		flags;
+
+#ifdef FEAT_EVAL
+    if (tfu_in_use)
+    {
+	emsg(_(e_cannot_modify_tag_stack_within_tagfunc));
+	return FALSE;
+    }
+#endif
+
+#ifdef EXITFREE
+    if (type == DT_FREE)
+    {
+	// remove the list of matches
+	FreeWild(num_matches, matches);
+# ifdef FEAT_CSCOPE
+	cs_free_tags();
+# endif
+	num_matches = 0;
+	return FALSE;
+    }
+#endif
+
+    if (type == DT_HELP)
+    {
+	type = DT_TAG;
+	no_regexp = TRUE;
+	use_tfu = 0;
+    }
+
+    prev_num_matches = num_matches;
+    free_string_option(nofile_fname);
+    nofile_fname = NULL;
+
+    CLEAR_POS(&saved_fmark.mark);	// shutup gcc 4.0
+    saved_fmark.fnum = 0;
+
+    /*
+     * Don't add a tag to the tagstack if 'tagstack' has been reset.
+     */
+    if ((!p_tgst && *tag != NUL))
+    {
+	use_tagstack = FALSE;
+	new_tag = TRUE;
+#if defined(FEAT_QUICKFIX)
+	if (g_do_tagpreview != 0)
+	{
+	    tagstack_clear_entry(&ptag_entry);
+	    if ((ptag_entry.tagname = vim_strsave(tag)) == NULL)
+		goto end_do_tag;
+	}
+#endif
+    }
+    else
+    {
+#if defined(FEAT_QUICKFIX)
+	if (g_do_tagpreview != 0)
+	    use_tagstack = FALSE;
+	else
+#endif
+	    use_tagstack = TRUE;
+
+	// new pattern, add to the tag stack
+	if (*tag != NUL
+		&& (type == DT_TAG || type == DT_SELECT || type == DT_JUMP
+#ifdef FEAT_QUICKFIX
+		    || type == DT_LTAG
+#endif
+#ifdef FEAT_CSCOPE
+		    || type == DT_CSCOPE
+#endif
+		    ))
+	{
+#if defined(FEAT_QUICKFIX)
+	    if (g_do_tagpreview != 0)
+	    {
+		if (ptag_entry.tagname != NULL
+			&& STRCMP(ptag_entry.tagname, tag) == 0)
+		{
+		    // Jumping to same tag: keep the current match, so that
+		    // the CursorHold autocommand example works.
+		    cur_match = ptag_entry.cur_match;
+		    cur_fnum = ptag_entry.cur_fnum;
+		}
+		else
+		{
+		    tagstack_clear_entry(&ptag_entry);
+		    if ((ptag_entry.tagname = vim_strsave(tag)) == NULL)
+			goto end_do_tag;
+		}
+	    }
+	    else
+#endif
+	    {
+		/*
+		 * If the last used entry is not at the top, delete all tag
+		 * stack entries above it.
+		 */
+		while (tagstackidx < tagstacklen)
+		    tagstack_clear_entry(&tagstack[--tagstacklen]);
+
+		// if the tagstack is full: remove oldest entry
+		if (++tagstacklen > TAGSTACKSIZE)
+		{
+		    tagstacklen = TAGSTACKSIZE;
+		    tagstack_clear_entry(&tagstack[0]);
+		    for (i = 1; i < tagstacklen; ++i)
+			tagstack[i - 1] = tagstack[i];
+		    --tagstackidx;
+		}
+
+		/*
+		 * put the tag name in the tag stack
+		 */
+		if ((tagstack[tagstackidx].tagname = vim_strsave(tag)) == NULL)
+		{
+		    curwin->w_tagstacklen = tagstacklen - 1;
+		    goto end_do_tag;
+		}
+		curwin->w_tagstacklen = tagstacklen;
+
+		save_pos = TRUE;	// save the cursor position below
+	    }
+
+	    new_tag = TRUE;
+	}
+	else
+	{
+	    if (
+#if defined(FEAT_QUICKFIX)
+		    g_do_tagpreview != 0 ? ptag_entry.tagname == NULL :
+#endif
+		    tagstacklen == 0)
+	    {
+		// empty stack
+		emsg(_(e_tag_stack_empty));
+		goto end_do_tag;
+	    }
+
+	    if (type == DT_POP)		// go to older position
+	    {
+#ifdef FEAT_FOLDING
+		int	old_KeyTyped = KeyTyped;
+#endif
+		if ((tagstackidx -= count) < 0)
+		{
+		    emsg(_(e_at_bottom_of_tag_stack));
+		    if (tagstackidx + count == 0)
+		    {
+			// We did [num]^T from the bottom of the stack
+			tagstackidx = 0;
+			goto end_do_tag;
+		    }
+		    // We weren't at the bottom of the stack, so jump all the
+		    // way to the bottom now.
+		    tagstackidx = 0;
+		}
+		else if (tagstackidx >= tagstacklen)    // count == 0?
+		{
+		    emsg(_(e_at_top_of_tag_stack));
+		    goto end_do_tag;
+		}
+
+		// Make a copy of the fmark, autocommands may invalidate the
+		// tagstack before it's used.
+		saved_fmark = tagstack[tagstackidx].fmark;
+		if (saved_fmark.fnum != curbuf->b_fnum)
+		{
+		    /*
+		     * Jump to other file. If this fails (e.g. because the
+		     * file was changed) keep original position in tag stack.
+		     */
+		    if (buflist_getfile(saved_fmark.fnum, saved_fmark.mark.lnum,
+					       GETF_SETMARK, forceit) == FAIL)
+		    {
+			tagstackidx = oldtagstackidx;  // back to old posn
+			goto end_do_tag;
+		    }
+		    // An BufReadPost autocommand may jump to the '" mark, but
+		    // we don't what that here.
+		    curwin->w_cursor.lnum = saved_fmark.mark.lnum;
+		}
+		else
+		{
+		    setpcmark();
+		    curwin->w_cursor.lnum = saved_fmark.mark.lnum;
+		}
+		curwin->w_cursor.col = saved_fmark.mark.col;
+		curwin->w_set_curswant = TRUE;
+		check_cursor();
+#ifdef FEAT_FOLDING
+		if ((fdo_flags & FDO_TAG) && old_KeyTyped)
+		    foldOpenCursor();
+#endif
+
+		// remove the old list of matches
+		FreeWild(num_matches, matches);
+#ifdef FEAT_CSCOPE
+		cs_free_tags();
+#endif
+		num_matches = 0;
+		tag_freematch();
+		goto end_do_tag;
+	    }
+
+	    if (type == DT_TAG
+#if defined(FEAT_QUICKFIX)
+		    || type == DT_LTAG
+#endif
+	       )
+	    {
+#if defined(FEAT_QUICKFIX)
+		if (g_do_tagpreview != 0)
+		{
+		    cur_match = ptag_entry.cur_match;
+		    cur_fnum = ptag_entry.cur_fnum;
+		}
+		else
+#endif
+		{
+		    // ":tag" (no argument): go to newer pattern
+		    save_pos = TRUE;	// save the cursor position below
+		    if ((tagstackidx += count - 1) >= tagstacklen)
+		    {
+			/*
+			 * Beyond the last one, just give an error message and
+			 * go to the last one.  Don't store the cursor
+			 * position.
+			 */
+			tagstackidx = tagstacklen - 1;
+			emsg(_(e_at_top_of_tag_stack));
+			save_pos = FALSE;
+		    }
+		    else if (tagstackidx < 0)	// must have been count == 0
+		    {
+			emsg(_(e_at_bottom_of_tag_stack));
+			tagstackidx = 0;
+			goto end_do_tag;
+		    }
+		    cur_match = tagstack[tagstackidx].cur_match;
+		    cur_fnum = tagstack[tagstackidx].cur_fnum;
+		}
+		new_tag = TRUE;
+	    }
+	    else				// go to other matching tag
+	    {
+		// Save index for when selection is cancelled.
+		prevtagstackidx = tagstackidx;
+
+#if defined(FEAT_QUICKFIX)
+		if (g_do_tagpreview != 0)
+		{
+		    cur_match = ptag_entry.cur_match;
+		    cur_fnum = ptag_entry.cur_fnum;
+		}
+		else
+#endif
+		{
+		    if (--tagstackidx < 0)
+			tagstackidx = 0;
+		    cur_match = tagstack[tagstackidx].cur_match;
+		    cur_fnum = tagstack[tagstackidx].cur_fnum;
+		}
+		switch (type)
+		{
+		    case DT_FIRST: cur_match = count - 1; break;
+		    case DT_SELECT:
+		    case DT_JUMP:
+#ifdef FEAT_CSCOPE
+		    case DT_CSCOPE:
+#endif
+		    case DT_LAST:  cur_match = MAXCOL - 1; break;
+		    case DT_NEXT:  cur_match += count; break;
+		    case DT_PREV:  cur_match -= count; break;
+		}
+		if (cur_match >= MAXCOL)
+		    cur_match = MAXCOL - 1;
+		else if (cur_match < 0)
+		{
+		    emsg(_(e_cannot_go_before_first_matching_tag));
+		    skip_msg = TRUE;
+		    cur_match = 0;
+		    cur_fnum = curbuf->b_fnum;
+		}
+	    }
+	}
+
+#if defined(FEAT_QUICKFIX)
+	if (g_do_tagpreview != 0)
+	{
+	    if (type != DT_SELECT && type != DT_JUMP)
+	    {
+		ptag_entry.cur_match = cur_match;
+		ptag_entry.cur_fnum = cur_fnum;
+	    }
+	}
+	else
+#endif
+	{
+	    /*
+	     * For ":tag [arg]" or ":tselect" remember position before the jump.
+	     */
+	    saved_fmark = tagstack[tagstackidx].fmark;
+	    if (save_pos)
+	    {
+		tagstack[tagstackidx].fmark.mark = curwin->w_cursor;
+		tagstack[tagstackidx].fmark.fnum = curbuf->b_fnum;
+	    }
+
+	    // Curwin will change in the call to jumpto_tag() if ":stag" was
+	    // used or an autocommand jumps to another window; store value of
+	    // tagstackidx now.
+	    curwin->w_tagstackidx = tagstackidx;
+	    if (type != DT_SELECT && type != DT_JUMP)
+	    {
+		curwin->w_tagstack[tagstackidx].cur_match = cur_match;
+		curwin->w_tagstack[tagstackidx].cur_fnum = cur_fnum;
+	    }
+	}
+    }
+
+    // When not using the current buffer get the name of buffer "cur_fnum".
+    // Makes sure that the tag order doesn't change when using a remembered
+    // position for "cur_match".
+    if (cur_fnum != curbuf->b_fnum)
+    {
+	buf_T *buf = buflist_findnr(cur_fnum);
+
+	if (buf != NULL)
+	    buf_ffname = buf->b_ffname;
+    }
+
+    /*
+     * Repeat searching for tags, when a file has not been found.
+     */
+    for (;;)
+    {
+	int	other_name;
+	char_u	*name;
+
+	/*
+	 * When desired match not found yet, try to find it (and others).
+	 */
+	if (use_tagstack)
+	{
+	    // make a copy, the tagstack may change in 'tagfunc'
+	    name = vim_strsave(tagstack[tagstackidx].tagname);
+	    vim_free(tofree);
+	    tofree = name;
+	}
+#if defined(FEAT_QUICKFIX)
+	else if (g_do_tagpreview != 0)
+	    name = ptag_entry.tagname;
+#endif
+	else
+	    name = tag;
+	other_name = (tagmatchname == NULL || STRCMP(tagmatchname, name) != 0);
+	if (new_tag
+		|| (cur_match >= num_matches && max_num_matches != MAXCOL)
+		|| other_name)
+	{
+	    if (other_name)
+	    {
+		vim_free(tagmatchname);
+		tagmatchname = vim_strsave(name);
+	    }
+
+	    if (type == DT_SELECT || type == DT_JUMP
+#if defined(FEAT_QUICKFIX)
+		|| type == DT_LTAG
+#endif
+		)
+		cur_match = MAXCOL - 1;
+	    if (type == DT_TAG)
+		max_num_matches = MAXCOL;
+	    else
+		max_num_matches = cur_match + 1;
+
+	    // when the argument starts with '/', use it as a regexp
+	    if (!no_regexp && *name == '/')
+	    {
+		flags = TAG_REGEXP;
+		++name;
+	    }
+	    else
+		flags = TAG_NOIC;
+
+#ifdef FEAT_CSCOPE
+	    if (type == DT_CSCOPE)
+		flags = TAG_CSCOPE;
+#endif
+	    if (verbose)
+		flags |= TAG_VERBOSE;
+
+	    if (!use_tfu)
+		flags |= TAG_NO_TAGFUNC;
+
+	    if (find_tags(name, &new_num_matches, &new_matches, flags,
+					    max_num_matches, buf_ffname) == OK
+		    && new_num_matches < max_num_matches)
+		max_num_matches = MAXCOL; // If less than max_num_matches
+					  // found: all matches found.
+
+	    // A tag function may do anything, which may cause various
+	    // information to become invalid.  At least check for the tagstack
+	    // to still be the same.
+	    if (tagstack != curwin->w_tagstack)
+	    {
+		emsg(_(e_window_unexpectedly_close_while_searching_for_tags));
+		FreeWild(new_num_matches, new_matches);
+		break;
+	    }
+
+	    // If there already were some matches for the same name, move them
+	    // to the start.  Avoids that the order changes when using
+	    // ":tnext" and jumping to another file.
+	    if (!new_tag && !other_name)
+	    {
+		int	    j, k;
+		int	    idx = 0;
+		tagptrs_T   tagp, tagp2;
+
+		// Find the position of each old match in the new list.  Need
+		// to use parse_match() to find the tag line.
+		for (j = 0; j < num_matches; ++j)
+		{
+		    parse_match(matches[j], &tagp);
+		    for (i = idx; i < new_num_matches; ++i)
+		    {
+			parse_match(new_matches[i], &tagp2);
+			if (STRCMP(tagp.tagname, tagp2.tagname) == 0)
+			{
+			    char_u *p = new_matches[i];
+			    for (k = i; k > idx; --k)
+				new_matches[k] = new_matches[k - 1];
+			    new_matches[idx++] = p;
+			    break;
+			}
+		    }
+		}
+	    }
+	    FreeWild(num_matches, matches);
+	    num_matches = new_num_matches;
+	    matches = new_matches;
+	}
+
+	if (num_matches <= 0)
+	{
+	    if (verbose)
+		semsg(_(e_tag_not_found_str), name);
+#if defined(FEAT_QUICKFIX)
+	    g_do_tagpreview = 0;
+#endif
+	}
+	else
+	{
+	    int ask_for_selection = FALSE;
+
+#ifdef FEAT_CSCOPE
+	    if (type == DT_CSCOPE && num_matches > 1)
+	    {
+		cs_print_tags();
+		ask_for_selection = TRUE;
+	    }
+	    else
+#endif
+	    if (type == DT_TAG && *tag != NUL)
+		// If a count is supplied to the ":tag <name>" command, then
+		// jump to count'th matching tag.
+		cur_match = count > 0 ? count - 1 : 0;
+	    else if (type == DT_SELECT || (type == DT_JUMP && num_matches > 1))
+	    {
+		print_tag_list(new_tag, use_tagstack, num_matches, matches);
+		ask_for_selection = TRUE;
+	    }
+#if defined(FEAT_QUICKFIX) && defined(FEAT_EVAL)
+	    else if (type == DT_LTAG)
+	    {
+		if (add_llist_tags(tag, num_matches, matches) == FAIL)
+		    goto end_do_tag;
+		cur_match = 0;		// Jump to the first tag
+	    }
+#endif
+
+	    if (ask_for_selection == TRUE)
+	    {
+		/*
+		 * Ask to select a tag from the list.
+		 */
+		i = prompt_for_number(NULL);
+		if (i <= 0 || i > num_matches || got_int)
+		{
+		    // no valid choice: don't change anything
+		    if (use_tagstack)
+		    {
+			tagstack[tagstackidx].fmark = saved_fmark;
+			tagstackidx = prevtagstackidx;
+		    }
+#ifdef FEAT_CSCOPE
+		    cs_free_tags();
+		    jumped_to_tag = TRUE;
+#endif
+		    break;
+		}
+		cur_match = i - 1;
+	    }
+
+	    if (cur_match >= num_matches)
+	    {
+		// Avoid giving this error when a file wasn't found and we're
+		// looking for a match in another file, which wasn't found.
+		// There will be an emsg("file doesn't exist") below then.
+		if ((type == DT_NEXT || type == DT_FIRST)
+						      && nofile_fname == NULL)
+		{
+		    if (num_matches == 1)
+			emsg(_(e_there_is_only_one_matching_tag));
+		    else
+			emsg(_(e_cannot_go_beyond_last_matching_tag));
+		    skip_msg = TRUE;
+		}
+		cur_match = num_matches - 1;
+	    }
+	    if (use_tagstack)
+	    {
+		tagptrs_T   tagp;
+
+		tagstack[tagstackidx].cur_match = cur_match;
+		tagstack[tagstackidx].cur_fnum = cur_fnum;
+
+		// store user-provided data originating from tagfunc
+		if (use_tfu && parse_match(matches[cur_match], &tagp) == OK
+			&& tagp.user_data)
+		{
+		    VIM_CLEAR(tagstack[tagstackidx].user_data);
+		    tagstack[tagstackidx].user_data = vim_strnsave(
+			  tagp.user_data, tagp.user_data_end - tagp.user_data);
+		}
+
+		++tagstackidx;
+	    }
+#if defined(FEAT_QUICKFIX)
+	    else if (g_do_tagpreview != 0)
+	    {
+		ptag_entry.cur_match = cur_match;
+		ptag_entry.cur_fnum = cur_fnum;
+	    }
+#endif
+
+	    /*
+	     * Only when going to try the next match, report that the previous
+	     * file didn't exist.  Otherwise an emsg() is given below.
+	     */
+	    if (nofile_fname != NULL && error_cur_match != cur_match)
+		smsg(_("File \"%s\" does not exist"), nofile_fname);
+
+
+	    ic = (matches[cur_match][0] & MT_IC_OFF);
+	    if (type != DT_TAG && type != DT_SELECT && type != DT_JUMP
+#ifdef FEAT_CSCOPE
+		&& type != DT_CSCOPE
+#endif
+		&& (num_matches > 1 || ic)
+		&& !skip_msg)
+	    {
+		// Give an indication of the number of matching tags
+		sprintf((char *)IObuff, _("tag %d of %d%s"),
+				cur_match + 1,
+				num_matches,
+				max_num_matches != MAXCOL ? _(" or more") : "");
+		if (ic)
+		    STRCAT(IObuff, _("  Using tag with different case!"));
+		if ((num_matches > prev_num_matches || new_tag)
+							   && num_matches > 1)
+		{
+		    if (ic)
+			msg_attr((char *)IObuff, HL_ATTR(HLF_W));
+		    else
+			msg((char *)IObuff);
+		    msg_scroll = TRUE;	// don't overwrite this message
+		}
+		else
+		    give_warning(IObuff, ic);
+		if (ic && !msg_scrolled && msg_silent == 0)
+		{
+		    out_flush();
+		    ui_delay(1007L, TRUE);
+		}
+	    }
+
+#if defined(FEAT_EVAL)
+	    // Let the SwapExists event know what tag we are jumping to.
+	    vim_snprintf((char *)IObuff, IOSIZE, ":ta %s\r", name);
+	    set_vim_var_string(VV_SWAPCOMMAND, IObuff, -1);
+#endif
+
+	    /*
+	     * Jump to the desired match.
+	     */
+	    i = jumpto_tag(matches[cur_match], forceit, type != DT_CSCOPE);
+
+#if defined(FEAT_EVAL)
+	    set_vim_var_string(VV_SWAPCOMMAND, NULL, -1);
+#endif
+
+	    if (i == NOTAGFILE)
+	    {
+		// File not found: try again with another matching tag
+		if ((type == DT_PREV && cur_match > 0)
+			|| ((type == DT_TAG || type == DT_NEXT
+							  || type == DT_FIRST)
+			    && (max_num_matches != MAXCOL
+					     || cur_match < num_matches - 1)))
+		{
+		    error_cur_match = cur_match;
+		    if (use_tagstack)
+			--tagstackidx;
+		    if (type == DT_PREV)
+			--cur_match;
+		    else
+		    {
+			type = DT_NEXT;
+			++cur_match;
+		    }
+		    continue;
+		}
+		semsg(_(e_file_str_does_not_exist), nofile_fname);
+	    }
+	    else
+	    {
+		// We may have jumped to another window, check that
+		// tagstackidx is still valid.
+		if (use_tagstack && tagstackidx > curwin->w_tagstacklen)
+		    tagstackidx = curwin->w_tagstackidx;
+#ifdef FEAT_CSCOPE
+		jumped_to_tag = TRUE;
+#endif
+	    }
+	}
+	break;
+    }
+
+end_do_tag:
+    // Only store the new index when using the tagstack and it's valid.
+    if (use_tagstack && tagstackidx <= curwin->w_tagstacklen)
+	curwin->w_tagstackidx = tagstackidx;
+    postponed_split = 0;	// don't split next time
+# ifdef FEAT_QUICKFIX
+    g_do_tagpreview = 0;	// don't do tag preview next time
+# endif
+
+    vim_free(tofree);
+#ifdef FEAT_CSCOPE
+    return jumped_to_tag;
+#else
+    return FALSE;
+#endif
+}
+
+/*
+ * List all the matching tags.
+ */
+    static void
+print_tag_list(
+    int		new_tag,
+    int		use_tagstack,
+    int		num_matches,
+    char_u	**matches)
+{
+    taggy_T	*tagstack = curwin->w_tagstack;
+    int		tagstackidx = curwin->w_tagstackidx;
+    int		i;
+    char_u	*p;
+    char_u	*command_end;
+    tagptrs_T	tagp;
+    int		taglen;
+    int		attr;
+
+    /*
+     * Assume that the first match indicates how long the tags can
+     * be, and align the file names to that.
+     */
+    parse_match(matches[0], &tagp);
+    taglen = (int)(tagp.tagname_end - tagp.tagname + 2);
+    if (taglen < 18)
+	taglen = 18;
+    if (taglen > Columns - 25)
+	taglen = MAXCOL;
+    if (msg_col == 0)
+	msg_didout = FALSE;	// overwrite previous message
+    msg_start();
+    msg_puts_attr(_("  # pri kind tag"), HL_ATTR(HLF_T));
+    msg_clr_eos();
+    taglen_advance(taglen);
+    msg_puts_attr(_("file\n"), HL_ATTR(HLF_T));
+
+    for (i = 0; i < num_matches && !got_int; ++i)
+    {
+	parse_match(matches[i], &tagp);
+	if (!new_tag && (
+#if defined(FEAT_QUICKFIX)
+		    (g_do_tagpreview != 0
+		     && i == ptag_entry.cur_match) ||
+#endif
+		    (use_tagstack
+		     && i == tagstack[tagstackidx].cur_match)))
+	    *IObuff = '>';
+	else
+	    *IObuff = ' ';
+	vim_snprintf((char *)IObuff + 1, IOSIZE - 1,
+		"%2d %s ", i + 1,
+			       mt_names[matches[i][0] & MT_MASK]);
+	msg_puts((char *)IObuff);
+	if (tagp.tagkind != NULL)
+	    msg_outtrans_len(tagp.tagkind,
+			  (int)(tagp.tagkind_end - tagp.tagkind));
+	msg_advance(13);
+	msg_outtrans_len_attr(tagp.tagname,
+			   (int)(tagp.tagname_end - tagp.tagname),
+						  HL_ATTR(HLF_T));
+	msg_putchar(' ');
+	taglen_advance(taglen);
+
+	// Find out the actual file name. If it is long, truncate
+	// it and put "..." in the middle
+	p = tag_full_fname(&tagp);
+	if (p != NULL)
+	{
+	    msg_outtrans_long_attr(p, HL_ATTR(HLF_D));
+	    vim_free(p);
+	}
+	if (msg_col > 0)
+	    msg_putchar('\n');
+	if (got_int)
+	    break;
+	msg_advance(15);
+
+	// print any extra fields
+	command_end = tagp.command_end;
+	if (command_end != NULL)
+	{
+	    p = command_end + 3;
+	    while (*p && *p != '\r' && *p != '\n')
+	    {
+		while (*p == TAB)
+		    ++p;
+
+		// skip "file:" without a value (static tag)
+		if (STRNCMP(p, "file:", 5) == 0
+					     && vim_isspace(p[5]))
+		{
+		    p += 5;
+		    continue;
+		}
+		// skip "kind:<kind>" and "<kind>"
+		if (p == tagp.tagkind
+			|| (p + 5 == tagp.tagkind
+				&& STRNCMP(p, "kind:", 5) == 0))
+		{
+		    p = tagp.tagkind_end;
+		    continue;
+		}
+		// print all other extra fields
+		attr = HL_ATTR(HLF_CM);
+		while (*p && *p != '\r' && *p != '\n')
+		{
+		    if (msg_col + ptr2cells(p) >= Columns)
+		    {
+			msg_putchar('\n');
+			if (got_int)
+			    break;
+			msg_advance(15);
+		    }
+		    p = msg_outtrans_one(p, attr);
+		    if (*p == TAB)
+		    {
+			msg_puts_attr(" ", attr);
+			break;
+		    }
+		    if (*p == ':')
+			attr = 0;
+		}
+	    }
+	    if (msg_col > 15)
+	    {
+		msg_putchar('\n');
+		if (got_int)
+		    break;
+		msg_advance(15);
+	    }
+	}
+	else
+	{
+	    for (p = tagp.command;
+			      *p && *p != '\r' && *p != '\n'; ++p)
+		;
+	    command_end = p;
+	}
+
+	// Put the info (in several lines) at column 15.
+	// Don't display "/^" and "?^".
+	p = tagp.command;
+	if (*p == '/' || *p == '?')
+	{
+	    ++p;
+	    if (*p == '^')
+		++p;
+	}
+	// Remove leading whitespace from pattern
+	while (p != command_end && vim_isspace(*p))
+	    ++p;
+
+	while (p != command_end)
+	{
+	    if (msg_col + (*p == TAB ? 1 : ptr2cells(p)) > Columns)
+		msg_putchar('\n');
+	    if (got_int)
+		break;
+	    msg_advance(15);
+
+	    // skip backslash used for escaping a command char or
+	    // a backslash
+	    if (*p == '\\' && (*(p + 1) == *tagp.command
+			    || *(p + 1) == '\\'))
+		++p;
+
+	    if (*p == TAB)
+	    {
+		msg_putchar(' ');
+		++p;
+	    }
+	    else
+		p = msg_outtrans_one(p, 0);
+
+	    // don't display the "$/;\"" and "$?;\""
+	    if (p == command_end - 2 && *p == '$'
+				     && *(p + 1) == *tagp.command)
+		break;
+	    // don't display matching '/' or '?'
+	    if (p == command_end - 1 && *p == *tagp.command
+				     && (*p == '/' || *p == '?'))
+		break;
+	}
+	if (msg_col)
+	    msg_putchar('\n');
+	ui_breakcheck();
+    }
+    if (got_int)
+	got_int = FALSE;	// only stop the listing
+}
+
+#if defined(FEAT_QUICKFIX) && defined(FEAT_EVAL)
+/*
+ * Add the matching tags to the location list for the current
+ * window.
+ */
+    static int
+add_llist_tags(
+    char_u	*tag,
+    int		num_matches,
+    char_u	**matches)
+{
+    list_T	*list;
+    char_u	tag_name[128 + 1];
+    char_u	*fname;
+    char_u	*cmd;
+    int		i;
+    char_u	*p;
+    tagptrs_T	tagp;
+
+    fname = alloc(MAXPATHL + 1);
+    cmd = alloc(CMDBUFFSIZE + 1);
+    list = list_alloc();
+    if (list == NULL || fname == NULL || cmd == NULL)
+    {
+	vim_free(cmd);
+	vim_free(fname);
+	if (list != NULL)
+	    list_free(list);
+	return FAIL;
+    }
+
+    for (i = 0; i < num_matches; ++i)
+    {
+	int	    len, cmd_len;
+	long    lnum;
+	dict_T  *dict;
+
+	parse_match(matches[i], &tagp);
+
+	// Save the tag name
+	len = (int)(tagp.tagname_end - tagp.tagname);
+	if (len > 128)
+	    len = 128;
+	vim_strncpy(tag_name, tagp.tagname, len);
+	tag_name[len] = NUL;
+
+	// Save the tag file name
+	p = tag_full_fname(&tagp);
+	if (p == NULL)
+	    continue;
+	vim_strncpy(fname, p, MAXPATHL);
+	vim_free(p);
+
+	// Get the line number or the search pattern used to locate
+	// the tag.
+	lnum = 0;
+	if (isdigit(*tagp.command))
+	    // Line number is used to locate the tag
+	    lnum = atol((char *)tagp.command);
+	else
+	{
+	    char_u *cmd_start, *cmd_end;
+
+	    // Search pattern is used to locate the tag
+
+	    // Locate the end of the command
+	    cmd_start = tagp.command;
+	    cmd_end = tagp.command_end;
+	    if (cmd_end == NULL)
+	    {
+		for (p = tagp.command;
+		     *p && *p != '\r' && *p != '\n'; ++p)
+		    ;
+		cmd_end = p;
+	    }
+
+	    // Now, cmd_end points to the character after the
+	    // command. Adjust it to point to the last
+	    // character of the command.
+	    cmd_end--;
+
+	    // Skip the '/' and '?' characters at the
+	    // beginning and end of the search pattern.
+	    if (*cmd_start == '/' || *cmd_start == '?')
+		cmd_start++;
+
+	    if (*cmd_end == '/' || *cmd_end == '?')
+		cmd_end--;
+
+	    len = 0;
+	    cmd[0] = NUL;
+
+	    // If "^" is present in the tag search pattern, then
+	    // copy it first.
+	    if (*cmd_start == '^')
+	    {
+		STRCPY(cmd, "^");
+		cmd_start++;
+		len++;
+	    }
+
+	    // Precede the tag pattern with \V to make it very
+	    // nomagic.
+	    STRCAT(cmd, "\\V");
+	    len += 2;
+
+	    cmd_len = (int)(cmd_end - cmd_start + 1);
+	    if (cmd_len > (CMDBUFFSIZE - 5))
+		cmd_len = CMDBUFFSIZE - 5;
+	    STRNCAT(cmd, cmd_start, cmd_len);
+	    len += cmd_len;
+
+	    if (cmd[len - 1] == '$')
+	    {
+		// Replace '$' at the end of the search pattern
+		// with '\$'
+		cmd[len - 1] = '\\';
+		cmd[len] = '$';
+		len++;
+	    }
+
+	    cmd[len] = NUL;
+	}
+
+	if ((dict = dict_alloc()) == NULL)
+	    continue;
+	if (list_append_dict(list, dict) == FAIL)
+	{
+	    vim_free(dict);
+	    continue;
+	}
+
+	dict_add_string(dict, "text", tag_name);
+	dict_add_string(dict, "filename", fname);
+	dict_add_number(dict, "lnum", lnum);
+	if (lnum == 0)
+	    dict_add_string(dict, "pattern", cmd);
+    }
+
+    vim_snprintf((char *)IObuff, IOSIZE, "ltag %s", tag);
+    set_errorlist(curwin, list, ' ', IObuff, NULL);
+
+    list_free(list);
+    vim_free(fname);
+    vim_free(cmd);
+
+    return OK;
+}
+#endif
+
+/*
+ * Free cached tags.
+ */
+    void
+tag_freematch(void)
+{
+    VIM_CLEAR(tagmatchname);
+}
+
+    static void
+taglen_advance(int l)
+{
+    if (l == MAXCOL)
+    {
+	msg_putchar('\n');
+	msg_advance(24);
+    }
+    else
+	msg_advance(13 + l);
+}
+
+/*
+ * Print the tag stack
+ */
+    void
+do_tags(exarg_T *eap UNUSED)
+{
+    int		i;
+    char_u	*name;
+    taggy_T	*tagstack = curwin->w_tagstack;
+    int		tagstackidx = curwin->w_tagstackidx;
+    int		tagstacklen = curwin->w_tagstacklen;
+
+    // Highlight title
+    msg_puts_title(_("\n  # TO tag         FROM line  in file/text"));
+    for (i = 0; i < tagstacklen; ++i)
+    {
+	if (tagstack[i].tagname != NULL)
+	{
+	    name = fm_getname(&(tagstack[i].fmark), 30);
+	    if (name == NULL)	    // file name not available
+		continue;
+
+	    msg_putchar('\n');
+	    vim_snprintf((char *)IObuff, IOSIZE, "%c%2d %2d %-15s %5ld  ",
+		i == tagstackidx ? '>' : ' ',
+		i + 1,
+		tagstack[i].cur_match + 1,
+		tagstack[i].tagname,
+		tagstack[i].fmark.mark.lnum);
+	    msg_outtrans(IObuff);
+	    msg_outtrans_attr(name, tagstack[i].fmark.fnum == curbuf->b_fnum
+							? HL_ATTR(HLF_D) : 0);
+	    vim_free(name);
+	}
+	out_flush();		    // show one line at a time
+    }
+    if (tagstackidx == tagstacklen)	// idx at top of stack
+	msg_puts("\n>");
+}
+
+/*
+ * Compare two strings, for length "len", ignoring case the ASCII way.
+ * return 0 for match, < 0 for smaller, > 0 for bigger
+ * Make sure case is folded to uppercase in comparison (like for 'sort -f')
+ */
+    static int
+tag_strnicmp(char_u *s1, char_u *s2, size_t len)
+{
+    int		i;
+
+    while (len > 0)
+    {
+	i = (int)TOUPPER_ASC(*s1) - (int)TOUPPER_ASC(*s2);
+	if (i != 0)
+	    return i;			// this character different
+	if (*s1 == NUL)
+	    break;			// strings match until NUL
+	++s1;
+	++s2;
+	--len;
+    }
+    return 0;				// strings match
+}
+
+/*
+ * Structure to hold info about the tag pattern being used.
+ */
+typedef struct
+{
+    char_u	*pat;		// the pattern
+    int		len;		// length of pat[]
+    char_u	*head;		// start of pattern head
+    int		headlen;	// length of head[]
+    regmatch_T	regmatch;	// regexp program, may be NULL
+} pat_T;
+
+/*
+ * Extract info from the tag search pattern "pats->pat".
+ */
+    static void
+prepare_pats(pat_T *pats, int has_re)
+{
+    pats->head = pats->pat;
+    pats->headlen = pats->len;
+    if (has_re)
+    {
+	// When the pattern starts with '^' or "\\<", binary searching can be
+	// used (much faster).
+	if (pats->pat[0] == '^')
+	    pats->head = pats->pat + 1;
+	else if (pats->pat[0] == '\\' && pats->pat[1] == '<')
+	    pats->head = pats->pat + 2;
+	if (pats->head == pats->pat)
+	    pats->headlen = 0;
+	else
+	    for (pats->headlen = 0; pats->head[pats->headlen] != NUL;
+							      ++pats->headlen)
+		if (vim_strchr((char_u *)(magic_isset() ? ".[~*\\$" : "\\$"),
+					   pats->head[pats->headlen]) != NULL)
+		    break;
+	if (p_tl != 0 && pats->headlen > p_tl)	// adjust for 'taglength'
+	    pats->headlen = p_tl;
+    }
+
+    if (has_re)
+	pats->regmatch.regprog = vim_regcomp(pats->pat,
+						 magic_isset() ? RE_MAGIC : 0);
+    else
+	pats->regmatch.regprog = NULL;
+}
+
+#ifdef FEAT_EVAL
+/*
+ * Call the user-defined function to generate a list of tags used by
+ * find_tags().
+ *
+ * Return OK if at least 1 tag has been successfully found,
+ * NOTDONE if the function returns v:null, and FAIL otherwise.
+ */
+    static int
+find_tagfunc_tags(
+    char_u	*pat,		// pattern supplied to the user-defined function
+    garray_T	*ga,		// the tags will be placed here
+    int		*match_count,	// here the number of tags found will be placed
+    int		flags,		// flags from find_tags (TAG_*)
+    char_u	*buf_ffname)	// name of buffer for priority
+{
+    pos_T       save_pos;
+    list_T      *taglist;
+    listitem_T  *item;
+    int		ntags = 0;
+    int		result = FAIL;
+    typval_T	args[4];
+    typval_T	rettv;
+    char_u      flagString[4];
+    dict_T	*d;
+    taggy_T	*tag = &curwin->w_tagstack[curwin->w_tagstackidx];
+
+    if (*curbuf->b_p_tfu == NUL || curbuf->b_tfu_cb.cb_name == NULL
+					   || *curbuf->b_tfu_cb.cb_name == NUL)
+	return FAIL;
+
+    args[0].v_type = VAR_STRING;
+    args[0].vval.v_string = pat;
+    args[1].v_type = VAR_STRING;
+    args[1].vval.v_string = flagString;
+
+    // create 'info' dict argument
+    if ((d = dict_alloc_lock(VAR_FIXED)) == NULL)
+	return FAIL;
+    if (tag->user_data != NULL)
+	dict_add_string(d, "user_data", tag->user_data);
+    if (buf_ffname != NULL)
+	dict_add_string(d, "buf_ffname", buf_ffname);
+
+    ++d->dv_refcount;
+    args[2].v_type = VAR_DICT;
+    args[2].vval.v_dict = d;
+
+    args[3].v_type = VAR_UNKNOWN;
+
+    vim_snprintf((char *)flagString, sizeof(flagString),
+		 "%s%s%s",
+		 g_tag_at_cursor      ? "c": "",
+		 flags & TAG_INS_COMP ? "i": "",
+		 flags & TAG_REGEXP   ? "r": "");
+
+    save_pos = curwin->w_cursor;
+    result = call_callback(&curbuf->b_tfu_cb, 0, &rettv, 3, args);
+    curwin->w_cursor = save_pos;	// restore the cursor position
+    --d->dv_refcount;
+
+    if (result == FAIL)
+	return FAIL;
+    if (rettv.v_type == VAR_SPECIAL && rettv.vval.v_number == VVAL_NULL)
+    {
+	clear_tv(&rettv);
+	return NOTDONE;
+    }
+    if (rettv.v_type != VAR_LIST || !rettv.vval.v_list)
+    {
+	clear_tv(&rettv);
+	emsg(_(e_invalid_return_value_from_tagfunc));
+	return FAIL;
+    }
+    taglist = rettv.vval.v_list;
+
+    FOR_ALL_LIST_ITEMS(taglist, item)
+    {
+	char_u		*mfp;
+	char_u		*res_name, *res_fname, *res_cmd, *res_kind;
+	int		len;
+	dict_iterator_T	iter;
+	char_u		*dict_key;
+	typval_T	*tv;
+	int		has_extra = 0;
+	int		name_only = flags & TAG_NAMES;
+
+	if (item->li_tv.v_type != VAR_DICT)
+	{
+	    emsg(_(e_invalid_return_value_from_tagfunc));
+	    break;
+	}
+
+#ifdef FEAT_EMACS_TAGS
+	len = 3;
+#else
+	len = 2;
+#endif
+	res_name = NULL;
+	res_fname = NULL;
+	res_cmd = NULL;
+	res_kind = NULL;
+
+	dict_iterate_start(&item->li_tv, &iter);
+	while (NULL != (dict_key = dict_iterate_next(&iter, &tv)))
+	{
+	    if (tv->v_type != VAR_STRING || tv->vval.v_string == NULL)
+		continue;
+
+	    len += (int)STRLEN(tv->vval.v_string) + 1;   // Space for "\tVALUE"
+	    if (!STRCMP(dict_key, "name"))
+	    {
+		res_name = tv->vval.v_string;
+		continue;
+	    }
+	    if (!STRCMP(dict_key, "filename"))
+	    {
+		res_fname = tv->vval.v_string;
+		continue;
+	    }
+	    if (!STRCMP(dict_key, "cmd"))
+	    {
+		res_cmd = tv->vval.v_string;
+		continue;
+	    }
+	    has_extra = 1;
+	    if (!STRCMP(dict_key, "kind"))
+	    {
+		res_kind = tv->vval.v_string;
+		continue;
+	    }
+	    // Other elements will be stored as "\tKEY:VALUE"
+	    // Allocate space for the key and the colon
+	    len += (int)STRLEN(dict_key) + 1;
+	}
+
+	if (has_extra)
+	    len += 2;	// need space for ;"
+
+	if (!res_name || !res_fname || !res_cmd)
+	{
+	    emsg(_(e_invalid_return_value_from_tagfunc));
+	    break;
+	}
+
+	if (name_only)
+	    mfp = vim_strsave(res_name);
+	else
+	    mfp = alloc(sizeof(char_u) + len + 1);
+
+	if (mfp == NULL)
+	    continue;
+
+	if (!name_only)
+	{
+	    char_u *p = mfp;
+
+	    *p++ = MT_GL_OTH + 1;   // mtt
+	    *p++ = TAG_SEP;	    // no tag file name
+#ifdef FEAT_EMACS_TAGS
+	    *p++ = TAG_SEP;
+#endif
+
+	    STRCPY(p, res_name);
+	    p += STRLEN(p);
+
+	    *p++ = TAB;
+	    STRCPY(p, res_fname);
+	    p += STRLEN(p);
+
+	    *p++ = TAB;
+	    STRCPY(p, res_cmd);
+	    p += STRLEN(p);
+
+	    if (has_extra)
+	    {
+		STRCPY(p, ";\"");
+		p += STRLEN(p);
+
+		if (res_kind)
+		{
+		    *p++ = TAB;
+		    STRCPY(p, res_kind);
+		    p += STRLEN(p);
+		}
+
+		dict_iterate_start(&item->li_tv, &iter);
+		while (NULL != (dict_key = dict_iterate_next(&iter, &tv)))
+		{
+		    if (tv->v_type != VAR_STRING || tv->vval.v_string == NULL)
+			continue;
+
+		    if (!STRCMP(dict_key, "name"))
+			continue;
+		    if (!STRCMP(dict_key, "filename"))
+			continue;
+		    if (!STRCMP(dict_key, "cmd"))
+			continue;
+		    if (!STRCMP(dict_key, "kind"))
+			continue;
+
+		    *p++ = TAB;
+		    STRCPY(p, dict_key);
+		    p += STRLEN(p);
+		    STRCPY(p, ":");
+		    p += STRLEN(p);
+		    STRCPY(p, tv->vval.v_string);
+		    p += STRLEN(p);
+		}
+	    }
+	}
+
+	// Add all matches because tagfunc should do filtering.
+	if (ga_grow(ga, 1) == OK)
+	{
+	    ((char_u **)(ga->ga_data))[ga->ga_len++] = mfp;
+	    ++ntags;
+	    result = OK;
+	}
+	else
+	{
+	    vim_free(mfp);
+	    break;
+	}
+    }
+
+    clear_tv(&rettv);
+
+    *match_count = ntags;
+    return result;
+}
+#endif
+
+/*
+ * State information used during a tag search
+ */
+typedef struct
+{
+    tagsearch_state_T	state;		// tag search state
+    int		stop_searching;		// stop when match found or error
+    pat_T	*orgpat;		// holds unconverted pattern info
+    char_u     *lbuf;			// line buffer
+    int		lbuf_size;		// length of lbuf
+    char_u	*tag_fname;		// name of the tag file
+    FILE	*fp;			// current tags file pointer
+    int		flags;			// flags used for tag search
+    int		tag_file_sorted;	// !_TAG_FILE_SORTED value
+    int		get_searchpat;		// used for 'showfulltag'
+    int		help_only;		// only search for help tags
+    int		did_open;		// did open a tag file
+    int		mincount;		// MAXCOL: find all matches
+					// other: minimal number of matches
+    int		linear;			// do a linear search
+    vimconv_T	vimconv;
+#ifdef FEAT_EMACS_TAGS
+    int		is_etag;		// current file is emacs style
+    char_u	*ebuf;			// additional buffer for etag fname
+#endif
+#ifdef FEAT_MULTI_LANG
+    char_u	help_lang[3];		// lang of current tags file
+    int		help_pri;		// help language priority
+    char_u	*help_lang_find;	// lang to be found
+    int		is_txt;			// flag of file extension
+#endif
+    int		match_count;		// number of matches found
+    garray_T	ga_match[MT_COUNT];	// stores matches in sequence
+    hashtab_T	ht_match[MT_COUNT];	// stores matches by key
+} findtags_state_T;
+
+/*
+ * Initialize the state used by find_tags().
+ * Returns OK on success and FAIL on memory allocation failure.
+ */
+    static int
+findtags_state_init(
+    findtags_state_T	*st,
+    char_u		*pat,
+    int			flags,
+    int			mincount)
+{
+    int		mtt;
+
+    st->tag_fname = alloc(MAXPATHL + 1);
+    st->fp = NULL;
+    st->orgpat = ALLOC_ONE(pat_T);
+    st->orgpat->pat = pat;
+    st->orgpat->len = (int)STRLEN(pat);
+    st->orgpat->regmatch.regprog = NULL;
+    st->flags = flags;
+    st->tag_file_sorted = NUL;
+    st->help_only = (flags & TAG_HELP);
+    st->get_searchpat = FALSE;
+#ifdef FEAT_MULTI_LANG
+    st->help_lang[0] = NUL;
+    st->help_pri = 0;
+    st->help_lang_find = NULL;
+    st->is_txt = FALSE;
+#endif
+    st->did_open = FALSE;
+    st->mincount = mincount;
+    st->lbuf_size = LSIZE;
+    st->lbuf = alloc(st->lbuf_size);
+#ifdef FEAT_EMACS_TAGS
+    st->ebuf = alloc(LSIZE);
+#endif
+    st->match_count = 0;
+    st->stop_searching = FALSE;
+
+    for (mtt = 0; mtt < MT_COUNT; ++mtt)
+    {
+	ga_init2(&st->ga_match[mtt], sizeof(char_u *), 100);
+	hash_init(&st->ht_match[mtt]);
+    }
+
+    // check for out of memory situation
+    if (st->tag_fname == NULL
+	    || st->lbuf == NULL
+#ifdef FEAT_EMACS_TAGS
+	    || st->ebuf == NULL
+#endif
+       )
+	return FAIL;
+
+    return OK;
+}
+
+/*
+ * Free the state used by find_tags()
+ */
+    static void
+findtags_state_free(findtags_state_T *st)
+{
+    vim_free(st->tag_fname);
+    vim_free(st->lbuf);
+    vim_regfree(st->orgpat->regmatch.regprog);
+    vim_free(st->orgpat);
+#ifdef FEAT_EMACS_TAGS
+    vim_free(st->ebuf);
+#endif
+}
+
+#ifdef FEAT_MULTI_LANG
+/*
+ * Initialize the language and priority used for searching tags in a Vim help
+ * file.
+ * Returns TRUE to process the help file for tags and FALSE to skip the file.
+ */
+    static int
+findtags_in_help_init(findtags_state_T *st)
+{
+    int		i;
+    char_u	*s;
+
+    // Keep "en" as the language if the file extension is ".txt"
+    if (st->is_txt)
+	STRCPY(st->help_lang, "en");
+    else
+    {
+	// Prefer help tags according to 'helplang'.  Put the two-letter
+	// language name in help_lang[].
+	i = (int)STRLEN(st->tag_fname);
+	if (i > 3 && st->tag_fname[i - 3] == '-')
+	    vim_strncpy(st->help_lang, st->tag_fname + i - 2, 2);
+	else
+	    STRCPY(st->help_lang, "en");
+    }
+    // When searching for a specific language skip tags files for other
+    // languages.
+    if (st->help_lang_find != NULL
+	    && STRICMP(st->help_lang, st->help_lang_find) != 0)
+	return FALSE;
+
+    // For CTRL-] in a help file prefer a match with the same language.
+    if ((st->flags & TAG_KEEP_LANG)
+	    && st->help_lang_find == NULL
+	    && curbuf->b_fname != NULL
+	    && (i = (int)STRLEN(curbuf->b_fname)) > 4
+	    && curbuf->b_fname[i - 1] == 'x'
+	    && curbuf->b_fname[i - 4] == '.'
+	    && STRNICMP(curbuf->b_fname + i - 3, st->help_lang, 2) == 0)
+	st->help_pri = 0;
+    else
+    {
+	// search for the language in 'helplang'
+	st->help_pri = 1;
+	for (s = p_hlg; *s != NUL; ++s)
+	{
+	    if (STRNICMP(s, st->help_lang, 2) == 0)
+		break;
+	    ++st->help_pri;
+	    if ((s = vim_strchr(s, ',')) == NULL)
+		break;
+	}
+	if (s == NULL || *s == NUL)
+	{
+	    // Language not in 'helplang': use last, prefer English, unless
+	    // found already.
+	    ++st->help_pri;
+	    if (STRICMP(st->help_lang, "en") != 0)
+		++st->help_pri;
+	}
+    }
+
+    return TRUE;
+}
+#endif
+
+#ifdef FEAT_EVAL
+/*
+ * Use the function set in 'tagfunc' (if configured and enabled) to get the
+ * tags.
+ * Return OK if at least 1 tag has been successfully found, NOTDONE if the
+ * 'tagfunc' is not used or the 'tagfunc' returns v:null and FAIL otherwise.
+ */
+    static int
+findtags_apply_tfu(findtags_state_T *st, char_u *pat, char_u *buf_ffname)
+{
+    int		use_tfu = ((st->flags & TAG_NO_TAGFUNC) == 0);
+    int		retval;
+
+    if (!use_tfu || tfu_in_use || *curbuf->b_p_tfu == NUL)
+	return NOTDONE;
+
+    tfu_in_use = TRUE;
+    retval = find_tagfunc_tags(pat, st->ga_match, &st->match_count,
+						st->flags, buf_ffname);
+    tfu_in_use = FALSE;
+
+    return retval;
+}
+#endif
+
+#ifdef FEAT_EMACS_TAGS
+/*
+ * Stack for included emacs-tags file.
+ * It has a fixed size, to truncate cyclic includes. jw
+ */
+# define INCSTACK_SIZE 42
+static struct
+{
+    FILE	*fp;
+    char_u	*etag_fname;
+} incstack[INCSTACK_SIZE];
+static int incstack_idx = 0;	// index in incstack
+
+/*
+ * Free the emacs include tags file stack.
+ */
+    static void
+emacs_tags_incstack_free(void)
+{
+    while (incstack_idx)
+    {
+	--incstack_idx;
+	fclose(incstack[incstack_idx].fp);
+	incstack[incstack_idx].fp = NULL;
+	VIM_CLEAR(incstack[incstack_idx].etag_fname);
+    }
+}
+
+/*
+ * Emacs tags line with CTRL-L: New file name on next line.
+ * The file name is followed by a ','.  Remember etag file name in ebuf.
+ * The FILE pointer to the tags file is stored in "st->fp".  If another tags
+ * file is included, then the FILE pointer to the new tags file is stored in
+ * "st->fp". The old file pointer is saved in incstack.
+ */
+    static void
+emacs_tags_new_filename(findtags_state_T *st)
+{
+    char_u	*p;
+    char_u	*fullpath_ebuf;
+
+    if (vim_fgets(st->ebuf, LSIZE, st->fp))
+	return;
+
+    for (p = st->ebuf; *p && *p != ','; p++)
+	;
+    *p = NUL;
+
+    // check for an included tags file.
+    // atoi(p+1) is the number of bytes before the next ^L unless it is an
+    // include statement. Skip the included tags file if it exceeds the
+    // maximum.
+    if (STRNCMP(p + 1, "include", 7) != 0 || incstack_idx >= INCSTACK_SIZE)
+	return;
+
+    // Save current "fp" and "tag_fname" in the stack.
+    incstack[incstack_idx].etag_fname = vim_strsave(st->tag_fname);
+    if (incstack[incstack_idx].etag_fname == NULL)
+	return;
+
+    incstack[incstack_idx].fp = st->fp;
+    st->fp = NULL;
+
+    // Figure out "tag_fname" and "fp" to use for
+    // included file.
+    fullpath_ebuf = expand_tag_fname(st->ebuf, st->tag_fname, FALSE);
+    if (fullpath_ebuf != NULL)
+    {
+	st->fp = mch_fopen((char *)fullpath_ebuf, "r");
+	if (st->fp != NULL)
+	{
+	    if (STRLEN(fullpath_ebuf) > LSIZE)
+		semsg(_(e_tag_file_path_truncated_for_str), st->ebuf);
+	    vim_strncpy(st->tag_fname, fullpath_ebuf, MAXPATHL);
+	    ++incstack_idx;
+	    st->is_etag = FALSE; // we can include anything
+	}
+	vim_free(fullpath_ebuf);
+    }
+    if (st->fp == NULL)
+    {
+	// Can't open the included file, skip it and
+	// restore old value of "fp".
+	st->fp = incstack[incstack_idx].fp;
+	vim_free(incstack[incstack_idx].etag_fname);
+    }
+}
+
+/*
+ * Reached the end of an emacs-style tags file. If this is an included tags
+ * file, then pop it from the incstack and continue processing the parent tags
+ * file. Otherwise, processed all the tags.
+ * Returns TRUE if an included tags file is popped and processing should
+ * continue with the parent tags file. Returns FALSE to stop processing tags.
+ */
+    static int
+emacs_tags_file_eof(findtags_state_T *st)
+{
+    if (!incstack_idx)	// reached end of file. stop processing.
+	return FALSE;
+
+    // reached the end of an included tags file. pop it.
+    --incstack_idx;
+    fclose(st->fp);	// end of this file ...
+    st->fp = incstack[incstack_idx].fp;
+    STRCPY(st->tag_fname, incstack[incstack_idx].etag_fname);
+    vim_free(incstack[incstack_idx].etag_fname);
+    st->is_etag = TRUE;	// (only etags can include)
+
+    return TRUE;
+}
+
+/*
+ * Parse a line from an emacs-style tags file.
+ * Returns OK if the line is parsed successfully, returns FAIL if the line is
+ * not terminated by a newline.
+ */
+    static int
+emacs_tags_parse_line(char_u *lbuf, tagptrs_T *tagp)
+{
+    char_u	*p_7f;
+    char_u	*p;
+
+    // There are two formats for an emacs tag line:
+    // 1:  struct EnvBase ^?EnvBase^A139,4627
+    // 2: #define	ARPB_WILD_WORLD ^?153,5194
+    p_7f = vim_strchr(lbuf, 0x7f);
+    if (p_7f == NULL)
+    {
+etag_fail:
+	if (vim_strchr(lbuf, '\n') != NULL)
+	    return FAIL;
+
+	// Truncated line.  Ignore it.
+	if (p_verbose >= 5)
+	{
+	    verbose_enter();
+	    msg(_("Ignoring long line in tags file"));
+	    verbose_leave();
+	}
+	tagp->command = lbuf;
+	tagp->tagname = lbuf;
+	tagp->tagname_end = lbuf;
+	return OK;
+    }
+
+    // Find ^A.  If not found the line number is after the 0x7f
+    p = vim_strchr(p_7f, Ctrl_A);
+    if (p == NULL)
+	p = p_7f + 1;
+    else
+	++p;
+
+    if (!VIM_ISDIGIT(*p))	    // check for start of line number
+	goto etag_fail;
+    tagp->command = p;
+
+    if (p[-1] == Ctrl_A)	    // first format: explicit tagname given
+    {
+	tagp->tagname = p_7f + 1;
+	tagp->tagname_end = p - 1;
+    }
+    else			    // second format: isolate tagname
+    {
+	// find end of tagname
+	for (p = p_7f - 1; !vim_iswordc(*p); --p)
+	    if (p == lbuf)
+		goto etag_fail;
+	tagp->tagname_end = p + 1;
+	while (p >= lbuf && vim_iswordc(*p))
+	    --p;
+	tagp->tagname = p + 1;
+    }
+
+    return OK;
+}
+#endif
+
+/*
+ * Read the next line from a tags file.
+ * Returns TAGS_READ_SUCCESS if a tags line is successfully read and should be
+ * processed.
+ * Returns TAGS_READ_EOF if the end of file is reached.
+ * Returns TAGS_READ_IGNORE if the current line should be ignored (used when
+ * reached end of a emacs included tags file)
+ */
+    static tags_read_status_T
+findtags_get_next_line(findtags_state_T *st, tagsearch_info_T *sinfo_p)
+{
+    int		eof;
+    off_T	offset;
+
+    // For binary search: compute the next offset to use.
+    if (st->state == TS_BINARY)
+    {
+	offset = sinfo_p->low_offset + ((sinfo_p->high_offset
+						- sinfo_p->low_offset) / 2);
+	if (offset == sinfo_p->curr_offset)
+	    return TAGS_READ_EOF; // End the binary search without a match.
+	else
+	    sinfo_p->curr_offset = offset;
+    }
+
+    // Skipping back (after a match during binary search).
+    else if (st->state == TS_SKIP_BACK)
+    {
+	sinfo_p->curr_offset -= st->lbuf_size * 2;
+	if (sinfo_p->curr_offset < 0)
+	{
+	    sinfo_p->curr_offset = 0;
+	    rewind(st->fp);
+	    st->state = TS_STEP_FORWARD;
+	}
+    }
+
+    // When jumping around in the file, first read a line to find the
+    // start of the next line.
+    if (st->state == TS_BINARY || st->state == TS_SKIP_BACK)
+    {
+	// Adjust the search file offset to the correct position
+	sinfo_p->curr_offset_used = sinfo_p->curr_offset;
+	vim_ignored = vim_fseek(st->fp, sinfo_p->curr_offset, SEEK_SET);
+	eof = vim_fgets(st->lbuf, st->lbuf_size, st->fp);
+	if (!eof && sinfo_p->curr_offset != 0)
+	{
+	    sinfo_p->curr_offset = vim_ftell(st->fp);
+	    if (sinfo_p->curr_offset == sinfo_p->high_offset)
+	    {
+		// oops, gone a bit too far; try from low offset
+		vim_ignored = vim_fseek(st->fp, sinfo_p->low_offset, SEEK_SET);
+		sinfo_p->curr_offset = sinfo_p->low_offset;
+	    }
+	    eof = vim_fgets(st->lbuf, st->lbuf_size, st->fp);
+	}
+	// skip empty and blank lines
+	while (!eof && vim_isblankline(st->lbuf))
+	{
+	    sinfo_p->curr_offset = vim_ftell(st->fp);
+	    eof = vim_fgets(st->lbuf, st->lbuf_size, st->fp);
+	}
+	if (eof)
+	{
+	    // Hit end of file.  Skip backwards.
+	    st->state = TS_SKIP_BACK;
+	    sinfo_p->match_offset = vim_ftell(st->fp);
+	    sinfo_p->curr_offset = sinfo_p->curr_offset_used;
+	    return TAGS_READ_IGNORE;
+	}
+    }
+    // Not jumping around in the file: Read the next line.
+    else
+    {
+	// skip empty and blank lines
+	do
+	{
+#ifdef FEAT_CSCOPE
+	    if (st->flags & TAG_CSCOPE)
+		eof = cs_fgets(st->lbuf, st->lbuf_size);
+	    else
+#endif
+		eof = vim_fgets(st->lbuf, st->lbuf_size, st->fp);
+	} while (!eof && vim_isblankline(st->lbuf));
+
+	if (eof)
+	{
+#ifdef FEAT_EMACS_TAGS
+	    if (emacs_tags_file_eof(st) == TRUE)
+		// an included tags file. Continue processing the parent
+		// tags file.
+		return TAGS_READ_IGNORE;
+#endif
+	    return TAGS_READ_EOF;
+	}
+    }
+
+    return TAGS_READ_SUCCESS;
+}
+
+/*
+ * Parse a tags file header line in "st->lbuf".
+ * Returns TRUE if the current line in st->lbuf is not a tags header line and
+ * should be parsed as a regular tag line. Returns FALSE if the line is a
+ * header line and the next header line should be read.
+ */
+    static int
+findtags_hdr_parse(findtags_state_T *st)
+{
+    char_u	*p;
+
+    // Header lines in a tags file start with "!_TAG_"
+    if (STRNCMP(st->lbuf, "!_TAG_", 6) != 0)
+	// Non-header item before the header, e.g. "!" itself.
+	return TRUE;
+
+    // Process the header line.
+    if (STRNCMP(st->lbuf, "!_TAG_FILE_SORTED\t", 18) == 0)
+	st->tag_file_sorted = st->lbuf[18];
+    if (STRNCMP(st->lbuf, "!_TAG_FILE_ENCODING\t", 20) == 0)
+    {
+	// Prepare to convert every line from the specified encoding to
+	// 'encoding'.
+	for (p = st->lbuf + 20; *p > ' ' && *p < 127; ++p)
+	    ;
+	*p = NUL;
+	convert_setup(&st->vimconv, st->lbuf + 20, p_enc);
+    }
+
+    // Read the next line.  Unrecognized flags are ignored.
+    return FALSE;
+}
+
+/*
+ * Handler to initialize the state when starting to process a new tags file.
+ * Called in the TS_START state when finding tags from a tags file.
+ * Returns TRUE if the line read from the tags file should be parsed and
+ * FALSE if the line should be ignored.
+ */
+    static int
+findtags_start_state_handler(
+    findtags_state_T	*st,
+    int			*sortic,
+    tagsearch_info_T	*sinfo_p)
+{
+#ifdef FEAT_CSCOPE
+    int		use_cscope = (st->flags & TAG_CSCOPE);
+#endif
+    int		noic = (st->flags & TAG_NOIC);
+    off_T	filesize;
+
+    // The header ends when the line sorts below "!_TAG_".  When case is
+    // folded lower case letters sort before "_".
+    if (STRNCMP(st->lbuf, "!_TAG_", 6) <= 0
+	    || (st->lbuf[0] == '!' && ASCII_ISLOWER(st->lbuf[1])))
+	return findtags_hdr_parse(st);
+
+    // Headers ends.
+
+    // When there is no tag head, or ignoring case, need to do a
+    // linear search.
+    // When no "!_TAG_" is found, default to binary search.  If
+    // the tag file isn't sorted, the second loop will find it.
+    // When "!_TAG_FILE_SORTED" found: start binary search if
+    // flag set.
+    // For cscope, it's always linear.
+# ifdef FEAT_CSCOPE
+    if (st->linear || use_cscope)
+# else
+    if (st->linear)
+# endif
+	st->state = TS_LINEAR;
+    else if (st->tag_file_sorted == NUL)
+	st->state = TS_BINARY;
+    else if (st->tag_file_sorted == '1')
+	st->state = TS_BINARY;
+    else if (st->tag_file_sorted == '2')
+    {
+	st->state = TS_BINARY;
+	*sortic = TRUE;
+	st->orgpat->regmatch.rm_ic = (p_ic || !noic);
+    }
+    else
+	st->state = TS_LINEAR;
+
+    if (st->state == TS_BINARY && st->orgpat->regmatch.rm_ic && !*sortic)
+    {
+	// Binary search won't work for ignoring case, use linear
+	// search.
+	st->linear = TRUE;
+	st->state = TS_LINEAR;
+    }
+
+    // When starting a binary search, get the size of the file and
+    // compute the first offset.
+    if (st->state == TS_BINARY)
+    {
+	if (vim_fseek(st->fp, 0L, SEEK_END) != 0)
+	    // can't seek, don't use binary search
+	    st->state = TS_LINEAR;
+	else
+	{
+	    // Get the tag file size (don't use mch_fstat(), it's
+	    // not portable).  Don't use lseek(), it doesn't work
+	    // properly on MacOS Catalina.
+	    filesize = vim_ftell(st->fp);
+	    vim_ignored = vim_fseek(st->fp, 0L, SEEK_SET);
+
+	    // Calculate the first read offset in the file.  Start
+	    // the search in the middle of the file.
+	    sinfo_p->low_offset = 0;
+	    sinfo_p->low_char = 0;
+	    sinfo_p->high_offset = filesize;
+	    sinfo_p->curr_offset = 0;
+	    sinfo_p->high_char = 0xff;
+	}
+	return FALSE;
+    }
+
+    return TRUE;
+}
+
+/*
+ * Parse a tag line read from a tags file.
+ * Also compares the tag name in "tagpp->tagname" with a search pattern in
+ * "st->orgpat->head" as a quick check if the tag may match.
+ * Returns:
+ * - TAG_MATCH_SUCCESS if the tag may match
+ * - TAG_MATCH_FAIL if the tag doesn't match
+ * - TAG_MATCH_NEXT to look for the next matching tag (used in a binary search)
+ * - TAG_MATCH_STOP if all the tags are processed without a match. Uses the
+ *   values in "margs" for doing the comparison.
+ */
+    static tagmatch_status_T
+findtags_parse_line(
+    findtags_state_T		*st,
+    tagptrs_T			*tagpp,
+    findtags_match_args_T	*margs,
+    tagsearch_info_T		*sinfo_p)
+{
+    int		status;
+    int		i;
+    int		cmplen;
+    int		tagcmp;
+
+    // Figure out where the different strings are in this line.
+    // For "normal" tags: Do a quick check if the tag matches.
+    // This speeds up tag searching a lot!
+    if (st->orgpat->headlen
+#ifdef FEAT_EMACS_TAGS
+	    && !st->is_etag
+#endif
+       )
+    {
+	CLEAR_FIELD(*tagpp);
+	tagpp->tagname = st->lbuf;
+	tagpp->tagname_end = vim_strchr(st->lbuf, TAB);
+	if (tagpp->tagname_end == NULL)
+	    // Corrupted tag line.
+	    return TAG_MATCH_FAIL;
+
+	// Skip this line if the length of the tag is different and
+	// there is no regexp, or the tag is too short.
+	cmplen = (int)(tagpp->tagname_end - tagpp->tagname);
+	if (p_tl != 0 && cmplen > p_tl)	    // adjust for 'taglength'
+	    cmplen = p_tl;
+	if ((st->flags & TAG_REGEXP) && st->orgpat->headlen < cmplen)
+	    cmplen = st->orgpat->headlen;
+	else if (st->state == TS_LINEAR && st->orgpat->headlen != cmplen)
+	    return TAG_MATCH_NEXT;
+
+	if (st->state == TS_BINARY)
+	{
+	    // Simplistic check for unsorted tags file.
+	    i = (int)tagpp->tagname[0];
+	    if (margs->sortic)
+		i = (int)TOUPPER_ASC(tagpp->tagname[0]);
+	    if (i < sinfo_p->low_char || i > sinfo_p->high_char)
+		margs->sort_error = TRUE;
+
+	    // Compare the current tag with the searched tag.
+	    if (margs->sortic)
+		tagcmp = tag_strnicmp(tagpp->tagname, st->orgpat->head,
+							(size_t)cmplen);
+	    else
+		tagcmp = STRNCMP(tagpp->tagname, st->orgpat->head, cmplen);
+
+	    // A match with a shorter tag means to search forward.
+	    // A match with a longer tag means to search backward.
+	    if (tagcmp == 0)
+	    {
+		if (cmplen < st->orgpat->headlen)
+		    tagcmp = -1;
+		else if (cmplen > st->orgpat->headlen)
+		    tagcmp = 1;
+	    }
+
+	    if (tagcmp == 0)
+	    {
+		// We've located the tag, now skip back and search
+		// forward until the first matching tag is found.
+		st->state = TS_SKIP_BACK;
+		sinfo_p->match_offset = sinfo_p->curr_offset;
+		return TAG_MATCH_NEXT;
+	    }
+	    if (tagcmp < 0)
+	    {
+		sinfo_p->curr_offset = vim_ftell(st->fp);
+		if (sinfo_p->curr_offset < sinfo_p->high_offset)
+		{
+		    sinfo_p->low_offset = sinfo_p->curr_offset;
+		    if (margs->sortic)
+			sinfo_p->low_char = TOUPPER_ASC(tagpp->tagname[0]);
+		    else
+			sinfo_p->low_char = tagpp->tagname[0];
+		    return TAG_MATCH_NEXT;
+		}
+	    }
+	    if (tagcmp > 0 && sinfo_p->curr_offset != sinfo_p->high_offset)
+	    {
+		sinfo_p->high_offset = sinfo_p->curr_offset;
+		if (margs->sortic)
+		    sinfo_p->high_char = TOUPPER_ASC(tagpp->tagname[0]);
+		else
+		    sinfo_p->high_char = tagpp->tagname[0];
+		return TAG_MATCH_NEXT;
+	    }
+
+	    // No match yet and are at the end of the binary search.
+	    return TAG_MATCH_STOP;
+	}
+	else if (st->state == TS_SKIP_BACK)
+	{
+	    if (MB_STRNICMP(tagpp->tagname, st->orgpat->head, cmplen) != 0)
+		st->state = TS_STEP_FORWARD;
+	    else
+		// Have to skip back more.  Restore the curr_offset
+		// used, otherwise we get stuck at a long line.
+		sinfo_p->curr_offset = sinfo_p->curr_offset_used;
+	    return TAG_MATCH_NEXT;
+	}
+	else if (st->state == TS_STEP_FORWARD)
+	{
+	    if (MB_STRNICMP(tagpp->tagname, st->orgpat->head, cmplen) != 0)
+	    {
+		if ((off_T)vim_ftell(st->fp) > sinfo_p->match_offset)
+		    return TAG_MATCH_STOP;	// past last match
+		else
+		    return TAG_MATCH_NEXT;	// before first match
+	    }
+	}
+	else
+	    // skip this match if it can't match
+	    if (MB_STRNICMP(tagpp->tagname, st->orgpat->head, cmplen) != 0)
+		return TAG_MATCH_NEXT;
+
+	// Can be a matching tag, isolate the file name and command.
+	tagpp->fname = tagpp->tagname_end + 1;
+	tagpp->fname_end = vim_strchr(tagpp->fname, TAB);
+	if (tagpp->fname_end == NULL)
+	    status = FAIL;
+	else
+	{
+	    tagpp->command = tagpp->fname_end + 1;
+	    status = OK;
+	}
+    }
+    else
+	status = parse_tag_line(st->lbuf,
+#ifdef FEAT_EMACS_TAGS
+		st->is_etag,
+#endif
+		tagpp);
+
+    if (status == FAIL)
+	return TAG_MATCH_FAIL;
+
+#ifdef FEAT_EMACS_TAGS
+    if (st->is_etag)
+	tagpp->fname = st->ebuf;
+#endif
+
+    return TAG_MATCH_SUCCESS;
+}
+
+/*
+ * Initialize the structure used for tag matching.
+ */
+    static void
+findtags_matchargs_init(findtags_match_args_T *margs, int flags)
+{
+    margs->matchoff = 0;			// match offset
+    margs->match_re = FALSE;			// match with regexp
+    margs->match_no_ic = FALSE;			// matches with case
+    margs->has_re = (flags & TAG_REGEXP);	// regexp used
+    margs->sortic = FALSE;			// tag file sorted in nocase
+    margs->sort_error = FALSE;			// tags file not sorted
+}
+
+/*
+ * Compares the tag name in "tagpp->tagname" with a search pattern in
+ * "st->orgpat->pat".
+ * Returns TRUE if the tag matches, FALSE if the tag doesn't match.
+ * Uses the values in "margs" for doing the comparison.
+ */
+    static int
+findtags_match_tag(
+    findtags_state_T	*st,
+    tagptrs_T		*tagpp,
+    findtags_match_args_T *margs)
+{
+    int		match = FALSE;
+    int		cmplen;
+
+    // First try matching with the pattern literally (also when it is
+    // a regexp).
+    cmplen = (int)(tagpp->tagname_end - tagpp->tagname);
+    if (p_tl != 0 && cmplen > p_tl)	    // adjust for 'taglength'
+	cmplen = p_tl;
+    // if tag length does not match, don't try comparing
+    if (st->orgpat->len != cmplen)
+	match = FALSE;
+    else
+    {
+	if (st->orgpat->regmatch.rm_ic)
+	{
+	    match =
+		(MB_STRNICMP(tagpp->tagname, st->orgpat->pat, cmplen) == 0);
+	    if (match)
+		margs->match_no_ic =
+		    (STRNCMP(tagpp->tagname, st->orgpat->pat, cmplen) == 0);
+	}
+	else
+	    match = (STRNCMP(tagpp->tagname, st->orgpat->pat, cmplen) == 0);
+    }
+
+    // Has a regexp: Also find tags matching regexp.
+    margs->match_re = FALSE;
+    if (!match && st->orgpat->regmatch.regprog != NULL)
+    {
+	int	cc;
+
+	cc = *tagpp->tagname_end;
+	*tagpp->tagname_end = NUL;
+	match = vim_regexec(&st->orgpat->regmatch, tagpp->tagname, (colnr_T)0);
+	if (match)
+	{
+	    margs->matchoff = (int)(st->orgpat->regmatch.startp[0] -
+							tagpp->tagname);
+	    if (st->orgpat->regmatch.rm_ic)
+	    {
+		st->orgpat->regmatch.rm_ic = FALSE;
+		margs->match_no_ic = vim_regexec(&st->orgpat->regmatch,
+			tagpp->tagname, (colnr_T)0);
+		st->orgpat->regmatch.rm_ic = TRUE;
+	    }
+	}
+	*tagpp->tagname_end = cc;
+	margs->match_re = TRUE;
+    }
+
+    return match;
+}
+
+/*
+ * Convert the encoding of a line read from a tags file in "st->lbuf".
+ * Converting the pattern from 'enc' to the tags file encoding doesn't work,
+ * because characters are not recognized. The converted line is saved in
+ * st->lbuf.
+ */
+    static void
+findtags_string_convert(findtags_state_T *st)
+{
+    char_u	*conv_line;
+    int		len;
+
+    conv_line = string_convert(&st->vimconv, st->lbuf, NULL);
+    if (conv_line == NULL)
+	return;
+
+    // Copy or swap lbuf and conv_line.
+    len = (int)STRLEN(conv_line) + 1;
+    if (len > st->lbuf_size)
+    {
+	vim_free(st->lbuf);
+	st->lbuf = conv_line;
+	st->lbuf_size = len;
+    }
+    else
+    {
+	STRCPY(st->lbuf, conv_line);
+	vim_free(conv_line);
+    }
+}
+
+/*
+ * Add a matching tag found in a tags file to st->ht_match and st->ga_match.
+ * Returns OK if successfully added the match and FAIL on memory allocation
+ * failure.
+ */
+    static int
+findtags_add_match(
+    findtags_state_T	*st,
+    tagptrs_T		*tagpp,
+    findtags_match_args_T   *margs,
+    char_u		*buf_ffname,
+    hash_T		*hash)
+{
+#ifdef FEAT_CSCOPE
+    int		use_cscope = (st->flags & TAG_CSCOPE);
+#endif
+    int		name_only = (st->flags & TAG_NAMES);
+    int		mtt;
+    int		len = 0;
+    int		is_current;		// file name matches
+    int		is_static;		// current tag line is static
+    char_u	*mfp;
+    char_u	*p;
+    char_u	*s;
+
+#ifdef FEAT_CSCOPE
+    if (use_cscope)
+    {
+	// Don't change the ordering, always use the same table.
+	mtt = MT_GL_OTH;
+    }
+    else
+#endif
+    {
+	// Decide in which array to store this match.
+	is_current = test_for_current(
+#ifdef FEAT_EMACS_TAGS
+		st->is_etag,
+#endif
+		tagpp->fname, tagpp->fname_end, st->tag_fname, buf_ffname);
+#ifdef FEAT_EMACS_TAGS
+	is_static = FALSE;
+	if (!st->is_etag)	// emacs tags are never static
+#endif
+	    is_static = test_for_static(tagpp);
+
+	// decide in which of the sixteen tables to store this
+	// match
+	if (is_static)
+	{
+	    if (is_current)
+		mtt = MT_ST_CUR;
+	    else
+		mtt = MT_ST_OTH;
+	}
+	else
+	{
+	    if (is_current)
+		mtt = MT_GL_CUR;
+	    else
+		mtt = MT_GL_OTH;
+	}
+	if (st->orgpat->regmatch.rm_ic && !margs->match_no_ic)
+	    mtt += MT_IC_OFF;
+	if (margs->match_re)
+	    mtt += MT_RE_OFF;
+    }
+
+    // Add the found match in ht_match[mtt] and ga_match[mtt].
+    // Store the info we need later, which depends on the kind of
+    // tags we are dealing with.
+    if (st->help_only)
+    {
+#ifdef FEAT_MULTI_LANG
+# define ML_EXTRA 3
+#else
+# define ML_EXTRA 0
+#endif
+	// Append the help-heuristic number after the tagname, for
+	// sorting it later.  The heuristic is ignored for
+	// detecting duplicates.
+	// The format is {tagname}@{lang}NUL{heuristic}NUL
+	*tagpp->tagname_end = NUL;
+	len = (int)(tagpp->tagname_end - tagpp->tagname);
+	mfp = alloc(sizeof(char_u) + len + 10 + ML_EXTRA + 1);
+	if (mfp != NULL)
+	{
+	    int heuristic;
+
+	    p = mfp;
+	    STRCPY(p, tagpp->tagname);
+#ifdef FEAT_MULTI_LANG
+	    p[len] = '@';
+	    STRCPY(p + len + 1, st->help_lang);
+#endif
+
+	    heuristic = help_heuristic(tagpp->tagname,
+				margs->match_re ? margs->matchoff : 0,
+				!margs->match_no_ic);
+#ifdef FEAT_MULTI_LANG
+	    heuristic += st->help_pri;
+#endif
+	    sprintf((char *)p + len + 1 + ML_EXTRA, "%06d",
+		    heuristic);
+	}
+	*tagpp->tagname_end = TAB;
+    }
+    else if (name_only)
+    {
+	if (st->get_searchpat)
+	{
+	    char_u *temp_end = tagpp->command;
+
+	    if (*temp_end == '/')
+		while (*temp_end && *temp_end != '\r'
+			&& *temp_end != '\n'
+			&& *temp_end != '$')
+		    temp_end++;
+
+	    if (tagpp->command + 2 < temp_end)
+	    {
+		len = (int)(temp_end - tagpp->command - 2);
+		mfp = alloc(len + 2);
+		if (mfp != NULL)
+		    vim_strncpy(mfp, tagpp->command + 2, len);
+	    }
+	    else
+		mfp = NULL;
+	    st->get_searchpat = FALSE;
+	}
+	else
+	{
+	    len = (int)(tagpp->tagname_end - tagpp->tagname);
+	    mfp = alloc(sizeof(char_u) + len + 1);
+	    if (mfp != NULL)
+		vim_strncpy(mfp, tagpp->tagname, len);
+
+	    // if wanted, re-read line to get long form too
+	    if (State & MODE_INSERT)
+		st->get_searchpat = p_sft;
+	}
+    }
+    else
+    {
+	size_t tag_fname_len = STRLEN(st->tag_fname);
+#ifdef FEAT_EMACS_TAGS
+	size_t ebuf_len = 0;
+#endif
+
+	// Save the tag in a buffer.
+	// Use 0x02 to separate fields (Can't use NUL because the
+	// hash key is terminated by NUL, or Ctrl_A because that is
+	// part of some Emacs tag files -- see parse_tag_line).
+	// Emacs tag: <mtt><tag_fname><0x02><ebuf><0x02><lbuf><NUL>
+	// other tag: <mtt><tag_fname><0x02><0x02><lbuf><NUL>
+	// without Emacs tags: <mtt><tag_fname><0x02><lbuf><NUL>
+	// Here <mtt> is the "mtt" value plus 1 to avoid NUL.
+	len = (int)tag_fname_len + (int)STRLEN(st->lbuf) + 3;
+#ifdef FEAT_EMACS_TAGS
+	if (st->is_etag)
+	{
+	    ebuf_len = STRLEN(st->ebuf);
+	    len += (int)ebuf_len + 1;
+	}
+	else
+	    ++len;
+#endif
+	mfp = alloc(sizeof(char_u) + len + 1);
+	if (mfp != NULL)
+	{
+	    p = mfp;
+	    p[0] = mtt + 1;
+	    STRCPY(p + 1, st->tag_fname);
+#ifdef BACKSLASH_IN_FILENAME
+	    // Ignore differences in slashes, avoid adding
+	    // both path/file and path\file.
+	    slash_adjust(p + 1);
+#endif
+	    p[tag_fname_len + 1] = TAG_SEP;
+	    s = p + 1 + tag_fname_len + 1;
+#ifdef FEAT_EMACS_TAGS
+	    if (st->is_etag)
+	    {
+		STRCPY(s, st->ebuf);
+		s[ebuf_len] = TAG_SEP;
+		s += ebuf_len + 1;
+	    }
+	    else
+		*s++ = TAG_SEP;
+#endif
+	    STRCPY(s, st->lbuf);
+	}
+    }
+
+    if (mfp != NULL)
+    {
+	hashitem_T	*hi;
+
+	// Don't add identical matches.
+	// Add all cscope tags, because they are all listed.
+	// "mfp" is used as a hash key, there is a NUL byte to end
+	// the part that matters for comparing, more bytes may
+	// follow after it.  E.g. help tags store the priority
+	// after the NUL.
+#ifdef FEAT_CSCOPE
+	if (use_cscope)
+	    ++*hash;
+	else
+#endif
+	    *hash = hash_hash(mfp);
+	hi = hash_lookup(&st->ht_match[mtt], mfp, *hash);
+	if (HASHITEM_EMPTY(hi))
+	{
+	    if (hash_add_item(&st->ht_match[mtt], hi, mfp, *hash) == FAIL
+		    || ga_grow(&st->ga_match[mtt], 1) == FAIL)
+	    {
+		// Out of memory! Just forget about the rest.
+		st->stop_searching = TRUE;
+		return FAIL;
+	    }
+
+	    ((char_u **)(st->ga_match[mtt].ga_data))
+		[st->ga_match[mtt].ga_len++] = mfp;
+	    st->match_count++;
+	}
+	else
+	    // duplicate tag, drop it
+	    vim_free(mfp);
+    }
+
+    return OK;
+}
+
+/*
+ * Read and get all the tags from file st->tag_fname.
+ * Sets "st->stop_searching" to TRUE to stop searching for additional tags.
+ */
+    static void
+findtags_get_all_tags(
+    findtags_state_T		*st,
+    findtags_match_args_T	*margs,
+    char_u			*buf_ffname)
+{
+    tagptrs_T		tagp;
+    tagsearch_info_T	search_info;
+    int			retval;
+#ifdef FEAT_CSCOPE
+    int			use_cscope = (st->flags & TAG_CSCOPE);
+#endif
+    hash_T		hash = 0;
+
+    // This is only to avoid a compiler warning for using search_info
+    // uninitialised.
+    CLEAR_FIELD(search_info);
+
+    // Read and parse the lines in the file one by one
+    for (;;)
+    {
+	// check for CTRL-C typed, more often when jumping around
+	if (st->state == TS_BINARY || st->state == TS_SKIP_BACK)
+	    line_breakcheck();
+	else
+	    fast_breakcheck();
+	if ((st->flags & TAG_INS_COMP))	// Double brackets for gcc
+	    ins_compl_check_keys(30, FALSE);
+	if (got_int || ins_compl_interrupted())
+	{
+	    st->stop_searching = TRUE;
+	    break;
+	}
+	// When mincount is TAG_MANY, stop when enough matches have been
+	// found (for completion).
+	if (st->mincount == TAG_MANY && st->match_count >= TAG_MANY)
+	{
+	    st->stop_searching = TRUE;
+	    break;
+	}
+	if (st->get_searchpat)
+	    goto line_read_in;
+
+	retval = findtags_get_next_line(st, &search_info);
+	if (retval == TAGS_READ_IGNORE)
+	    continue;
+	if (retval == TAGS_READ_EOF)
+	    break;
+
+line_read_in:
+
+	if (st->vimconv.vc_type != CONV_NONE)
+	    findtags_string_convert(st);
+
+#ifdef FEAT_EMACS_TAGS
+	// Emacs tags line with CTRL-L: New file name on next line.
+	// The file name is followed by a ','.
+	// Remember etag file name in ebuf.
+	if (*st->lbuf == Ctrl_L
+# ifdef FEAT_CSCOPE
+		&& !use_cscope
+# endif
+	   )
+	{
+	    st->is_etag = TRUE;		// in case at the start
+	    st->state = TS_LINEAR;
+	    emacs_tags_new_filename(st);
+	    continue;
+	}
+#endif
+
+	// When still at the start of the file, check for Emacs tags file
+	// format, and for "not sorted" flag.
+	if (st->state == TS_START)
+	{
+	    if (findtags_start_state_handler(st, &margs->sortic, &search_info) == FALSE)
+		continue;
+	}
+
+	// When the line is too long the NUL will not be in the
+	// last-but-one byte (see vim_fgets()).
+	// Has been reported for Mozilla JS with extremely long names.
+	// In that case we need to increase lbuf_size.
+	if (st->lbuf[st->lbuf_size - 2] != NUL
+#ifdef FEAT_CSCOPE
+		&& !use_cscope
+#endif
+	   )
+	{
+	    st->lbuf_size *= 2;
+	    vim_free(st->lbuf);
+	    st->lbuf = alloc(st->lbuf_size);
+	    if (st->lbuf == NULL)
+	    {
+		if (st->fp != NULL)
+		    fclose(st->fp);
+		st->fp = NULL;
+		st->stop_searching = TRUE;
+		return;
+	    }
+
+	    if (st->state == TS_STEP_FORWARD || st->state == TS_LINEAR)
+		// Seek to the same position to read the same line again
+		vim_ignored = vim_fseek(st->fp, search_info.curr_offset,
+								     SEEK_SET);
+	    // this will try the same thing again, make sure the offset is
+	    // different
+	    search_info.curr_offset = 0;
+	    continue;
+	}
+
+	retval = findtags_parse_line(st, &tagp, margs, &search_info);
+	if (retval == TAG_MATCH_NEXT)
+	    continue;
+	if (retval == TAG_MATCH_STOP)
+	    break;
+	if (retval == TAG_MATCH_FAIL)
+	{
+	    semsg(_(e_format_error_in_tags_file_str), st->tag_fname);
+#ifdef FEAT_CSCOPE
+	    if (!use_cscope)
+#endif
+		semsg(_("Before byte %ld"), (long)vim_ftell(st->fp));
+	    st->stop_searching = TRUE;
+	    return;
+	}
+
+	// If a match is found, add it to ht_match[] and ga_match[].
+	if (findtags_match_tag(st, &tagp, margs))
+	{
+	    if (findtags_add_match(st, &tagp, margs, buf_ffname, &hash)
+								== FAIL)
+		break;
+	}
+    } // forever
+}
+
+/*
+ * Search for tags matching "st->orgpat->pat" in the "st->tag_fname" tags file.
+ * Information needed to search for the tags is in the "st" state structure.
+ * The matching tags are returned in "st". If an error is encountered, then
+ * "st->stop_searching" is set to TRUE.
+ */
+    static void
+findtags_in_file(findtags_state_T *st, char_u *buf_ffname)
+{
+    findtags_match_args_T margs;
+#ifdef FEAT_CSCOPE
+    int		use_cscope = (st->flags & TAG_CSCOPE);
+#endif
+
+    st->vimconv.vc_type = CONV_NONE;
+    st->tag_file_sorted = NUL;
+    st->fp = NULL;
+    findtags_matchargs_init(&margs, st->flags);
+
+    // A file that doesn't exist is silently ignored.  Only when not a
+    // single file is found, an error message is given (further on).
+#ifdef FEAT_CSCOPE
+    if (use_cscope)
+	st->fp = NULL;	    // avoid GCC warning
+    else
+#endif
+    {
+#ifdef FEAT_MULTI_LANG
+	if (curbuf->b_help)
+	{
+	    if (!findtags_in_help_init(st))
+		return;
+	}
+#endif
+
+	st->fp = mch_fopen((char *)st->tag_fname, "r");
+	if (st->fp == NULL)
+	    return;
+
+	if (p_verbose >= 5)
+	{
+	    verbose_enter();
+	    smsg(_("Searching tags file %s"), st->tag_fname);
+	    verbose_leave();
+	}
+    }
+    st->did_open = TRUE;	// remember that we found at least one file
+
+    st->state = TS_START;	// we're at the start of the file
+#ifdef FEAT_EMACS_TAGS
+    st->is_etag = FALSE;	// default is: not emacs style
+#endif
+
+    // Read and parse the lines in the file one by one
+    findtags_get_all_tags(st, &margs, buf_ffname);
+
+    if (st->fp != NULL)
+    {
+	fclose(st->fp);
+	st->fp = NULL;
+    }
+#ifdef FEAT_EMACS_TAGS
+    emacs_tags_incstack_free();
+#endif
+    if (st->vimconv.vc_type != CONV_NONE)
+	convert_setup(&st->vimconv, NULL, NULL);
+
+    if (margs.sort_error)
+	semsg(_(e_tags_file_not_sorted_str), st->tag_fname);
+
+    // Stop searching if sufficient tags have been found.
+    if (st->match_count >= st->mincount)
+	st->stop_searching = TRUE;
+}
+
+/*
+ * Copy the tags found by find_tags() to "matchesp".
+ * Returns the number of matches copied.
+ */
+    static int
+findtags_copy_matches(findtags_state_T *st, char_u ***matchesp)
+{
+    int		name_only = (st->flags & TAG_NAMES);
+    char_u	**matches;
+    int		mtt;
+    int		i;
+    char_u	*mfp;
+    char_u	*p;
+
+    if (st->match_count > 0)
+	matches = ALLOC_MULT(char_u *, st->match_count);
+    else
+	matches = NULL;
+    st->match_count = 0;
+    for (mtt = 0; mtt < MT_COUNT; ++mtt)
+    {
+	for (i = 0; i < st->ga_match[mtt].ga_len; ++i)
+	{
+	    mfp = ((char_u **)(st->ga_match[mtt].ga_data))[i];
+	    if (matches == NULL)
+		vim_free(mfp);
+	    else
+	    {
+		if (!name_only)
+		{
+		    // Change mtt back to zero-based.
+		    *mfp = *mfp - 1;
+
+		    // change the TAG_SEP back to NUL
+		    for (p = mfp + 1; *p != NUL; ++p)
+			if (*p == TAG_SEP)
+			    *p = NUL;
+		}
+		matches[st->match_count++] = mfp;
+	    }
+	}
+
+	ga_clear(&st->ga_match[mtt]);
+	hash_clear(&st->ht_match[mtt]);
+    }
+
+    *matchesp = matches;
+    return st->match_count;
+}
+
+/*
+ * find_tags() - search for tags in tags files
+ *
+ * Return FAIL if search completely failed (*num_matches will be 0, *matchesp
+ * will be NULL), OK otherwise.
+ *
+ * Priority depending on which type of tag is recognized:
+ *  6.	A static or global tag with a full matching tag for the current file.
+ *  5.	A global tag with a full matching tag for another file.
+ *  4.	A static tag with a full matching tag for another file.
+ *  3.	A static or global tag with an ignore-case matching tag for the
+ *	current file.
+ *  2.	A global tag with an ignore-case matching tag for another file.
+ *  1.	A static tag with an ignore-case matching tag for another file.
+ *
+ * Tags in an emacs-style tags file are always global.
+ *
+ * flags:
+ * TAG_HELP	  only search for help tags
+ * TAG_NAMES	  only return name of tag
+ * TAG_REGEXP	  use "pat" as a regexp
+ * TAG_NOIC	  don't always ignore case
+ * TAG_KEEP_LANG  keep language
+ * TAG_CSCOPE	  use cscope results for tags
+ * TAG_NO_TAGFUNC do not call the 'tagfunc' function
+ */
+    int
+find_tags(
+    char_u	*pat,			// pattern to search for
+    int		*num_matches,		// return: number of matches found
+    char_u	***matchesp,		// return: array of matches found
+    int		flags,
+    int		mincount,		// MAXCOL: find all matches
+					// other: minimal number of matches
+    char_u	*buf_ffname)		// name of buffer for priority
+{
+    findtags_state_T	st;
+    tagname_T	tn;			// info for get_tagfname()
+    int		first_file;		// trying first tag file
+    int		retval = FAIL;		// return value
+    int		round;
+
+    int		save_emsg_off;
+
+    int		help_save;
+#ifdef FEAT_MULTI_LANG
+    int		i;
+    char_u	*saved_pat = NULL;		// copy of pat[]
+#endif
+
+    int		findall = (mincount == MAXCOL || mincount == TAG_MANY);
+						// find all matching tags
+    int		has_re = (flags & TAG_REGEXP);	// regexp used
+    int		noic = (flags & TAG_NOIC);
+#ifdef FEAT_CSCOPE
+    int		use_cscope = (flags & TAG_CSCOPE);
+#endif
+    int		verbose = (flags & TAG_VERBOSE);
+    int		save_p_ic = p_ic;
+
+    /*
+     * Change the value of 'ignorecase' according to 'tagcase' for the
+     * duration of this function.
+     */
+    switch (curbuf->b_tc_flags ? curbuf->b_tc_flags : tc_flags)
+    {
+	case TC_FOLLOWIC:		 break;
+	case TC_IGNORE:    p_ic = TRUE;  break;
+	case TC_MATCH:     p_ic = FALSE; break;
+	case TC_FOLLOWSCS: p_ic = ignorecase(pat); break;
+	case TC_SMART:     p_ic = ignorecase_opt(pat, TRUE, TRUE); break;
+    }
+
+    help_save = curbuf->b_help;
+
+    if (findtags_state_init(&st, pat, flags, mincount) == FAIL)
+	goto findtag_end;
+
+#ifdef FEAT_CSCOPE
+    STRCPY(st.tag_fname, "from cscope");	// for error messages
+#endif
+
+    /*
+     * Initialize a few variables
+     */
+    if (st.help_only)				// want tags from help file
+	curbuf->b_help = TRUE;			// will be restored later
+#ifdef FEAT_CSCOPE
+    else if (use_cscope)
+    {
+	// Make sure we don't mix help and cscope, confuses Coverity.
+	st.help_only = FALSE;
+	curbuf->b_help = FALSE;
+    }
+#endif
+
+#ifdef FEAT_MULTI_LANG
+    if (curbuf->b_help)
+    {
+	// When "@ab" is specified use only the "ab" language, otherwise
+	// search all languages.
+	if (st.orgpat->len > 3 && pat[st.orgpat->len - 3] == '@'
+				&& ASCII_ISALPHA(pat[st.orgpat->len - 2])
+				&& ASCII_ISALPHA(pat[st.orgpat->len - 1]))
+	{
+	    saved_pat = vim_strnsave(pat, st.orgpat->len - 3);
+	    if (saved_pat != NULL)
+	    {
+		st.help_lang_find = &pat[st.orgpat->len - 2];
+		st.orgpat->pat = saved_pat;
+		st.orgpat->len -= 3;
+	    }
+	}
+    }
+#endif
+    if (p_tl != 0 && st.orgpat->len > p_tl)	// adjust for 'taglength'
+	st.orgpat->len = p_tl;
+
+    save_emsg_off = emsg_off;
+    emsg_off = TRUE;  // don't want error for invalid RE here
+    prepare_pats(st.orgpat, has_re);
+    emsg_off = save_emsg_off;
+    if (has_re && st.orgpat->regmatch.regprog == NULL)
+	goto findtag_end;
+
+#ifdef FEAT_EVAL
+    retval = findtags_apply_tfu(&st, pat, buf_ffname);
+    if (retval != NOTDONE)
+	goto findtag_end;
+
+    // re-initialize the default return value
+    retval = FAIL;
+#endif
+
+#ifdef FEAT_MULTI_LANG
+    // Set a flag if the file extension is .txt
+    if ((flags & TAG_KEEP_LANG)
+	    && st.help_lang_find == NULL
+	    && curbuf->b_fname != NULL
+	    && (i = (int)STRLEN(curbuf->b_fname)) > 4
+	    && STRICMP(curbuf->b_fname + i - 4, ".txt") == 0)
+	st.is_txt = TRUE;
+#endif
+
+    /*
+     * When finding a specified number of matches, first try with matching
+     * case, so binary search can be used, and try ignore-case matches in a
+     * second loop.
+     * When finding all matches, 'tagbsearch' is off, or there is no fixed
+     * string to look for, ignore case right away to avoid going though the
+     * tags files twice.
+     * When the tag file is case-fold sorted, it is either one or the other.
+     * Only ignore case when TAG_NOIC not used or 'ignorecase' set.
+     */
+    st.orgpat->regmatch.rm_ic = ((p_ic || !noic)
+			&& (findall || st.orgpat->headlen == 0 || !p_tbs));
+    for (round = 1; round <= 2; ++round)
+    {
+	st.linear = (st.orgpat->headlen == 0 || !p_tbs || round == 2);
+
+      /*
+       * Try tag file names from tags option one by one.
+       */
+      for (first_file = TRUE;
+#ifdef FEAT_CSCOPE
+	    use_cscope ||
+#endif
+		get_tagfname(&tn, first_file, st.tag_fname) == OK;
+							   first_file = FALSE)
+      {
+	  findtags_in_file(&st, buf_ffname);
+	  if (st.stop_searching
+#ifdef FEAT_CSCOPE
+		  || use_cscope
+#endif
+	     )
+	  {
+	      retval = OK;
+	      break;
+	  }
+      } // end of for-each-file loop
+
+#ifdef FEAT_CSCOPE
+	if (!use_cscope)
+#endif
+	    tagname_free(&tn);
+
+	// stop searching when already did a linear search, or when TAG_NOIC
+	// used, and 'ignorecase' not set or already did case-ignore search
+	if (st.stop_searching || st.linear || (!p_ic && noic) ||
+						st.orgpat->regmatch.rm_ic)
+	    break;
+# ifdef FEAT_CSCOPE
+	if (use_cscope)
+	    break;
+# endif
+
+	// try another time while ignoring case
+	st.orgpat->regmatch.rm_ic = TRUE;
+    }
+
+    if (!st.stop_searching)
+    {
+	if (!st.did_open && verbose)	// never opened any tags file
+	    emsg(_(e_no_tags_file));
+	retval = OK;		// It's OK even when no tag found
+    }
+
+findtag_end:
+    findtags_state_free(&st);
+
+    /*
+     * Move the matches from the ga_match[] arrays into one list of
+     * matches.  When retval == FAIL, free the matches.
+     */
+    if (retval == FAIL)
+	st.match_count = 0;
+
+    *num_matches = findtags_copy_matches(&st, matchesp);
+
+    curbuf->b_help = help_save;
+#ifdef FEAT_MULTI_LANG
+    vim_free(saved_pat);
+#endif
+
+    p_ic = save_p_ic;
+
+    return retval;
+}
+
+static garray_T tag_fnames = GA_EMPTY;
+
+/*
+ * Callback function for finding all "tags" and "tags-??" files in
+ * 'runtimepath' doc directories.
+ */
+    static void
+found_tagfile_cb(char_u *fname, void *cookie UNUSED)
+{
+    if (ga_grow(&tag_fnames, 1) == FAIL)
+	return;
+
+    char_u	*tag_fname = vim_strsave(fname);
+
+#ifdef BACKSLASH_IN_FILENAME
+    slash_adjust(tag_fname);
+#endif
+    simplify_filename(tag_fname);
+    ((char_u **)(tag_fnames.ga_data))[tag_fnames.ga_len++] = tag_fname;
+}
+
+#if defined(EXITFREE) || defined(PROTO)
+    void
+free_tag_stuff(void)
+{
+    ga_clear_strings(&tag_fnames);
+    if (curwin != NULL)
+	do_tag(NULL, DT_FREE, 0, 0, 0);
+    tag_freematch();
+
+# if defined(FEAT_QUICKFIX)
+    tagstack_clear_entry(&ptag_entry);
+# endif
+}
+#endif
+
+/*
+ * Get the next name of a tag file from the tag file list.
+ * For help files, use "tags" file only.
+ *
+ * Return FAIL if no more tag file names, OK otherwise.
+ */
+    int
+get_tagfname(
+    tagname_T	*tnp,	// holds status info
+    int		first,	// TRUE when first file name is wanted
+    char_u	*buf)	// pointer to buffer of MAXPATHL chars
+{
+    char_u		*fname = NULL;
+    char_u		*r_ptr;
+    int			i;
+
+    if (first)
+	CLEAR_POINTER(tnp);
+
+    if (curbuf->b_help)
+    {
+	/*
+	 * For help files it's done in a completely different way:
+	 * Find "doc/tags" and "doc/tags-??" in all directories in
+	 * 'runtimepath'.
+	 */
+	if (first)
+	{
+	    ga_clear_strings(&tag_fnames);
+	    ga_init2(&tag_fnames, sizeof(char_u *), 10);
+	    do_in_runtimepath((char_u *)
+#ifdef FEAT_MULTI_LANG
+# ifdef VMS
+		    // Functions decc$to_vms() and decc$translate_vms() crash
+		    // on some VMS systems with wildcards "??".  Seems ECO
+		    // patches do fix the problem in C RTL, but we can't use
+		    // an #ifdef for that.
+		    "doc/tags doc/tags-*"
+# else
+		    "doc/tags doc/tags-??"
+# endif
+#else
+		    "doc/tags"
+#endif
+					   , DIP_ALL, found_tagfile_cb, NULL);
+	}
+
+	if (tnp->tn_hf_idx >= tag_fnames.ga_len)
+	{
+	    // Not found in 'runtimepath', use 'helpfile', if it exists and
+	    // wasn't used yet, replacing "help.txt" with "tags".
+	    if (tnp->tn_hf_idx > tag_fnames.ga_len || *p_hf == NUL)
+		return FAIL;
+	    ++tnp->tn_hf_idx;
+	    STRCPY(buf, p_hf);
+	    STRCPY(gettail(buf), "tags");
+#ifdef BACKSLASH_IN_FILENAME
+	    slash_adjust(buf);
+#endif
+	    simplify_filename(buf);
+
+	    for (i = 0; i < tag_fnames.ga_len; ++i)
+		if (STRCMP(buf, ((char_u **)(tag_fnames.ga_data))[i]) == 0)
+		    return FAIL; // avoid duplicate file names
+	}
+	else
+	    vim_strncpy(buf, ((char_u **)(tag_fnames.ga_data))[
+					     tnp->tn_hf_idx++], MAXPATHL - 1);
+	return OK;
+    }
+
+    if (first)
+    {
+	// Init.  We make a copy of 'tags', because autocommands may change
+	// the value without notifying us.
+	tnp->tn_tags = vim_strsave((*curbuf->b_p_tags != NUL)
+						 ? curbuf->b_p_tags : p_tags);
+	if (tnp->tn_tags == NULL)
+	    return FAIL;
+	tnp->tn_np = tnp->tn_tags;
+    }
+
+    /*
+     * Loop until we have found a file name that can be used.
+     * There are two states:
+     * tnp->tn_did_filefind_init == FALSE: setup for next part in 'tags'.
+     * tnp->tn_did_filefind_init == TRUE: find next file in this part.
+     */
+    for (;;)
+    {
+	if (tnp->tn_did_filefind_init)
+	{
+	    fname = vim_findfile(tnp->tn_search_ctx);
+	    if (fname != NULL)
+		break;
+
+	    tnp->tn_did_filefind_init = FALSE;
+	}
+	else
+	{
+	    char_u  *filename = NULL;
+
+	    // Stop when used all parts of 'tags'.
+	    if (*tnp->tn_np == NUL)
+	    {
+		vim_findfile_cleanup(tnp->tn_search_ctx);
+		tnp->tn_search_ctx = NULL;
+		return FAIL;
+	    }
+
+	    /*
+	     * Copy next file name into buf.
+	     */
+	    buf[0] = NUL;
+	    (void)copy_option_part(&tnp->tn_np, buf, MAXPATHL - 1, " ,");
+
+	    r_ptr = vim_findfile_stopdir(buf);
+	    // move the filename one char forward and truncate the
+	    // filepath with a NUL
+	    filename = gettail(buf);
+	    STRMOVE(filename + 1, filename);
+	    *filename++ = NUL;
+
+	    tnp->tn_search_ctx = vim_findfile_init(buf, filename,
+		    r_ptr, 100,
+		    FALSE,	   // don't free visited list
+		    FINDFILE_FILE, // we search for a file
+		    tnp->tn_search_ctx, TRUE, curbuf->b_ffname);
+	    if (tnp->tn_search_ctx != NULL)
+		tnp->tn_did_filefind_init = TRUE;
+	}
+    }
+
+    STRCPY(buf, fname);
+    vim_free(fname);
+    return OK;
+}
+
+/*
+ * Free the contents of a tagname_T that was filled by get_tagfname().
+ */
+    void
+tagname_free(tagname_T *tnp)
+{
+    vim_free(tnp->tn_tags);
+    vim_findfile_cleanup(tnp->tn_search_ctx);
+    tnp->tn_search_ctx = NULL;
+    ga_clear_strings(&tag_fnames);
+}
+
+/*
+ * Parse one line from the tags file. Find start/end of tag name, start/end of
+ * file name and start of search pattern.
+ *
+ * If is_etag is TRUE, tagp->fname and tagp->fname_end are not set.
+ *
+ * Return FAIL if there is a format error in this line, OK otherwise.
+ */
+    static int
+parse_tag_line(
+    char_u	*lbuf,		// line to be parsed
+#ifdef FEAT_EMACS_TAGS
+    int		is_etag,
+#endif
+    tagptrs_T	*tagp)
+{
+    char_u	*p;
+
+#ifdef FEAT_EMACS_TAGS
+    if (is_etag)
+	// emacs-style tag file
+	return emacs_tags_parse_line(lbuf, tagp);
+#endif
+
+    // Isolate the tagname, from lbuf up to the first white
+    tagp->tagname = lbuf;
+    p = vim_strchr(lbuf, TAB);
+    if (p == NULL)
+	return FAIL;
+    tagp->tagname_end = p;
+
+    // Isolate file name, from first to second white space
+    if (*p != NUL)
+	++p;
+    tagp->fname = p;
+    p = vim_strchr(p, TAB);
+    if (p == NULL)
+	return FAIL;
+    tagp->fname_end = p;
+
+    // find start of search command, after second white space
+    if (*p != NUL)
+	++p;
+    if (*p == NUL)
+	return FAIL;
+    tagp->command = p;
+
+    return OK;
+}
+
+/*
+ * Check if tagname is a static tag
+ *
+ * Static tags produced by the older ctags program have the format:
+ *	'file:tag  file  /pattern'.
+ * This is only recognized when both occurrence of 'file' are the same, to
+ * avoid recognizing "string::string" or ":exit".
+ *
+ * Static tags produced by the new ctags program have the format:
+ *	'tag  file  /pattern/;"<Tab>file:'	    "
+ *
+ * Return TRUE if it is a static tag and adjust *tagname to the real tag.
+ * Return FALSE if it is not a static tag.
+ */
+    static int
+test_for_static(tagptrs_T *tagp)
+{
+    char_u	*p;
+
+    /*
+     * Check for new style static tag ":...<Tab>file:[<Tab>...]"
+     */
+    p = tagp->command;
+    while ((p = vim_strchr(p, '\t')) != NULL)
+    {
+	++p;
+	if (STRNCMP(p, "file:", 5) == 0)
+	    return TRUE;
+    }
+
+    return FALSE;
+}
+
+/*
+ * Returns the length of a matching tag line.
+ */
+    static size_t
+matching_line_len(char_u *lbuf)
+{
+    char_u	*p = lbuf + 1;
+
+    // does the same thing as parse_match()
+    p += STRLEN(p) + 1;
+#ifdef FEAT_EMACS_TAGS
+    p += STRLEN(p) + 1;
+#endif
+    return (p - lbuf) + STRLEN(p);
+}
+
+/*
+ * Parse a line from a matching tag.  Does not change the line itself.
+ *
+ * The line that we get looks like this:
+ * Emacs tag: <mtt><tag_fname><NUL><ebuf><NUL><lbuf>
+ * other tag: <mtt><tag_fname><NUL><NUL><lbuf>
+ * without Emacs tags: <mtt><tag_fname><NUL><lbuf>
+ *
+ * Return OK or FAIL.
+ */
+    static int
+parse_match(
+    char_u	*lbuf,	    // input: matching line
+    tagptrs_T	*tagp)	    // output: pointers into the line
+{
+    int		retval;
+    char_u	*p;
+    char_u	*pc, *pt;
+
+    tagp->tag_fname = lbuf + 1;
+    lbuf += STRLEN(tagp->tag_fname) + 2;
+#ifdef FEAT_EMACS_TAGS
+    if (*lbuf)
+    {
+	tagp->is_etag = TRUE;
+	tagp->fname = lbuf;
+	lbuf += STRLEN(lbuf);
+	tagp->fname_end = lbuf++;
+    }
+    else
+    {
+	tagp->is_etag = FALSE;
+	++lbuf;
+    }
+#endif
+
+    // Find search pattern and the file name for non-etags.
+    retval = parse_tag_line(lbuf,
+#ifdef FEAT_EMACS_TAGS
+			tagp->is_etag,
+#endif
+			tagp);
+
+    tagp->tagkind = NULL;
+    tagp->user_data = NULL;
+    tagp->tagline = 0;
+    tagp->command_end = NULL;
+
+    if (retval != OK)
+	return retval;
+
+    // Try to find a kind field: "kind:<kind>" or just "<kind>"
+    p = tagp->command;
+    if (find_extra(&p) == OK)
+    {
+	if (p > tagp->command && p[-1] == '|')
+	    tagp->command_end = p - 1;  // drop trailing bar
+	else
+	    tagp->command_end = p;
+	p += 2;	// skip ";\""
+	if (*p++ == TAB)
+	    // Accept ASCII alphabetic kind characters and any multi-byte
+	    // character.
+	    while (ASCII_ISALPHA(*p) || mb_ptr2len(p) > 1)
+	    {
+		if (STRNCMP(p, "kind:", 5) == 0)
+		    tagp->tagkind = p + 5;
+		else if (STRNCMP(p, "user_data:", 10) == 0)
+		    tagp->user_data = p + 10;
+		else if (STRNCMP(p, "line:", 5) == 0)
+		    tagp->tagline = atoi((char *)p + 5);
+		if (tagp->tagkind != NULL && tagp->user_data != NULL)
+		    break;
+		pc = vim_strchr(p, ':');
+		pt = vim_strchr(p, '\t');
+		if (pc == NULL || (pt != NULL && pc > pt))
+		    tagp->tagkind = p;
+		if (pt == NULL)
+		    break;
+		p = pt;
+		MB_PTR_ADV(p);
+	    }
+    }
+    if (tagp->tagkind != NULL)
+    {
+	for (p = tagp->tagkind;
+		*p && *p != '\t' && *p != '\r' && *p != '\n'; MB_PTR_ADV(p))
+	    ;
+	tagp->tagkind_end = p;
+    }
+    if (tagp->user_data != NULL)
+    {
+	for (p = tagp->user_data;
+		*p && *p != '\t' && *p != '\r' && *p != '\n'; MB_PTR_ADV(p))
+	    ;
+	tagp->user_data_end = p;
+    }
+    return retval;
+}
+
+/*
+ * Find out the actual file name of a tag.  Concatenate the tags file name
+ * with the matching tag file name.
+ * Returns an allocated string or NULL (out of memory).
+ */
+    static char_u *
+tag_full_fname(tagptrs_T *tagp)
+{
+    char_u	*fullname;
+    int		c;
+
+#ifdef FEAT_EMACS_TAGS
+    if (tagp->is_etag)
+	c = 0;	    // to shut up GCC
+    else
+#endif
+    {
+	c = *tagp->fname_end;
+	*tagp->fname_end = NUL;
+    }
+    fullname = expand_tag_fname(tagp->fname, tagp->tag_fname, FALSE);
+
+#ifdef FEAT_EMACS_TAGS
+    if (!tagp->is_etag)
+#endif
+	*tagp->fname_end = c;
+
+    return fullname;
+}
+
+/*
+ * Jump to a tag that has been found in one of the tag files
+ *
+ * returns OK for success, NOTAGFILE when file not found, FAIL otherwise.
+ */
+    static int
+jumpto_tag(
+    char_u	*lbuf_arg,	// line from the tags file for this tag
+    int		forceit,	// :ta with !
+    int		keep_help)	// keep help flag (FALSE for cscope)
+{
+    optmagic_T	save_magic_overruled;
+    int		save_p_ws, save_p_scs, save_p_ic;
+    linenr_T	save_lnum;
+    char_u	*str;
+    char_u	*pbuf;			// search pattern buffer
+    char_u	*pbuf_end;
+    char_u	*tofree_fname = NULL;
+    char_u	*fname;
+    tagptrs_T	tagp;
+    int		retval = FAIL;
+    int		getfile_result = GETFILE_UNUSED;
+    int		search_options;
+#ifdef FEAT_SEARCH_EXTRA
+    int		save_no_hlsearch;
+#endif
+#if defined(FEAT_QUICKFIX)
+    win_T	*curwin_save = NULL;
+#endif
+    char_u	*full_fname = NULL;
+#ifdef FEAT_FOLDING
+    int		old_KeyTyped = KeyTyped;    // getting the file may reset it
+#endif
+    size_t	len;
+    char_u	*lbuf;
+
+    // Make a copy of the line, it can become invalid when an autocommand calls
+    // back here recursively.
+    len = matching_line_len(lbuf_arg) + 1;
+    lbuf = alloc(len);
+    if (lbuf != NULL)
+	mch_memmove(lbuf, lbuf_arg, len);
+
+    pbuf = alloc(LSIZE);
+
+    // parse the match line into the tagp structure
+    if (pbuf == NULL || lbuf == NULL || parse_match(lbuf, &tagp) == FAIL)
+    {
+	tagp.fname_end = NULL;
+	goto erret;
+    }
+
+    // truncate the file name, so it can be used as a string
+    *tagp.fname_end = NUL;
+    fname = tagp.fname;
+
+    // copy the command to pbuf[], remove trailing CR/NL
+    str = tagp.command;
+    for (pbuf_end = pbuf; *str && *str != '\n' && *str != '\r'; )
+    {
+#ifdef FEAT_EMACS_TAGS
+	if (tagp.is_etag && *str == ',')// stop at ',' after line number
+	    break;
+#endif
+	*pbuf_end++ = *str++;
+	if (pbuf_end - pbuf + 1 >= LSIZE)
+	    break;
+    }
+    *pbuf_end = NUL;
+
+#ifdef FEAT_EMACS_TAGS
+    if (!tagp.is_etag)
+#endif
+    {
+	/*
+	 * Remove the "<Tab>fieldname:value" stuff; we don't need it here.
+	 */
+	str = pbuf;
+	if (find_extra(&str) == OK)
+	{
+	    pbuf_end = str;
+	    *pbuf_end = NUL;
+	}
+    }
+
+    /*
+     * Expand file name, when needed (for environment variables).
+     * If 'tagrelative' option set, may change file name.
+     */
+    fname = expand_tag_fname(fname, tagp.tag_fname, TRUE);
+    if (fname == NULL)
+	goto erret;
+    tofree_fname = fname;	// free() it later
+
+    /*
+     * Check if the file with the tag exists before abandoning the current
+     * file.  Also accept a file name for which there is a matching BufReadCmd
+     * autocommand event (e.g., http://sys/file).
+     */
+    if (mch_getperm(fname) < 0 && !has_autocmd(EVENT_BUFREADCMD, fname, NULL))
+    {
+	retval = NOTAGFILE;
+	vim_free(nofile_fname);
+	nofile_fname = vim_strsave(fname);
+	if (nofile_fname == NULL)
+	    nofile_fname = empty_option;
+	goto erret;
+    }
+
+    ++RedrawingDisabled;
+
+#ifdef FEAT_GUI
+    need_mouse_correct = TRUE;
+#endif
+
+#if defined(FEAT_QUICKFIX)
+    if (g_do_tagpreview != 0)
+    {
+	postponed_split = 0;	// don't split again below
+	curwin_save = curwin;	// Save current window
+
+	/*
+	 * If we are reusing a window, we may change dir when
+	 * entering it (autocommands) so turn the tag filename
+	 * into a fullpath
+	 */
+	if (!curwin->w_p_pvw)
+	{
+	    full_fname = FullName_save(fname, FALSE);
+	    fname = full_fname;
+
+	    /*
+	     * Make the preview window the current window.
+	     * Open a preview window when needed.
+	     */
+	    prepare_tagpreview(TRUE, TRUE, FALSE);
+	}
+    }
+
+    // If it was a CTRL-W CTRL-] command split window now.  For ":tab tag"
+    // open a new tab page.
+    if (postponed_split && (swb_flags & (SWB_USEOPEN | SWB_USETAB)))
+    {
+	buf_T *existing_buf = buflist_findname_exp(fname);
+
+	if (existing_buf != NULL)
+	{
+	    win_T *wp = NULL;
+
+	    if (swb_flags & SWB_USEOPEN)
+		wp = buf_jump_open_win(existing_buf);
+
+	    // If 'switchbuf' contains "usetab": jump to first window in any tab
+	    // page containing "existing_buf" if one exists
+	    if (wp == NULL && (swb_flags & SWB_USETAB))
+		wp = buf_jump_open_tab(existing_buf);
+	    // We've switched to the buffer, the usual loading of the file must
+	    // be skipped.
+	    if (wp != NULL)
+		getfile_result = GETFILE_SAME_FILE;
+	}
+    }
+    if (getfile_result == GETFILE_UNUSED
+				  && (postponed_split || cmdmod.cmod_tab != 0))
+    {
+	if (win_split(postponed_split > 0 ? postponed_split : 0,
+						postponed_split_flags) == FAIL)
+	{
+	    --RedrawingDisabled;
+	    goto erret;
+	}
+	RESET_BINDING(curwin);
+    }
+#endif
+
+    if (keep_help)
+    {
+	// A :ta from a help file will keep the b_help flag set.  For ":ptag"
+	// we need to use the flag from the window where we came from.
+#if defined(FEAT_QUICKFIX)
+	if (g_do_tagpreview != 0)
+	    keep_help_flag = bt_help(curwin_save->w_buffer);
+	else
+#endif
+	    keep_help_flag = curbuf->b_help;
+    }
+
+    if (getfile_result == GETFILE_UNUSED)
+	// Careful: getfile() may trigger autocommands and call jumpto_tag()
+	// recursively.
+	getfile_result = getfile(0, fname, NULL, TRUE, (linenr_T)0, forceit);
+    keep_help_flag = FALSE;
+
+    if (GETFILE_SUCCESS(getfile_result))	// got to the right file
+    {
+	curwin->w_set_curswant = TRUE;
+	postponed_split = 0;
+
+	save_magic_overruled = magic_overruled;
+	magic_overruled = OPTION_MAGIC_OFF;	// always execute with 'nomagic'
+#ifdef FEAT_SEARCH_EXTRA
+	// Save value of no_hlsearch, jumping to a tag is not a real search
+	save_no_hlsearch = no_hlsearch;
+#endif
+#if defined(FEAT_PROP_POPUP) && defined(FEAT_QUICKFIX)
+	// getfile() may have cleared options, apply 'previewpopup' again.
+	if (g_do_tagpreview != 0 && *p_pvp != NUL)
+	    parse_previewpopup(curwin);
+#endif
+
+	/*
+	 * If 'cpoptions' contains 't', store the search pattern for the "n"
+	 * command.  If 'cpoptions' does not contain 't', the search pattern
+	 * is not stored.
+	 */
+	if (vim_strchr(p_cpo, CPO_TAGPAT) != NULL)
+	    search_options = 0;
+	else
+	    search_options = SEARCH_KEEP;
+
+	/*
+	 * If the command is a search, try here.
+	 *
+	 * Reset 'smartcase' for the search, since the search pattern was not
+	 * typed by the user.
+	 * Only use do_search() when there is a full search command, without
+	 * anything following.
+	 */
+	str = pbuf;
+	if (pbuf[0] == '/' || pbuf[0] == '?')
+	    str = skip_regexp(pbuf + 1, pbuf[0], FALSE) + 1;
+	if (str > pbuf_end - 1)	// search command with nothing following
+	{
+	    save_p_ws = p_ws;
+	    save_p_ic = p_ic;
+	    save_p_scs = p_scs;
+	    p_ws = TRUE;	// need 'wrapscan' for backward searches
+	    p_ic = FALSE;	// don't ignore case now
+	    p_scs = FALSE;
+	    save_lnum = curwin->w_cursor.lnum;
+	    if (tagp.tagline > 0)
+		// start search before line from "line:" field
+		curwin->w_cursor.lnum = tagp.tagline - 1;
+	    else
+		// start search before first line
+		curwin->w_cursor.lnum = 0;
+	    if (do_search(NULL, pbuf[0], pbuf[0], pbuf + 1, (long)1,
+							 search_options, NULL))
+		retval = OK;
+	    else
+	    {
+		int	found = 1;
+		int	cc;
+
+		/*
+		 * try again, ignore case now
+		 */
+		p_ic = TRUE;
+		if (!do_search(NULL, pbuf[0], pbuf[0], pbuf + 1, (long)1,
+							 search_options, NULL))
+		{
+		    /*
+		     * Failed to find pattern, take a guess: "^func  ("
+		     */
+		    found = 2;
+		    (void)test_for_static(&tagp);
+		    cc = *tagp.tagname_end;
+		    *tagp.tagname_end = NUL;
+		    sprintf((char *)pbuf, "^%s\\s\\*(", tagp.tagname);
+		    if (!do_search(NULL, '/', '/', pbuf, (long)1,
+							 search_options, NULL))
+		    {
+			// Guess again: "^char * \<func  ("
+			sprintf((char *)pbuf, "^\\[#a-zA-Z_]\\.\\*\\<%s\\s\\*(",
+								tagp.tagname);
+			if (!do_search(NULL, '/', '/', pbuf, (long)1,
+							 search_options, NULL))
+			    found = 0;
+		    }
+		    *tagp.tagname_end = cc;
+		}
+		if (found == 0)
+		{
+		    emsg(_(e_cannot_find_tag_pattern));
+		    curwin->w_cursor.lnum = save_lnum;
+		}
+		else
+		{
+		    /*
+		     * Only give a message when really guessed, not when 'ic'
+		     * is set and match found while ignoring case.
+		     */
+		    if (found == 2 || !save_p_ic)
+		    {
+			msg(_(e_couldnt_find_tag_just_guessing));
+			if (!msg_scrolled && msg_silent == 0)
+			{
+			    out_flush();
+			    ui_delay(1010L, TRUE);
+			}
+		    }
+		    retval = OK;
+		}
+	    }
+	    p_ws = save_p_ws;
+	    p_ic = save_p_ic;
+	    p_scs = save_p_scs;
+
+	    // A search command may have positioned the cursor beyond the end
+	    // of the line.  May need to correct that here.
+	    check_cursor();
+	}
+	else
+	{
+	    int		save_secure = secure;
+
+	    // Setup the sandbox for executing the command from the tags file.
+	    secure = 1;
+#ifdef HAVE_SANDBOX
+	    ++sandbox;
+#endif
+	    curwin->w_cursor.lnum = 1;		// start command in line 1
+	    do_cmdline_cmd(pbuf);
+	    retval = OK;
+
+	    // When the command has done something that is not allowed make
+	    // sure the error message can be seen.
+	    if (secure == 2)
+		wait_return(TRUE);
+	    secure = save_secure;
+#ifdef HAVE_SANDBOX
+	    --sandbox;
+#endif
+	}
+
+	magic_overruled = save_magic_overruled;
+#ifdef FEAT_SEARCH_EXTRA
+	// restore no_hlsearch when keeping the old search pattern
+	if (search_options)
+	    set_no_hlsearch(save_no_hlsearch);
+#endif
+
+	// Return OK if jumped to another file (at least we found the file!).
+	if (getfile_result == GETFILE_OPEN_OTHER)
+	    retval = OK;
+
+	if (retval == OK)
+	{
+	    /*
+	     * For a help buffer: Put the cursor line at the top of the window,
+	     * the help subject will be below it.
+	     */
+	    if (curbuf->b_help)
+		set_topline(curwin, curwin->w_cursor.lnum);
+#ifdef FEAT_FOLDING
+	    if ((fdo_flags & FDO_TAG) && old_KeyTyped)
+		foldOpenCursor();
+#endif
+	}
+
+#if defined(FEAT_QUICKFIX)
+	if (g_do_tagpreview != 0
+			   && curwin != curwin_save && win_valid(curwin_save))
+	{
+	    // Return cursor to where we were
+	    validate_cursor();
+	    redraw_later(UPD_VALID);
+	    win_enter(curwin_save, TRUE);
+	}
+#endif
+
+	--RedrawingDisabled;
+    }
+    else
+    {
+	--RedrawingDisabled;
+	got_int = FALSE;  // don't want entering window to fail
+
+	if (postponed_split)		// close the window
+	{
+	    win_close(curwin, FALSE);
+	    postponed_split = 0;
+	}
+#if defined(FEAT_QUICKFIX) && defined(FEAT_PROP_POPUP)
+	else if (WIN_IS_POPUP(curwin))
+	{
+	    win_T   *wp = curwin;
+
+	    if (win_valid(curwin_save))
+		win_enter(curwin_save, TRUE);
+	    popup_close(wp->w_id, FALSE);
+	}
+#endif
+    }
+#if defined(FEAT_QUICKFIX) && defined(FEAT_PROP_POPUP)
+    if (WIN_IS_POPUP(curwin))
+	// something went wrong, still in popup, but it can't have focus
+	win_enter(firstwin, TRUE);
+#endif
+
+erret:
+#if defined(FEAT_QUICKFIX)
+    g_do_tagpreview = 0; // For next time
+#endif
+    vim_free(lbuf);
+    vim_free(pbuf);
+    vim_free(tofree_fname);
+    vim_free(full_fname);
+
+    return retval;
+}
+
+/*
+ * If "expand" is TRUE, expand wildcards in fname.
+ * If 'tagrelative' option set, change fname (name of file containing tag)
+ * according to tag_fname (name of tag file containing fname).
+ * Returns a pointer to allocated memory (or NULL when out of memory).
+ */
+    static char_u *
+expand_tag_fname(char_u *fname, char_u *tag_fname, int expand)
+{
+    char_u	*p;
+    char_u	*retval;
+    char_u	*expanded_fname = NULL;
+    expand_T	xpc;
+
+    /*
+     * Expand file name (for environment variables) when needed.
+     */
+    if (expand && mch_has_wildcard(fname))
+    {
+	ExpandInit(&xpc);
+	xpc.xp_context = EXPAND_FILES;
+	expanded_fname = ExpandOne(&xpc, fname, NULL,
+			    WILD_LIST_NOTFOUND|WILD_SILENT, WILD_EXPAND_FREE);
+	if (expanded_fname != NULL)
+	    fname = expanded_fname;
+    }
+
+    if ((p_tr || curbuf->b_help)
+	    && !vim_isAbsName(fname)
+	    && (p = gettail(tag_fname)) != tag_fname)
+    {
+	retval = alloc(MAXPATHL);
+	if (retval != NULL)
+	{
+	    STRCPY(retval, tag_fname);
+	    vim_strncpy(retval + (p - tag_fname), fname,
+					      MAXPATHL - (p - tag_fname) - 1);
+	    /*
+	     * Translate names like "src/a/../b/file.c" into "src/b/file.c".
+	     */
+	    simplify_filename(retval);
+	}
+    }
+    else
+	retval = vim_strsave(fname);
+
+    vim_free(expanded_fname);
+
+    return retval;
+}
+
+/*
+ * Check if we have a tag for the buffer with name "buf_ffname".
+ * This is a bit slow, because of the full path compare in fullpathcmp().
+ * Return TRUE if tag for file "fname" if tag file "tag_fname" is for current
+ * file.
+ */
+    static int
+test_for_current(
+#ifdef FEAT_EMACS_TAGS
+    int	    is_etag,
+#endif
+    char_u  *fname,
+    char_u  *fname_end,
+    char_u  *tag_fname,
+    char_u  *buf_ffname)
+{
+    int	    c;
+    int	    retval = FALSE;
+    char_u  *fullname;
+
+    if (buf_ffname != NULL)	// if the buffer has a name
+    {
+#ifdef FEAT_EMACS_TAGS
+	if (is_etag)
+	    c = 0;	    // to shut up GCC
+	else
+#endif
+	{
+	    c = *fname_end;
+	    *fname_end = NUL;
+	}
+	fullname = expand_tag_fname(fname, tag_fname, TRUE);
+	if (fullname != NULL)
+	{
+	    retval = (fullpathcmp(fullname, buf_ffname, TRUE, TRUE) & FPC_SAME);
+	    vim_free(fullname);
+	}
+#ifdef FEAT_EMACS_TAGS
+	if (!is_etag)
+#endif
+	    *fname_end = c;
+    }
+
+    return retval;
+}
+
+/*
+ * Find the end of the tagaddress.
+ * Return OK if ";\"" is following, FAIL otherwise.
+ */
+    static int
+find_extra(char_u **pp)
+{
+    char_u	*str = *pp;
+    char_u	first_char = **pp;
+
+    // Repeat for addresses separated with ';'
+    for (;;)
+    {
+	if (VIM_ISDIGIT(*str))
+	    str = skipdigits(str + 1);
+	else if (*str == '/' || *str == '?')
+	{
+	    str = skip_regexp(str + 1, *str, FALSE);
+	    if (*str != first_char)
+		str = NULL;
+	    else
+		++str;
+	}
+	else
+	{
+	    // not a line number or search string, look for terminator.
+	    str = (char_u *)strstr((char *)str, "|;\"");
+	    if (str != NULL)
+	    {
+		++str;
+		break;
+	    }
+
+	}
+	if (str == NULL || *str != ';'
+		  || !(VIM_ISDIGIT(str[1]) || str[1] == '/' || str[1] == '?'))
+	    break;
+	++str;	// skip ';'
+	first_char = *str;
+    }
+
+    if (str != NULL && STRNCMP(str, ";\"", 2) == 0)
+    {
+	*pp = str;
+	return OK;
+    }
+    return FAIL;
+}
+
+/*
+ * Free a single entry in a tag stack
+ */
+    static void
+tagstack_clear_entry(taggy_T *item)
+{
+    VIM_CLEAR(item->tagname);
+    VIM_CLEAR(item->user_data);
+}
+
+    int
+expand_tags(
+    int		tagnames,	// expand tag names
+    char_u	*pat,
+    int		*num_file,
+    char_u	***file)
+{
+    int		i;
+    int		extra_flag;
+    char_u	*name_buf;
+    size_t	name_buf_size = 100;
+    tagptrs_T	t_p;
+    int		ret;
+
+    name_buf = alloc(name_buf_size);
+    if (name_buf == NULL)
+	return FAIL;
+
+    if (tagnames)
+	extra_flag = TAG_NAMES;
+    else
+	extra_flag = 0;
+    if (pat[0] == '/')
+	ret = find_tags(pat + 1, num_file, file,
+		TAG_REGEXP | extra_flag | TAG_VERBOSE | TAG_NO_TAGFUNC,
+		TAG_MANY, curbuf->b_ffname);
+    else
+	ret = find_tags(pat, num_file, file,
+	      TAG_REGEXP | extra_flag | TAG_VERBOSE | TAG_NO_TAGFUNC | TAG_NOIC,
+		TAG_MANY, curbuf->b_ffname);
+    if (ret == OK && !tagnames)
+    {
+	 // Reorganize the tags for display and matching as strings of:
+	 // "<tagname>\0<kind>\0<filename>\0"
+	 for (i = 0; i < *num_file; i++)
+	 {
+	     size_t	len;
+
+	     parse_match((*file)[i], &t_p);
+	     len = t_p.tagname_end - t_p.tagname;
+	     if (len > name_buf_size - 3)
+	     {
+		 char_u *buf;
+
+		 name_buf_size = len + 3;
+		 buf = vim_realloc(name_buf, name_buf_size);
+		 if (buf == NULL)
+		 {
+		     vim_free(name_buf);
+		     return FAIL;
+		 }
+		 name_buf = buf;
+	     }
+
+	     mch_memmove(name_buf, t_p.tagname, len);
+	     name_buf[len++] = 0;
+	     name_buf[len++] = (t_p.tagkind != NULL && *t_p.tagkind)
+							  ? *t_p.tagkind : 'f';
+	     name_buf[len++] = 0;
+	     mch_memmove((*file)[i] + len, t_p.fname,
+						    t_p.fname_end - t_p.fname);
+	     (*file)[i][len + (t_p.fname_end - t_p.fname)] = 0;
+	     mch_memmove((*file)[i], name_buf, len);
+	}
+    }
+
+    vim_free(name_buf);
+    return ret;
+}
+
+#if defined(FEAT_EVAL) || defined(PROTO)
+/*
+ * Add a tag field to the dictionary "dict".
+ * Return OK or FAIL.
+ */
+    static int
+add_tag_field(
+    dict_T  *dict,
+    char    *field_name,
+    char_u  *start,		// start of the value
+    char_u  *end)		// after the value; can be NULL
+{
+    char_u	*buf;
+    int		len = 0;
+    int		retval;
+
+    // check that the field name doesn't exist yet
+    if (dict_has_key(dict, field_name))
+    {
+	if (p_verbose > 0)
+	{
+	    verbose_enter();
+	    smsg(_("Duplicate field name: %s"), field_name);
+	    verbose_leave();
+	}
+	return FAIL;
+    }
+    buf = alloc(MAXPATHL);
+    if (buf == NULL)
+	return FAIL;
+    if (start != NULL)
+    {
+	if (end == NULL)
+	{
+	    end = start + STRLEN(start);
+	    while (end > start && (end[-1] == '\r' || end[-1] == '\n'))
+		--end;
+	}
+	len = (int)(end - start);
+	if (len > MAXPATHL - 1)
+	    len = MAXPATHL - 1;
+	vim_strncpy(buf, start, len);
+    }
+    buf[len] = NUL;
+    retval = dict_add_string(dict, field_name, buf);
+    vim_free(buf);
+    return retval;
+}
+
+/*
+ * Add the tags matching the specified pattern "pat" to the list "list"
+ * as a dictionary. Use "buf_fname" for priority, unless NULL.
+ */
+    int
+get_tags(list_T *list, char_u *pat, char_u *buf_fname)
+{
+    int		num_matches, i, ret;
+    char_u	**matches, *p;
+    char_u	*full_fname;
+    dict_T	*dict;
+    tagptrs_T	tp;
+    long	is_static;
+
+    ret = find_tags(pat, &num_matches, &matches,
+				TAG_REGEXP | TAG_NOIC, (int)MAXCOL, buf_fname);
+    if (ret != OK || num_matches <= 0)
+	return ret;
+
+    for (i = 0; i < num_matches; ++i)
+    {
+	if (parse_match(matches[i], &tp) == FAIL)
+	{
+	    vim_free(matches[i]);
+	    continue;
+	}
+
+	is_static = test_for_static(&tp);
+
+	// Skip pseudo-tag lines.
+	if (STRNCMP(tp.tagname, "!_TAG_", 6) == 0)
+	{
+	    vim_free(matches[i]);
+	    continue;
+	}
+
+	if ((dict = dict_alloc()) == NULL)
+	{
+	    ret = FAIL;
+	    vim_free(matches[i]);
+	    break;
+	}
+	if (list_append_dict(list, dict) == FAIL)
+	    ret = FAIL;
+
+	full_fname = tag_full_fname(&tp);
+	if (add_tag_field(dict, "name", tp.tagname, tp.tagname_end) == FAIL
+		|| add_tag_field(dict, "filename", full_fname,
+		    NULL) == FAIL
+		|| add_tag_field(dict, "cmd", tp.command,
+		    tp.command_end) == FAIL
+		|| add_tag_field(dict, "kind", tp.tagkind,
+		    tp.tagkind_end) == FAIL
+		|| dict_add_number(dict, "static", is_static) == FAIL)
+	    ret = FAIL;
+
+	vim_free(full_fname);
+
+	if (tp.command_end != NULL)
+	{
+	    for (p = tp.command_end + 3;
+		    *p != NUL && *p != '\n' && *p != '\r'; MB_PTR_ADV(p))
+	    {
+		if (p == tp.tagkind || (p + 5 == tp.tagkind
+			    && STRNCMP(p, "kind:", 5) == 0))
+		    // skip "kind:<kind>" and "<kind>"
+		    p = tp.tagkind_end - 1;
+		else if (STRNCMP(p, "file:", 5) == 0)
+		    // skip "file:" (static tag)
+		    p += 4;
+		else if (!VIM_ISWHITE(*p))
+		{
+		    char_u	*s, *n;
+		    int	len;
+
+		    // Add extra field as a dict entry.  Fields are
+		    // separated by Tabs.
+		    n = p;
+		    while (*p != NUL && *p >= ' ' && *p < 127 && *p != ':')
+			++p;
+		    len = (int)(p - n);
+		    if (*p == ':' && len > 0)
+		    {
+			s = ++p;
+			while (*p != NUL && *p >= ' ')
+			    ++p;
+			n[len] = NUL;
+			if (add_tag_field(dict, (char *)n, s, p) == FAIL)
+			    ret = FAIL;
+			n[len] = ':';
+		    }
+		    else
+			// Skip field without colon.
+			while (*p != NUL && *p >= ' ')
+			    ++p;
+		    if (*p == NUL)
+			break;
+		}
+	    }
+	}
+
+	vim_free(matches[i]);
+    }
+    vim_free(matches);
+    return ret;
+}
+
+/*
+ * Return information about 'tag' in dict 'retdict'.
+ */
+    static void
+get_tag_details(taggy_T *tag, dict_T *retdict)
+{
+    list_T	*pos;
+    fmark_T	*fmark;
+
+    dict_add_string(retdict, "tagname", tag->tagname);
+    dict_add_number(retdict, "matchnr", tag->cur_match + 1);
+    dict_add_number(retdict, "bufnr", tag->cur_fnum);
+    if (tag->user_data)
+	dict_add_string(retdict, "user_data", tag->user_data);
+
+    if ((pos = list_alloc_id(aid_tagstack_from)) == NULL)
+	return;
+    dict_add_list(retdict, "from", pos);
+
+    fmark = &tag->fmark;
+    list_append_number(pos,
+			(varnumber_T)(fmark->fnum != -1 ? fmark->fnum : 0));
+    list_append_number(pos, (varnumber_T)fmark->mark.lnum);
+    list_append_number(pos, (varnumber_T)(fmark->mark.col == MAXCOL ?
+					MAXCOL : fmark->mark.col + 1));
+    list_append_number(pos, (varnumber_T)fmark->mark.coladd);
+}
+
+/*
+ * Return the tag stack entries of the specified window 'wp' in dictionary
+ * 'retdict'.
+ */
+    void
+get_tagstack(win_T *wp, dict_T *retdict)
+{
+    list_T	*l;
+    int		i;
+    dict_T	*d;
+
+    dict_add_number(retdict, "length", wp->w_tagstacklen);
+    dict_add_number(retdict, "curidx", wp->w_tagstackidx + 1);
+    l = list_alloc_id(aid_tagstack_items);
+    if (l == NULL)
+	return;
+    dict_add_list(retdict, "items", l);
+
+    for (i = 0; i < wp->w_tagstacklen; i++)
+    {
+	if ((d = dict_alloc_id(aid_tagstack_details)) == NULL)
+	    return;
+	list_append_dict(l, d);
+
+	get_tag_details(&wp->w_tagstack[i], d);
+    }
+}
+
+/*
+ * Free all the entries in the tag stack of the specified window
+ */
+    static void
+tagstack_clear(win_T *wp)
+{
+    int i;
+
+    // Free the current tag stack
+    for (i = 0; i < wp->w_tagstacklen; ++i)
+	tagstack_clear_entry(&wp->w_tagstack[i]);
+    wp->w_tagstacklen = 0;
+    wp->w_tagstackidx = 0;
+}
+
+/*
+ * Remove the oldest entry from the tag stack and shift the rest of
+ * the entries to free up the top of the stack.
+ */
+    static void
+tagstack_shift(win_T *wp)
+{
+    taggy_T	*tagstack = wp->w_tagstack;
+    int		i;
+
+    tagstack_clear_entry(&tagstack[0]);
+    for (i = 1; i < wp->w_tagstacklen; ++i)
+	tagstack[i - 1] = tagstack[i];
+    wp->w_tagstacklen--;
+}
+
+/*
+ * Push a new item to the tag stack
+ */
+    static void
+tagstack_push_item(
+	win_T	*wp,
+	char_u	*tagname,
+	int	cur_fnum,
+	int	cur_match,
+	pos_T	mark,
+	int	fnum,
+	char_u  *user_data)
+{
+    taggy_T	*tagstack = wp->w_tagstack;
+    int		idx = wp->w_tagstacklen;	// top of the stack
+
+    // if the tagstack is full: remove the oldest entry
+    if (idx >= TAGSTACKSIZE)
+    {
+	tagstack_shift(wp);
+	idx = TAGSTACKSIZE - 1;
+    }
+
+    wp->w_tagstacklen++;
+    tagstack[idx].tagname = tagname;
+    tagstack[idx].cur_fnum = cur_fnum;
+    tagstack[idx].cur_match = cur_match;
+    if (tagstack[idx].cur_match < 0)
+	tagstack[idx].cur_match = 0;
+    tagstack[idx].fmark.mark = mark;
+    tagstack[idx].fmark.fnum = fnum;
+    tagstack[idx].user_data = user_data;
+}
+
+/*
+ * Add a list of items to the tag stack in the specified window
+ */
+    static void
+tagstack_push_items(win_T *wp, list_T *l)
+{
+    listitem_T	*li;
+    dictitem_T	*di;
+    dict_T	*itemdict;
+    char_u	*tagname;
+    pos_T	mark;
+    int		fnum;
+
+    // Add one entry at a time to the tag stack
+    FOR_ALL_LIST_ITEMS(l, li)
+    {
+	if (li->li_tv.v_type != VAR_DICT || li->li_tv.vval.v_dict == NULL)
+	    continue;				// Skip non-dict items
+	itemdict = li->li_tv.vval.v_dict;
+
+	// parse 'from' for the cursor position before the tag jump
+	if ((di = dict_find(itemdict, (char_u *)"from", -1)) == NULL)
+	    continue;
+	if (list2fpos(&di->di_tv, &mark, &fnum, NULL, FALSE) != OK)
+	    continue;
+	if ((tagname = dict_get_string(itemdict, "tagname", TRUE)) == NULL)
+	    continue;
+
+	if (mark.col > 0)
+	    mark.col--;
+	tagstack_push_item(wp, tagname,
+		(int)dict_get_number(itemdict, "bufnr"),
+		(int)dict_get_number(itemdict, "matchnr") - 1,
+		mark, fnum,
+		dict_get_string(itemdict, "user_data", TRUE));
+    }
+}
+
+/*
+ * Set the current index in the tag stack. Valid values are between 0
+ * and the stack length (inclusive).
+ */
+    static void
+tagstack_set_curidx(win_T *wp, int curidx)
+{
+    wp->w_tagstackidx = curidx;
+    if (wp->w_tagstackidx < 0)			// sanity check
+	wp->w_tagstackidx = 0;
+    if (wp->w_tagstackidx > wp->w_tagstacklen)
+	wp->w_tagstackidx = wp->w_tagstacklen;
+}
+
+/*
+ * Set the tag stack entries of the specified window.
+ * 'action' is set to one of:
+ *	'a' for append
+ *	'r' for replace
+ *	't' for truncate
+ */
+    int
+set_tagstack(win_T *wp, dict_T *d, int action)
+{
+    dictitem_T	*di;
+    list_T	*l = NULL;
+
+#ifdef FEAT_EVAL
+    // not allowed to alter the tag stack entries from inside tagfunc
+    if (tfu_in_use)
+    {
+	emsg(_(e_cannot_modify_tag_stack_within_tagfunc));
+	return FAIL;
+    }
+#endif
+
+    if ((di = dict_find(d, (char_u *)"items", -1)) != NULL)
+    {
+	if (di->di_tv.v_type != VAR_LIST)
+	{
+	    emsg(_(e_list_required));
+	    return FAIL;
+	}
+	l = di->di_tv.vval.v_list;
+    }
+
+    if ((di = dict_find(d, (char_u *)"curidx", -1)) != NULL)
+	tagstack_set_curidx(wp, (int)tv_get_number(&di->di_tv) - 1);
+
+    if (action == 't')		    // truncate the stack
+    {
+	taggy_T	*tagstack = wp->w_tagstack;
+	int	tagstackidx = wp->w_tagstackidx;
+	int	tagstacklen = wp->w_tagstacklen;
+
+	// delete all the tag stack entries above the current entry
+	while (tagstackidx < tagstacklen)
+	    tagstack_clear_entry(&tagstack[--tagstacklen]);
+	wp->w_tagstacklen = tagstacklen;
+    }
+
+    if (l != NULL)
+    {
+	if (action == 'r')		// replace the stack
+	    tagstack_clear(wp);
+
+	tagstack_push_items(wp, l);
+	// set the current index after the last entry
+	wp->w_tagstackidx = wp->w_tagstacklen;
+    }
+
+    return OK;
+}
+#endif