view src/textprop.c @ 32936:c517845bd10e v9.0.1776

patch 9.0.1776: No support for stable Python 3 ABI Commit: https://github.com/vim/vim/commit/c13b3d1350b60b94fe87f0761ea31c0e7fb6ebf3 Author: Yee Cheng Chin <ychin.git@gmail.com> Date: Sun Aug 20 21:18:38 2023 +0200 patch 9.0.1776: No support for stable Python 3 ABI Problem: No support for stable Python 3 ABI Solution: Support Python 3 stable ABI Commits: 1) Support Python 3 stable ABI to allow mixed version interoperatbility Vim currently supports embedding Python for use with plugins, and the "dynamic" linking option allows the user to specify a locally installed version of Python by setting `pythonthreedll`. However, one caveat is that the Python 3 libs are not binary compatible across minor versions, and mixing versions can potentially be dangerous (e.g. let's say Vim was linked against the Python 3.10 SDK, but the user sets `pythonthreedll` to a 3.11 lib). Usually, nothing bad happens, but in theory this could lead to crashes, memory corruption, and other unpredictable behaviors. It's also difficult for the user to tell something is wrong because Vim has no way of reporting what Python 3 version Vim was linked with. For Vim installed via a package manager, this usually isn't an issue because all the dependencies would already be figured out. For prebuilt Vim binaries like MacVim (my motivation for working on this), AppImage, and Win32 installer this could potentially be an issue as usually a single binary is distributed. This is more tricky when a new Python version is released, as there's a chicken-and-egg issue with deciding what Python version to build against and hard to keep in sync when a new Python version just drops and we have a mix of users of different Python versions, and a user just blindly upgrading to a new Python could lead to bad interactions with Vim. Python 3 does have a solution for this problem: stable ABI / limited API (see https://docs.python.org/3/c-api/stable.html). The C SDK limits the API to a set of functions that are promised to be stable across versions. This pull request adds an ifdef config that allows us to turn it on when building Vim. Vim binaries built with this option should be safe to freely link with any Python 3 libraies without having the constraint of having to use the same minor version. Note: Python 2 has no such concept and this doesn't change how Python 2 integration works (not that there is going to be a new version of Python 2 that would cause compatibility issues in the future anyway). --- Technical details: ====== The stable ABI can be accessed when we compile with the Python 3 limited API (by defining `Py_LIMITED_API`). The Python 3 code (in `if_python3.c` and `if_py_both.h`) would now handle this and switch to limited API mode. Without it set, Vim will still use the full API as before so this is an opt-in change. The main difference is that `PyType_Object` is now an opaque struct that we can't directly create "static types" out of, and we have to create type objects as "heap types" instead. This is because the struct is not stable and changes from version to version (e.g. 3.8 added a `tp_vectorcall` field to it). I had to change all the types to be allocated on the heap instead with just a pointer to them. Other functions are also simply missing in limited API, or they are introduced too late (e.g. `PyUnicode_AsUTF8AndSize` in 3.10) to it that we need some other ways to do the same thing, so I had to abstract a few things into macros, and sometimes re-implement functions like `PyObject_NEW`. One caveat is that in limited API, `OutputType` (used for replacing `sys.stdout`) no longer inherits from `PyStdPrinter_Type` which I don't think has any real issue other than minor differences in how they convert to a string and missing a couple functions like `mode()` and `fileno()`. Also fixed an existing bug where `tp_basicsize` was set incorrectly for `BufferObject`, `TabListObject, `WinListObject`. Technically, there could be a small performance drop, there is a little more indirection with accessing type objects, and some APIs like `PyUnicode_AsUTF8AndSize` are missing, but in practice I didn't see any difference, and any well-written Python plugin should try to avoid excessing callbacks to the `vim` module in Python anyway. I only tested limited API mode down to Python 3.7, which seemes to compile and work fine. I haven't tried earlier Python versions. 2) Fix PyIter_Check on older Python vers / type##Ptr unused warning For PyIter_Check, older versions exposed them as either macros (used in full API), or a function (for use in limited API). A previous change exposed PyIter_Check to the dynamic build because Python just moved it to function-only in 3.10 anyway. Because of that, just make sure we always grab the function in dynamic builds in earlier versions since that's what Python eventually did anyway. 3) Move Py_LIMITED_API define to configure script Can now use --with-python-stable-abi flag to customize what stable ABI version to target. Can also use an env var to do so as well. 4) Show +python/dyn-stable in :version, and allow has() feature query Not sure if the "/dyn-stable" suffix would break things, or whether we should do it another way. Or just don't show it in version and rely on has() feature checking. 5) Documentation first draft. Still need to implement v:python3_version 6) Fix PyIter_Check build breaks when compiling against Python 3.8 7) Add CI coverage stable ABI on Linux/Windows / make configurable on Windows This adds configurable options for Windows make files (both MinGW and MSVC). CI will also now exercise both traditional full API and stable ABI for Linux and Windows in the matrix for coverage. Also added a "dynamic" option to Linux matrix as a drive-by change to make other scripting languages like Ruby / Perl testable under both static and dynamic builds. 8) Fix inaccuracy in Windows docs Python's own docs are confusing but you don't actually want to use `python3.dll` for the dynamic linkage. 9) Add generated autoconf file 10) Add v:python3_version support This variable indicates the version of Python3 that Vim was built against (PY_VERSION_HEX), and will be useful to check whether the Python library you are loading in dynamically actually fits it. When built with stable ABI, it will be the limited ABI version instead (`Py_LIMITED_API`), which indicates the minimum version of Python 3 the user should have, rather than the exact match. When stable ABI is used, we won't be exposing PY_VERSION_HEX in this var because it just doesn't seem necessary to do so (the whole point of stable ABI is the promise that it will work across versions), and I don't want to confuse the user with too many variables. Also, cleaned up some documentation, and added help tags. 11) Fix Python 3.7 compat issues Fix a couple issues when using limited API < 3.8 - Crash on exit: In Python 3.7, if a heap-allocated type is destroyed before all instances are, it would cause a crash later. This happens when we destroyed `OptionsType` before calling `Py_Finalize` when using the limited API. To make it worse, later versions changed the semantics and now each instance has a strong reference to its own type and the recommendation has changed to have each instance de-ref its own type and have its type in GC traversal. To avoid dealing with these cross-version variations, we just don't free the heap type. They are static types in non-limited-API anyway and are designed to last through the entirety of the app, and we also don't restart the Python runtime and therefore do not need it to have absolutely 0 leaks. See: - https://docs.python.org/3/whatsnew/3.8.html#changes-in-the-c-api - https://docs.python.org/3/whatsnew/3.9.html#changes-in-the-c-api - PyIter_Check: This function is not provided in limited APIs older than 3.8. Previously I was trying to mock it out using manual PyType_GetSlot() but it was brittle and also does not actually work properly for static types (it will generate a Python error). Just return false. It does mean using limited API < 3.8 is not recommended as you lose the functionality to handle iterators, but from playing with plugins I couldn't find it to be an issue. - Fix loading of PyIter_Check so it will be done when limited API < 3.8. Otherwise loading a 3.7 Python lib will fail even if limited API was specified to use it. 12) Make sure to only load `PyUnicode_AsUTF8AndSize` in needed in limited API We don't use this function unless limited API >= 3.10, but we were loading it regardless. Usually it's ok in Unix-like systems where Python just has a single lib that we load from, but in Windows where there is a separate python3.dll this would not work as the symbol would not have been exposed in this more limited DLL file. This makes it much clearer under what condition is this function needed. closes: #12032 Signed-off-by: Christian Brabandt <cb@256bit.org> Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
author Christian Brabandt <cb@256bit.org>
date Sun, 20 Aug 2023 21:30:04 +0200
parents 98db7b3b4827
children c7a332a34fe7
line wrap: on
line source

/* 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.
 */

/*
 * Text properties implementation.  See ":help text-properties".
 */

#include "vim.h"

#if defined(FEAT_PROP_POPUP) || defined(PROTO)

/*
 * In a hashtable item "hi_key" points to "pt_name" in a proptype_T.
 * This avoids adding a pointer to the hashtable item.
 * PT2HIKEY() converts a proptype pointer to a hashitem key pointer.
 * HIKEY2PT() converts a hashitem key pointer to a proptype pointer.
 * HI2PT() converts a hashitem pointer to a proptype pointer.
 */
#define PT2HIKEY(p)  ((p)->pt_name)
#define HIKEY2PT(p)   ((proptype_T *)((p) - offsetof(proptype_T, pt_name)))
#define HI2PT(hi)      HIKEY2PT((hi)->hi_key)

// The global text property types.
static hashtab_T *global_proptypes = NULL;
static proptype_T **global_proparray = NULL;

// The last used text property type ID.
static int proptype_id = 0;

/*
 * Find a property type by name, return the hashitem.
 * Returns NULL if the item can't be found.
 */
    static hashitem_T *
find_prop_type_hi(char_u *name, buf_T *buf)
{
    hashtab_T	*ht;
    hashitem_T	*hi;

    if (*name == NUL)
	return NULL;
    if (buf == NULL)
	ht = global_proptypes;
    else
	ht = buf->b_proptypes;

    if (ht == NULL)
	return NULL;
    hi = hash_find(ht, name);
    if (HASHITEM_EMPTY(hi))
	return NULL;
    return hi;
}

/*
 * Like find_prop_type_hi() but return the property type.
 */
    static proptype_T *
find_prop_type(char_u *name, buf_T *buf)
{
    hashitem_T	*hi = find_prop_type_hi(name, buf);

    if (hi == NULL)
	return NULL;
    return HI2PT(hi);
}

/*
 * Get the prop type ID of "name".
 * When not found return zero.
 */
    int
find_prop_type_id(char_u *name, buf_T *buf)
{
    proptype_T *pt = find_prop_type(name, buf);

    if (pt == NULL)
	return 0;
    return pt->pt_id;
}

/*
 * Lookup a property type by name.  First in "buf" and when not found in the
 * global types.
 * When not found gives an error message and returns NULL.
 */
    static proptype_T *
lookup_prop_type(char_u *name, buf_T *buf)
{
    proptype_T *type = find_prop_type(name, buf);

    if (type == NULL)
	type = find_prop_type(name, NULL);
    if (type == NULL)
	semsg(_(e_property_type_str_does_not_exist), name);
    return type;
}

/*
 * Get an optional "bufnr" item from the dict in "arg".
 * When the argument is not used or "bufnr" is not present then "buf" is
 * unchanged.
 * If "bufnr" is valid or not present return OK.
 * When "arg" is not a dict or "bufnr" is invalid return FAIL.
 */
    static int
get_bufnr_from_arg(typval_T *arg, buf_T **buf)
{
    dictitem_T	*di;

    if (arg->v_type != VAR_DICT)
    {
	emsg(_(e_dictionary_required));
	return FAIL;
    }
    if (arg->vval.v_dict == NULL)
	return OK;  // NULL dict is like an empty dict
    di = dict_find(arg->vval.v_dict, (char_u *)"bufnr", -1);
    if (di != NULL && (di->di_tv.v_type != VAR_NUMBER
					      || di->di_tv.vval.v_number != 0))
    {
	*buf = get_buf_arg(&di->di_tv);
	if (*buf == NULL)
	    return FAIL;
    }
    return OK;
}

/*
 * prop_add({lnum}, {col}, {props})
 */
    void
f_prop_add(typval_T *argvars, typval_T *rettv)
{
    linenr_T	start_lnum;
    colnr_T	start_col;

    if (in_vim9script()
	    && (check_for_number_arg(argvars, 0) == FAIL
		|| check_for_number_arg(argvars, 1) == FAIL
		|| check_for_dict_arg(argvars, 2) == FAIL))
	return;

    start_lnum = tv_get_number(&argvars[0]);
    start_col = tv_get_number(&argvars[1]);
    if (check_for_dict_arg(argvars, 2) == FAIL)
	return;

    rettv->vval.v_number = prop_add_common(start_lnum, start_col,
				 argvars[2].vval.v_dict, curbuf, &argvars[2]);
}

/*
 * Attach a text property 'type_name' to the text starting
 * at [start_lnum, start_col] and ending at [end_lnum, end_col] in
 * the buffer "buf" and assign identifier "id".
 * When "text" is not NULL add it to buf->b_textprop_text[-id - 1].
 */
    static int
prop_add_one(
	buf_T		*buf,
	char_u		*type_name,
	int		id,
	char_u		*text_arg,
	int		text_padding_left,
	int		text_flags,
	linenr_T	start_lnum,
	linenr_T	end_lnum,
	colnr_T		start_col,
	colnr_T		end_col)
{
    proptype_T	*type;
    linenr_T	lnum;
    int		proplen;
    char_u	*props = NULL;
    char_u	*newprops;
    size_t	textlen;
    char_u	*newtext;
    int		i;
    textprop_T	tmp_prop;
    char_u	*text = text_arg;
    int		res = FAIL;

    type = lookup_prop_type(type_name, buf);
    if (type == NULL)
	goto theend;

    if (start_lnum < 1 || start_lnum > buf->b_ml.ml_line_count)
    {
	semsg(_(e_invalid_line_number_nr), (long)start_lnum);
	goto theend;
    }
    if (end_lnum < start_lnum || end_lnum > buf->b_ml.ml_line_count)
    {
	semsg(_(e_invalid_line_number_nr), (long)end_lnum);
	goto theend;
    }

    if (buf->b_ml.ml_mfp == NULL)
    {
	emsg(_(e_cannot_add_text_property_to_unloaded_buffer));
	goto theend;
    }

    if (text != NULL)
    {
	garray_T    *gap = &buf->b_textprop_text;
	char_u	    *p;

	// double check we got the right ID
	if (-id - 1 != gap->ga_len)
	    iemsg("text prop ID mismatch");
	if (gap->ga_growsize == 0)
	    ga_init2(gap, sizeof(char *), 50);
	if (ga_grow(gap, 1) == FAIL)
	    goto theend;
	((char_u **)gap->ga_data)[gap->ga_len++] = text;

	// change any control character (Tab, Newline, etc.) to a Space to make
	// it simpler to compute the size
	for (p = text; *p != NUL; MB_PTR_ADV(p))
	    if (*p < ' ')
		*p = ' ';
	text = NULL;
    }

    for (lnum = start_lnum; lnum <= end_lnum; ++lnum)
    {
	colnr_T col;	    // start column use in tp_col
	colnr_T sort_col;   // column where it appears
	long	length;	    // in bytes

	// Fetch the line to get the ml_line_len field updated.
	proplen = get_text_props(buf, lnum, &props, TRUE);
	textlen = buf->b_ml.ml_line_len - proplen * sizeof(textprop_T);

	if (lnum == start_lnum)
	    col = start_col;
	else
	    col = 1;
	if (col - 1 > (colnr_T)textlen && !(col == 0 && text_arg != NULL))
	{
	    semsg(_(e_invalid_column_number_nr), (long)start_col);
	    goto theend;
	}
	sort_col = col;

	if (lnum == end_lnum)
	    length = end_col - col;
	else
	    length = (int)textlen - col + 1;
	if (length > (long)textlen)
	    length = (int)textlen;	// can include the end-of-line
	if (length < 0)
	    length = 0;		// zero-width property

	if (text_arg != NULL)
	{
	    length = 1;		// text is placed on one character
	    if (col == 0)
	    {
		col = MAXCOL;	// before or after the line
		if ((text_flags & TP_FLAG_ALIGN_ABOVE) == 0)
		    sort_col = MAXCOL;
		length += text_padding_left;
	    }
	}

	// Allocate the new line with space for the new property.
	newtext = alloc(buf->b_ml.ml_line_len + sizeof(textprop_T));
	if (newtext == NULL)
	    goto theend;
	// Copy the text, including terminating NUL.
	mch_memmove(newtext, buf->b_ml.ml_line_ptr, textlen);

	// Find the index where to insert the new property.
	// Since the text properties are not aligned properly when stored with
	// the text, we need to copy them as bytes before using it as a struct.
	for (i = 0; i < proplen; ++i)
	{
	    colnr_T prop_col;

	    mch_memmove(&tmp_prop, props + i * sizeof(textprop_T),
							   sizeof(textprop_T));
	    // col is MAXCOL when the text goes above or after the line, when
	    // above we should use column zero for sorting
	    prop_col = (tmp_prop.tp_flags & TP_FLAG_ALIGN_ABOVE)
				? 0 : tmp_prop.tp_col;
	    if (prop_col >= sort_col)
		break;
	}
	newprops = newtext + textlen;
	if (i > 0)
	    mch_memmove(newprops, props, sizeof(textprop_T) * i);

	tmp_prop.tp_col = col;
	tmp_prop.tp_len = length;
	tmp_prop.tp_id = id;
	tmp_prop.tp_type = type->pt_id;
	tmp_prop.tp_flags = text_flags
			    | (lnum > start_lnum ? TP_FLAG_CONT_PREV : 0)
			    | (lnum < end_lnum ? TP_FLAG_CONT_NEXT : 0)
			    | ((type->pt_flags & PT_FLAG_INS_START_INCL)
						     ? TP_FLAG_START_INCL : 0);
	mch_memmove(newprops + i * sizeof(textprop_T), &tmp_prop,
							   sizeof(textprop_T));

	if (i < proplen)
	    mch_memmove(newprops + (i + 1) * sizeof(textprop_T),
					    props + i * sizeof(textprop_T),
					    sizeof(textprop_T) * (proplen - i));

	if (buf->b_ml.ml_flags & (ML_LINE_DIRTY | ML_ALLOCATED))
	    vim_free(buf->b_ml.ml_line_ptr);
	buf->b_ml.ml_line_ptr = newtext;
	buf->b_ml.ml_line_len += sizeof(textprop_T);
	buf->b_ml.ml_flags |= ML_LINE_DIRTY;
    }

    changed_line_display_buf(buf);
    changed_lines_buf(buf, start_lnum, end_lnum + 1, 0);
    res = OK;

theend:
    vim_free(text);
    return res;
}

/*
 * prop_add_list()
 * First argument specifies the text property:
 *   {'type': <str>, 'id': <num>, 'bufnr': <num>}
 * Second argument is a List where each item is a List with the following
 * entries: [lnum, start_col, end_col]
 */
    void
f_prop_add_list(typval_T *argvars, typval_T *rettv UNUSED)
{
    dict_T	*dict;
    char_u	*type_name;
    buf_T	*buf = curbuf;
    int		id = 0;
    listitem_T	*li;
    list_T	*pos_list;
    linenr_T	start_lnum;
    colnr_T	start_col;
    linenr_T	end_lnum;
    colnr_T	end_col;
    int		error = FALSE;
    int		prev_did_emsg = did_emsg;

    if (check_for_dict_arg(argvars, 0) == FAIL
	    || check_for_list_arg(argvars, 1) == FAIL)
	return;

    if (check_for_nonnull_list_arg(argvars, 1) == FAIL)
	return;

    dict = argvars[0].vval.v_dict;
    if (dict == NULL || !dict_has_key(dict, "type"))
    {
	emsg(_(e_missing_property_type_name));
	return;
    }
    type_name = dict_get_string(dict, "type", FALSE);

    if (dict_has_key(dict, "id"))
	id = dict_get_number(dict, "id");

    if (get_bufnr_from_arg(&argvars[0], &buf) == FAIL)
	return;

    // This must be done _before_ we start adding properties because property
    // changes trigger buffer (memline) reorganisation, which needs this flag
    // to be correctly set.
    buf->b_has_textprop = TRUE;  // this is never reset
    FOR_ALL_LIST_ITEMS(argvars[1].vval.v_list, li)
    {
	if (li->li_tv.v_type != VAR_LIST || li->li_tv.vval.v_list == NULL)
	{
	    emsg(_(e_list_required));
	    return;
	}

	pos_list = li->li_tv.vval.v_list;
	start_lnum = list_find_nr(pos_list, 0L, &error);
	if (!error)
	    start_col = list_find_nr(pos_list, 1L, &error);
	if (!error)
	    end_lnum = list_find_nr(pos_list, 2L, &error);
	if (!error)
	    end_col = list_find_nr(pos_list, 3L, &error);
	int this_id = id;
	if (!error && pos_list->lv_len > 4)
	    this_id = list_find_nr(pos_list, 4L, &error);
	if (error || start_lnum <= 0 || start_col <= 0
		  || end_lnum <= 0 || end_col <= 0)
	{
	    if (prev_did_emsg == did_emsg)
		emsg(_(e_invalid_argument));
	    return;
	}
	if (prop_add_one(buf, type_name, this_id, NULL, 0, 0,
			     start_lnum, end_lnum, start_col, end_col) == FAIL)
	    return;
    }

    redraw_buf_later(buf, UPD_VALID);
}

/*
 * Get the next ID to use for a textprop with text in buffer "buf".
 */
    static int
get_textprop_id(buf_T *buf)
{
    // TODO: recycle deleted entries
    return -(buf->b_textprop_text.ga_len + 1);
}

// Flag that is set when a negative ID isused for a normal text property.
// It is then impossible to use virtual text properties.
static int did_use_negative_pop_id = FALSE;

/*
 * Shared between prop_add() and popup_create().
 * "dict_arg" is the function argument of a dict containing "bufnr".
 * it is NULL for popup_create().
 * Returns the "id" used for "text" or zero.
 */
    int
prop_add_common(
	linenr_T    start_lnum,
	colnr_T	    start_col,
	dict_T	    *dict,
	buf_T	    *default_buf,
	typval_T    *dict_arg)
{
    linenr_T	end_lnum;
    colnr_T	end_col;
    char_u	*type_name;
    buf_T	*buf = default_buf;
    int		id = 0;
    char_u	*text = NULL;
    int		text_padding_left = 0;
    int		flags = 0;

    if (dict == NULL || !dict_has_key(dict, "type"))
    {
	emsg(_(e_missing_property_type_name));
	goto theend;
    }
    type_name = dict_get_string(dict, "type", FALSE);

    if (dict_has_key(dict, "end_lnum"))
    {
	end_lnum = dict_get_number(dict, "end_lnum");
	if (end_lnum < start_lnum)
	{
	    semsg(_(e_invalid_value_for_argument_str), "end_lnum");
	    goto theend;
	}
    }
    else
	end_lnum = start_lnum;

    if (dict_has_key(dict, "length"))
    {
	long length = dict_get_number(dict, "length");

	if (length < 0 || end_lnum > start_lnum)
	{
	    semsg(_(e_invalid_value_for_argument_str), "length");
	    goto theend;
	}
	end_col = start_col + length;
    }
    else if (dict_has_key(dict, "end_col"))
    {
	end_col = dict_get_number(dict, "end_col");
	if (end_col <= 0)
	{
	    semsg(_(e_invalid_value_for_argument_str), "end_col");
	    goto theend;
	}
    }
    else if (start_lnum == end_lnum)
	end_col = start_col;
    else
	end_col = 1;

    if (dict_has_key(dict, "id"))
	id = dict_get_number(dict, "id");

    if (dict_has_key(dict, "text"))
    {
	if (dict_has_key(dict, "length")
		|| dict_has_key(dict, "end_col")
		|| dict_has_key(dict, "end_lnum"))
	{
	    emsg(_(e_cannot_use_length_endcol_and_endlnum_with_text));
	    goto theend;
	}

	text = dict_get_string(dict, "text", TRUE);
	if (text == NULL)
	    goto theend;
	// use a default length of 1 to make multiple props show up
	end_col = start_col + 1;

	if (dict_has_key(dict, "text_align"))
	{
	    char_u *p = dict_get_string(dict, "text_align", FALSE);

	    if (p == NULL)
		goto theend;
	    if (start_col != 0)
	    {
		emsg(_(e_can_only_use_text_align_when_column_is_zero));
		goto theend;
	    }
	    if (STRCMP(p, "right") == 0)
		flags |= TP_FLAG_ALIGN_RIGHT;
	    else if (STRCMP(p, "above") == 0)
		flags |= TP_FLAG_ALIGN_ABOVE;
	    else if (STRCMP(p, "below") == 0)
		flags |= TP_FLAG_ALIGN_BELOW;
	    else if (STRCMP(p, "after") != 0)
	    {
		semsg(_(e_invalid_value_for_argument_str_str), "text_align", p);
		goto theend;
	    }
	}

	if (dict_has_key(dict, "text_padding_left"))
	{
	    text_padding_left = dict_get_number(dict, "text_padding_left");
	    if (text_padding_left < 0)
	    {
		semsg(_(e_argument_must_be_positive_str), "text_padding_left");
		goto theend;
	    }
	}

	if (dict_has_key(dict, "text_wrap"))
	{
	    char_u *p = dict_get_string(dict, "text_wrap", FALSE);

	    if (p == NULL)
		goto theend;
	    if (STRCMP(p, "wrap") == 0)
		flags |= TP_FLAG_WRAP;
	    else if (STRCMP(p, "truncate") != 0)
	    {
		semsg(_(e_invalid_value_for_argument_str_str), "text_wrap", p);
		goto theend;
	    }
	}
    }

    // Column must be 1 or more for a normal text property; when "text" is
    // present zero means it goes after the line.
    if (start_col < (text == NULL ? 1 : 0))
    {
	semsg(_(e_invalid_column_number_nr), (long)start_col);
	goto theend;
    }
    if (start_col > 0 && text_padding_left > 0)
    {
	emsg(_(e_can_only_use_left_padding_when_column_is_zero));
	goto theend;
    }

    if (dict_arg != NULL && get_bufnr_from_arg(dict_arg, &buf) == FAIL)
	goto theend;

    if (id < 0)
    {
	if (buf->b_textprop_text.ga_len > 0)
	{
	    emsg(_(e_cannot_use_negative_id_after_adding_textprop_with_text));
	    goto theend;
	}
	did_use_negative_pop_id = TRUE;
    }

    if (text != NULL)
    {
	if (did_use_negative_pop_id)
	{
	    emsg(_(e_cannot_add_textprop_with_text_after_using_textprop_with_negative_id));
	    goto theend;
	}
	id = get_textprop_id(buf);
    }

    // This must be done _before_ we add the property because property changes
    // trigger buffer (memline) reorganisation, which needs this flag to be
    // correctly set.
    buf->b_has_textprop = TRUE;  // this is never reset

    prop_add_one(buf, type_name, id, text, text_padding_left, flags,
				    start_lnum, end_lnum, start_col, end_col);
    text = NULL;

    redraw_buf_later(buf, UPD_VALID);

theend:
    vim_free(text);
    return id;
}

/*
 * Fetch the text properties for line "lnum" in buffer "buf".
 * Returns the number of text properties and, when non-zero, a pointer to the
 * first one in "props" (note that it is not aligned, therefore the char_u
 * pointer).
 */
    int
get_text_props(buf_T *buf, linenr_T lnum, char_u **props, int will_change)
{
    char_u *text;
    size_t textlen;
    size_t proplen;

    // Be quick when no text property types have been defined or the buffer,
    // unless we are adding one.
    if ((!buf->b_has_textprop && !will_change) || buf->b_ml.ml_mfp == NULL)
	return 0;

    // Fetch the line to get the ml_line_len field updated.
    text = ml_get_buf(buf, lnum, will_change);
    textlen = STRLEN(text) + 1;
    proplen = buf->b_ml.ml_line_len - textlen;
    if (proplen == 0)
	return 0;
    if (proplen % sizeof(textprop_T) != 0)
    {
	iemsg(e_text_property_info_corrupted);
	return 0;
    }
    *props = text + textlen;
    return (int)(proplen / sizeof(textprop_T));
}

/*
 * Return the number of text properties with "above" or "below" alignment in
 * line "lnum".  A "right" aligned property also goes below after a "below" or
 * other "right" aligned property.
 */
    int
prop_count_above_below(buf_T *buf, linenr_T lnum)
{
    char_u	*props;
    int		count = get_text_props(buf, lnum, &props, FALSE);
    int		result = 0;
    textprop_T	prop;
    int		i;
    int		next_right_goes_below = FALSE;

    if (count == 0)
	return 0;
    for (i = 0; i < count; ++i)
    {
	mch_memmove(&prop, props + i * sizeof(prop), sizeof(prop));
	if (prop.tp_col == MAXCOL && text_prop_type_valid(buf, &prop))
	{
	    if ((prop.tp_flags & TP_FLAG_ALIGN_BELOW)
		    || (next_right_goes_below
				     && (prop.tp_flags & TP_FLAG_ALIGN_RIGHT)))
	    {
		next_right_goes_below = TRUE;
		++result;
	    }
	    else if (prop.tp_flags & TP_FLAG_ALIGN_ABOVE)
	    {
		next_right_goes_below = FALSE;
		++result;
	    }
	    else if (prop.tp_flags & TP_FLAG_ALIGN_RIGHT)
		next_right_goes_below = TRUE;
	}
    }
    return result;
}

/**
 * Return the number of text properties on line "lnum" in the current buffer.
 * When "only_starting" is true only text properties starting in this line will
 * be considered.
 * When "last_line" is FALSE then text properties after the line are not
 * counted.
 */
    int
count_props(linenr_T lnum, int only_starting, int last_line)
{
    char_u	*props;
    int		proplen = get_text_props(curbuf, lnum, &props, 0);
    int		result = proplen;
    int		i;
    textprop_T	prop;

    for (i = 0; i < proplen; ++i)
    {
	mch_memmove(&prop, props + i * sizeof(prop), sizeof(prop));
	// A prop is dropped when in the first line and it continues from the
	// previous line, or when not in the last line and it is virtual text
	// after the line.
	if ((only_starting && (prop.tp_flags & TP_FLAG_CONT_PREV))
		|| (!last_line && prop.tp_col == MAXCOL)
		|| !text_prop_type_valid(curbuf, &prop))
	    --result;
    }
    return result;
}

static textprop_T	*text_prop_compare_props;
static buf_T		*text_prop_compare_buf;

/*
 * Score for sorting on position of the text property: 0: above,
 * 1: after (default), 2: right, 3: below (comes last)
 */
    static int
text_prop_order(int flags)
{
    if (flags & TP_FLAG_ALIGN_ABOVE)
	return 0;
    if (flags & TP_FLAG_ALIGN_RIGHT)
	return 2;
    if (flags & TP_FLAG_ALIGN_BELOW)
	return 3;
    return 1;
}

/*
 * Function passed to qsort() to sort text properties.
 * Return 1 if "s1" has priority over "s2", -1 if the other way around, zero if
 * both have the same priority.
 */
    static int
text_prop_compare(const void *s1, const void *s2)
{
    int  idx1, idx2;
    textprop_T	*tp1, *tp2;
    proptype_T  *pt1, *pt2;
    colnr_T col1, col2;

    idx1 = *(int *)s1;
    idx2 = *(int *)s2;
    tp1 = &text_prop_compare_props[idx1];
    tp2 = &text_prop_compare_props[idx2];
    col1 = tp1->tp_col;
    col2 = tp2->tp_col;
    if (col1 == MAXCOL && col2 == MAXCOL)
    {
	int order1 = text_prop_order(tp1->tp_flags);
	int order2 = text_prop_order(tp2->tp_flags);

	// both props add text before or after the line, sort on order where it
	// is added
	if (order1 != order2)
	    return order1 < order2 ? 1 : -1;
    }

    // property that inserts text has priority over one that doesn't
    if ((tp1->tp_id < 0) != (tp2->tp_id < 0))
	return tp1->tp_id < 0 ? 1 : -1;

    // check highest priority, defined by the type
    pt1 = text_prop_type_by_id(text_prop_compare_buf, tp1->tp_type);
    pt2 = text_prop_type_by_id(text_prop_compare_buf, tp2->tp_type);
    if (pt1 != pt2)
    {
	if (pt1 == NULL)
	    return -1;
	if (pt2 == NULL)
	    return 1;
	if (pt1->pt_priority != pt2->pt_priority)
	    return pt1->pt_priority > pt2->pt_priority ? 1 : -1;
    }

    // same priority, one that starts first wins
    if (col1 != col2)
	return col1 < col2 ? 1 : -1;

    // for a property with text the id can be used as tie breaker
    if (tp1->tp_id < 0)
	return tp1->tp_id > tp2->tp_id ? 1 : -1;

    return 0;
}

/*
 * Sort "count" text properties using an array if indexes "idxs" into the list
 * of text props "props" for buffer "buf".
 */
    void
sort_text_props(
	buf_T	    *buf,
	textprop_T  *props,
	int	    *idxs,
	int	    count)
{
    text_prop_compare_buf = buf;
    text_prop_compare_props = props;
    qsort((void *)idxs, (size_t)count, sizeof(int), text_prop_compare);
}

/*
 * Find text property "type_id" in the visible lines of window "wp".
 * Match "id" when it is > 0.
 * Returns FAIL when not found.
 */
    int
find_visible_prop(
	win_T	    *wp,
	int	    type_id,
	int	    id,
	textprop_T  *prop,
	linenr_T    *found_lnum)
{
    // return when "type_id" no longer exists
    if (text_prop_type_by_id(wp->w_buffer, type_id) == NULL)
	return FAIL;

    // w_botline may not have been updated yet.
    validate_botline_win(wp);
    for (linenr_T lnum = wp->w_topline; lnum < wp->w_botline; ++lnum)
    {
	char_u	*props;
	int	count = get_text_props(wp->w_buffer, lnum, &props, FALSE);
	for (int i = 0; i < count; ++i)
	{
	    mch_memmove(prop, props + i * sizeof(textprop_T),
							   sizeof(textprop_T));
	    if (prop->tp_type == type_id && (id <= 0 || prop->tp_id == id))
	    {
		*found_lnum = lnum;
		return OK;
	    }
	}
    }
    return FAIL;
}

/*
 * Set the text properties for line "lnum" to "props" with length "len".
 * If "len" is zero text properties are removed, "props" is not used.
 * Any existing text properties are dropped.
 * Only works for the current buffer.
 */
    static void
set_text_props(linenr_T lnum, char_u *props, int len)
{
    char_u  *text;
    char_u  *newtext;
    int	    textlen;

    text = ml_get(lnum);
    textlen = (int)STRLEN(text) + 1;
    newtext = alloc(textlen + len);
    if (newtext == NULL)
	return;
    mch_memmove(newtext, text, textlen);
    if (len > 0)
	mch_memmove(newtext + textlen, props, len);
    if (curbuf->b_ml.ml_flags & (ML_LINE_DIRTY | ML_ALLOCATED))
	vim_free(curbuf->b_ml.ml_line_ptr);
    curbuf->b_ml.ml_line_ptr = newtext;
    curbuf->b_ml.ml_line_len = textlen + len;
    curbuf->b_ml.ml_flags |= ML_LINE_DIRTY;
}

/*
 * Add "text_props" with "text_prop_count" text properties to line "lnum".
 */
    void
add_text_props(linenr_T lnum, textprop_T *text_props, int text_prop_count)
{
    char_u  *text;
    char_u  *newtext;
    int	    proplen = text_prop_count * (int)sizeof(textprop_T);

    text = ml_get(lnum);
    newtext = alloc(curbuf->b_ml.ml_line_len + proplen);
    if (newtext == NULL)
	return;
    mch_memmove(newtext, text, curbuf->b_ml.ml_line_len);
    mch_memmove(newtext + curbuf->b_ml.ml_line_len, text_props, proplen);
    if (curbuf->b_ml.ml_flags & (ML_LINE_DIRTY | ML_ALLOCATED))
	vim_free(curbuf->b_ml.ml_line_ptr);
    curbuf->b_ml.ml_line_ptr = newtext;
    curbuf->b_ml.ml_line_len += proplen;
    curbuf->b_ml.ml_flags |= ML_LINE_DIRTY;
}

/*
 * Function passed to qsort() for sorting proptype_T on pt_id.
 */
    static int
compare_pt(const void *s1, const void *s2)
{
    proptype_T	*tp1 = *(proptype_T **)s1;
    proptype_T	*tp2 = *(proptype_T **)s2;

    return tp1->pt_id == tp2->pt_id ? 0 : tp1->pt_id < tp2->pt_id ? -1 : 1;
}

    static proptype_T *
find_type_by_id(hashtab_T *ht, proptype_T ***array, int id)
{
    int low = 0;
    int high;

    if (ht == NULL || ht->ht_used == 0)
	return NULL;

    // Make the lookup faster by creating an array with pointers to
    // hashtable entries, sorted on pt_id.
    if (*array == NULL)
    {
	long	    todo;
	hashitem_T  *hi;
	int	    i = 0;

	*array = ALLOC_MULT(proptype_T *, ht->ht_used);
	if (*array == NULL)
	    return NULL;
	todo = (long)ht->ht_used;
	FOR_ALL_HASHTAB_ITEMS(ht, hi, todo)
	{
	    if (!HASHITEM_EMPTY(hi))
	    {
		(*array)[i++] = HI2PT(hi);
		--todo;
	    }
	}
	qsort((void *)*array, ht->ht_used, sizeof(proptype_T *), compare_pt);
    }

    // binary search in the sorted array
    high = ht->ht_used;
    while (high > low)
    {
	int m = (high + low) / 2;

	if ((*array)[m]->pt_id == id)
	    return (*array)[m];
	if ((*array)[m]->pt_id > id)
	    high = m;
	else
	    low = m + 1;
    }
    return NULL;
}

/*
 * Fill 'dict' with text properties in 'prop'.
 */
    static void
prop_fill_dict(dict_T *dict, textprop_T *prop, buf_T *buf)
{
    proptype_T *pt;
    int buflocal = TRUE;

    dict_add_number(dict, "col", prop->tp_col);
    dict_add_number(dict, "length", prop->tp_len);
    dict_add_number(dict, "id", prop->tp_id);
    dict_add_number(dict, "start", !(prop->tp_flags & TP_FLAG_CONT_PREV));
    dict_add_number(dict, "end", !(prop->tp_flags & TP_FLAG_CONT_NEXT));

    pt = find_type_by_id(buf->b_proptypes, &buf->b_proparray, prop->tp_type);
    if (pt == NULL)
    {
	pt = find_type_by_id(global_proptypes, &global_proparray,
								prop->tp_type);
	buflocal = FALSE;
    }
    if (pt != NULL)
	dict_add_string(dict, "type", pt->pt_name);

    if (buflocal)
	dict_add_number(dict, "type_bufnr", buf->b_fnum);
    else
	dict_add_number(dict, "type_bufnr", 0);
    if (prop->tp_id < 0)
    {
	// virtual text property
	garray_T    *gap = &buf->b_textprop_text;
	char_u	    *text;

	// negate the property id to get the string index
	text = ((char_u **)gap->ga_data)[-prop->tp_id - 1];
	dict_add_string(dict, "text", text);

	// text_align
	char_u	    *text_align = NULL;
	if (prop->tp_flags & TP_FLAG_ALIGN_RIGHT)
	    text_align = (char_u *)"right";
	else if (prop->tp_flags & TP_FLAG_ALIGN_ABOVE)
	    text_align = (char_u *)"above";
	else if (prop->tp_flags & TP_FLAG_ALIGN_BELOW)
	    text_align = (char_u *)"below";
	if (text_align != NULL)
	    dict_add_string(dict, "text_align", text_align);

	// text_wrap
	if (prop->tp_flags & TP_FLAG_WRAP)
	    dict_add_string(dict, "text_wrap", (char_u *)"wrap");
    }
}

/*
 * Find a property type by ID in "buf" or globally.
 * Returns NULL if not found.
 */
    proptype_T *
text_prop_type_by_id(buf_T *buf, int id)
{
    proptype_T *type;

    type = find_type_by_id(buf->b_proptypes, &buf->b_proparray, id);
    if (type == NULL)
	type = find_type_by_id(global_proptypes, &global_proparray, id);
    return type;
}

/*
 * Return TRUE if "prop" is a valid text property type.
 */
    int
text_prop_type_valid(buf_T *buf, textprop_T *prop)
{
    return text_prop_type_by_id(buf, prop->tp_type) != NULL;
}

/*
 * prop_clear({lnum} [, {lnum_end} [, {bufnr}]])
 */
    void
f_prop_clear(typval_T *argvars, typval_T *rettv UNUSED)
{
    linenr_T start;
    linenr_T end;
    linenr_T lnum;
    buf_T    *buf = curbuf;
    int	    did_clear = FALSE;

    if (in_vim9script()
	    && (check_for_number_arg(argvars, 0) == FAIL
		|| check_for_opt_number_arg(argvars, 1) == FAIL
		|| (argvars[1].v_type != VAR_UNKNOWN
		    && check_for_opt_dict_arg(argvars, 2) == FAIL)))
	return;

    start = tv_get_number(&argvars[0]);
    end = start;
    if (argvars[1].v_type != VAR_UNKNOWN)
    {
	end = tv_get_number(&argvars[1]);
	if (argvars[2].v_type != VAR_UNKNOWN)
	{
	    if (get_bufnr_from_arg(&argvars[2], &buf) == FAIL)
		return;
	}
    }
    if (start < 1 || end < 1)
    {
	emsg(_(e_invalid_range));
	return;
    }

    for (lnum = start; lnum <= end; ++lnum)
    {
	char_u *text;
	size_t len;

	if (lnum > buf->b_ml.ml_line_count)
	    break;
	text = ml_get_buf(buf, lnum, FALSE);
	len = STRLEN(text) + 1;
	if ((size_t)buf->b_ml.ml_line_len > len)
	{
	    did_clear = TRUE;
	    if (!(buf->b_ml.ml_flags & ML_LINE_DIRTY))
	    {
		char_u *newtext = vim_strsave(text);

		// need to allocate the line now
		if (newtext == NULL)
		    return;
		if (buf->b_ml.ml_flags & ML_ALLOCATED)
		    vim_free(buf->b_ml.ml_line_ptr);
		buf->b_ml.ml_line_ptr = newtext;
		buf->b_ml.ml_flags |= ML_LINE_DIRTY;
	    }
	    buf->b_ml.ml_line_len = (int)len;
	}
    }
    if (did_clear)
	redraw_buf_later(buf, UPD_NOT_VALID);
}

/*
 * prop_find({props} [, {direction}])
 */
    void
f_prop_find(typval_T *argvars, typval_T *rettv)
{
    pos_T       *cursor = &curwin->w_cursor;
    dict_T      *dict;
    buf_T       *buf = curbuf;
    dictitem_T  *di;
    int		lnum_start;
    int		start_pos_has_prop = 0;
    int		seen_end = FALSE;
    int		id = 0;
    int		id_found = FALSE;
    int		type_id = -1;
    int		skipstart = FALSE;
    int		lnum = -1;
    int		col = -1;
    int		dir = FORWARD;    // FORWARD == 1, BACKWARD == -1
    int		both;

    if (in_vim9script()
	    && (check_for_dict_arg(argvars, 0) == FAIL
		|| check_for_opt_string_arg(argvars, 1) == FAIL))
	return;

    if (check_for_nonnull_dict_arg(argvars, 0) == FAIL)
	return;
    dict = argvars[0].vval.v_dict;

    if (get_bufnr_from_arg(&argvars[0], &buf) == FAIL)
	return;
    if (buf->b_ml.ml_mfp == NULL)
	return;

    if (argvars[1].v_type != VAR_UNKNOWN)
    {
	char_u      *dir_s = tv_get_string(&argvars[1]);

	if (*dir_s == 'b')
	    dir = BACKWARD;
	else if (*dir_s != 'f')
	{
	    emsg(_(e_invalid_argument));
	    return;
	}
    }

    di = dict_find(dict, (char_u *)"lnum", -1);
    if (di != NULL)
	lnum = tv_get_number(&di->di_tv);

    di = dict_find(dict, (char_u *)"col", -1);
    if (di != NULL)
	col = tv_get_number(&di->di_tv);

    if (lnum == -1)
    {
	lnum = cursor->lnum;
	col = cursor->col + 1;
    }
    else if (col == -1)
	col = 1;

    if (lnum < 1 || lnum > buf->b_ml.ml_line_count)
    {
	emsg(_(e_invalid_range));
	return;
    }

    skipstart = dict_get_bool(dict, "skipstart", 0);

    if (dict_has_key(dict, "id"))
    {
	id = dict_get_number(dict, "id");
	id_found = TRUE;
    }
    if (dict_has_key(dict, "type"))
    {
	char_u	    *name = dict_get_string(dict, "type", FALSE);
	proptype_T  *type = lookup_prop_type(name, buf);

	if (type == NULL)
	    return;
	type_id = type->pt_id;
    }
    both = dict_get_bool(dict, "both", FALSE);
    if (!id_found && type_id == -1)
    {
	emsg(_(e_need_at_least_one_of_id_or_type));
	return;
    }
    if (both && (!id_found || type_id == -1))
    {
	emsg(_(e_need_id_and_type_or_types_with_both));
	return;
    }

    lnum_start = lnum;

    if (rettv_dict_alloc(rettv) == FAIL)
	return;

    while (1)
    {
	char_u	*text = ml_get_buf(buf, lnum, FALSE);
	size_t	textlen = STRLEN(text) + 1;
	int	count = (int)((buf->b_ml.ml_line_len - textlen)
							 / sizeof(textprop_T));
	int	    i;
	textprop_T  prop;
	int	    prop_start;
	int	    prop_end;

	for (i = dir == BACKWARD ? count - 1 : 0; i >= 0 && i < count; i += dir)
	{
	    mch_memmove(&prop, text + textlen + i * sizeof(textprop_T),
							   sizeof(textprop_T));

	    // For the very first line try to find the first property before or
	    // after `col`, depending on the search direction.
	    if (lnum == lnum_start)
	    {
		if (dir == BACKWARD)
		{
		    if (prop.tp_col > col)
			continue;
		}
		else if (prop.tp_col + prop.tp_len - (prop.tp_len != 0) < col)
		    continue;
	    }
	    if (both ? prop.tp_id == id && prop.tp_type == type_id
		     : (id_found && prop.tp_id == id)
						    || prop.tp_type == type_id)
	    {
		// Check if the starting position has text props.
		if (lnum_start == lnum
			&& col >= prop.tp_col
			&& (col <= prop.tp_col + prop.tp_len
							 - (prop.tp_len != 0)))
		    start_pos_has_prop = 1;

		// The property was not continued from last line, it starts on
		// this line.
		prop_start = !(prop.tp_flags & TP_FLAG_CONT_PREV);
		// The property does not continue on the next line, it ends on
		// this line.
		prop_end = !(prop.tp_flags & TP_FLAG_CONT_NEXT);
		if (!prop_start && prop_end && dir == FORWARD)
		    seen_end = 1;

		// Skip lines without the start flag.
		if (!prop_start)
		{
		    // Always search backwards for start when search started
		    // on a prop and we're not skipping.
		    if (start_pos_has_prop && !skipstart)
			dir = BACKWARD;
		    continue;
		}

		// If skipstart is true, skip the prop at start pos (even if
		// continued from another line).
		if (start_pos_has_prop && skipstart && !seen_end)
		{
		    start_pos_has_prop = 0;
		    continue;
		}

		prop_fill_dict(rettv->vval.v_dict, &prop, buf);
		dict_add_number(rettv->vval.v_dict, "lnum", lnum);

		return;
	    }
	}

	if (dir > 0)
	{
	    if (lnum >= buf->b_ml.ml_line_count)
		break;
	    lnum++;
	}
	else
	{
	    if (lnum <= 1)
		break;
	    lnum--;
	}
    }
}

/*
 * Returns TRUE if 'type_or_id' is in the 'types_or_ids' list.
 */
    static int
prop_type_or_id_in_list(int *types_or_ids, int len, int type_or_id)
{
    int i;

    for (i = 0; i < len; i++)
	if (types_or_ids[i] == type_or_id)
	    return TRUE;

    return FALSE;
}

/*
 * Return all the text properties in line 'lnum' in buffer 'buf' in 'retlist'.
 * If 'prop_types' is not NULL, then return only the text properties with
 * matching property type in the 'prop_types' array.
 * If 'prop_ids' is not NULL, then return only the text properties with
 * an identifier in the 'props_ids' array.
 * If 'add_lnum' is TRUE, then add the line number also to the text property
 * dictionary.
 */
    static void
get_props_in_line(
	buf_T		*buf,
	linenr_T	lnum,
	int		*prop_types,
	int		prop_types_len,
	int		*prop_ids,
	int		prop_ids_len,
	list_T		*retlist,
	int		add_lnum)
{
    char_u	*text = ml_get_buf(buf, lnum, FALSE);
    size_t	textlen = STRLEN(text) + 1;
    int		count;
    int		i;
    textprop_T	prop;

    count = (int)((buf->b_ml.ml_line_len - textlen) / sizeof(textprop_T));
    for (i = 0; i < count; ++i)
    {
	mch_memmove(&prop, text + textlen + i * sizeof(textprop_T),
		sizeof(textprop_T));
	if ((prop_types == NULL
		    || prop_type_or_id_in_list(prop_types, prop_types_len,
			prop.tp_type))
		&& (prop_ids == NULL
		    || prop_type_or_id_in_list(prop_ids, prop_ids_len,
								 prop.tp_id)))
	{
	    dict_T *d = dict_alloc();

	    if (d == NULL)
		break;
	    prop_fill_dict(d, &prop, buf);
	    if (add_lnum)
		dict_add_number(d, "lnum", lnum);
	    list_append_dict(retlist, d);
	}
    }
}

/*
 * Convert a List of property type names into an array of property type
 * identifiers. Returns a pointer to the allocated array. Returns NULL on
 * error. 'num_types' is set to the number of returned property types.
 */
    static int *
get_prop_types_from_names(list_T *l, buf_T *buf, int *num_types)
{
    int		*prop_types;
    listitem_T	*li;
    int		i;
    char_u	*name;
    proptype_T	*type;

    *num_types = 0;

    prop_types = ALLOC_MULT(int, list_len(l));
    if (prop_types == NULL)
	return NULL;

    i = 0;
    FOR_ALL_LIST_ITEMS(l, li)
    {
	if (li->li_tv.v_type != VAR_STRING)
	{
	    emsg(_(e_string_required));
	    goto errret;
	}
	name = li->li_tv.vval.v_string;
	if (name == NULL)
	    goto errret;

	type = lookup_prop_type(name, buf);
	if (type == NULL)
	    goto errret;
	prop_types[i++] = type->pt_id;
    }

    *num_types = i;
    return prop_types;

errret:
    VIM_CLEAR(prop_types);
    return NULL;
}

/*
 * Convert a List of property identifiers into an array of property
 * identifiers.  Returns a pointer to the allocated array. Returns NULL on
 * error. 'num_ids' is set to the number of returned property identifiers.
 */
    static int *
get_prop_ids_from_list(list_T *l, int *num_ids)
{
    int		*prop_ids;
    listitem_T	*li;
    int		i;
    int		id;
    int		error;

    *num_ids = 0;

    prop_ids = ALLOC_MULT(int, list_len(l));
    if (prop_ids == NULL)
	return NULL;

    i = 0;
    FOR_ALL_LIST_ITEMS(l, li)
    {
	error = FALSE;
	id = tv_get_number_chk(&li->li_tv, &error);
	if (error)
	    goto errret;

	prop_ids[i++] = id;
    }

    *num_ids = i;
    return prop_ids;

errret:
    VIM_CLEAR(prop_ids);
    return NULL;
}

/*
 * prop_list({lnum} [, {bufnr}])
 */
    void
f_prop_list(typval_T *argvars, typval_T *rettv)
{
    linenr_T	lnum;
    linenr_T	start_lnum;
    linenr_T	end_lnum;
    buf_T	*buf = curbuf;
    int		add_lnum = FALSE;
    int		*prop_types = NULL;
    int		prop_types_len = 0;
    int		*prop_ids = NULL;
    int		prop_ids_len = 0;
    list_T	*l;
    dictitem_T	*di;

    if (in_vim9script()
	    && (check_for_number_arg(argvars, 0) == FAIL
		|| check_for_opt_dict_arg(argvars, 1) == FAIL))
	return;

    if (rettv_list_alloc(rettv) == FAIL)
	return;

    // default: get text properties on current line
    start_lnum = tv_get_number(&argvars[0]);
    end_lnum = start_lnum;
    if (argvars[1].v_type != VAR_UNKNOWN)
    {
	dict_T *d;

	if (check_for_dict_arg(argvars, 1) == FAIL)
	    return;
	d = argvars[1].vval.v_dict;

	if (get_bufnr_from_arg(&argvars[1], &buf) == FAIL)
	    return;

	if (d != NULL && (di = dict_find(d, (char_u *)"end_lnum", -1)) != NULL)
	{
	    if (di->di_tv.v_type != VAR_NUMBER)
	    {
		emsg(_(e_number_required));
		return;
	    }
	    end_lnum = tv_get_number(&di->di_tv);
	    if (end_lnum < 0)
		// negative end_lnum is used as an offset from the last buffer
		// line
		end_lnum = buf->b_ml.ml_line_count + end_lnum + 1;
	    else if (end_lnum > buf->b_ml.ml_line_count)
		end_lnum = buf->b_ml.ml_line_count;
	    add_lnum = TRUE;
	}
	if (d != NULL && (di = dict_find(d, (char_u *)"types", -1)) != NULL)
	{
	    if (di->di_tv.v_type != VAR_LIST)
	    {
		emsg(_(e_list_required));
		return;
	    }

	    l = di->di_tv.vval.v_list;
	    if (l != NULL && list_len(l) > 0)
	    {
		prop_types = get_prop_types_from_names(l, buf, &prop_types_len);
		if (prop_types == NULL)
		    return;
	    }
	}
	if (d != NULL && (di = dict_find(d, (char_u *)"ids", -1)) != NULL)
	{
	    if (di->di_tv.v_type != VAR_LIST)
	    {
		emsg(_(e_list_required));
		goto errret;
	    }

	    l = di->di_tv.vval.v_list;
	    if (l != NULL && list_len(l) > 0)
	    {
		prop_ids = get_prop_ids_from_list(l, &prop_ids_len);
		if (prop_ids == NULL)
		    goto errret;
	    }
	}
    }
    if (start_lnum < 1 || start_lnum > buf->b_ml.ml_line_count
		|| end_lnum < 1 || end_lnum < start_lnum)
	emsg(_(e_invalid_range));
    else
	for (lnum = start_lnum; lnum <= end_lnum; lnum++)
	    get_props_in_line(buf, lnum, prop_types, prop_types_len,
		    prop_ids, prop_ids_len,
		    rettv->vval.v_list, add_lnum);

errret:
    VIM_CLEAR(prop_types);
    VIM_CLEAR(prop_ids);
}

/*
 * prop_remove({props} [, {lnum} [, {lnum_end}]])
 */
    void
f_prop_remove(typval_T *argvars, typval_T *rettv)
{
    linenr_T	start = 1;
    linenr_T	end = 0;
    linenr_T	lnum;
    linenr_T	first_changed = 0;
    linenr_T	last_changed = 0;
    dict_T	*dict;
    buf_T	*buf = curbuf;
    int		do_all;
    int		id = -MAXCOL;
    int		type_id = -1;	    // for a single "type"
    int		*type_ids = NULL;   // array, for a list of "types", allocated
    int		num_type_ids = 0;   // number of elements in "type_ids"
    int		both;
    int		did_remove_text = FALSE;

    rettv->vval.v_number = 0;

    if (in_vim9script()
	    && (check_for_dict_arg(argvars, 0) == FAIL
		|| check_for_opt_number_arg(argvars, 1) == FAIL
		|| (argvars[1].v_type != VAR_UNKNOWN
		    && check_for_opt_number_arg(argvars, 2) == FAIL)))
	return;

    if (check_for_nonnull_dict_arg(argvars, 0) == FAIL)
	return;

    if (argvars[1].v_type != VAR_UNKNOWN)
    {
	start = tv_get_number(&argvars[1]);
	end = start;
	if (argvars[2].v_type != VAR_UNKNOWN)
	    end = tv_get_number(&argvars[2]);
	if (start < 1 || end < 1)
	{
	    emsg(_(e_invalid_range));
	    return;
	}
    }

    dict = argvars[0].vval.v_dict;
    if (get_bufnr_from_arg(&argvars[0], &buf) == FAIL)
	return;
    if (buf->b_ml.ml_mfp == NULL)
	return;

    do_all = dict_get_bool(dict, "all", FALSE);

    if (dict_has_key(dict, "id"))
	id = dict_get_number(dict, "id");

    // if a specific type was supplied "type": check that (and ignore "types".
    // Otherwise check against the list of "types".
    if (dict_has_key(dict, "type"))
    {
	char_u	    *name = dict_get_string(dict, "type", FALSE);
	proptype_T  *type = lookup_prop_type(name, buf);

	if (type == NULL)
	    return;
	type_id = type->pt_id;
    }
    if (dict_has_key(dict, "types"))
    {
	typval_T types;
	listitem_T *li = NULL;

	dict_get_tv(dict, "types", &types);
	if (types.v_type == VAR_LIST && types.vval.v_list->lv_len > 0)
	{
	    type_ids = alloc( sizeof(int) * types.vval.v_list->lv_len );

	    FOR_ALL_LIST_ITEMS(types.vval.v_list, li)
	    {
		proptype_T *prop_type;

		if (li->li_tv.v_type != VAR_STRING)
		    continue;

		prop_type = lookup_prop_type(li->li_tv.vval.v_string, buf);

		if (!prop_type)
		    goto cleanup_prop_remove;

		type_ids[num_type_ids++] = prop_type->pt_id;
	    }
	}
    }
    both = dict_get_bool(dict, "both", FALSE);

    if (id == -MAXCOL && (type_id == -1 && num_type_ids == 0))
    {
	emsg(_(e_need_at_least_one_of_id_or_type));
	goto cleanup_prop_remove;
    }
    if (both && (id == -MAXCOL || (type_id == -1 && num_type_ids == 0)))
    {
	emsg(_(e_need_id_and_type_or_types_with_both));
	goto cleanup_prop_remove;
    }
    if (type_id != -1 && num_type_ids > 0)
    {
	emsg(_(e_cannot_specify_both_type_and_types));
	goto cleanup_prop_remove;
    }

    if (end == 0)
	end = buf->b_ml.ml_line_count;
    for (lnum = start; lnum <= end; ++lnum)
    {
	char_u *text;
	size_t len;

	if (lnum > buf->b_ml.ml_line_count)
	    break;
	text = ml_get_buf(buf, lnum, FALSE);
	len = STRLEN(text) + 1;
	if ((size_t)buf->b_ml.ml_line_len > len)
	{
	    static textprop_T	textprop;  // static because of alignment
	    unsigned		idx;

	    for (idx = 0; idx < (buf->b_ml.ml_line_len - len)
						   / sizeof(textprop_T); ++idx)
	    {
		char_u *cur_prop = buf->b_ml.ml_line_ptr + len
						    + idx * sizeof(textprop_T);
		size_t	taillen;
		int matches_id = 0;
		int matches_type = 0;

		mch_memmove(&textprop, cur_prop, sizeof(textprop_T));

		matches_id = textprop.tp_id == id;
		if (num_type_ids > 0)
		{
		    int idx2;

		    for (idx2 = 0; !matches_type && idx2 < num_type_ids; ++idx2)
			matches_type = textprop.tp_type == type_ids[idx2];
		}
		else
		{
		    matches_type = textprop.tp_type == type_id;
		}

		if (both ? matches_id && matches_type
			 : matches_id || matches_type)
		{
		    if (!(buf->b_ml.ml_flags & ML_LINE_DIRTY))
		    {
			char_u *newptr = alloc(buf->b_ml.ml_line_len);

			// need to allocate the line to be able to change it
			if (newptr == NULL)
			    goto cleanup_prop_remove;
			mch_memmove(newptr, buf->b_ml.ml_line_ptr,
							buf->b_ml.ml_line_len);
			if (buf->b_ml.ml_flags & ML_ALLOCATED)
			    vim_free(buf->b_ml.ml_line_ptr);
			buf->b_ml.ml_line_ptr = newptr;
			buf->b_ml.ml_flags |= ML_LINE_DIRTY;

			cur_prop = buf->b_ml.ml_line_ptr + len
						    + idx * sizeof(textprop_T);
		    }

		    taillen = buf->b_ml.ml_line_len - len
					      - (idx + 1) * sizeof(textprop_T);
		    if (taillen > 0)
			mch_memmove(cur_prop, cur_prop + sizeof(textprop_T),
								      taillen);
		    buf->b_ml.ml_line_len -= sizeof(textprop_T);
		    --idx;

		    if (textprop.tp_id < 0)
		    {
			garray_T    *gap = &buf->b_textprop_text;
			int	    ii = -textprop.tp_id - 1;

			// negative ID: property with text - free the text
			if (ii < gap->ga_len)
			{
			    char_u **p = ((char_u **)gap->ga_data) + ii;
			    VIM_CLEAR(*p);
			    did_remove_text = TRUE;
			}
		    }

		    if (first_changed == 0)
			first_changed = lnum;
		    last_changed = lnum;
		    ++rettv->vval.v_number;
		    if (!do_all)
			break;
		}
	    }
	}
    }

    if (first_changed > 0)
    {
	changed_line_display_buf(buf);
	changed_lines_buf(buf, first_changed, last_changed + 1, 0);
	redraw_buf_later(buf, UPD_VALID);
    }

    if (did_remove_text)
    {
	garray_T    *gap = &buf->b_textprop_text;

	// Reduce the growarray size for NULL pointers at the end.
	while (gap->ga_len > 0
			 && ((char_u **)gap->ga_data)[gap->ga_len - 1] == NULL)
	    --gap->ga_len;
    }

cleanup_prop_remove:
    vim_free(type_ids);
}

/*
 * Common for f_prop_type_add() and f_prop_type_change().
 */
    static void
prop_type_set(typval_T *argvars, int add)
{
    char_u	*name;
    buf_T	*buf = NULL;
    dict_T	*dict;
    dictitem_T  *di;
    proptype_T	*prop;

    if (in_vim9script()
	    && (check_for_string_arg(argvars, 0) == FAIL
		|| check_for_dict_arg(argvars, 1) == FAIL))
	return;

    name = tv_get_string(&argvars[0]);
    if (*name == NUL)
    {
	semsg(_(e_invalid_argument_str), "\"\"");
	return;
    }

    if (get_bufnr_from_arg(&argvars[1], &buf) == FAIL)
	return;
    dict = argvars[1].vval.v_dict;

    prop = find_prop_type(name, buf);
    if (add)
    {
	hashtab_T **htp;

	if (prop != NULL)
	{
	    semsg(_(e_property_type_str_already_defined), name);
	    return;
	}
	prop = alloc_clear(offsetof(proptype_T, pt_name) + STRLEN(name) + 1);
	if (prop == NULL)
	    return;
	STRCPY(prop->pt_name, name);
	prop->pt_id = ++proptype_id;
	prop->pt_flags = PT_FLAG_COMBINE;
	if (buf == NULL)
	{
	    htp = &global_proptypes;
	    VIM_CLEAR(global_proparray);
	}
	else
	{
	    htp = &buf->b_proptypes;
	    VIM_CLEAR(buf->b_proparray);
	}
	if (*htp == NULL)
	{
	    *htp = ALLOC_ONE(hashtab_T);
	    if (*htp == NULL)
	    {
		vim_free(prop);
		return;
	    }
	    hash_init(*htp);
	}
	hash_add(*htp, PT2HIKEY(prop), "prop type");
    }
    else
    {
	if (prop == NULL)
	{
	    semsg(_(e_property_type_str_does_not_exist), name);
	    return;
	}
    }

    if (dict != NULL)
    {
	di = dict_find(dict, (char_u *)"highlight", -1);
	if (di != NULL)
	{
	    char_u	*highlight;
	    int		hl_id = 0;

	    highlight = dict_get_string(dict, "highlight", FALSE);
	    if (highlight != NULL && *highlight != NUL)
		hl_id = syn_name2id(highlight);
	    if (hl_id <= 0)
	    {
		semsg(_(e_unknown_highlight_group_name_str),
			highlight == NULL ? (char_u *)"" : highlight);
		return;
	    }
	    prop->pt_hl_id = hl_id;
	}

	di = dict_find(dict, (char_u *)"combine", -1);
	if (di != NULL)
	{
	    if (tv_get_bool(&di->di_tv))
		prop->pt_flags |= PT_FLAG_COMBINE;
	    else
		prop->pt_flags &= ~PT_FLAG_COMBINE;
	}

	di = dict_find(dict, (char_u *)"override", -1);
	if (di != NULL)
	{
	    if (tv_get_bool(&di->di_tv))
		prop->pt_flags |= PT_FLAG_OVERRIDE;
	    else
		prop->pt_flags &= ~PT_FLAG_OVERRIDE;
	}

	di = dict_find(dict, (char_u *)"priority", -1);
	if (di != NULL)
	    prop->pt_priority = tv_get_number(&di->di_tv);

	di = dict_find(dict, (char_u *)"start_incl", -1);
	if (di != NULL)
	{
	    if (tv_get_bool(&di->di_tv))
		prop->pt_flags |= PT_FLAG_INS_START_INCL;
	    else
		prop->pt_flags &= ~PT_FLAG_INS_START_INCL;
	}

	di = dict_find(dict, (char_u *)"end_incl", -1);
	if (di != NULL)
	{
	    if (tv_get_bool(&di->di_tv))
		prop->pt_flags |= PT_FLAG_INS_END_INCL;
	    else
		prop->pt_flags &= ~PT_FLAG_INS_END_INCL;
	}
    }
}

/*
 * prop_type_add({name}, {props})
 */
    void
f_prop_type_add(typval_T *argvars, typval_T *rettv UNUSED)
{
    prop_type_set(argvars, TRUE);
}

/*
 * prop_type_change({name}, {props})
 */
    void
f_prop_type_change(typval_T *argvars, typval_T *rettv UNUSED)
{
    prop_type_set(argvars, FALSE);
}

/*
 * prop_type_delete({name} [, {bufnr}])
 */
    void
f_prop_type_delete(typval_T *argvars, typval_T *rettv UNUSED)
{
    char_u	*name;
    buf_T	*buf = NULL;
    hashitem_T	*hi;

    if (in_vim9script()
	    && (check_for_string_arg(argvars, 0) == FAIL
		|| check_for_opt_dict_arg(argvars, 1) == FAIL))
	return;

    name = tv_get_string(&argvars[0]);
    if (*name == NUL)
    {
	semsg(_(e_invalid_argument_str), "\"\"");
	return;
    }

    if (argvars[1].v_type != VAR_UNKNOWN)
    {
	if (get_bufnr_from_arg(&argvars[1], &buf) == FAIL)
	    return;
    }

    hi = find_prop_type_hi(name, buf);
    if (hi == NULL)
	return;

    hashtab_T	*ht;
    proptype_T	*prop = HI2PT(hi);

    if (buf == NULL)
    {
	ht = global_proptypes;
	VIM_CLEAR(global_proparray);
    }
    else
    {
	ht = buf->b_proptypes;
	VIM_CLEAR(buf->b_proparray);
    }
    hash_remove(ht, hi, "prop type delete");
    vim_free(prop);

    // currently visible text properties will disappear
    redraw_all_later(UPD_CLEAR);
    changed_window_setting_buf(buf == NULL ? curbuf : buf);
}

/*
 * prop_type_get({name} [, {props}])
 */
    void
f_prop_type_get(typval_T *argvars, typval_T *rettv)
{
    char_u *name;

    if (in_vim9script()
	    && (check_for_string_arg(argvars, 0) == FAIL
		|| check_for_opt_dict_arg(argvars, 1) == FAIL))
	return;

    name = tv_get_string(&argvars[0]);
    if (*name == NUL)
    {
	semsg(_(e_invalid_argument_str), "\"\"");
	return;
    }

    if (rettv_dict_alloc(rettv) == FAIL)
	return;

    proptype_T  *prop = NULL;
    buf_T	    *buf = NULL;

    if (argvars[1].v_type != VAR_UNKNOWN)
    {
	if (get_bufnr_from_arg(&argvars[1], &buf) == FAIL)
	    return;
    }

    prop = find_prop_type(name, buf);
    if (prop == NULL)
	return;

    dict_T *d = rettv->vval.v_dict;

    if (prop->pt_hl_id > 0)
	dict_add_string(d, "highlight", syn_id2name(prop->pt_hl_id));
    dict_add_number(d, "priority", prop->pt_priority);
    dict_add_number(d, "combine",
	    (prop->pt_flags & PT_FLAG_COMBINE) ? 1 : 0);
    dict_add_number(d, "start_incl",
	    (prop->pt_flags & PT_FLAG_INS_START_INCL) ? 1 : 0);
    dict_add_number(d, "end_incl",
	    (prop->pt_flags & PT_FLAG_INS_END_INCL) ? 1 : 0);
    if (buf != NULL)
	dict_add_number(d, "bufnr", buf->b_fnum);
}

    static void
list_types(hashtab_T *ht, list_T *l)
{
    long	todo;
    hashitem_T	*hi;

    todo = (long)ht->ht_used;
    FOR_ALL_HASHTAB_ITEMS(ht, hi, todo)
    {
	if (!HASHITEM_EMPTY(hi))
	{
	    proptype_T *prop = HI2PT(hi);

	    list_append_string(l, prop->pt_name, -1);
	    --todo;
	}
    }
}

/*
 * prop_type_list([{bufnr}])
 */
    void
f_prop_type_list(typval_T *argvars, typval_T *rettv UNUSED)
{
    buf_T *buf = NULL;

    if (rettv_list_alloc(rettv) == FAIL)
	return;

    if (in_vim9script() && check_for_opt_dict_arg(argvars, 0) == FAIL)
	return;

    if (argvars[0].v_type != VAR_UNKNOWN)
    {
	if (get_bufnr_from_arg(&argvars[0], &buf) == FAIL)
	    return;
    }
    if (buf == NULL)
    {
	if (global_proptypes != NULL)
	    list_types(global_proptypes, rettv->vval.v_list);
    }
    else if (buf->b_proptypes != NULL)
	list_types(buf->b_proptypes, rettv->vval.v_list);
}

/*
 * Free all property types in "ht".
 */
    static void
clear_ht_prop_types(hashtab_T *ht)
{
    long	todo;
    hashitem_T	*hi;

    if (ht == NULL)
	return;

    todo = (long)ht->ht_used;
    FOR_ALL_HASHTAB_ITEMS(ht, hi, todo)
    {
	if (!HASHITEM_EMPTY(hi))
	{
	    proptype_T *prop = HI2PT(hi);

	    vim_free(prop);
	    --todo;
	}
    }

    hash_clear(ht);
    vim_free(ht);
}

#if defined(EXITFREE) || defined(PROTO)
/*
 * Free all global property types.
 */
    void
clear_global_prop_types(void)
{
    clear_ht_prop_types(global_proptypes);
    global_proptypes = NULL;
    VIM_CLEAR(global_proparray);
}
#endif

/*
 * Free all property types for "buf".
 */
    void
clear_buf_prop_types(buf_T *buf)
{
    clear_ht_prop_types(buf->b_proptypes);
    buf->b_proptypes = NULL;
    VIM_CLEAR(buf->b_proparray);
}

// Struct used to return two values from adjust_prop().
typedef struct
{
    int dirty;	    // if the property was changed
    int can_drop;   // whether after this change, the prop may be removed
} adjustres_T;

/*
 * Adjust the property for "added" bytes (can be negative) inserted at "col".
 *
 * Note that "col" is zero-based, while tp_col is one-based.
 * Only for the current buffer.
 * "flags" can have:
 * APC_SUBSTITUTE:	Text is replaced, not inserted.
 * APC_INDENT:		Text is inserted before virtual text prop
 */
    static adjustres_T
adjust_prop(
	textprop_T  *prop,
	colnr_T	    col,
	int	    added,
	int	    flags)
{
    proptype_T	*pt;
    int		start_incl;
    int		end_incl;
    int		droppable;
    adjustres_T res = {TRUE, FALSE};

    // prop after end of the line doesn't move
    if (prop->tp_col == MAXCOL)
    {
	res.dirty = FALSE;
	return res;
    }

    pt = text_prop_type_by_id(curbuf, prop->tp_type);
    start_incl = (pt != NULL && (pt->pt_flags & PT_FLAG_INS_START_INCL))
				|| (flags & APC_SUBSTITUTE)
				|| (prop->tp_flags & TP_FLAG_CONT_PREV);
    if (prop->tp_id < 0 && (flags & APC_INDENT))
	// when inserting indent just before a character with virtual text
	// shift the text property
	start_incl = FALSE;
    end_incl = (pt != NULL && (pt->pt_flags & PT_FLAG_INS_END_INCL))
				|| (prop->tp_flags & TP_FLAG_CONT_NEXT);
    // do not drop zero-width props if they later can increase in size
    droppable = !(start_incl || end_incl);

    if (added > 0)
    {
	if (col + 1 <= prop->tp_col
		- (start_incl || (prop->tp_len == 0 && end_incl)))
	    // Change is entirely before the text property: Only shift
	    prop->tp_col += added;
	else if (col + 1 < prop->tp_col + prop->tp_len + end_incl)
	    // Insertion was inside text property
	    prop->tp_len += added;
    }
    else if (prop->tp_col > col + 1)
    {
	if (prop->tp_col + added < col + 1)
	{
	    prop->tp_len += (prop->tp_col - 1 - col) + added;
	    prop->tp_col = col + 1;
	    if (prop->tp_len <= 0)
	    {
		prop->tp_len = 0;
		res.can_drop = droppable;
	    }
	}
	else
	    prop->tp_col += added;
    }
    else if (prop->tp_len > 0 && prop->tp_col + prop->tp_len > col
	    && prop->tp_id >= 0)  // don't change length for virtual text
    {
	int after = col - added - (prop->tp_col - 1 + prop->tp_len);

	prop->tp_len += after > 0 ? added + after : added;
	res.can_drop = prop->tp_len <= 0 && droppable;
    }
    else
	res.dirty = FALSE;

    return res;
}

/*
 * Adjust the columns of text properties in line "lnum" after position "col" to
 * shift by "bytes_added" (can be negative).
 * Note that "col" is zero-based, while tp_col is one-based.
 * Only for the current buffer.
 * "flags" can have:
 * APC_SAVE_FOR_UNDO:	Call u_savesub() before making changes to the line.
 * APC_SUBSTITUTE:	Text is replaced, not inserted.
 * APC_INDENT:		Text is inserted before virtual text prop
 * Caller is expected to check b_has_textprop and "bytes_added" being non-zero.
 * Returns TRUE when props were changed.
 */
    int
adjust_prop_columns(
	linenr_T    lnum,
	colnr_T	    col,
	int	    bytes_added,
	int	    flags)
{
    int		proplen;
    char_u	*props;
    int		dirty = FALSE;
    int		ri, wi;
    size_t	textlen;

    if (text_prop_frozen > 0)
	return FALSE;

    proplen = get_text_props(curbuf, lnum, &props, TRUE);
    if (proplen == 0)
	return FALSE;
    textlen = curbuf->b_ml.ml_line_len - proplen * sizeof(textprop_T);

    wi = 0; // write index
    for (ri = 0; ri < proplen; ++ri)
    {
	textprop_T	prop;
	adjustres_T	res;

	mch_memmove(&prop, props + ri * sizeof(prop), sizeof(prop));
	res = adjust_prop(&prop, col, bytes_added, flags);
	if (res.dirty)
	{
	    // Save for undo if requested and not done yet.
	    if ((flags & APC_SAVE_FOR_UNDO) && !dirty
						    && u_savesub(lnum) == FAIL)
		return FALSE;
	    dirty = TRUE;

	    // u_savesub() may have updated curbuf->b_ml, fetch it again
	    if (curbuf->b_ml.ml_line_lnum != lnum)
		proplen = get_text_props(curbuf, lnum, &props, TRUE);
	}
	if (res.can_drop)
	    continue; // Drop this text property
	mch_memmove(props + wi * sizeof(textprop_T), &prop, sizeof(textprop_T));
	++wi;
    }
    if (dirty)
    {
	colnr_T newlen = (int)textlen + wi * (colnr_T)sizeof(textprop_T);

	if ((curbuf->b_ml.ml_flags & ML_LINE_DIRTY) == 0)
	{
	    char_u *p = vim_memsave(curbuf->b_ml.ml_line_ptr, newlen);

	    if (curbuf->b_ml.ml_flags & ML_ALLOCATED)
		vim_free(curbuf->b_ml.ml_line_ptr);
	    curbuf->b_ml.ml_line_ptr = p;
	}
	curbuf->b_ml.ml_flags |= ML_LINE_DIRTY;
	curbuf->b_ml.ml_line_len = newlen;
    }
    return dirty;
}

/*
 * Adjust text properties for a line that was split in two.
 * "lnum_props" is the line that has the properties from before the split.
 * "lnum_top" is the top line.
 * "kept" is the number of bytes kept in the first line, while
 * "deleted" is the number of bytes deleted.
 * "at_eol" is true if the split is after the end of the line.
 */
    void
adjust_props_for_split(
	linenr_T    lnum_props,
	linenr_T    lnum_top,
	int	    kept,
	int	    deleted,
	int	    at_eol)
{
    char_u	*props;
    int		count;
    garray_T    prevprop;
    garray_T    nextprop;
    int		i;
    int		skipped = kept + deleted;

    if (!curbuf->b_has_textprop)
	return;

    // Get the text properties from "lnum_props".
    count = get_text_props(curbuf, lnum_props, &props, FALSE);
    ga_init2(&prevprop, sizeof(textprop_T), 10);
    ga_init2(&nextprop, sizeof(textprop_T), 10);

    // Keep the relevant ones in the first line, reducing the length if needed.
    // Copy the ones that include the split to the second line.
    // Move the ones after the split to the second line.
    for (i = 0; i < count; ++i)
    {
	textprop_T  prop;
	proptype_T *pt;
	int	    start_incl, end_incl;
	int	    cont_prev, cont_next;
	int	    prop_col;

	// copy the prop to an aligned structure
	mch_memmove(&prop, props + i * sizeof(textprop_T), sizeof(textprop_T));

	pt = text_prop_type_by_id(curbuf, prop.tp_type);
	start_incl = (pt != NULL && (pt->pt_flags & PT_FLAG_INS_START_INCL));
	end_incl = (pt != NULL && (pt->pt_flags & PT_FLAG_INS_END_INCL));

	// a text prop "above" behaves like it is on the first text column
	prop_col = (prop.tp_flags & TP_FLAG_ALIGN_ABOVE) ? 1 : prop.tp_col;

	if (prop_col == MAXCOL)
	{
	    cont_prev = at_eol;
	    cont_next = !at_eol;
	}
	else
	{
	    cont_prev = prop_col + !start_incl <= kept;
	    cont_next = skipped <= prop_col + prop.tp_len - !end_incl;
	}
	// when a prop has text it is never copied
	if (prop.tp_id < 0 && cont_next)
	    cont_prev = FALSE;

	if (cont_prev && ga_grow(&prevprop, 1) == OK)
	{
	    textprop_T *p = ((textprop_T *)prevprop.ga_data) + prevprop.ga_len;

	    *p = prop;
	    ++prevprop.ga_len;
	    if (p->tp_col != MAXCOL && p->tp_col + p->tp_len >= kept)
		p->tp_len = kept - p->tp_col;
	    if (cont_next)
		p->tp_flags |= TP_FLAG_CONT_NEXT;
	}

	// Only add the property to the next line if the length is bigger than
	// zero.
	if (cont_next && ga_grow(&nextprop, 1) == OK)
	{
	    textprop_T *p = ((textprop_T *)nextprop.ga_data) + nextprop.ga_len;

	    *p = prop;
	    ++nextprop.ga_len;
	    if (p->tp_col != MAXCOL)
	    {
		if (p->tp_col > skipped)
		    p->tp_col -= skipped - 1;
		else
		{
		    p->tp_len -= skipped - p->tp_col;
		    p->tp_col = 1;
		}
	    }
	    if (cont_prev)
		p->tp_flags |= TP_FLAG_CONT_PREV;
	}
    }

    set_text_props(lnum_top, prevprop.ga_data,
					 prevprop.ga_len * sizeof(textprop_T));
    ga_clear(&prevprop);
    set_text_props(lnum_top + 1, nextprop.ga_data,
					 nextprop.ga_len * sizeof(textprop_T));
    ga_clear(&nextprop);
}

/*
 * Prepend properties of joined line "lnum" to "new_props".
 */
    void
prepend_joined_props(
	char_u	    *new_props,
	int	    propcount,
	int	    *props_remaining,
	linenr_T    lnum,
	int	    last_line,
	long	    col,
	int	    removed)
{
    char_u *props;
    int	    proplen = get_text_props(curbuf, lnum, &props, FALSE);
    int	    i;

    for (i = proplen; i-- > 0; )
    {
	textprop_T  prop;
	int	    end;

	mch_memmove(&prop, props + i * sizeof(prop), sizeof(prop));
	if (prop.tp_col == MAXCOL && !last_line)
	    continue;  // drop property with text after the line
	end = !(prop.tp_flags & TP_FLAG_CONT_NEXT);

	adjust_prop(&prop, 0, -removed, 0); // Remove leading spaces
	adjust_prop(&prop, -1, col, 0); // Make line start at its final column

	if (last_line || end)
	    mch_memmove(new_props + --(*props_remaining) * sizeof(prop),
							  &prop, sizeof(prop));
	else
	{
	    int j;
	    int	found = FALSE;

	    // Search for continuing prop.
	    for (j = *props_remaining; j < propcount; ++j)
	    {
		textprop_T op;

		mch_memmove(&op, new_props + j * sizeof(op), sizeof(op));
		if ((op.tp_flags & TP_FLAG_CONT_PREV)
			&& op.tp_id == prop.tp_id && op.tp_type == prop.tp_type)
		{
		    found = TRUE;
		    op.tp_len += op.tp_col - prop.tp_col;
		    op.tp_col = prop.tp_col;
		    // Start/end is taken care of when deleting joined lines
		    op.tp_flags = prop.tp_flags;
		    mch_memmove(new_props + j * sizeof(op), &op, sizeof(op));
		    break;
		}
	    }
	    if (!found)
		internal_error("text property above joined line not found");
	}
    }
}

#endif // FEAT_PROP_POPUP