harfbuzz/src/hb-subset-cff1.cc

/*
 * Copyright © 2018 Adobe Inc.
 *
 *  This is part of HarfBuzz, a text shaping library.
 *
 * Permission is hereby granted, without written agreement and without
 * license or royalty fees, to use, copy, modify, and distribute this
 * software and its documentation for any purpose, provided that the
 * above copyright notice and the following two paragraphs appear in
 * all copies of this software.
 *
 * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
 * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
 * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
 * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
 * DAMAGE.
 *
 * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
 * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
 * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
 * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
 * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
 *
 * Adobe Author(s): Michiharu Ariza
 */

#include "hb.hh"

#ifndef HB_NO_SUBSET_CFF

#include "hb-open-type.hh"
#include "hb-ot-cff1-table.hh"
#include "hb-set.h"
#include "hb-bimap.hh"
#include "hb-subset-cff1.hh"
#include "hb-subset-plan.hh"
#include "hb-subset-cff-common.hh"
#include "hb-cff1-interp-cs.hh"

using namespace CFF;

struct remap_sid_t : hb_inc_bimap_t
{
  unsigned int add (unsigned int sid)
  {
    if ((sid != CFF_UNDEF_SID) && !is_std_std (sid))
      return offset_sid (hb_inc_bimap_t::add (unoffset_sid (sid)));
    else
      return sid;
  }

  unsigned int operator[] (unsigned int sid) const
  {
    if (is_std_std (sid) || (sid == CFF_UNDEF_SID))
      return sid;
    else
      return offset_sid (get (unoffset_sid (sid)));
  }

  static const unsigned int num_std_strings = 391;

  static bool is_std_std (unsigned int sid) { return sid < num_std_strings; }
  static unsigned int offset_sid (unsigned int sid) { return sid + num_std_strings; }
  static unsigned int unoffset_sid (unsigned int sid) { return sid - num_std_strings; }
};

struct cff1_sub_table_info_t : cff_sub_table_info_t
{
  cff1_sub_table_info_t ()
    : cff_sub_table_info_t (),
      encoding_link (0),
      charset_link (0)
   {
    privateDictInfo.init ();
  }

  objidx_t	encoding_link;
  objidx_t	charset_link;
  table_info_t	privateDictInfo;
};

/* a copy of a parsed out cff1_top_dict_values_t augmented with additional operators */
struct cff1_top_dict_values_mod_t : cff1_top_dict_values_t
{
  void init (const cff1_top_dict_values_t *base_= &Null (cff1_top_dict_values_t))
  {
    SUPER::init ();
    base = base_;
  }

  void fini () { SUPER::fini (); }

  unsigned get_count () const { return base->get_count () + SUPER::get_count (); }
  const cff1_top_dict_val_t &get_value (unsigned int i) const
  {
    if (i < base->get_count ())
      return (*base)[i];
    else
      return SUPER::values[i - base->get_count ()];
  }
  const cff1_top_dict_val_t &operator [] (unsigned int i) const { return get_value (i); }

  void reassignSIDs (const remap_sid_t& sidmap)
  {
    for (unsigned int i = 0; i < name_dict_values_t::ValCount; i++)
      nameSIDs[i] = sidmap[base->nameSIDs[i]];
  }

  protected:
  typedef cff1_top_dict_values_t SUPER;
  const cff1_top_dict_values_t *base;
};

struct top_dict_modifiers_t
{
  top_dict_modifiers_t (const cff1_sub_table_info_t &info_,
			const unsigned int (&nameSIDs_)[name_dict_values_t::ValCount])
    : info (info_),
      nameSIDs (nameSIDs_)
  {}

  const cff1_sub_table_info_t &info;
  const unsigned int	(&nameSIDs)[name_dict_values_t::ValCount];
};

struct cff1_top_dict_op_serializer_t : cff_top_dict_op_serializer_t<cff1_top_dict_val_t>
{
  bool serialize (hb_serialize_context_t *c,
		  const cff1_top_dict_val_t &opstr,
		  const top_dict_modifiers_t &mod) const
  {
    TRACE_SERIALIZE (this);

    op_code_t op = opstr.op;
    switch (op)
    {
      case OpCode_charset:
	if (mod.info.charset_link)
	  return_trace (FontDict::serialize_link4_op(c, op, mod.info.charset_link, whence_t::Absolute));
	else
	  goto fall_back;

      case OpCode_Encoding:
	if (mod.info.encoding_link)
	  return_trace (FontDict::serialize_link4_op(c, op, mod.info.encoding_link, whence_t::Absolute));
	else
	  goto fall_back;

      case OpCode_Private:
	return_trace (UnsizedByteStr::serialize_int2 (c, mod.info.privateDictInfo.size) &&
		      Dict::serialize_link4_op (c, op, mod.info.privateDictInfo.link, whence_t::Absolute));

      case OpCode_version:
      case OpCode_Notice:
      case OpCode_Copyright:
      case OpCode_FullName:
      case OpCode_FamilyName:
      case OpCode_Weight:
      case OpCode_PostScript:
      case OpCode_BaseFontName:
      case OpCode_FontName:
	return_trace (FontDict::serialize_int2_op (c, op, mod.nameSIDs[name_dict_values_t::name_op_to_index (op)]));

      case OpCode_ROS:
	{
	  /* for registry & ordering, reassigned SIDs are serialized
	   * for supplement, the original byte string is copied along with the op code */
	  op_str_t supp_op;
	  supp_op.op = op;
	  if ( unlikely (!(opstr.str.length >= opstr.last_arg_offset + 3)))
	    return_trace (false);
	  supp_op.str = byte_str_t (&opstr.str + opstr.last_arg_offset, opstr.str.length - opstr.last_arg_offset);
	  return_trace (UnsizedByteStr::serialize_int2 (c, mod.nameSIDs[name_dict_values_t::registry]) &&
			UnsizedByteStr::serialize_int2 (c, mod.nameSIDs[name_dict_values_t::ordering]) &&
			copy_opstr (c, supp_op));
	}
      fall_back:
      default:
	return_trace (cff_top_dict_op_serializer_t<cff1_top_dict_val_t>::serialize (c, opstr, mod.info));
    }
    return_trace (true);
  }

};

struct cff1_font_dict_op_serializer_t : cff_font_dict_op_serializer_t
{
  bool serialize (hb_serialize_context_t *c,
		  const op_str_t &opstr,
		  const cff1_font_dict_values_mod_t &mod) const
  {
    TRACE_SERIALIZE (this);

    if (opstr.op == OpCode_FontName)
      return_trace (FontDict::serialize_int2_op (c, opstr.op, mod.fontName));
    else
      return_trace (SUPER::serialize (c, opstr, mod.privateDictInfo));
  }

  private:
  typedef cff_font_dict_op_serializer_t SUPER;
};

struct cff1_cs_opset_flatten_t : cff1_cs_opset_t<cff1_cs_opset_flatten_t, flatten_param_t>
{
  static void flush_args_and_op (op_code_t op, cff1_cs_interp_env_t &env, flatten_param_t& param)
  {
    if (env.arg_start > 0)
      flush_width (env, param);

    switch (op)
    {
      case OpCode_hstem:
      case OpCode_hstemhm:
      case OpCode_vstem:
      case OpCode_vstemhm:
      case OpCode_hintmask:
      case OpCode_cntrmask:
      case OpCode_dotsection:
	if (param.drop_hints)
	{
	  env.clear_args ();
	  return;
	}
	HB_FALLTHROUGH;

      default:
	SUPER::flush_args_and_op (op, env, param);
	break;
    }
  }
  static void flush_args (cff1_cs_interp_env_t &env, flatten_param_t& param)
  {
    str_encoder_t  encoder (param.flatStr);
    for (unsigned int i = env.arg_start; i < env.argStack.get_count (); i++)
      encoder.encode_num (env.eval_arg (i));
    SUPER::flush_args (env, param);
  }

  static void flush_op (op_code_t op, cff1_cs_interp_env_t &env, flatten_param_t& param)
  {
    str_encoder_t  encoder (param.flatStr);
    encoder.encode_op (op);
  }

  static void flush_width (cff1_cs_interp_env_t &env, flatten_param_t& param)
  {
    assert (env.has_width);
    str_encoder_t  encoder (param.flatStr);
    encoder.encode_num (env.width);
  }

  static void flush_hintmask (op_code_t op, cff1_cs_interp_env_t &env, flatten_param_t& param)
  {
    SUPER::flush_hintmask (op, env, param);
    if (!param.drop_hints)
    {
      str_encoder_t  encoder (param.flatStr);
      for (unsigned int i = 0; i < env.hintmask_size; i++)
	encoder.encode_byte (env.str_ref[i]);
    }
  }

  private:
  typedef cff1_cs_opset_t<cff1_cs_opset_flatten_t, flatten_param_t> SUPER;
};

struct range_list_t : hb_vector_t<code_pair_t>
{
  /* replace the first glyph ID in the "glyph" field each range with a nLeft value */
  bool complete (unsigned int last_glyph)
  {
    bool  two_byte = false;
    for (unsigned int i = (*this).length; i > 0; i--)
    {
      code_pair_t &pair = (*this)[i - 1];
      unsigned int  nLeft = last_glyph - pair.glyph - 1;
      if (nLeft >= 0x100)
	two_byte = true;
      last_glyph = pair.glyph;
      pair.glyph = nLeft;
    }
    return two_byte;
  }
};

struct cff1_cs_opset_subr_subset_t : cff1_cs_opset_t<cff1_cs_opset_subr_subset_t, subr_subset_param_t>
{
  static void process_op (op_code_t op, cff1_cs_interp_env_t &env, subr_subset_param_t& param)
  {
    switch (op) {

      case OpCode_return:
	param.current_parsed_str->add_op (op, env.str_ref);
	param.current_parsed_str->set_parsed ();
	env.return_from_subr ();
	param.set_current_str (env, false);
	break;

      case OpCode_endchar:
	param.current_parsed_str->add_op (op, env.str_ref);
	param.current_parsed_str->set_parsed ();
	SUPER::process_op (op, env, param);
	break;

      case OpCode_callsubr:
	process_call_subr (op, CSType_LocalSubr, env, param, env.localSubrs, param.local_closure);
	break;

      case OpCode_callgsubr:
	process_call_subr (op, CSType_GlobalSubr, env, param, env.globalSubrs, param.global_closure);
	break;

      default:
	SUPER::process_op (op, env, param);
	param.current_parsed_str->add_op (op, env.str_ref);
	break;
    }
  }

  protected:
  static void process_call_subr (op_code_t op, cs_type_t type,
				 cff1_cs_interp_env_t &env, subr_subset_param_t& param,
				 cff1_biased_subrs_t& subrs, hb_set_t *closure)
  {
    byte_str_ref_t    str_ref = env.str_ref;
    env.call_subr (subrs, type);
    param.current_parsed_str->add_call_op (op, str_ref, env.context.subr_num);
    closure->add (env.context.subr_num);
    param.set_current_str (env, true);
  }

  private:
  typedef cff1_cs_opset_t<cff1_cs_opset_subr_subset_t, subr_subset_param_t> SUPER;
};

struct cff1_subr_subsetter_t : subr_subsetter_t<cff1_subr_subsetter_t, CFF1Subrs, const OT::cff1::accelerator_subset_t, cff1_cs_interp_env_t, cff1_cs_opset_subr_subset_t, OpCode_endchar>
{
  cff1_subr_subsetter_t (const OT::cff1::accelerator_subset_t &acc_, const hb_subset_plan_t *plan_)
    : subr_subsetter_t (acc_, plan_) {}

  static void complete_parsed_str (cff1_cs_interp_env_t &env, subr_subset_param_t& param, parsed_cs_str_t &charstring)
  {
    /* insert width at the beginning of the charstring as necessary */
    if (env.has_width)
      charstring.set_prefix (env.width);

    /* subroutines/charstring left on the call stack are legally left unmarked
     * unmarked when a subroutine terminates with endchar. mark them.
     */
    param.current_parsed_str->set_parsed ();
    for (unsigned int i = 0; i < env.callStack.get_count (); i++)
    {
      parsed_cs_str_t *parsed_str = param.get_parsed_str_for_context (env.callStack[i]);
      if (likely (parsed_str))
	parsed_str->set_parsed ();
      else
	env.set_error ();
    }
  }
};

struct cff_subset_plan {
  cff_subset_plan ()
    : info (),
      orig_fdcount (0),
      subset_fdcount (1),
      subset_fdselect_format (0),
      drop_hints (false),
      desubroutinize(false)
  {
    topdict_mod.init ();
    subset_fdselect_ranges.init ();
    fdmap.init ();
    subset_charstrings.init ();
    subset_globalsubrs.init ();
    subset_localsubrs.init ();
    fontdicts_mod.init ();
    subset_enc_code_ranges.init ();
    subset_enc_supp_codes.init ();
    subset_charset_ranges.init ();
    sidmap.init ();
    for (unsigned int i = 0; i < name_dict_values_t::ValCount; i++)
      topDictModSIDs[i] = CFF_UNDEF_SID;
  }

  ~cff_subset_plan ()
  {
    topdict_mod.fini ();
    subset_fdselect_ranges.fini ();
    fdmap.fini ();
    subset_charstrings.fini_deep ();
    subset_globalsubrs.fini_deep ();
    subset_localsubrs.fini_deep ();
    fontdicts_mod.fini ();
    subset_enc_code_ranges.fini ();
    subset_enc_supp_codes.fini ();
    subset_charset_ranges.fini ();
    sidmap.fini ();
  }

  void plan_subset_encoding (const OT::cff1::accelerator_subset_t &acc, hb_subset_plan_t *plan)
  {
    const Encoding *encoding = acc.encoding;
    unsigned int  size0, size1;
    hb_codepoint_t  code, last_code = CFF_UNDEF_CODE;
    hb_vector_t<hb_codepoint_t> supp_codes;

    if (unlikely (!subset_enc_code_ranges.resize (0)))
    {
      plan->check_success (false);
      return;
    }

    supp_codes.init ();

    subset_enc_num_codes = plan->num_output_glyphs () - 1;
    unsigned int glyph;
    for (glyph = 1; glyph < plan->num_output_glyphs (); glyph++)
    {
      hb_codepoint_t  old_glyph;
      if (!plan->old_gid_for_new_gid (glyph, &old_glyph))
      {
	/* Retain the code for the old missing glyph ID */
	old_glyph = glyph;
      }
      code = acc.glyph_to_code (old_glyph);
      if (code == CFF_UNDEF_CODE)
      {
	subset_enc_num_codes = glyph - 1;
	break;
      }

      if ((last_code == CFF_UNDEF_CODE) || (code != last_code + 1))
      {
	code_pair_t pair = { code, glyph };
	subset_enc_code_ranges.push (pair);
      }
      last_code = code;

      if (encoding != &Null (Encoding))
      {
	hb_codepoint_t  sid = acc.glyph_to_sid (old_glyph);
	encoding->get_supplement_codes (sid, supp_codes);
	for (unsigned int i = 0; i < supp_codes.length; i++)
	{
	  code_pair_t pair = { supp_codes[i], sid };
	  subset_enc_supp_codes.push (pair);
	}
      }
    }
    supp_codes.fini ();

    subset_enc_code_ranges.complete (glyph);

    assert (subset_enc_num_codes <= 0xFF);
    size0 = Encoding0::min_size + HBUINT8::static_size * subset_enc_num_codes;
    size1 = Encoding1::min_size + Encoding1_Range::static_size * subset_enc_code_ranges.length;

    if (size0 < size1)
      subset_enc_format = 0;
    else
      subset_enc_format = 1;
  }

  void plan_subset_charset (const OT::cff1::accelerator_subset_t &acc, hb_subset_plan_t *plan)
  {
    unsigned int  size0, size_ranges;
    hb_codepoint_t  sid, last_sid = CFF_UNDEF_CODE;

    if (unlikely (!subset_charset_ranges.resize (0)))
    {
      plan->check_success (false);
      return;
    }

    unsigned int glyph;
    for (glyph = 1; glyph < plan->num_output_glyphs (); glyph++)
    {
      hb_codepoint_t  old_glyph;
      if (!plan->old_gid_for_new_gid (glyph, &old_glyph))
      {
	/* Retain the SID for the old missing glyph ID */
	old_glyph = glyph;
      }
      sid = acc.glyph_to_sid (old_glyph);

      if (!acc.is_CID ())
	sid = sidmap.add (sid);

      if ((last_sid == CFF_UNDEF_CODE) || (sid != last_sid + 1))
      {
	code_pair_t pair = { sid, glyph };
	subset_charset_ranges.push (pair);
      }
      last_sid = sid;
    }

    bool two_byte = subset_charset_ranges.complete (glyph);

    size0 = Charset0::min_size + HBUINT16::static_size * (plan->num_output_glyphs () - 1);
    if (!two_byte)
      size_ranges = Charset1::min_size + Charset1_Range::static_size * subset_charset_ranges.length;
    else
      size_ranges = Charset2::min_size + Charset2_Range::static_size * subset_charset_ranges.length;

    if (size0 < size_ranges)
      subset_charset_format = 0;
    else if (!two_byte)
      subset_charset_format = 1;
    else
      subset_charset_format = 2;
  }

  bool collect_sids_in_dicts (const OT::cff1::accelerator_subset_t &acc)
  {
    sidmap.reset ();

    for (unsigned int i = 0; i < name_dict_values_t::ValCount; i++)
    {
      unsigned int sid = acc.topDict.nameSIDs[i];
      if (sid != CFF_UNDEF_SID)
      {
	(void)sidmap.add (sid);
	topDictModSIDs[i] = sidmap[sid];
      }
    }

    if (acc.fdArray != &Null (CFF1FDArray))
      for (unsigned int i = 0; i < orig_fdcount; i++)
	if (fdmap.has (i))
	  (void)sidmap.add (acc.fontDicts[i].fontName);

    return true;
  }

  bool create (const OT::cff1::accelerator_subset_t &acc,
	       hb_subset_plan_t *plan)
  {
    /* make sure notdef is first */
    hb_codepoint_t old_glyph;
    if (!plan->old_gid_for_new_gid (0, &old_glyph) || (old_glyph != 0)) return false;

    num_glyphs = plan->num_output_glyphs ();
    orig_fdcount = acc.fdCount;
    drop_hints = plan->drop_hints;
    desubroutinize = plan->desubroutinize;

    /* check whether the subset renumbers any glyph IDs */
    gid_renum = false;
    for (hb_codepoint_t new_glyph = 0; new_glyph < plan->num_output_glyphs (); new_glyph++)
    {
      if (!plan->old_gid_for_new_gid(new_glyph, &old_glyph))
	continue;
      if (new_glyph != old_glyph) {
	gid_renum = true;
	break;
      }
    }

    subset_charset = gid_renum || !acc.is_predef_charset ();
    subset_encoding = !acc.is_CID() && !acc.is_predef_encoding ();

    /* top dict INDEX */
    {
      /* Add encoding/charset to a (copy of) top dict as necessary */
      topdict_mod.init (&acc.topDict);
      bool need_to_add_enc = (subset_encoding && !acc.topDict.has_op (OpCode_Encoding));
      bool need_to_add_set = (subset_charset && !acc.topDict.has_op (OpCode_charset));
      if (need_to_add_enc || need_to_add_set)
      {
	if (need_to_add_enc)
	  topdict_mod.add_op (OpCode_Encoding);
	if (need_to_add_set)
	  topdict_mod.add_op (OpCode_charset);
      }
    }

    /* Determine re-mapping of font index as fdmap among other info */
    if (acc.fdSelect != &Null (CFF1FDSelect))
    {
	if (unlikely (!hb_plan_subset_cff_fdselect (plan,
				  orig_fdcount,
				  *acc.fdSelect,
				  subset_fdcount,
				  info.fd_select.size,
				  subset_fdselect_format,
				  subset_fdselect_ranges,
				  fdmap)))
	return false;
    }
    else
      fdmap.identity (1);

    /* remove unused SIDs & reassign SIDs */
    {
      /* SIDs for name strings in dicts are added before glyph names so they fit in 16-bit int range */
      if (unlikely (!collect_sids_in_dicts (acc)))
	return false;
      if (unlikely (sidmap.get_population () > 0x8000))	/* assumption: a dict won't reference that many strings */
	return false;

      if (subset_charset) plan_subset_charset (acc, plan);

      topdict_mod.reassignSIDs (sidmap);
    }

    if (desubroutinize)
    {
      /* Flatten global & local subrs */
      subr_flattener_t<const OT::cff1::accelerator_subset_t, cff1_cs_interp_env_t, cff1_cs_opset_flatten_t, OpCode_endchar>
		    flattener(acc, plan);
      if (!flattener.flatten (subset_charstrings))
	return false;
    }
    else
    {
      cff1_subr_subsetter_t       subr_subsetter (acc, plan);

      /* Subset subrs: collect used subroutines, leaving all unused ones behind */
      if (!subr_subsetter.subset ())
	return false;

      /* encode charstrings, global subrs, local subrs with new subroutine numbers */
      if (!subr_subsetter.encode_charstrings (subset_charstrings))
	return false;

      if (!subr_subsetter.encode_globalsubrs (subset_globalsubrs))
	return false;

      /* local subrs */
      if (!subset_localsubrs.resize (orig_fdcount))
	return false;
      for (unsigned int fd = 0; fd < orig_fdcount; fd++)
      {
	subset_localsubrs[fd].init ();
	if (fdmap.has (fd))
	{
	  if (!subr_subsetter.encode_localsubrs (fd, subset_localsubrs[fd]))
	    return false;
	}
      }
    }

    /* Encoding */
    if (subset_encoding)
      plan_subset_encoding (acc, plan);

    /* private dicts & local subrs */
    if (!acc.is_CID ())
      fontdicts_mod.push (cff1_font_dict_values_mod_t ());
    else
    {
      + hb_iter (acc.fontDicts)
      | hb_filter ([&] (const cff1_font_dict_values_t &_)
	{ return fdmap.has (&_ - &acc.fontDicts[0]); } )
      | hb_map ([&] (const cff1_font_dict_values_t &_)
	{
	  cff1_font_dict_values_mod_t mod;
	  mod.init (&_, sidmap[_.fontName]);
	  return mod;
	})
      | hb_sink (fontdicts_mod)
      ;
    }

    return ((subset_charstrings.length == plan->num_output_glyphs ())
	   && (fontdicts_mod.length == subset_fdcount));
  }

  cff1_top_dict_values_mod_t	topdict_mod;
  cff1_sub_table_info_t		info;

  unsigned int    num_glyphs;
  unsigned int    orig_fdcount;
  unsigned int    subset_fdcount;
  unsigned int    subset_fdselect_format;
  hb_vector_t<code_pair_t>   subset_fdselect_ranges;

  /* font dict index remap table from fullset FDArray to subset FDArray.
   * set to CFF_UNDEF_CODE if excluded from subset */
  hb_inc_bimap_t   fdmap;

  str_buff_vec_t		subset_charstrings;
  str_buff_vec_t		subset_globalsubrs;
  hb_vector_t<str_buff_vec_t>	subset_localsubrs;
  hb_vector_t<cff1_font_dict_values_mod_t>  fontdicts_mod;

  bool		drop_hints;

  bool		gid_renum;
  bool		subset_encoding;
  uint8_t	subset_enc_format;
  unsigned int	subset_enc_num_codes;
  range_list_t	subset_enc_code_ranges;
  hb_vector_t<code_pair_t>  subset_enc_supp_codes;

  uint8_t	subset_charset_format;
  range_list_t	subset_charset_ranges;
  bool		subset_charset;

  remap_sid_t	sidmap;
  unsigned int	topDictModSIDs[name_dict_values_t::ValCount];

  bool		desubroutinize;
};

static bool _serialize_cff1 (hb_serialize_context_t *c,
			     cff_subset_plan &plan,
			     const OT::cff1::accelerator_subset_t  &acc,
			     unsigned int num_glyphs)
{
  /* private dicts & local subrs */
  for (int i = (int)acc.privateDicts.length; --i >= 0 ;)
  {
    if (plan.fdmap.has (i))
    {
      objidx_t	subrs_link = 0;
      if (plan.subset_localsubrs[i].length > 0)
      {
	CFF1Subrs *dest = c->start_embed <CFF1Subrs> ();
	if (unlikely (!dest)) return false;
	c->push ();
	if (likely (dest && dest->serialize (c, plan.subset_localsubrs[i])))
	  subrs_link = c->pop_pack ();
	else
	{
	  c->pop_discard ();
	  return false;
	}
      }

      PrivateDict *pd = c->start_embed<PrivateDict> ();
      if (unlikely (!pd)) return false;
      c->push ();
      cff_private_dict_op_serializer_t privSzr (plan.desubroutinize, plan.drop_hints);
      /* N.B. local subrs immediately follows its corresponding private dict. i.e., subr offset == private dict size */
      if (likely (pd->serialize (c, acc.privateDicts[i], privSzr, subrs_link)))
      {
	unsigned fd = plan.fdmap[i];
	plan.fontdicts_mod[fd].privateDictInfo.size = c->length ();
	plan.fontdicts_mod[fd].privateDictInfo.link = c->pop_pack ();
      }
      else
      {
	c->pop_discard ();
	return false;
      }
    }
  }

  if (!acc.is_CID ())
    plan.info.privateDictInfo = plan.fontdicts_mod[0].privateDictInfo;

  /* CharStrings */
  {
    CFF1CharStrings  *cs = c->start_embed<CFF1CharStrings> ();
    if (unlikely (!cs)) return false;
    c->push ();
    if (likely (cs->serialize (c, plan.subset_charstrings)))
      plan.info.char_strings_link = c->pop_pack ();
    else
    {
      c->pop_discard ();
      return false;
    }
  }

  /* FDArray (FD Index) */
  if (acc.fdArray != &Null (CFF1FDArray))
  {
    CFF1FDArray *fda = c->start_embed<CFF1FDArray> ();
    if (unlikely (!fda)) return false;
    c->push ();
    cff1_font_dict_op_serializer_t  fontSzr;
    auto it = + hb_zip (+ hb_iter (plan.fontdicts_mod), + hb_iter (plan.fontdicts_mod));
    if (likely (fda->serialize (c, it, fontSzr)))
      plan.info.fd_array_link = c->pop_pack (false);
    else
    {
      c->pop_discard ();
      return false;
    }
  }

  /* FDSelect */
  if (acc.fdSelect != &Null (CFF1FDSelect))
  {
    c->push ();
    if (likely (hb_serialize_cff_fdselect (c, num_glyphs, *acc.fdSelect, acc.fdCount,
					   plan.subset_fdselect_format, plan.info.fd_select.size,
					   plan.subset_fdselect_ranges)))
      plan.info.fd_select.link = c->pop_pack ();
    else
    {
      c->pop_discard ();
      return false;
    }
  }

  /* Charset */
  if (plan.subset_charset)
  {
    Charset *dest = c->start_embed<Charset> ();
    if (unlikely (!dest)) return false;
    c->push ();
    if (likely (dest->serialize (c,
				 plan.subset_charset_format,
				 plan.num_glyphs,
				 plan.subset_charset_ranges)))
      plan.info.charset_link = c->pop_pack ();
    else
    {
      c->pop_discard ();
      return false;
    }
  }

  /* Encoding */
  if (plan.subset_encoding)
  {
    Encoding *dest = c->start_embed<Encoding> ();
    if (unlikely (!dest)) return false;
    c->push ();
    if (likely (dest->serialize (c,
				 plan.subset_enc_format,
				 plan.subset_enc_num_codes,
				 plan.subset_enc_code_ranges,
				 plan.subset_enc_supp_codes)))
      plan.info.encoding_link = c->pop_pack ();
    else
    {
      c->pop_discard ();
      return false;
    }
  }

  /* global subrs */
  {
    c->push ();
    CFF1Subrs *dest = c->start_embed <CFF1Subrs> ();
    if (unlikely (!dest)) return false;
    if (likely (dest->serialize (c, plan.subset_globalsubrs)))
      c->pop_pack ();
    else
    {
      c->pop_discard ();
      return false;
    }
  }

  /* String INDEX */
  {
    CFF1StringIndex *dest = c->start_embed<CFF1StringIndex> ();
    if (unlikely (!dest)) return false;
    c->push ();
    if (likely (dest->serialize (c, *acc.stringIndex, plan.sidmap)))
      c->pop_pack ();
    else
    {
      c->pop_discard ();
      return false;
    }
  }

  OT::cff1 *cff = c->allocate_min<OT::cff1> ();
  if (unlikely (!cff))
    return false;

  /* header */
  cff->version.major = 0x01;
  cff->version.minor = 0x00;
  cff->nameIndex = cff->min_size;
  cff->offSize = 4; /* unused? */

  /* name INDEX */
  if (unlikely (!(*acc.nameIndex).copy (c))) return false;

  /* top dict INDEX */
  {
    /* serialize singleton TopDict */
    TopDict *top = c->start_embed<TopDict> ();
    if (!top) return false;
    c->push ();
    cff1_top_dict_op_serializer_t topSzr;
    unsigned top_size = 0;
    top_dict_modifiers_t  modifier (plan.info, plan.topDictModSIDs);
    if (likely (top->serialize (c, plan.topdict_mod, topSzr, modifier)))
    {
      top_size = c->length ();
      c->pop_pack (false);
    }
    else
    {
      c->pop_discard ();
      return false;
    }
    /* serialize INDEX header for above */
    CFF1Index *dest = c->start_embed<CFF1Index> ();
    if (!dest) return false;
    return dest->serialize_header (c, hb_iter (hb_array_t<unsigned> (&top_size, 1)));
  }
}

static bool
_hb_subset_cff1 (const OT::cff1::accelerator_subset_t  &acc,
		hb_subset_context_t	*c)
{
  cff_subset_plan cff_plan;

  if (unlikely (!cff_plan.create (acc, c->plan)))
  {
    DEBUG_MSG(SUBSET, nullptr, "Failed to generate a cff subsetting plan.");
    return false;
  }

  return _serialize_cff1 (c->serializer, cff_plan, acc, c->plan->num_output_glyphs ());
}

/**
 * hb_subset_cff1:
 * Subsets the CFF table according to a provided plan.
 *
 * Return value: subsetted cff table.
 **/
bool
hb_subset_cff1 (hb_subset_context_t *c)
{
  OT::cff1::accelerator_subset_t acc;
  acc.init (c->plan->source);
  bool result = likely (acc.is_valid ()) && _hb_subset_cff1 (acc, c);
  acc.fini ();

  return result;
}


#endif