1#!/usr/bin/env python
2#===- gen_std.py -  ------------------------------------------*- python -*--===#
3#
4# Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
5# See https://llvm.org/LICENSE.txt for license information.
6# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
7#
8#===------------------------------------------------------------------------===#
9
10"""gen_std.py is a tool to generate a lookup table (from qualified names to
11include headers) for C/C++ Standard Library symbols by parsing archieved HTML
12files from cppreference.
13
14Caveats and FIXMEs:
15  - only symbols directly in "std" namespace are added, we should also add std's
16    subnamespace symbols (e.g. chrono).
17  - symbols with multiple variants or defined in multiple headers aren't added,
18    e.g. std::move, std::swap
19
20Usage:
21  1. Install BeautifulSoup dependency, see instruction:
22       https://www.crummy.com/software/BeautifulSoup/bs4/doc/#installing-beautiful-soup
23  2. Download cppreference offline HTML files (e.g. html_book_20181028.zip) at
24       https://en.cppreference.com/w/Cppreference:Archives
25  3. Unzip the zip file from step 2 to directory </cppreference>, you should
26     get a "reference" directory in </cppreference>
27  4. Run the command:
28       // Generate C++ symbols
29       gen_std.py -cppreference </cppreference/reference> -language=cpp > StdSymbolMap.inc
30       // Generate C symbols
31       gen_std.py -cppreference </cppreference/reference> -language=c > CSymbolMap.inc
32"""
33
34
35import cppreference_parser
36import argparse
37import datetime
38import os
39import sys
40
41CODE_PREFIX = """\
42//===-- gen_std.py generated file -------------------------------*- C++ -*-===//
43//
44// Used to build a lookup table (qualified names => include headers) for %s
45// Standard Library symbols.
46//
47// Automatically generated file, DO NOT EDIT!
48//
49// Generated from cppreference offline HTML book (modified on %s).
50//===----------------------------------------------------------------------===//
51"""
52
53def ParseArg():
54  parser = argparse.ArgumentParser(description='Generate StdGen file')
55  parser.add_argument('-cppreference', metavar='PATH',
56                      default='',
57                      help='path to the cppreference offline HTML directory',
58                      required=True
59                     )
60  parser.add_argument('-language',
61                      default='cpp',
62                      help='Generate c or cpp symbols',
63                      required=True)
64  return parser.parse_args()
65
66
67def main():
68  args = ParseArg()
69  if args.language == 'cpp':
70    page_root = os.path.join(args.cppreference, "en", "cpp")
71    symbol_index_root = os.path.join(page_root, "symbol_index")
72    parse_pages =  [
73      (page_root, "symbol_index.html", "std::"),
74      # std sub-namespace symbols have separated pages.
75      # We don't index std literal operators (e.g.
76      # std::literals::chrono_literals::operator""d), these symbols can't be
77      # accessed by std::<symbol_name>.
78      # FIXME: index std::placeholders symbols, placeholders.html page is
79      # different (which contains one entry for _1, _2, ..., _N), we need special
80      # handling.
81      (symbol_index_root, "chrono.html", "std::chrono::"),
82      (symbol_index_root, "filesystem.html", "std::filesystem::"),
83      (symbol_index_root, "pmr.html", "std::pmr::"),
84      (symbol_index_root, "regex_constants.html", "std::regex_constants::"),
85      (symbol_index_root, "this_thread.html", "std::this_thread::"),
86    ]
87  elif args.language == 'c':
88    page_root = os.path.join(args.cppreference, "en", "c")
89    symbol_index_root = page_root
90    parse_pages = [(page_root, "index.html", None)]
91
92  if not os.path.exists(symbol_index_root):
93    exit("Path %s doesn't exist!" % symbol_index_root)
94
95  symbols = cppreference_parser.GetSymbols(parse_pages)
96
97  # We don't have version information from the unzipped offline HTML files.
98  # so we use the modified time of the symbol_index.html as the version.
99  index_page_path = os.path.join(page_root, "index.html")
100  cppreference_modified_date = datetime.datetime.fromtimestamp(
101    os.stat(index_page_path).st_mtime).strftime('%Y-%m-%d')
102  print CODE_PREFIX % (args.language.upper(), cppreference_modified_date)
103  for symbol in symbols:
104    if len(symbol.headers) == 1:
105      # SYMBOL(unqualified_name, namespace, header)
106      print "SYMBOL(%s, %s, %s)" % (symbol.name, symbol.namespace,
107                                    symbol.headers[0])
108    elif len(symbol.headers) == 0:
109      sys.stderr.write("No header found for symbol %s\n" % symbol.name)
110    else:
111      # FIXME: support symbols with multiple headers (e.g. std::move).
112      sys.stderr.write("Ambiguous header for symbol %s: %s\n" % (
113          symbol.name, ', '.join(symbol.headers)))
114
115
116if __name__ == '__main__':
117  main()
118