Victor Stinner | 376658f | 2015-03-18 14:16:50 +0100 | [diff] [blame] | 1 | """Cache lines from Python source files. |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 2 | |
| 3 | This is intended to read lines from modules imported -- hence if a filename |
| 4 | is not found, it will look down the module search path for a file by |
| 5 | that name. |
| 6 | """ |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 7 | |
Robert Collins | 6bc2c1e | 2015-03-05 12:07:57 +1300 | [diff] [blame] | 8 | import functools |
Guido van Rossum | c341c62 | 1992-03-27 15:12:43 +0000 | [diff] [blame] | 9 | import sys |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 10 | import os |
Benjamin Peterson | 9b8d24b | 2009-03-24 22:30:15 +0000 | [diff] [blame] | 11 | import tokenize |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 12 | |
Jeremy Hylton | 97b2e84 | 2003-06-29 16:59:43 +0000 | [diff] [blame] | 13 | __all__ = ["getline", "clearcache", "checkcache"] |
Skip Montanaro | 17ab123 | 2001-01-24 06:27:27 +0000 | [diff] [blame] | 14 | |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 15 | def getline(filename, lineno, module_globals=None): |
| 16 | lines = getlines(filename, module_globals) |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 17 | if 1 <= lineno <= len(lines): |
| 18 | return lines[lineno-1] |
| 19 | else: |
| 20 | return '' |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 21 | |
| 22 | |
| 23 | # The cache |
| 24 | |
Robert Collins | 6bc2c1e | 2015-03-05 12:07:57 +1300 | [diff] [blame] | 25 | # The cache. Maps filenames to either a thunk which will provide source code, |
| 26 | # or a tuple (size, mtime, lines, fullname) once loaded. |
| 27 | cache = {} |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 28 | |
| 29 | |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 30 | def clearcache(): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 31 | """Clear the cache entirely.""" |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 32 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 33 | global cache |
| 34 | cache = {} |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 35 | |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 36 | |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 37 | def getlines(filename, module_globals=None): |
Victor Stinner | 376658f | 2015-03-18 14:16:50 +0100 | [diff] [blame] | 38 | """Get the lines for a Python source file from the cache. |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 39 | Update the cache if it doesn't contain an entry for this file already.""" |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 40 | |
Raymond Hettinger | 54f0222 | 2002-06-01 14:18:47 +0000 | [diff] [blame] | 41 | if filename in cache: |
Robert Collins | 6bc2c1e | 2015-03-05 12:07:57 +1300 | [diff] [blame] | 42 | entry = cache[filename] |
Serhiy Storchaka | 05ddbf0 | 2015-04-01 16:56:13 +0300 | [diff] [blame] | 43 | if len(entry) != 1: |
| 44 | return cache[filename][2] |
Serhiy Storchaka | c512adc | 2015-04-01 16:54:05 +0300 | [diff] [blame] | 45 | |
| 46 | try: |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 47 | return updatecache(filename, module_globals) |
Serhiy Storchaka | c512adc | 2015-04-01 16:54:05 +0300 | [diff] [blame] | 48 | except MemoryError: |
| 49 | clearcache() |
| 50 | return [] |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 51 | |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 52 | |
Hye-Shik Chang | 182ac85 | 2004-10-26 09:16:42 +0000 | [diff] [blame] | 53 | def checkcache(filename=None): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 54 | """Discard cache entries that are out of date. |
| 55 | (This is not checked upon each call!)""" |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 56 | |
Hye-Shik Chang | 182ac85 | 2004-10-26 09:16:42 +0000 | [diff] [blame] | 57 | if filename is None: |
Guido van Rossum | f543348 | 2007-02-26 22:21:25 +0000 | [diff] [blame] | 58 | filenames = list(cache.keys()) |
Hye-Shik Chang | 182ac85 | 2004-10-26 09:16:42 +0000 | [diff] [blame] | 59 | else: |
| 60 | if filename in cache: |
| 61 | filenames = [filename] |
| 62 | else: |
| 63 | return |
| 64 | |
| 65 | for filename in filenames: |
Robert Collins | 6bc2c1e | 2015-03-05 12:07:57 +1300 | [diff] [blame] | 66 | entry = cache[filename] |
| 67 | if len(entry) == 1: |
| 68 | # lazy cache entry, leave it lazy. |
| 69 | continue |
| 70 | size, mtime, lines, fullname = entry |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 71 | if mtime is None: |
| 72 | continue # no-op for files loaded via a __loader__ |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 73 | try: |
| 74 | stat = os.stat(fullname) |
Andrew Svetlov | ad28c7f | 2012-12-18 22:02:39 +0200 | [diff] [blame] | 75 | except OSError: |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 76 | del cache[filename] |
| 77 | continue |
Raymond Hettinger | 32200ae | 2002-06-01 19:51:15 +0000 | [diff] [blame] | 78 | if size != stat.st_size or mtime != stat.st_mtime: |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 79 | del cache[filename] |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 80 | |
Guido van Rossum | 921c824 | 1992-01-10 14:54:42 +0000 | [diff] [blame] | 81 | |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 82 | def updatecache(filename, module_globals=None): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 83 | """Update a cache entry and return its list of lines. |
| 84 | If something's wrong, print a message, discard the cache entry, |
| 85 | and return an empty list.""" |
| 86 | |
Raymond Hettinger | 54f0222 | 2002-06-01 14:18:47 +0000 | [diff] [blame] | 87 | if filename in cache: |
Robert Collins | 6bc2c1e | 2015-03-05 12:07:57 +1300 | [diff] [blame] | 88 | if len(cache[filename]) != 1: |
| 89 | del cache[filename] |
Benjamin Peterson | aada7b8 | 2010-05-21 21:45:06 +0000 | [diff] [blame] | 90 | if not filename or (filename.startswith('<') and filename.endswith('>')): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 91 | return [] |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 92 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 93 | fullname = filename |
| 94 | try: |
| 95 | stat = os.stat(fullname) |
Benjamin Peterson | aada7b8 | 2010-05-21 21:45:06 +0000 | [diff] [blame] | 96 | except OSError: |
Georg Brandl | 991f920 | 2009-05-05 08:31:54 +0000 | [diff] [blame] | 97 | basename = filename |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 98 | |
Robert Collins | 6bc2c1e | 2015-03-05 12:07:57 +1300 | [diff] [blame] | 99 | # Realise a lazy loader based lookup if there is one |
| 100 | # otherwise try to lookup right now. |
| 101 | if lazycache(filename, module_globals): |
| 102 | try: |
| 103 | data = cache[filename][0]() |
| 104 | except (ImportError, OSError): |
| 105 | pass |
| 106 | else: |
| 107 | if data is None: |
| 108 | # No luck, the PEP302 loader cannot find the source |
| 109 | # for this module. |
| 110 | return [] |
| 111 | cache[filename] = ( |
| 112 | len(data), None, |
| 113 | [line+'\n' for line in data.splitlines()], fullname |
| 114 | ) |
| 115 | return cache[filename][2] |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 116 | |
Georg Brandl | 991f920 | 2009-05-05 08:31:54 +0000 | [diff] [blame] | 117 | # Try looking through the module search path, which is only useful |
| 118 | # when handling a relative filename. |
| 119 | if os.path.isabs(filename): |
| 120 | return [] |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 121 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 122 | for dirname in sys.path: |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 123 | try: |
Tim Peters | 12f21ae | 2001-05-29 04:27:01 +0000 | [diff] [blame] | 124 | fullname = os.path.join(dirname, basename) |
| 125 | except (TypeError, AttributeError): |
| 126 | # Not sufficiently string-like to do anything useful with. |
Benjamin Peterson | aada7b8 | 2010-05-21 21:45:06 +0000 | [diff] [blame] | 127 | continue |
| 128 | try: |
| 129 | stat = os.stat(fullname) |
| 130 | break |
Andrew Svetlov | ad28c7f | 2012-12-18 22:02:39 +0200 | [diff] [blame] | 131 | except OSError: |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 132 | pass |
| 133 | else: |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 134 | return [] |
Benjamin Peterson | aada7b8 | 2010-05-21 21:45:06 +0000 | [diff] [blame] | 135 | try: |
Victor Stinner | 58c0752 | 2010-11-09 01:08:59 +0000 | [diff] [blame] | 136 | with tokenize.open(fullname) as fp: |
Benjamin Peterson | aada7b8 | 2010-05-21 21:45:06 +0000 | [diff] [blame] | 137 | lines = fp.readlines() |
Andrew Svetlov | f7a17b4 | 2012-12-25 16:47:37 +0200 | [diff] [blame] | 138 | except OSError: |
Victor Stinner | 41a64a5 | 2010-09-29 01:30:45 +0000 | [diff] [blame] | 139 | return [] |
Benjamin Peterson | aada7b8 | 2010-05-21 21:45:06 +0000 | [diff] [blame] | 140 | if lines and not lines[-1].endswith('\n'): |
| 141 | lines[-1] += '\n' |
Raymond Hettinger | 32200ae | 2002-06-01 19:51:15 +0000 | [diff] [blame] | 142 | size, mtime = stat.st_size, stat.st_mtime |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 143 | cache[filename] = size, mtime, lines, fullname |
| 144 | return lines |
Robert Collins | 6bc2c1e | 2015-03-05 12:07:57 +1300 | [diff] [blame] | 145 | |
| 146 | |
| 147 | def lazycache(filename, module_globals): |
| 148 | """Seed the cache for filename with module_globals. |
| 149 | |
| 150 | The module loader will be asked for the source only when getlines is |
| 151 | called, not immediately. |
| 152 | |
| 153 | If there is an entry in the cache already, it is not altered. |
| 154 | |
| 155 | :return: True if a lazy load is registered in the cache, |
| 156 | otherwise False. To register such a load a module loader with a |
| 157 | get_source method must be found, the filename must be a cachable |
| 158 | filename, and the filename must not be already cached. |
| 159 | """ |
| 160 | if filename in cache: |
| 161 | if len(cache[filename]) == 1: |
| 162 | return True |
| 163 | else: |
| 164 | return False |
| 165 | if not filename or (filename.startswith('<') and filename.endswith('>')): |
| 166 | return False |
| 167 | # Try for a __loader__, if available |
| 168 | if module_globals and '__loader__' in module_globals: |
| 169 | name = module_globals.get('__name__') |
| 170 | loader = module_globals['__loader__'] |
| 171 | get_source = getattr(loader, 'get_source', None) |
| 172 | |
| 173 | if name and get_source: |
| 174 | get_lines = functools.partial(get_source, name) |
| 175 | cache[filename] = (get_lines,) |
| 176 | return True |
| 177 | return False |