1beam_lib(3) Erlang Module Definition beam_lib(3)
2
6 beam_lib - An interface to the BEAM file format.
7
9 This module provides an interface to files created by the BEAM Compiler
10 ("BEAM files"). The format used, a variant of "EA IFF 1985" Standard
11 for Interchange Format Files, divides data into chunks.
12 Chunk data can be returned as binaries or as compound terms. Compound
14 terms are returned when chunks are referenced by names (atoms) rather
15 than identifiers (strings). The recognized names and the corresponding
16 identifiers are as follows:
17 * atoms ("Atom")
19 * attributes ("Attr")
21 * compile_info ("CInf")
23 * debug_info ("Dbgi")
25 * exports ("ExpT")
27 * imports ("ImpT")
29 * indexed_imports ("ImpT")
31 * labeled_exports ("ExpT")
33 * labeled_locals ("LocT")
35 * locals ("LocT")
37
39 Option debug_info can be specified to the Compiler (see compile(3)) to
40 have debug information, such as Erlang Abstract Format, stored in the
41 debug_info chunk. Tools such as Debugger and Xref require the debug
42 information to be included.
43 Warning:
45 Source code can be reconstructed from the debug information. To prevent
46 this, use encrypted debug information (see below).
47 The debug information can also be removed from BEAM files using
50 strip/1, strip_files/1, and/or strip_release/1.
51
53 The following example shows how to reconstruct Erlang source code from
54 the debug information in a BEAM file Beam:
55 {ok,{_,[{abstract_code,{_,AC}}]}} = beam_lib:chunks(Beam,[abstract_code]).
57 io:fwrite("~s~n", [erl_prettypr:format(erl_syntax:form_list(AC))]).
58
60 The debug information can be encrypted to keep the source code secret,
61 but still be able to use tools such as Debugger or Xref.
62 To use encrypted debug information, a key must be provided to the com‐
64 piler and beam_lib. The key is specified as a string. It is recommended
65 that the string contains at least 32 characters and that both upper and
66 lower case letters as well as digits and special characters are used.
67 The default type (and currently the only type) of crypto algorithm is
69 des3_cbc, three rounds of DES. The key string is scrambled using
70 erlang:md5/1 to generate the keys used for des3_cbc.
71 Note:
73 As far as we know by the time of writing, it is infeasible to break
74 des3_cbc encryption without any knowledge of the key. Therefore, as
75 long as the key is kept safe and is unguessable, the encrypted debug
76 information should be safe from intruders.
77 The key can be provided in the following two ways:
80 * Use Compiler option {debug_info_key,Key}, see compile(3) and func‐
82 tion crypto_key_fun/1 to register a fun that returns the key when‐
83 ever beam_lib must decrypt the debug information.
84 If no such fun is registered, beam_lib instead searches for an
86 .erlang.crypt file, see the next section.
87 * Store the key in a text file named .erlang.crypt.
89 In this case, Compiler option encrypt_debug_info can be used, see
91 compile(3).
92
94 beam_lib searches for .erlang.crypt in the current directory and then
95 the home directory for the current user. If the file is found and con‐
96 tains a key, beam_lib implicitly creates a crypto key fun and registers
97 it.
98 File .erlang.crypt is to contain a single list of tuples:
100 {debug_info, Mode, Module, Key}
102 Mode is the type of crypto algorithm; currently, the only allowed value
104 is des3_cbc. Module is either an atom, in which case Key is only used
105 for the module Module, or [], in which case Key is used for all mod‐
106 ules. Key is the non-empty key string.
107 Key in the first tuple where both Mode and Module match is used.
109 The following is an example of an .erlang.crypt file that returns the
111 same key for all modules:
112 [{debug_info, des3_cbc, [], "%>7}|pc/DM6Cga*68$Mw]L#&_Gejr]G^"}].
114 The following is a slightly more complicated example of an
116 .erlang.crypt providing one key for module t and another key for all
117 other modules:
118 [{debug_info, des3_cbc, t, "My KEY"},
120 {debug_info, des3_cbc, [], "%>7}|pc/DM6Cga*68$Mw]L#&_Gejr]G^"}].
121 Note:
123 Do not use any of the keys in these examples. Use your own keys.
124
127 beam() = file:filename() | binary()
128 Each of the functions described below accept either the filename
130 (as a string) or a binary containing the BEAM module.
131 chunkdata() =
133 {chunkid(), dataB()} |
134 {abstract_code, abst_code()} |
135 {debug_info, debug_info()} |
136 {attributes, [attrib_entry()]} |
137 {compile_info, [compinfo_entry()]} |
138 {exports, [{atom(), arity()}]} |
139 {labeled_exports, [labeled_entry()]} |
140 {imports, [mfa()]} |
141 {indexed_imports,
142 [{index(), module(), Function :: atom(), arity()}]} |
143 {locals, [{atom(), arity()}]} |
144 {labeled_locals, [labeled_entry()]} |
145 {atoms, [{integer(), atom()}]}
146 The list of attributes is sorted on Attribute (in
148 attrib_entry()) and each attribute name occurs once in the list.
149 The attribute values occur in the same order as in the file. The
150 lists of functions are also sorted.
151 chunkid() = nonempty_string()
153 "Attr" | "CInf" | "Dbgi" | "ExpT" | "ImpT" | "LocT" | "AtU8"
155 dataB() = binary()
157 debug_info() =
159 {DbgiVersion :: atom(), Backend :: module(), Data :: term()} |
160 no_debug_info
161 The format stored in the debug_info chunk. To retrieve particu‐
163 lar code representation from the backend, Back‐
164 end:debug_info(Format, Module, Data, Opts) must be invoked. For‐
165 mat is an atom, such as erlang_v1 for the Erlang Abstract Format
166 or core_v1 for Core Erlang. Module is the module represented by
167 the beam file and Data is the value stored in the debug info
168 chunk. Opts is any list of values supported by the Backend.
169 Backend:debug_info/4 must return {ok, Code} or {error, Term}.
170 Developers must always invoke the debug_info/4 function and
172 never rely on the Data stored in the debug_info chunk, as it is
173 opaque and may change at any moment. no_debug_info means that
174 chunk "Dbgi" is present, but empty.
175 abst_code() =
177 {AbstVersion :: atom(), forms()} | no_abstract_code
178 It is not checked that the forms conform to the abstract format
180 indicated by AbstVersion. no_abstract_code means that chunk
181 "Abst" is present, but empty.
182 For modules compiled with OTP 20 onwards, the abst_code chunk is
184 automatically computed from the debug_info chunk.
185 forms() = [erl_parse:abstract_form() | erl_parse:form_info()]
187 compinfo_entry() = {InfoKey :: atom(), term()}
189 attrib_entry() =
191 {Attribute :: atom(), [AttributeValue :: term()]}
192 labeled_entry() = {Function :: atom(), arity(), label()}
194 index() = integer() >= 0
196 label() = integer()
198 chunkref() = chunkname() | chunkid()
200 chunkname() =
202 abstract_code | debug_info | attributes | compile_info |
203 exports | labeled_exports | imports | indexed_imports |
204 locals | labeled_locals | atoms
205 chnk_rsn() =
207 {unknown_chunk, file:filename(), atom()} |
208 {key_missing_or_invalid,
209 file:filename(),
210 abstract_code | debug_info} |
211 info_rsn()
212 info_rsn() =
214 {chunk_too_big,
215 file:filename(),
216 chunkid(),
217 ChunkSize :: integer() >= 0,
218 FileSize :: integer() >= 0} |
219 {invalid_beam_file,
220 file:filename(),
221 Position :: integer() >= 0} |
222 {invalid_chunk, file:filename(), chunkid()} |
223 {missing_chunk, file:filename(), chunkid()} |
224 {not_a_beam_file, file:filename()} |
225 {file_error, file:filename(), file:posix()}
226
228 all_chunks(File :: beam()) ->
229 {ok, beam_lib, [{chunkid(), dataB()}]} |
230 {error, beam_lib, info_rsn()}
231 Reads chunk data for all chunks.
233 build_module(Chunks) -> {ok, Binary}
235 Types:
237 Chunks = [{chunkid(), dataB()}]
239 Binary = binary()
240 Builds a BEAM module (as a binary) from a list of chunks.
242 chunks(Beam, ChunkRefs) ->
244 {ok, {module(), [chunkdata()]}} |
245 {error, beam_lib, chnk_rsn()}
246 Types:
248 Beam = beam()
250 ChunkRefs = [chunkref()]
251 Reads chunk data for selected chunks references. The order of
253 the returned list of chunk data is determined by the order of
254 the list of chunks references.
255 chunks(Beam, ChunkRefs, Options) ->
257 {ok, {module(), [ChunkResult]}} |
258 {error, beam_lib, chnk_rsn()}
259 Types:
261 Beam = beam()
263 ChunkRefs = [chunkref()]
264 Options = [allow_missing_chunks]
265 ChunkResult =
266 chunkdata() | {ChunkRef :: chunkref(), missing_chunk}
267 Reads chunk data for selected chunks references. The order of
269 the returned list of chunk data is determined by the order of
270 the list of chunks references.
271 By default, if any requested chunk is missing in Beam, an error
273 tuple is returned. However, if option allow_missing_chunks is
274 specified, a result is returned even if chunks are missing. In
275 the result list, any missing chunks are represented as
276 {ChunkRef,missing_chunk}. Notice however that if chunk "Atom" is
277 missing, that is considered a fatal error and the return value
278 is an error tuple.
279 clear_crypto_key_fun() -> undefined | {ok, Result}
281 Types:
283 Result = undefined | term()
285 Unregisters the crypto key fun and terminates the process hold‐
287 ing it, started by crypto_key_fun/1.
288 Returns either {ok, undefined} if no crypto key fun is regis‐
290 tered, or {ok, Term}, where Term is the return value from Cryp‐
291 toKeyFun(clear), see crypto_key_fun/1.
292 cmp(Beam1, Beam2) -> ok | {error, beam_lib, cmp_rsn()}
294 Types:
296 Beam1 = Beam2 = beam()
298 cmp_rsn() =
299 {modules_different, module(), module()} |
300 {chunks_different, chunkid()} |
301 different_chunks |
302 info_rsn()
303 Compares the contents of two BEAM files. If the module names are
305 the same, and all chunks except for chunk "CInf" (the chunk con‐
306 taining the compilation information that is returned by Mod‐
307 ule:module_info(compile)) have the same contents in both files,
308 ok is returned. Otherwise an error message is returned.
309 cmp_dirs(Dir1, Dir2) ->
311 {Only1, Only2, Different} | {error, beam_lib, Reason}
312 Types:
314 Dir1 = Dir2 = atom() | file:filename()
316 Only1 = Only2 = [file:filename()]
317 Different =
318 [{Filename1 :: file:filename(), Filename2 :: file:file‐
319 name()}]
320 Reason = {not_a_directory, term()} | info_rsn()
321 Compares the BEAM files in two directories. Only files with
323 extension ".beam" are compared. BEAM files that exist only in
324 directory Dir1 (Dir2) are returned in Only1 (Only2). BEAM files
325 that exist in both directories but are considered different by
326 cmp/2 are returned as pairs {Filename1, Filename2}, where File‐
327 name1 (Filename2) exists in directory Dir1 (Dir2).
328 crypto_key_fun(CryptoKeyFun) -> ok | {error, Reason}
330 Types:
332 CryptoKeyFun = crypto_fun()
334 Reason = badfun | exists | term()
335 crypto_fun() = fun((crypto_fun_arg()) -> term())
336 crypto_fun_arg() =
337 init | clear | {debug_info, mode(), module(), file:filename()}
338 mode() = des3_cbc
339 Registers an unary fun that is called if beam_lib must read an
341 debug_info chunk that has been encrypted. The fun is held in a
342 process that is started by the function.
343 If a fun is already registered when attempting to register a
345 fun, {error, exists} is returned.
346 The fun must handle the following arguments:
348 CryptoKeyFun(init) -> ok | {ok, NewCryptoKeyFun} | {error, Term}
350 Called when the fun is registered, in the process that holds the
352 fun. Here the crypto key fun can do any necessary initializa‐
353 tions. If {ok, NewCryptoKeyFun} is returned, NewCryptoKeyFun is
354 registered instead of CryptoKeyFun. If {error, Term} is
355 returned, the registration is aborted and crypto_key_fun/1 also
356 returns {error, Term}.
357 CryptoKeyFun({debug_info, Mode, Module, Filename}) -> Key
359 Called when the key is needed for module Module in the file
361 named Filename. Mode is the type of crypto algorithm; currently,
362 the only possible value is des3_cbc. The call is to fail (raise
363 an exception) if no key is available.
364 CryptoKeyFun(clear) -> term()
366 Called before the fun is unregistered. Here any cleaning up can
368 be done. The return value is not important, but is passed back
369 to the caller of clear_crypto_key_fun/0 as part of its return
370 value.
371 diff_dirs(Dir1, Dir2) -> ok | {error, beam_lib, Reason}
373 Types:
375 Dir1 = Dir2 = atom() | file:filename()
377 Reason = {not_a_directory, term()} | info_rsn()
378 Compares the BEAM files in two directories as cmp_dirs/2, but
380 the names of files that exist in only one directory or are dif‐
381 ferent are presented on standard output.
382 format_error(Reason) -> io_lib:chars()
384 Types:
386 Reason = term()
388 For a specified error returned by any function in this module,
390 this function returns a descriptive string of the error in Eng‐
391 lish. For file errors, function file:format_error(Posix) is to
392 be called.
393 info(Beam) -> [InfoPair] | {error, beam_lib, info_rsn()}
395 Types:
397 Beam = beam()
399 InfoPair =
400 {file, Filename :: file:filename()} |
401 {binary, Binary :: binary()} |
402 {module, Module :: module()} |
403 {chunks,
404 [{ChunkId :: chunkid(),
405 Pos :: integer() >= 0,
406 Size :: integer() >= 0}]}
407 Returns a list containing some information about a BEAM file as
409 tuples {Item, Info}:
410 {file, Filename} | {binary, Binary}:
412 The name (string) of the BEAM file, or the binary from which
413 the information was extracted.
414 {module, Module}:
416 The name (atom) of the module.
417 {chunks, [{ChunkId, Pos, Size}]}:
419 For each chunk, the identifier (string) and the position and
420 size of the chunk data, in bytes.
421 md5(Beam) -> {ok, {module(), MD5}} | {error, beam_lib, chnk_rsn()}
423 Types:
425 Beam = beam()
427 MD5 = binary()
428 Calculates an MD5 redundancy check for the code of the module
430 (compilation date and other attributes are not included).
431 strip(Beam1) ->
433 {ok, {module(), Beam2}} | {error, beam_lib, info_rsn()}
434 Types:
436 Beam1 = Beam2 = beam()
438 Removes all chunks from a BEAM file except those needed by the
440 loader. In particular, the debug information (chunk debug_info
441 and abstract_code) is removed.
442 strip(Beam1, AdditionalChunks) ->
444 {ok, {module(), Beam2}} | {error, beam_lib, info_rsn()}
445 Types:
447 Beam1 = beam()
449 AdditionalChunks = [chunkid()]
450 Beam2 = beam()
451 Removes all chunks from a BEAM file except those needed by the
453 loader or passed in. In particular, the debug information (chunk
454 debug_info and abstract_code) is removed.
455 strip_files(Files) ->
457 {ok, [{module(), Beam}]} |
458 {error, beam_lib, info_rsn()}
459 Types:
461 Files = [beam()]
463 Beam = beam()
464 Removes all chunks except those needed by the loader from BEAM
466 files. In particular, the debug information (chunk debug_info
467 and abstract_code) is removed. The returned list contains one
468 element for each specified filename, in the same order as in
469 Files.
470 strip_files(Files, AdditionalChunks) ->
472 {ok, [{module(), Beam}]} |
473 {error, beam_lib, info_rsn()}
474 Types:
476 Files = [beam()]
478 AdditionalChunks = [chunkid()]
479 Beam = beam()
480 Removes all chunks except those needed by the loader or passed
482 in from BEAM files. In particular, the debug information (chunk
483 debug_info and abstract_code) is removed. The returned list con‐
484 tains one element for each specified filename, in the same order
485 as in Files.
486 strip_release(Dir) ->
488 {ok, [{module(), file:filename()}]} |
489 {error, beam_lib, Reason}
490 Types:
492 Dir = atom() | file:filename()
494 Reason = {not_a_directory, term()} | info_rsn()
495 Removes all chunks except those needed by the loader from the
497 BEAM files of a release. Dir is to be the installation root
498 directory. For example, the current OTP release can be stripped
499 with the call beam_lib:strip_release(code:root_dir()).
500 strip_release(Dir, AdditionalChunks) ->
502 {ok, [{module(), file:filename()}]} |
503 {error, beam_lib, Reason}
504 Types:
506 Dir = atom() | file:filename()
508 AdditionalChunks = [chunkid()]
509 Reason = {not_a_directory, term()} | info_rsn()
510 Removes all chunks except those needed by the loader or passed
512 in from the BEAM files of a release. Dir is to be the installa‐
513 tion root directory. For example, the current OTP release can be
514 stripped with the call beam_lib:strip_release(code:root_dir()).
515 version(Beam) ->
517 {ok, {module(), [Version :: term()]}} |
518 {error, beam_lib, chnk_rsn()}
519 Types:
521 Beam = beam()
523 Returns the module version or versions. A version is defined by
525 module attribute -vsn(Vsn). If this attribute is not specified,
526 the version defaults to the checksum of the module. Notice that
527 if version Vsn is not a list, it is made into one, that is
528 {ok,{Module,[Vsn]}} is returned. If there are many -vsn module
529 attributes, the result is the concatenated list of versions.
530 Examples:
532 1> beam_lib:version(a). % -vsn(1).
534 {ok,{a,[1]}}
535 2> beam_lib:version(b). % -vsn([1]).
536 {ok,{b,[1]}}
537 3> beam_lib:version(c). % -vsn([1]). -vsn(2).
538 {ok,{c,[1,2]}}
539 4> beam_lib:version(d). % no -vsn attribute
540 {ok,{d,[275613208176997377698094100858909383631]}}
541Ericsson AB stdlib 3.12.1 beam_lib(3)