3 * Copyright (C) 2009-2010 Red Hat Inc.
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 (* This script generates language bindings and some documentation for
23 * After editing this file, run it (./generator/generator.ml) to
24 * regenerate all the output files. 'make' will rerun this
25 * automatically when necessary. Note that if you are using a separate
26 * build directory you must run generator.ml from the _source_
29 * IMPORTANT: This script should NOT print any warnings. If it prints
30 * warnings, you should treat them as errors.
32 * OCaml tips: (1) In emacs, install tuareg-mode to display and format
33 * OCaml code correctly. 'vim' comes with a good OCaml editing mode by
34 * default. (2) Read the resources at http://ocaml-tutorial.org/
43 type style = ret * args
45 | RErr (* 0 = ok, -1 = error *)
46 | RErrDispose (* Disposes handle, see hivex_close. *)
47 | RHive (* Returns a hive_h or NULL. *)
48 | RNode (* Returns hive_node_h or 0. *)
49 | RNodeNotFound (* See hivex_node_get_child. *)
50 | RNodeList (* Returns hive_node_h* or NULL. *)
51 | RValue (* Returns hive_value_h or 0. *)
52 | RValueList (* Returns hive_value_h* or NULL. *)
53 | RString (* Returns char* or NULL. *)
54 | RStringList (* Returns char** or NULL. *)
55 | RLenType (* See hivex_value_type. *)
56 | RLenTypeVal (* See hivex_value_value. *)
57 | RInt32 (* Returns int32. *)
58 | RInt64 (* Returns int64. *)
60 and args = argt list (* List of parameters. *)
62 and argt = (* Note, cannot be NULL/0 unless it
63 says so explicitly below. *)
65 | ANode of string (* hive_node_h *)
66 | AValue of string (* hive_value_h *)
67 | AString of string (* char* *)
68 | AStringNullable of string (* char* (can be NULL) *)
69 | AOpenFlags (* HIVEX_OPEN_* flags list. *)
70 | AUnusedFlags (* Flags arg that is always 0 *)
71 | ASetValues (* See hivex_node_set_values. *)
72 | ASetValue (* See hivex_node_set_value. *)
75 * https://secure.wikimedia.org/wikipedia/en/wiki/Windows_Registry#Keys_and_values
77 * It's unfortunate that in our original C binding we strayed away from
78 * the names that Windows uses (eg. REG_SZ for strings). We include
79 * both our names and the Windows names.
83 "Just a key without a value";
85 "A Windows string (encoding is unknown, but often UTF16-LE)";
86 2, "expand_string", "EXPAND_SZ",
87 "A Windows string that contains %env% (environment variable expansion)";
88 3, "binary", "BINARY",
91 "DWORD (32 bit integer), little endian";
92 5, "dword_be", "DWORD_BIG_ENDIAN",
93 "DWORD (32 bit integer), big endian";
95 "Symbolic link to another part of the registry tree";
96 7, "multiple_strings", "MULTI_SZ",
97 "Multiple Windows strings. See http://blogs.msdn.com/oldnewthing/archive/2009/10/08/9904646.aspx";
98 8, "resource_list", "RESOURCE_LIST",
100 9, "full_resource_description", "FULL_RESOURCE_DESCRIPTOR",
101 "Resource descriptor";
102 10, "resource_requirements_list", "RESOURCE_REQUIREMENTS_LIST",
103 "Resouce requirements list";
104 11, "qword", "QWORD",
105 "QWORD (64 bit integer), unspecified endianness but usually little endian"
107 let max_hive_type = 11
109 (* Open flags (bitmask passed to AOpenFlags) *)
111 1, "VERBOSE", "Verbose messages";
112 2, "DEBUG", "Debug messages";
113 4, "WRITE", "Enable writes to the hive";
118 "open", (RHive, [AString "filename"; AOpenFlags]),
121 Opens the hive named C<filename> for reading.
123 Flags is an ORed list of the open flags (or C<0> if you don't
124 want to pass any flags). These flags are defined:
128 =item HIVEX_OPEN_VERBOSE
132 =item HIVEX_OPEN_DEBUG
134 Very verbose messages, suitable for debugging problems in the library
137 This is also selected if the C<HIVEX_DEBUG> environment variable
140 =item HIVEX_OPEN_WRITE
142 Open the hive for writing. If omitted, the hive is read-only.
144 See L<hivex(3)/WRITING TO HIVE FILES>.
148 "close", (RErrDispose, [AHive]),
149 "close a hive handle",
151 Close a hive handle and free all associated resources.
153 Note that any uncommitted writes are I<not> committed by this call,
154 but instead are lost. See L<hivex(3)/WRITING TO HIVE FILES>.";
156 "root", (RNode, [AHive]),
157 "return the root node of the hive",
159 Return root node of the hive. All valid hives must contain a root node.";
161 "node_name", (RString, [AHive; ANode "node"]),
162 "return the name of the node",
164 Return the name of the node.
166 Note that the name of the root node is a dummy, such as
167 C<$$$PROTO.HIV> (other names are possible: it seems to depend on the
168 tool or program that created the hive in the first place). You can
169 only know the \"real\" name of the root node by knowing which registry
170 file this hive originally comes from, which is knowledge that is
171 outside the scope of this library.";
173 "node_children", (RNodeList, [AHive; ANode "node"]),
174 "return children of node",
176 Return an array of nodes which are the subkeys
177 (children) of C<node>.";
179 "node_get_child", (RNodeNotFound, [AHive; ANode "node"; AString "name"]),
180 "return named child of node",
182 Return the child of node with the name C<name>, if it exists.
184 The name is matched case insensitively.";
186 "node_parent", (RNode, [AHive; ANode "node"]),
187 "return the parent of node",
189 Return the parent of C<node>.
191 The parent pointer of the root node in registry files that we
192 have examined seems to be invalid, and so this function will
193 return an error if called on the root node.";
195 "node_values", (RValueList, [AHive; ANode "node"]),
196 "return (key, value) pairs attached to a node",
198 Return the array of (key, value) pairs attached to this node.";
200 "node_get_value", (RValue, [AHive; ANode "node"; AString "key"]),
201 "return named key at node",
203 Return the value attached to this node which has the name C<key>,
206 The key name is matched case insensitively.
208 Note that to get the default key, you should pass the empty
209 string C<\"\"> here. The default key is often written C<\"@\">, but
210 inside hives that has no meaning and won't give you the
213 "value_key", (RString, [AHive; AValue "val"]),
214 "return the key of a (key, value) pair",
216 Return the key (name) of a (key, value) pair. The name
217 is reencoded as UTF-8 and returned as a string.
219 The string should be freed by the caller when it is no longer needed.
221 Note that this function can return a zero-length string. In the
222 context of Windows Registries, this means that this value is the
223 default key for this node in the tree. This is usually written
226 "value_type", (RLenType, [AHive; AValue "val"]),
227 "return data length and data type of a value",
229 Return the data length and data type of the value in this (key, value)
230 pair. See also C<hivex_value_value> which returns all this
231 information, and the value itself. Also, C<hivex_value_*> functions
232 below which can be used to return the value in a more useful form when
233 you know the type in advance.";
235 "value_value", (RLenTypeVal, [AHive; AValue "val"]),
236 "return data length, data type and data of a value",
238 Return the value of this (key, value) pair. The value should
239 be interpreted according to its type (see C<hive_type>).";
241 "value_string", (RString, [AHive; AValue "val"]),
242 "return value as a string",
244 If this value is a string, return the string reencoded as UTF-8
245 (as a C string). This only works for values which have type
246 C<hive_t_string>, C<hive_t_expand_string> or C<hive_t_link>.";
248 "value_multiple_strings", (RStringList, [AHive; AValue "val"]),
249 "return value as multiple strings",
251 If this value is a multiple-string, return the strings reencoded
252 as UTF-8 (in C, as a NULL-terminated array of C strings, in other
253 language bindings, as a list of strings). This only
254 works for values which have type C<hive_t_multiple_strings>.";
256 "value_dword", (RInt32, [AHive; AValue "val"]),
257 "return value as a DWORD",
259 If this value is a DWORD (Windows int32), return it. This only works
260 for values which have type C<hive_t_dword> or C<hive_t_dword_be>.";
262 "value_qword", (RInt64, [AHive; AValue "val"]),
263 "return value as a QWORD",
265 If this value is a QWORD (Windows int64), return it. This only
266 works for values which have type C<hive_t_qword>.";
268 "commit", (RErr, [AHive; AStringNullable "filename"; AUnusedFlags]),
269 "commit (write) changes to file",
271 Commit (write) any changes which have been made.
273 C<filename> is the new file to write. If C<filename> is null/undefined
274 then we overwrite the original file (ie. the file name that was passed to
277 Note this does not close the hive handle. You can perform further
278 operations on the hive after committing, including making more
279 modifications. If you no longer wish to use the hive, then you
280 should close the handle after committing.";
282 "node_add_child", (RNode, [AHive; ANode "parent"; AString "name"]),
285 Add a new child node named C<name> to the existing node C<parent>.
286 The new child initially has no subnodes and contains no keys or
287 values. The sk-record (security descriptor) is inherited from
290 The parent must not have an existing child called C<name>, so if you
291 want to overwrite an existing child, call C<hivex_node_delete_child>
294 "node_delete_child", (RErr, [AHive; ANode "node"]),
297 Delete the node C<node>. All values at the node and all subnodes are
298 deleted (recursively). The C<node> handle and the handles of all
299 subnodes become invalid. You cannot delete the root node.";
301 "node_set_values", (RErr, [AHive; ANode "node"; ASetValues; AUnusedFlags]),
302 "set (key, value) pairs at a node",
304 This call can be used to set all the (key, value) pairs
307 C<node> is the node to modify.";
309 "node_set_value", (RErr, [AHive; ANode "node"; ASetValue; AUnusedFlags]),
310 "set a single (key, value) pair at a given node",
312 This call can be used to replace a single C<(key, value)> pair
313 stored in C<node>. If the key does not already exist, then a
314 new key is added. Key matching is case insensitive.
316 C<node> is the node to modify.";
319 (* Used to memoize the result of pod2text. *)
320 let pod2text_memo_filename = "generator/.pod2text.data"
321 let pod2text_memo : ((int * string * string), string list) Hashtbl.t =
323 let chan = open_in pod2text_memo_filename in
324 let v = input_value chan in
328 _ -> Hashtbl.create 13
329 let pod2text_memo_updated () =
330 let chan = open_out pod2text_memo_filename in
331 output_value chan pod2text_memo;
335 * Note we don't want to use any external OCaml libraries which
336 * makes this a bit harder than it should be.
338 module StringMap = Map.Make (String)
340 let failwithf fs = ksprintf failwith fs
342 let unique = let i = ref 0 in fun () -> incr i; !i
344 let replace_char s c1 c2 =
345 let s2 = String.copy s in
347 for i = 0 to String.length s2 - 1 do
348 if String.unsafe_get s2 i = c1 then (
349 String.unsafe_set s2 i c2;
353 if not !r then s else s2
357 (* || c = '\f' *) || c = '\n' || c = '\r' || c = '\t' (* || c = '\v' *)
359 let triml ?(test = isspace) str =
361 let n = ref (String.length str) in
362 while !n > 0 && test str.[!i]; do
367 else String.sub str !i !n
369 let trimr ?(test = isspace) str =
370 let n = ref (String.length str) in
371 while !n > 0 && test str.[!n-1]; do
374 if !n = String.length str then str
375 else String.sub str 0 !n
377 let trim ?(test = isspace) str =
378 trimr ~test (triml ~test str)
381 let len = String.length s in
382 let sublen = String.length sub in
384 if i <= len-sublen then (
387 if s.[i+j] = sub.[j] then loop2 (j+1)
393 if r = -1 then loop (i+1) else r
399 let rec replace_str s s1 s2 =
400 let len = String.length s in
401 let sublen = String.length s1 in
405 let s' = String.sub s 0 i in
406 let s'' = String.sub s (i+sublen) (len-i-sublen) in
407 s' ^ s2 ^ replace_str s'' s1 s2
410 let rec string_split sep str =
411 let len = String.length str in
412 let seplen = String.length sep in
413 let i = find str sep in
416 let s' = String.sub str 0 i in
417 let s'' = String.sub str (i+seplen) (len-i-seplen) in
418 s' :: string_split sep s''
421 let files_equal n1 n2 =
422 let cmd = sprintf "cmp -s %s %s" (Filename.quote n1) (Filename.quote n2) in
423 match Sys.command cmd with
426 | i -> failwithf "%s: failed with error code %d" cmd i
428 let rec filter_map f = function
432 | Some y -> y :: filter_map f xs
433 | None -> filter_map f xs
435 let rec find_map f = function
436 | [] -> raise Not_found
440 | None -> find_map f xs
443 let rec loop i = function
445 | x :: xs -> f i x; loop (i+1) xs
450 let rec loop i = function
452 | x :: xs -> let r = f i x in r :: loop (i+1) xs
456 let count_chars c str =
458 for i = 0 to String.length str - 1 do
459 if c = String.unsafe_get str i then incr count
463 let name_of_argt = function
465 | ANode n | AValue n | AString n | AStringNullable n -> n
466 | AOpenFlags | AUnusedFlags -> "flags"
467 | ASetValues -> "values"
470 (* Check function names etc. for consistency. *)
471 let check_functions () =
472 let contains_uppercase str =
473 let len = String.length str in
475 if i >= len then false
478 if c >= 'A' && c <= 'Z' then true
485 (* Check function names. *)
487 fun (name, _, _, _) ->
488 if String.length name >= 7 && String.sub name 0 7 = "hivex" then
489 failwithf "function name %s does not need 'hivex' prefix" name;
491 failwithf "function name is empty";
492 if name.[0] < 'a' || name.[0] > 'z' then
493 failwithf "function name %s must start with lowercase a-z" name;
494 if String.contains name '-' then
495 failwithf "function name %s should not contain '-', use '_' instead."
499 (* Check function parameter/return names. *)
501 fun (name, style, _, _) ->
502 let check_arg_ret_name n =
503 if contains_uppercase n then
504 failwithf "%s param/ret %s should not contain uppercase chars"
506 if String.contains n '-' || String.contains n '_' then
507 failwithf "%s param/ret %s should not contain '-' or '_'"
510 failwithf "%s has a param/ret called 'value', which causes conflicts in the OCaml bindings, use something like 'val' or a more descriptive name" name;
511 if n = "int" || n = "char" || n = "short" || n = "long" then
512 failwithf "%s has a param/ret which conflicts with a C type (eg. 'int', 'char' etc.)" name;
513 if n = "i" || n = "n" then
514 failwithf "%s has a param/ret called 'i' or 'n', which will cause some conflicts in the generated code" name;
515 if n = "argv" || n = "args" then
516 failwithf "%s has a param/ret called 'argv' or 'args', which will cause some conflicts in the generated code" name;
518 (* List Haskell, OCaml and C keywords here.
519 * http://www.haskell.org/haskellwiki/Keywords
520 * http://caml.inria.fr/pub/docs/manual-ocaml/lex.html#operator-char
521 * http://en.wikipedia.org/wiki/C_syntax#Reserved_keywords
522 * Formatted via: cat c haskell ocaml|sort -u|grep -vE '_|^val$' \
523 * |perl -pe 's/(.+)/"$1";/'|fmt -70
524 * Omitting _-containing words, since they're handled above.
525 * Omitting the OCaml reserved word, "val", is ok,
526 * and saves us from renaming several parameters.
529 "and"; "as"; "asr"; "assert"; "auto"; "begin"; "break"; "case";
530 "char"; "class"; "const"; "constraint"; "continue"; "data";
531 "default"; "deriving"; "do"; "done"; "double"; "downto"; "else";
532 "end"; "enum"; "exception"; "extern"; "external"; "false"; "float";
533 "for"; "forall"; "foreign"; "fun"; "function"; "functor"; "goto";
534 "hiding"; "if"; "import"; "in"; "include"; "infix"; "infixl";
535 "infixr"; "inherit"; "initializer"; "inline"; "instance"; "int";
537 "land"; "lazy"; "let"; "long"; "lor"; "lsl"; "lsr"; "lxor";
538 "match"; "mdo"; "method"; "mod"; "module"; "mutable"; "new";
539 "newtype"; "object"; "of"; "open"; "or"; "private"; "qualified";
540 "rec"; "register"; "restrict"; "return"; "short"; "sig"; "signed";
541 "sizeof"; "static"; "struct"; "switch"; "then"; "to"; "true"; "try";
542 "type"; "typedef"; "union"; "unsigned"; "virtual"; "void";
543 "volatile"; "when"; "where"; "while";
545 if List.mem n reserved then
546 failwithf "%s has param/ret using reserved word %s" name n;
549 List.iter (fun arg -> check_arg_ret_name (name_of_argt arg)) (snd style)
552 (* Check short descriptions. *)
554 fun (name, _, shortdesc, _) ->
555 if shortdesc.[0] <> Char.lowercase shortdesc.[0] then
556 failwithf "short description of %s should begin with lowercase." name;
557 let c = shortdesc.[String.length shortdesc-1] in
558 if c = '\n' || c = '.' then
559 failwithf "short description of %s should not end with . or \\n." name
562 (* Check long dscriptions. *)
564 fun (name, _, _, longdesc) ->
565 if longdesc.[String.length longdesc-1] = '\n' then
566 failwithf "long description of %s should not end with \\n." name
569 (* 'pr' prints to the current output file. *)
570 let chan = ref Pervasives.stdout
575 let i = count_chars '\n' str in
577 output_string !chan str
580 let copyright_years =
581 let this_year = 1900 + (localtime (time ())).tm_year in
582 if this_year > 2009 then sprintf "2009-%04d" this_year else "2009"
584 (* Generate a header block in a number of standard styles. *)
586 | CStyle | CPlusPlusStyle | HashStyle | OCamlStyle | HaskellStyle
588 type license = GPLv2plus | LGPLv2plus | GPLv2 | LGPLv2
590 let generate_header ?(extra_inputs = []) comment license =
591 let inputs = "generator/generator.ml" :: extra_inputs in
592 let c = match comment with
593 | CStyle -> pr "/* "; " *"
594 | CPlusPlusStyle -> pr "// "; "//"
595 | HashStyle -> pr "# "; "#"
596 | OCamlStyle -> pr "(* "; " *"
597 | HaskellStyle -> pr "{- "; " "
598 | PODCommentStyle -> pr "=begin comment\n\n "; "" in
599 pr "hivex generated file\n";
600 pr "%s WARNING: THIS FILE IS GENERATED FROM:\n" c;
601 List.iter (pr "%s %s\n" c) inputs;
602 pr "%s ANY CHANGES YOU MAKE TO THIS FILE WILL BE LOST.\n" c;
604 pr "%s Copyright (C) %s Red Hat Inc.\n" c copyright_years;
605 pr "%s Derived from code by Petter Nordahl-Hagen under a compatible license:\n" c;
606 pr "%s Copyright (c) 1997-2007 Petter Nordahl-Hagen.\n" c;
607 pr "%s Derived from code by Markus Stephany under a compatible license:\n" c;
608 pr "%s Copyright (c)2000-2004, Markus Stephany.\n" c;
612 pr "%s This program is free software; you can redistribute it and/or modify\n" c;
613 pr "%s it under the terms of the GNU General Public License as published by\n" c;
614 pr "%s the Free Software Foundation; either version 2 of the License, or\n" c;
615 pr "%s (at your option) any later version.\n" c;
617 pr "%s This program is distributed in the hope that it will be useful,\n" c;
618 pr "%s but WITHOUT ANY WARRANTY; without even the implied warranty of\n" c;
619 pr "%s MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n" c;
620 pr "%s GNU General Public License for more details.\n" c;
622 pr "%s You should have received a copy of the GNU General Public License along\n" c;
623 pr "%s with this program; if not, write to the Free Software Foundation, Inc.,\n" c;
624 pr "%s 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.\n" c;
627 pr "%s This library is free software; you can redistribute it and/or\n" c;
628 pr "%s modify it under the terms of the GNU Lesser General Public\n" c;
629 pr "%s License as published by the Free Software Foundation; either\n" c;
630 pr "%s version 2 of the License, or (at your option) any later version.\n" c;
632 pr "%s This library is distributed in the hope that it will be useful,\n" c;
633 pr "%s but WITHOUT ANY WARRANTY; without even the implied warranty of\n" c;
634 pr "%s MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU\n" c;
635 pr "%s Lesser General Public License for more details.\n" c;
637 pr "%s You should have received a copy of the GNU Lesser General Public\n" c;
638 pr "%s License along with this library; if not, write to the Free Software\n" c;
639 pr "%s Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA\n" c;
642 pr "%s This program is free software; you can redistribute it and/or modify\n" c;
643 pr "%s it under the terms of the GNU General Public License as published by\n" c;
644 pr "%s the Free Software Foundation; version 2 of the License only.\n" c;
646 pr "%s This program is distributed in the hope that it will be useful,\n" c;
647 pr "%s but WITHOUT ANY WARRANTY; without even the implied warranty of\n" c;
648 pr "%s MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n" c;
649 pr "%s GNU General Public License for more details.\n" c;
651 pr "%s You should have received a copy of the GNU General Public License along\n" c;
652 pr "%s with this program; if not, write to the Free Software Foundation, Inc.,\n" c;
653 pr "%s 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.\n" c;
656 pr "%s This library is free software; you can redistribute it and/or\n" c;
657 pr "%s modify it under the terms of the GNU Lesser General Public\n" c;
658 pr "%s License as published by the Free Software Foundation;\n" c;
659 pr "%s version 2.1 of the License only.\n" c;
661 pr "%s This library is distributed in the hope that it will be useful,\n" c;
662 pr "%s but WITHOUT ANY WARRANTY; without even the implied warranty of\n" c;
663 pr "%s MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU\n" c;
664 pr "%s Lesser General Public License for more details.\n" c;
666 pr "%s You should have received a copy of the GNU Lesser General Public\n" c;
667 pr "%s License along with this library; if not, write to the Free Software\n" c;
668 pr "%s Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA\n" c;
671 | CStyle -> pr " */\n"
674 | OCamlStyle -> pr " *)\n"
675 | HaskellStyle -> pr "-}\n"
676 | PODCommentStyle -> pr "\n=end comment\n"
680 (* Start of main code generation functions below this line. *)
682 let rec generate_c_header () =
683 generate_header CStyle LGPLv2;
695 /* NOTE: This API is documented in the man page hivex(3). */
698 typedef struct hive_h hive_h;
700 /* Nodes and values. */
701 typedef size_t hive_node_h;
702 typedef size_t hive_value_h;
704 /* Pre-defined types. */
708 fun (t, old_style, new_style, description) ->
709 pr " /* %s */\n" description;
710 pr " hive_t_REG_%s,\n" new_style;
711 pr "#define hive_t_%s hive_t_REG_%s\n" old_style new_style;
717 typedef enum hive_type hive_type;
719 /* Bitmask of flags passed to hivex_open. */
722 fun (v, flag, description) ->
723 pr " /* %s */\n" description;
724 pr "#define HIVEX_OPEN_%-10s %d\n" flag v;
729 /* Array of (key, value) pairs passed to hivex_node_set_values. */
730 struct hive_set_value {
736 typedef struct hive_set_value hive_set_value;
740 pr "/* Functions. */\n";
742 (* Function declarations. *)
744 fun (shortname, style, _, _) ->
745 let name = "hivex_" ^ shortname in
746 generate_c_prototype ~extern:true name style
749 (* The visitor pattern. *)
751 /* Visit all nodes. This is specific to the C API and is not made
752 * available to other languages. This is because of the complexity
753 * of binding callbacks in other languages, but also because other
754 * languages make it much simpler to iterate over a tree.
756 struct hivex_visitor {
757 int (*node_start) (hive_h *, void *opaque, hive_node_h, const char *name);
758 int (*node_end) (hive_h *, void *opaque, hive_node_h, const char *name);
759 int (*value_string) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *str);
760 int (*value_multiple_strings) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, char **argv);
761 int (*value_string_invalid_utf16) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *str);
762 int (*value_dword) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, int32_t);
763 int (*value_qword) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, int64_t);
764 int (*value_binary) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *value);
765 int (*value_none) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *value);
766 int (*value_other) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *value);
767 int (*value_any) (hive_h *, void *opaque, hive_node_h, hive_value_h, hive_type t, size_t len, const char *key, const char *value);
770 #define HIVEX_VISIT_SKIP_BAD 1
772 extern int hivex_visit (hive_h *h, const struct hivex_visitor *visitor, size_t len, void *opaque, int flags);
773 extern int hivex_visit_node (hive_h *h, hive_node_h node, const struct hivex_visitor *visitor, size_t len, void *opaque, int flags);
777 (* Finish the header file. *)
783 #endif /* HIVEX_H_ */
786 and generate_c_prototype ?(extern = false) name style =
787 if extern then pr "extern ";
788 (match fst style with
790 | RErrDispose -> pr "int "
791 | RHive -> pr "hive_h *"
792 | RNode -> pr "hive_node_h "
793 | RNodeNotFound -> pr "hive_node_h "
794 | RNodeList -> pr "hive_node_h *"
795 | RValue -> pr "hive_value_h "
796 | RValueList -> pr "hive_value_h *"
797 | RString -> pr "char *"
798 | RStringList -> pr "char **"
799 | RLenType -> pr "int "
800 | RLenTypeVal -> pr "char *"
801 | RInt32 -> pr "int32_t "
802 | RInt64 -> pr "int64_t "
805 let comma = ref false in
808 if !comma then pr ", "; comma := true;
810 | AHive -> pr "hive_h *h"
811 | ANode n -> pr "hive_node_h %s" n
812 | AValue n -> pr "hive_value_h %s" n
813 | AString n | AStringNullable n -> pr "const char *%s" n
814 | AOpenFlags | AUnusedFlags -> pr "int flags"
815 | ASetValues -> pr "size_t nr_values, const hive_set_value *values"
816 | ASetValue -> pr "const hive_set_value *val"
818 (match fst style with
819 | RLenType | RLenTypeVal -> pr ", hive_type *t, size_t *len"
824 and generate_c_pod () =
825 generate_header PODCommentStyle GPLv2;
832 hivex - Windows Registry \"hive\" extraction library
840 fun (shortname, style, _, _) ->
841 let name = "hivex_" ^ shortname in
843 generate_c_prototype ~extern:false name style;
848 Link with I<-lhivex>.
852 Hivex is a library for extracting the contents of Windows Registry
853 \"hive\" files. It is designed to be secure against buggy or malicious
856 Unlike other tools in this area, it doesn't use the textual .REG
857 format, because parsing that is as much trouble as parsing the
858 original binary format. Instead it makes the file available
859 through a C API, and then wraps this API in higher level scripting
862 There is a separate program to export the hive as XML
863 (see L<hivexml(1)>), or to navigate the file (see L<hivexsh(1)>).
864 There is also a Perl script to export and merge the
865 file as a textual .REG (regedit) file, see L<hivexregedit(1)>.
867 If you just want to export or modify the Registry of a Windows
868 virtual machine, you should look at L<virt-win-reg(1)>.
870 Hivex is also comes with language bindings for
871 OCaml, Perl and Python.
877 This handle describes an open hive file.
879 =head2 C<hive_node_h>
881 This is a node handle, an integer but opaque outside the library.
882 Valid node handles cannot be 0. The library returns 0 in some
883 situations to indicate an error.
887 The enum below describes the possible types for the value(s)
888 stored at each node. Note that you should not trust the
889 type field in a Windows Registry, as it very often has no
890 relationship to reality. Some applications use their own
891 types. The encoding of strings is not specified. Some
892 programs store everything (including strings) in binary blobs.
897 fun (t, _, new_style, description) ->
898 pr " /* %s */\n" description;
899 pr " hive_t_REG_%s = %d,\n" new_style t
904 =head2 C<hive_value_h>
906 This is a value handle, an integer but opaque outside the library.
907 Valid value handles cannot be 0. The library returns 0 in some
908 situations to indicate an error.
910 =head2 C<hive_set_value>
912 The typedef C<hive_set_value> is used in conjunction with the
913 C<hivex_node_set_values> call described below.
915 struct hive_set_value {
916 char *key; /* key - a UTF-8 encoded ASCIIZ string */
917 hive_type t; /* type of value field */
918 size_t len; /* length of value field in bytes */
919 char *value; /* value field */
921 typedef struct hive_set_value hive_set_value;
923 To set the default value for a node, you have to pass C<key = \"\">.
925 Note that the C<value> field is just treated as a list of bytes, and
926 is stored directly in the hive. The caller has to ensure correct
927 encoding and endianness, for example converting dwords to little
930 The correct type and encoding for values depends on the node and key
931 in the registry, the version of Windows, and sometimes even changes
932 between versions of Windows for the same key. We don't document it
933 here. Often it's not documented at all.
939 fun (shortname, style, _, longdesc) ->
940 let name = "hivex_" ^ shortname in
941 pr "=head2 %s\n" name;
943 generate_c_prototype ~extern:false name style;
948 if List.mem AUnusedFlags (snd style) then
949 pr "The flags parameter is unused. Always pass 0.\n\n";
951 if List.mem ASetValues (snd style) then
952 pr "C<values> is an array of (key, value) pairs. There
953 should be C<nr_values> elements in this array.
955 Any existing values stored at the node are discarded, and their
956 C<hive_value_h> handles become invalid. Thus you can remove all
957 values stored at C<node> by passing C<nr_values = 0>.\n\n";
959 if List.mem ASetValue (snd style) then
960 pr "C<value> is a single (key, value) pair.
962 Existing C<hive_value_h> handles become invalid.\n\n";
964 (match fst style with
967 Returns 0 on success.
968 On error this returns -1 and sets errno.\n\n"
971 Returns 0 on success.
972 On error this returns -1 and sets errno.
974 This function frees the hive handle (even if it returns an error).
975 The hive handle must not be used again after calling this function.\n\n"
978 Returns a new hive handle.
979 On error this returns NULL and sets errno.\n\n"
982 Returns a node handle.
983 On error this returns 0 and sets errno.\n\n"
986 Returns a node handle.
987 If the node was not found, this returns 0 without setting errno.
988 On error this returns 0 and sets errno.\n\n"
991 Returns a 0-terminated array of nodes.
992 The array must be freed by the caller when it is no longer needed.
993 On error this returns NULL and sets errno.\n\n"
996 Returns a value handle.
997 On error this returns 0 and sets errno.\n\n"
1000 Returns a 0-terminated array of values.
1001 The array must be freed by the caller when it is no longer needed.
1002 On error this returns NULL and sets errno.\n\n"
1006 The string must be freed by the caller when it is no longer needed.
1007 On error this returns NULL and sets errno.\n\n"
1010 Returns a NULL-terminated array of C strings.
1011 The strings and the array must all be freed by the caller when
1012 they are no longer needed.
1013 On error this returns NULL and sets errno.\n\n"
1016 Returns 0 on success.
1017 On error this returns -1 and sets errno.\n\n"
1020 The value is returned as an array of bytes (of length C<len>).
1021 The value must be freed by the caller when it is no longer needed.
1022 On error this returns NULL and sets errno.\n\n"
1023 | RInt32 | RInt64 -> ()
1028 =head1 WRITING TO HIVE FILES
1030 The hivex library supports making limited modifications to hive files.
1031 We have tried to implement this very conservatively in order to reduce
1032 the chance of corrupting your registry. However you should be careful
1033 and take back-ups, since Microsoft has never documented the hive
1034 format, and so it is possible there are nuances in the
1035 reverse-engineered format that we do not understand.
1037 To be able to modify a hive, you must pass the C<HIVEX_OPEN_WRITE>
1038 flag to C<hivex_open>, otherwise any write operation will return with
1041 The write operations shown below do not modify the on-disk file
1042 immediately. You must call C<hivex_commit> in order to write the
1043 changes to disk. If you call C<hivex_close> without committing then
1044 any writes are discarded.
1046 Hive files internally consist of a \"memory dump\" of binary blocks
1047 (like the C heap), and some of these blocks can be unused. The hivex
1048 library never reuses these unused blocks. Instead, to ensure
1049 robustness in the face of the partially understood on-disk format,
1050 hivex only allocates new blocks after the end of the file, and makes
1051 minimal modifications to existing structures in the file to point to
1052 these new blocks. This makes hivex slightly less disk-efficient than
1053 it could be, but disk is cheap, and registry modifications tend to be
1056 When deleting nodes, it is possible that this library may leave
1057 unreachable live blocks in the hive. This is because certain parts of
1058 the hive disk format such as security (sk) records and big data (db)
1059 records and classname fields are not well understood (and not
1060 documented at all) and we play it safe by not attempting to modify
1061 them. Apart from wasting a little bit of disk space, it is not
1062 thought that unreachable blocks are a problem.
1064 =head2 WRITE OPERATIONS WHICH ARE NOT SUPPORTED
1070 Changing the root node.
1074 Creating a new hive file from scratch. This is impossible at present
1075 because not all fields in the header are understood. In the hivex
1076 source tree is a file called C<images/minimal> which could be used as
1077 the basis for a new hive (but I<caveat emptor>).
1081 Modifying or deleting single values at a node.
1085 Modifying security key (sk) records or classnames.
1086 Previously we did not understand these records. However now they
1087 are well-understood and we could add support if it was required
1088 (but nothing much really uses them).
1092 =head1 VISITING ALL NODES
1094 The visitor pattern is useful if you want to visit all nodes
1095 in the tree or all nodes below a certain point in the tree.
1097 First you set up your own C<struct hivex_visitor> with your
1100 Each of these callback functions should return 0 on success or -1
1101 on error. If any callback returns -1, then the entire visit
1102 terminates immediately. If you don't need a callback function at
1103 all, set the function pointer to NULL.
1105 struct hivex_visitor {
1106 int (*node_start) (hive_h *, void *opaque, hive_node_h, const char *name);
1107 int (*node_end) (hive_h *, void *opaque, hive_node_h, const char *name);
1108 int (*value_string) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1109 hive_type t, size_t len, const char *key, const char *str);
1110 int (*value_multiple_strings) (hive_h *, void *opaque, hive_node_h,
1111 hive_value_h, hive_type t, size_t len, const char *key, char **argv);
1112 int (*value_string_invalid_utf16) (hive_h *, void *opaque, hive_node_h,
1113 hive_value_h, hive_type t, size_t len, const char *key,
1115 int (*value_dword) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1116 hive_type t, size_t len, const char *key, int32_t);
1117 int (*value_qword) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1118 hive_type t, size_t len, const char *key, int64_t);
1119 int (*value_binary) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1120 hive_type t, size_t len, const char *key, const char *value);
1121 int (*value_none) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1122 hive_type t, size_t len, const char *key, const char *value);
1123 int (*value_other) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1124 hive_type t, size_t len, const char *key, const char *value);
1125 /* If value_any callback is not NULL, then the other value_*
1126 * callbacks are not used, and value_any is called on all values.
1128 int (*value_any) (hive_h *, void *opaque, hive_node_h, hive_value_h,
1129 hive_type t, size_t len, const char *key, const char *value);
1136 int hivex_visit (hive_h *h, const struct hivex_visitor *visitor, size_t len, void *opaque, int flags);
1138 Visit all the nodes recursively in the hive C<h>.
1140 C<visitor> should be a C<hivex_visitor> structure with callback
1141 fields filled in as required (unwanted callbacks can be set to
1142 NULL). C<len> must be the length of the 'visitor' struct (you
1143 should pass C<sizeof (struct hivex_visitor)> for this).
1145 This returns 0 if the whole recursive visit was completed
1146 successfully. On error this returns -1. If one of the callback
1147 functions returned an error than we don't touch errno. If the
1148 error was generated internally then we set errno.
1150 You can skip bad registry entries by setting C<flag> to
1151 C<HIVEX_VISIT_SKIP_BAD>. If this flag is not set, then a bad registry
1152 causes the function to return an error immediately.
1154 This function is robust if the registry contains cycles or
1155 pointers which are invalid or outside the registry. It detects
1156 these cases and returns an error.
1158 =item hivex_visit_node
1160 int hivex_visit_node (hive_h *h, hive_node_h node, const struct hivex_visitor *visitor, size_t len, void *opaque);
1162 Same as C<hivex_visit> but instead of starting out at the root, this
1167 =head1 THE STRUCTURE OF THE WINDOWS REGISTRY
1169 Note: To understand the relationship between hives and the common
1170 Windows Registry keys (like C<HKEY_LOCAL_MACHINE>) please see the
1171 Wikipedia page on the Windows Registry.
1173 The Windows Registry is split across various binary files, each
1174 file being known as a \"hive\". This library only handles a single
1175 hive file at a time.
1177 Hives are n-ary trees with a single root. Each node in the tree
1180 Each node in the tree (including non-leaf nodes) may have an
1181 arbitrary list of (key, value) pairs attached to it. It may
1182 be the case that one of these pairs has an empty key. This
1183 is referred to as the default key for the node.
1185 The (key, value) pairs are the place where the useful data is
1186 stored in the registry. The key is always a string (possibly the
1187 empty string for the default key). The value is a typed object
1188 (eg. string, int32, binary, etc.).
1190 =head2 RELATIONSHIP TO .REG FILES
1192 The hivex C library does not care about or deal with Windows .REG
1193 files. Instead we push this complexity up to the Perl
1194 L<Win::Hivex(3)> library and the Perl programs
1195 L<hivexregedit(1)> and L<virt-win-reg(1)>.
1196 Nevertheless it is useful to look at the relationship between the
1197 Registry and .REG files because they are so common.
1199 A .REG file is a textual representation of the registry, or part of the
1200 registry. The actual registry hives that Windows uses are binary
1201 files. There are a number of Windows and Linux tools that let you
1202 generate .REG files, or merge .REG files back into the registry hives.
1203 Notable amongst them is Microsoft's REGEDIT program (formerly known as
1206 A typical .REG file will contain many sections looking like this:
1208 [HKEY_LOCAL_MACHINE\\SOFTWARE\\Classes\\Stack]
1209 \"@\"=\"Generic Stack\"
1210 \"TileInfo\"=\"prop:System.FileCount\"
1211 \"TilePath\"=str(2):\"%%systemroot%%\\\\system32\"
1212 \"ThumbnailCutoff\"=dword:00000000
1213 \"FriendlyTypeName\"=hex(2):40,00,25,00,53,00,79,00,73,00,74,00,65,00,6d,00,52,00,6f,00,\\
1214 6f,00,74,00,25,00,5c,00,53,00,79,00,73,00,74,00,65,00,6d,00,\\
1215 33,00,32,00,5c,00,73,00,65,00,61,00,72,00,63,00,68,00,66,00,\\
1216 6f,00,6c,00,64,00,65,00,72,00,2e,00,64,00,6c,00,6c,00,2c,00,\\
1217 2d,00,39,00,30,00,32,00,38,00,00,00,d8
1219 Taking this one piece at a time:
1221 [HKEY_LOCAL_MACHINE\\SOFTWARE\\Classes\\Stack]
1223 This is the path to this node in the registry tree. The first part,
1224 C<HKEY_LOCAL_MACHINE\\SOFTWARE> means that this comes from a hive
1225 file called C<C:\\WINDOWS\\SYSTEM32\\CONFIG\\SOFTWARE>.
1226 C<\\Classes\\Stack> is the real path part,
1227 starting at the root node of the C<SOFTWARE> hive.
1229 Below the node name is a list of zero or more key-value pairs. Any
1230 interior or leaf node in the registry may have key-value pairs
1233 \"@\"=\"Generic Stack\"
1235 This is the \"default key\". In reality (ie. inside the binary hive)
1236 the key string is the empty string. In .REG files this is written as
1237 C<@> but this has no meaning either in the hives themselves or in this
1238 library. The value is a string (type 1 - see C<enum hive_type>
1241 \"TileInfo\"=\"prop:System.FileCount\"
1243 This is a regular (key, value) pair, with the value being a type 1
1244 string. Note that inside the binary file the string is likely to be
1245 UTF-16LE encoded. This library converts to and from UTF-8 strings
1246 transparently in some cases.
1248 \"TilePath\"=str(2):\"%%systemroot%%\\\\system32\"
1250 The value in this case has type 2 (expanded string) meaning that some
1251 %%...%% variables get expanded by Windows. (This library doesn't know
1252 or care about variable expansion).
1254 \"ThumbnailCutoff\"=dword:00000000
1256 The value in this case is a dword (type 4).
1258 \"FriendlyTypeName\"=hex(2):40,00,....
1260 This value is an expanded string (type 2) represented in the .REG file
1261 as a series of hex bytes. In this case the string appears to be a
1264 =head1 NOTE ON THE USE OF ERRNO
1266 Many functions in this library set errno to indicate errors. These
1267 are the values of errno you may encounter (this list is not
1274 Corrupt or unsupported Registry file format.
1282 Passed an invalid argument to the function.
1286 Followed a Registry pointer which goes outside
1287 the registry or outside a registry block.
1291 Registry contains cycles.
1295 Field in the registry out of range.
1299 Registry key already exists.
1303 Tried to write to a registry which is not opened for writing.
1307 =head1 ENVIRONMENT VARIABLES
1313 Setting HIVEX_DEBUG=1 will enable very verbose messages. This is
1314 useful for debugging problems with the library itself.
1327 L<http://libguestfs.org/>,
1330 L<http://en.wikipedia.org/wiki/Windows_Registry>.
1334 Richard W.M. Jones (C<rjones at redhat dot com>)
1338 Copyright (C) 2009-2010 Red Hat Inc.
1340 Derived from code by Petter Nordahl-Hagen under a compatible license:
1341 Copyright (C) 1997-2007 Petter Nordahl-Hagen.
1343 Derived from code by Markus Stephany under a compatible license:
1344 Copyright (C) 2000-2004 Markus Stephany.
1346 This library is free software; you can redistribute it and/or
1347 modify it under the terms of the GNU Lesser General Public
1348 License as published by the Free Software Foundation;
1349 version 2.1 of the License only.
1351 This library is distributed in the hope that it will be useful,
1352 but WITHOUT ANY WARRANTY; without even the implied warranty of
1353 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
1354 Lesser General Public License for more details.
1357 (* Generate the linker script which controls the visibility of
1358 * symbols in the public ABI and ensures no other symbols get
1359 * exported accidentally.
1361 and generate_linker_script () =
1362 generate_header HashStyle GPLv2plus;
1370 List.map (fun (name, _, _, _) -> "hivex_" ^ name)
1372 let globals = List.sort compare (globals @ functions) in
1376 List.iter (pr " %s;\n") globals;
1383 and generate_ocaml_interface () =
1384 generate_header OCamlStyle LGPLv2plus;
1388 (** A [hive_h] hive file handle. *)
1392 (** Nodes and values. *)
1394 exception Error of string * Unix.error * string
1395 (** Error raised by a function.
1397 The first parameter is the name of the function which raised the error.
1398 The second parameter is the errno (see the [Unix] module). The third
1399 parameter is a human-readable string corresponding to the errno.
1401 See hivex(3) for a partial list of interesting errno values that
1402 can be generated by the library. *)
1403 exception Handle_closed of string
1404 (** This exception is raised if you call a function on a closed handle. *)
1410 fun (t, _, new_style, description) ->
1412 pr " | REG_%s (** %s *)\n" new_style description
1416 | REG_UNKNOWN of int32 (** unknown type *)
1417 (** Hive type field. *)
1423 fun (v, flag, description) ->
1424 assert (1 lsl i = v);
1425 pr " | OPEN_%s (** %s *)\n" flag description
1429 (** Open flags for {!open_file} call. *)
1436 (** (key, value) pair passed (as an array) to {!node_set_values}. *)
1440 fun (name, style, shortdesc, _) ->
1442 generate_ocaml_prototype name style;
1443 pr "(** %s *)\n" shortdesc
1446 and generate_ocaml_implementation () =
1447 generate_header OCamlStyle LGPLv2plus;
1454 exception Error of string * Unix.error * string
1455 exception Handle_closed of string
1457 (* Give the exceptions names, so they can be raised from the C code. *)
1459 Callback.register_exception \"ocaml_hivex_error\"
1460 (Error (\"\", Unix.EUNKNOWNERR 0, \"\"));
1461 Callback.register_exception \"ocaml_hivex_closed\" (Handle_closed \"\")
1467 fun (t, _, new_style, _) ->
1469 pr " | REG_%s\n" new_style
1473 | REG_UNKNOWN of int32
1479 fun (v, flag, description) ->
1480 assert (1 lsl i = v);
1481 pr " | OPEN_%s (** %s *)\n" flag description
1495 fun (name, style, _, _) ->
1496 generate_ocaml_prototype ~is_external:true name style
1499 and generate_ocaml_prototype ?(is_external = false) name style =
1500 let ocaml_name = if name = "open" then "open_file" else name in
1502 if is_external then pr "external " else pr "val ";
1503 pr "%s : " ocaml_name;
1506 | AHive -> pr "t -> "
1507 | ANode _ -> pr "node -> "
1508 | AValue _ -> pr "value -> "
1509 | AString _ -> pr "string -> "
1510 | AStringNullable _ -> pr "string option -> "
1511 | AOpenFlags -> pr "open_flag list -> "
1512 | AUnusedFlags -> ()
1513 | ASetValues -> pr "set_value array -> "
1514 | ASetValue -> pr "set_value -> "
1516 (match fst style with
1517 | RErr -> pr "unit" (* all errors are turned into exceptions *)
1518 | RErrDispose -> pr "unit"
1520 | RNode -> pr "node"
1521 | RNodeNotFound -> pr "node"
1522 | RNodeList -> pr "node array"
1523 | RValue -> pr "value"
1524 | RValueList -> pr "value array"
1525 | RString -> pr "string"
1526 | RStringList -> pr "string array"
1527 | RLenType -> pr "hive_type * int"
1528 | RLenTypeVal -> pr "hive_type * string"
1529 | RInt32 -> pr "int32"
1530 | RInt64 -> pr "int64"
1533 pr " = \"ocaml_hivex_%s\"" name;
1536 and generate_ocaml_c () =
1537 generate_header CStyle LGPLv2plus;
1548 #include <caml/config.h>
1549 #include <caml/alloc.h>
1550 #include <caml/callback.h>
1551 #include <caml/custom.h>
1552 #include <caml/fail.h>
1553 #include <caml/memory.h>
1554 #include <caml/mlvalues.h>
1555 #include <caml/signals.h>
1557 #ifdef HAVE_CAML_UNIXSUPPORT_H
1558 #include <caml/unixsupport.h>
1560 extern value unix_error_of_code (int errcode);
1563 #ifndef HAVE_CAML_RAISE_WITH_ARGS
1565 caml_raise_with_args (value tag, int nargs, value args[])
1568 CAMLxparamN (args, nargs);
1572 bucket = caml_alloc_small (1 + nargs, 0);
1573 Field(bucket, 0) = tag;
1574 for (i = 0; i < nargs; i++) Field(bucket, 1 + i) = args[i];
1582 #define Hiveh_val(v) (*((hive_h **)Data_custom_val(v)))
1583 static value Val_hiveh (hive_h *);
1584 static int HiveOpenFlags_val (value);
1585 static hive_set_value *HiveSetValue_val (value);
1586 static hive_set_value *HiveSetValues_val (value);
1587 static hive_type HiveType_val (value);
1588 static value Val_hive_type (hive_type);
1589 static value copy_int_array (size_t *);
1590 static value copy_type_len (size_t, hive_type);
1591 static value copy_type_value (const char *, size_t, hive_type);
1592 static void raise_error (const char *) Noreturn;
1593 static void raise_closed (const char *) Noreturn;
1599 fun (name, style, _, _) ->
1600 pr "/* Automatically generated wrapper for function\n";
1601 pr " * "; generate_ocaml_prototype name style;
1607 | ASetValues -> ["nrvalues"; "values"]
1608 | AUnusedFlags -> ["0"]
1609 | arg -> [name_of_argt arg]) (snd style) in
1611 match fst style with
1612 | RLenType | RLenTypeVal -> c_params @ [["&t"; "&len"]]
1614 let c_params = List.concat c_params in
1617 filter_map (function
1618 | AUnusedFlags -> None
1619 | arg -> Some (name_of_argt arg ^ "v")) (snd style) in
1621 pr "/* Emit prototype to appease gcc's -Wmissing-prototypes. */\n";
1622 pr "CAMLprim value ocaml_hivex_%s (value %s" name (List.hd params);
1623 List.iter (pr ", value %s") (List.tl params); pr ");\n";
1626 pr "CAMLprim value\n";
1627 pr "ocaml_hivex_%s (value %s" name (List.hd params);
1628 List.iter (pr ", value %s") (List.tl params);
1632 pr " CAMLparam%d (%s);\n"
1633 (List.length params) (String.concat ", " params);
1634 pr " CAMLlocal1 (rv);\n";
1640 pr " hive_h *h = Hiveh_val (hv);\n";
1641 pr " if (h == NULL)\n";
1642 pr " raise_closed (\"%s\");\n" name
1644 pr " hive_node_h %s = Int_val (%sv);\n" n n
1646 pr " hive_value_h %s = Int_val (%sv);\n" n n
1648 pr " const char *%s = String_val (%sv);\n" n n
1649 | AStringNullable n ->
1650 pr " const char *%s =\n" n;
1651 pr " %sv != Val_int (0) ? String_val (Field (%sv, 0)) : NULL;\n"
1654 pr " int flags = HiveOpenFlags_val (flagsv);\n"
1655 | AUnusedFlags -> ()
1657 pr " int nrvalues = Wosize_val (valuesv);\n";
1658 pr " hive_set_value *values = HiveSetValues_val (valuesv);\n"
1660 pr " hive_set_value *val = HiveSetValue_val (valv);\n"
1665 match fst style with
1666 | RErr -> pr " int r;\n"; "-1"
1667 | RErrDispose -> pr " int r;\n"; "-1"
1668 | RHive -> pr " hive_h *r;\n"; "NULL"
1669 | RNode -> pr " hive_node_h r;\n"; "0"
1672 pr " hive_node_h r;\n";
1674 | RNodeList -> pr " hive_node_h *r;\n"; "NULL"
1675 | RValue -> pr " hive_value_h r;\n"; "0"
1676 | RValueList -> pr " hive_value_h *r;\n"; "NULL"
1677 | RString -> pr " char *r;\n"; "NULL"
1678 | RStringList -> pr " char **r;\n"; "NULL"
1681 pr " size_t len;\n";
1682 pr " hive_type t;\n";
1686 pr " size_t len;\n";
1687 pr " hive_type t;\n";
1696 "-1 && errno != 0" in
1698 (* The libguestfs OCaml bindings call enter_blocking_section
1699 * here. However I don't think that is safe, because we are
1700 * holding pointers to caml strings during the call, and these
1701 * could be moved or freed by other threads. In any case, there
1702 * is very little reason to enter_blocking_section for any hivex
1703 * call, so don't do it. XXX
1705 (*pr " caml_enter_blocking_section ();\n";*)
1706 pr " r = hivex_%s (%s" name (List.hd c_params);
1707 List.iter (pr ", %s") (List.tl c_params);
1709 (*pr " caml_leave_blocking_section ();\n";*)
1712 (* Dispose of the hive handle (even if hivex_close returns error). *)
1713 (match fst style with
1715 pr " /* So we don't double-free in the finalizer. */\n";
1716 pr " Hiveh_val (hv) = NULL;\n";
1723 | AHive | ANode _ | AValue _ | AString _ | AStringNullable _
1724 | AOpenFlags | AUnusedFlags -> ()
1726 pr " free (values);\n";
1729 pr " free (val);\n";
1733 (* Check for errors. *)
1734 pr " if (r == %s)\n" error_code;
1735 pr " raise_error (\"%s\");\n" name;
1738 (match fst style with
1739 | RErr -> pr " rv = Val_unit;\n"
1740 | RErrDispose -> pr " rv = Val_unit;\n"
1741 | RHive -> pr " rv = Val_hiveh (r);\n"
1742 | RNode -> pr " rv = Val_int (r);\n"
1744 pr " if (r == 0)\n";
1745 pr " caml_raise_not_found ();\n";
1747 pr " rv = Val_int (r);\n"
1749 pr " rv = copy_int_array (r);\n";
1751 | RValue -> pr " rv = Val_int (r);\n"
1753 pr " rv = copy_int_array (r);\n";
1756 pr " rv = caml_copy_string (r);\n";
1759 pr " rv = caml_copy_string_array ((const char **) r);\n";
1760 pr " for (int i = 0; r[i] != NULL; ++i) free (r[i]);\n";
1762 | RLenType -> pr " rv = copy_type_len (len, t);\n"
1764 pr " rv = copy_type_value (r, len, t);\n";
1766 | RInt32 -> pr " rv = caml_copy_int32 (r);\n"
1767 | RInt64 -> pr " rv = caml_copy_int32 (r);\n"
1770 pr " CAMLreturn (rv);\n";
1778 HiveOpenFlags_val (value v)
1783 while (v != Val_int (0)) {
1785 flags |= 1 << Int_val (v2);
1792 static hive_set_value *
1793 HiveSetValue_val (value v)
1795 hive_set_value *val = malloc (sizeof (hive_set_value));
1797 val->key = String_val (Field (v, 0));
1798 val->t = HiveType_val (Field (v, 1));
1799 val->len = caml_string_length (Field (v, 2));
1800 val->value = String_val (Field (v, 2));
1805 static hive_set_value *
1806 HiveSetValues_val (value v)
1808 size_t nr_values = Wosize_val (v);
1809 hive_set_value *values = malloc (nr_values * sizeof (hive_set_value));
1813 for (i = 0; i < nr_values; ++i) {
1815 values[i].key = String_val (Field (v2, 0));
1816 values[i].t = HiveType_val (Field (v2, 1));
1817 values[i].len = caml_string_length (Field (v2, 2));
1818 values[i].value = String_val (Field (v2, 2));
1825 HiveType_val (value v)
1828 return Int_val (v); /* REG_NONE etc. */
1830 return Int32_val (Field (v, 0)); /* REG_UNKNOWN of int32 */
1834 Val_hive_type (hive_type t)
1840 CAMLreturn (Val_int (t));
1842 rv = caml_alloc (1, 0); /* REG_UNKNOWN of int32 */
1843 v = caml_copy_int32 (t);
1844 caml_modify (&Field (rv, 0), v);
1850 copy_int_array (size_t *xs)
1856 for (nr = 0; xs[nr] != 0; ++nr)
1859 CAMLreturn (Atom (0));
1861 rv = caml_alloc (nr, 0);
1862 for (i = 0; i < nr; ++i) {
1863 v = Val_int (xs[i]);
1864 Store_field (rv, i, v); /* Safe because v is not a block. */
1871 copy_type_len (size_t len, hive_type t)
1876 rv = caml_alloc (2, 0);
1877 v = Val_hive_type (t);
1878 Store_field (rv, 0, v);
1880 Store_field (rv, 1, len);
1885 copy_type_value (const char *r, size_t len, hive_type t)
1890 rv = caml_alloc (2, 0);
1891 v = Val_hive_type (t);
1892 Store_field (rv, 0, v);
1893 v = caml_alloc_string (len);
1894 memcpy (String_val (v), r, len);
1895 caml_modify (&Field (rv, 1), v);
1899 /* Raise exceptions. */
1901 raise_error (const char *function)
1903 /* Save errno early in case it gets trashed. */
1907 CAMLlocal3 (v1, v2, v3);
1909 v1 = caml_copy_string (function);
1910 v2 = unix_error_of_code (err);
1911 v3 = caml_copy_string (strerror (err));
1912 value vvv[] = { v1, v2, v3 };
1913 caml_raise_with_args (*caml_named_value (\"ocaml_hivex_error\"), 3, vvv);
1919 raise_closed (const char *function)
1924 v = caml_copy_string (function);
1925 caml_raise_with_arg (*caml_named_value (\"ocaml_hivex_closed\"), v);
1930 /* Allocate handles and deal with finalization. */
1932 hivex_finalize (value hv)
1934 hive_h *h = Hiveh_val (hv);
1935 if (h) hivex_close (h);
1938 static struct custom_operations hivex_custom_operations = {
1939 (char *) \"hivex_custom_operations\",
1941 custom_compare_default,
1942 custom_hash_default,
1943 custom_serialize_default,
1944 custom_deserialize_default
1948 Val_hiveh (hive_h *h)
1953 rv = caml_alloc_custom (&hivex_custom_operations,
1954 sizeof (hive_h *), 0, 1);
1961 and generate_perl_pm () =
1962 generate_header HashStyle LGPLv2plus;
1969 Win::Hivex - Perl bindings for reading and writing Windows Registry hive files
1975 $h = Win::Hivex->open ('SOFTWARE');
1976 $root_node = $h->root ();
1977 print $h->node_name ($root_node);
1981 The C<Win::Hivex> module provides a Perl XS binding to the
1982 L<hivex(3)> API for reading and writing Windows Registry binary
1987 All errors turn into calls to C<croak> (see L<Carp(3)>).
2001 XSLoader::load ('Win::Hivex');
2005 $h = Win::Hivex->open ($filename,";
2009 pr "\n [%s => 1,]" (String.lowercase flag)
2014 Open a Windows Registry binary hive file.
2016 The C<verbose> and C<debug> flags enable different levels of
2019 The C<write> flag is required if you will be modifying the
2020 hive file (see L<hivex(3)/WRITING TO HIVE FILES>).
2022 This function returns a hive handle. The hive handle is
2023 closed automatically when its reference count drops to 0.
2029 my $class = ref ($proto) || $proto;
2030 my $filename = shift;
2037 fun (n, flag, description) ->
2038 pr " # %s\n" description;
2039 pr " $flags += %d if $flags{%s};\n" n (String.lowercase flag)
2044 my $self = Win::Hivex::_open ($filename, $flags);
2045 bless $self, $class;
2052 fun (name, style, _, longdesc) ->
2053 (* The close call isn't explicit in Perl: handles are closed
2054 * when their reference count drops to 0.
2056 * The open call is coded specially in Perl.
2058 * Therefore we don't generate prototypes for these two calls:
2060 if fst style <> RErrDispose && List.hd (snd style) = AHive then (
2061 let longdesc = replace_str longdesc "C<hivex_" "C<" in
2062 pr "=item %s\n\n " name;
2063 generate_perl_prototype name style;
2065 pr "%s\n\n" longdesc;
2067 (match fst style with
2079 This returns a node handle.\n\n"
2082 This returns a node handle, or C<undef> if the node was not found.\n\n"
2085 This returns a list of node handles.\n\n"
2088 This returns a value handle.\n\n"
2091 This returns a list of value handles.\n\n"
2094 if List.mem ASetValues (snd style) then
2095 pr "C<@values> is an array of (keys, value) pairs.
2096 Each element should be a hashref containing C<key>, C<t> (type)
2099 Any existing values stored at the node are discarded, and their
2100 C<value> handles become invalid. Thus you can remove all
2101 values stored at C<node> by passing C<@values = []>.\n\n"
2114 Copyright (C) %s Red Hat Inc.
2118 Please see the file COPYING.LIB for the full license.
2124 L<http://libguestfs.org>,
2130 and generate_perl_prototype name style =
2132 (match fst style with
2135 | RHive -> pr "$h = "
2137 | RNodeNotFound -> pr "$node = "
2138 | RNodeList -> pr "@nodes = "
2139 | RValue -> pr "$value = "
2140 | RValueList -> pr "@values = "
2141 | RString -> pr "$string = "
2142 | RStringList -> pr "@strings = "
2143 | RLenType -> pr "($type, $len) = "
2144 | RLenTypeVal -> pr "($type, $data) = "
2145 | RInt32 -> pr "$int32 = "
2146 | RInt64 -> pr "$int64 = "
2149 let args = List.tl (snd style) in
2151 (* AUnusedFlags is dropped in the bindings. *)
2152 let args = List.filter ((<>) AUnusedFlags) args in
2156 let comma = ref false in
2159 if !comma then pr ", "; comma := true;
2164 | AString n -> pr "$%s" n
2165 | AStringNullable n -> pr "[$%s|undef]" n
2166 | AOpenFlags -> pr "[flags]"
2167 | AUnusedFlags -> assert false
2168 | ASetValues -> pr "\\@values"
2169 | ASetValue -> pr "$val"
2174 and generate_perl_xs () =
2175 generate_header CStyle LGPLv2plus;
2178 #include \"EXTERN.h\"
2186 #define PRId64 \"lld\"
2190 my_newSVll(long long val) {
2191 #ifdef USE_64_BIT_ALL
2192 return newSViv(val);
2196 len = snprintf(buf, 100, \"%%\" PRId64, val);
2197 return newSVpv(buf, len);
2202 #define PRIu64 \"llu\"
2207 my_newSVull(unsigned long long val) {
2208 #ifdef USE_64_BIT_ALL
2209 return newSVuv(val);
2213 len = snprintf(buf, 100, \"%%\" PRIu64, val);
2214 return newSVpv(buf, len);
2220 /* http://www.perlmonks.org/?node_id=680842 */
2222 XS_unpack_charPtrPtr (SV *arg) {
2227 if (!arg || !SvOK (arg) || !SvROK (arg) || SvTYPE (SvRV (arg)) != SVt_PVAV)
2228 croak (\"array reference expected\");
2230 av = (AV *)SvRV (arg);
2231 ret = malloc ((av_len (av) + 1 + 1) * sizeof (char *));
2233 croak (\"malloc failed\");
2235 for (i = 0; i <= av_len (av); i++) {
2236 SV **elem = av_fetch (av, i, 0);
2238 if (!elem || !*elem)
2239 croak (\"missing element in list\");
2241 ret[i] = SvPV_nolen (*elem);
2250 /* Handle set_values parameter. */
2251 typedef struct pl_set_values {
2253 hive_set_value *values;
2256 static pl_set_values
2257 unpack_pl_set_values (SV *sv)
2263 if (!sv || !SvOK (sv) || !SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVAV)
2264 croak (\"array reference expected\");
2266 av = (AV *)SvRV(sv);
2267 ret.nr_values = av_len (av) + 1;
2268 ret.values = malloc (ret.nr_values * sizeof (hive_set_value));
2270 croak (\"malloc failed\");
2272 for (i = 0; i <= av_len (av); i++) {
2273 SV **hvp = av_fetch (av, i, 0);
2275 if (!hvp || !*hvp || !SvROK (*hvp) || SvTYPE (SvRV (*hvp)) != SVt_PVHV)
2276 croak (\"missing element in list or not a hash ref\");
2278 HV *hv = (HV *)SvRV(*hvp);
2281 svp = hv_fetch (hv, \"key\", 3, 0);
2283 croak (\"missing 'key' in hash\");
2284 ret.values[i].key = SvPV_nolen (*svp);
2286 svp = hv_fetch (hv, \"t\", 1, 0);
2288 croak (\"missing 't' in hash\");
2289 ret.values[i].t = SvIV (*svp);
2291 svp = hv_fetch (hv, \"value\", 5, 0);
2293 croak (\"missing 'value' in hash\");
2294 ret.values[i].value = SvPV (*svp, ret.values[i].len);
2300 static hive_set_value *
2301 unpack_set_value (SV *sv)
2303 hive_set_value *ret;
2305 if (!sv || !SvROK (sv) || SvTYPE (SvRV (sv)) != SVt_PVHV)
2306 croak (\"not a hash ref\");
2308 ret = malloc (sizeof (hive_set_value));
2310 croak (\"malloc failed\");
2312 HV *hv = (HV *)SvRV(sv);
2315 svp = hv_fetch (hv, \"key\", 3, 0);
2317 croak (\"missing 'key' in hash\");
2318 ret->key = SvPV_nolen (*svp);
2320 svp = hv_fetch (hv, \"t\", 1, 0);
2322 croak (\"missing 't' in hash\");
2323 ret->t = SvIV (*svp);
2325 svp = hv_fetch (hv, \"value\", 5, 0);
2327 croak (\"missing 'value' in hash\");
2328 ret->value = SvPV (*svp, ret->len);
2333 MODULE = Win::Hivex PACKAGE = Win::Hivex
2338 _open (filename, flags)
2342 RETVAL = hivex_open (filename, flags);
2344 croak (\"hivex_open: %%s: %%s\", filename, strerror (errno));
2352 if (hivex_close (h) == -1)
2353 croak (\"hivex_close: %%s\", strerror (errno));
2358 fun (name, style, _, longdesc) ->
2359 (* The close and open calls are handled specially above. *)
2360 if fst style <> RErrDispose && List.hd (snd style) = AHive then (
2361 (match fst style with
2362 | RErr -> pr "void\n"
2363 | RErrDispose -> failwith "perl bindings cannot handle a call which disposes of the handle"
2364 | RHive -> failwith "perl bindings cannot handle a call which returns a handle"
2368 | RString -> pr "SV *\n"
2373 | RLenTypeVal -> pr "void\n"
2374 | RInt32 -> pr "SV *\n"
2375 | RInt64 -> pr "SV *\n"
2378 (* Call and arguments. *)
2380 filter_map (function
2381 | AUnusedFlags -> None
2382 | arg -> Some (name_of_argt arg)) (snd style) in
2386 | AUnusedFlags -> "0"
2387 | ASetValues -> "values.nr_values, values.values"
2388 | arg -> name_of_argt arg) (snd style) in
2390 pr "%s (%s)\n" name (String.concat ", " perl_params);
2401 | AStringNullable n ->
2402 (* http://www.perlmonks.org/?node_id=554277 *)
2403 pr " char *%s = SvOK(ST(%d)) ? SvPV_nolen(ST(%d)) : NULL;\n" n i i
2406 | AUnusedFlags -> ()
2408 pr " pl_set_values values = unpack_pl_set_values (ST(%d));\n" i
2410 pr " hive_set_value *val = unpack_set_value (ST(%d));\n" i
2417 pr " free (values.values);\n"
2420 | AHive | ANode _ | AValue _ | AString _ | AStringNullable _
2421 | AOpenFlags | AUnusedFlags -> ()
2426 (match fst style with
2431 pr " r = hivex_%s (%s);\n"
2432 name (String.concat ", " c_params);
2434 pr " if (r == -1)\n";
2435 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2438 | RErrDispose -> assert false
2439 | RHive -> assert false
2444 pr " /* hive_node_h = hive_value_h = size_t so we cheat\n";
2445 pr " here to simplify the generator */\n";
2448 pr " r = hivex_%s (%s);\n"
2449 name (String.concat ", " c_params);
2451 pr " if (r == 0)\n";
2452 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2454 pr " RETVAL = newSViv (r);\n";
2460 pr " hive_node_h r;\n";
2463 pr " r = hivex_%s (%s);\n"
2464 name (String.concat ", " c_params);
2466 pr " if (r == 0 && errno != 0)\n";
2467 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2469 pr " if (r == 0)\n";
2470 pr " RETVAL = &PL_sv_undef;\n";
2472 pr " RETVAL = newSViv (r);\n";
2480 pr " r = hivex_%s (%s);\n"
2481 name (String.concat ", " c_params);
2483 pr " if (r == NULL)\n";
2484 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2486 pr " RETVAL = newSVpv (r, 0);\n";
2497 pr " r = hivex_%s (%s);\n"
2498 name (String.concat ", " c_params);
2500 pr " if (r == NULL)\n";
2501 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2503 pr " for (n = 0; r[n] != 0; ++n) /**/;\n";
2504 pr " EXTEND (SP, n);\n";
2505 pr " for (i = 0; i < n; ++i)\n";
2506 pr " PUSHs (sv_2mortal (newSViv (r[i])));\n";
2514 pr " r = hivex_%s (%s);\n"
2515 name (String.concat ", " c_params);
2517 pr " if (r == NULL)\n";
2518 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2520 pr " for (n = 0; r[n] != NULL; ++n) /**/;\n";
2521 pr " EXTEND (SP, n);\n";
2522 pr " for (i = 0; i < n; ++i) {\n";
2523 pr " PUSHs (sv_2mortal (newSVpv (r[i], 0)));\n";
2524 pr " free (r[i]);\n";
2531 pr " size_t len;\n";
2532 pr " hive_type type;\n";
2534 pr " r = hivex_%s (%s, &type, &len);\n"
2535 name (String.concat ", " c_params);
2537 pr " if (r == -1)\n";
2538 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2540 pr " EXTEND (SP, 2);\n";
2541 pr " PUSHs (sv_2mortal (newSViv (type)));\n";
2542 pr " PUSHs (sv_2mortal (newSViv (len)));\n";
2547 pr " size_t len;\n";
2548 pr " hive_type type;\n";
2550 pr " r = hivex_%s (%s, &type, &len);\n"
2551 name (String.concat ", " c_params);
2553 pr " if (r == NULL)\n";
2554 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2556 pr " EXTEND (SP, 2);\n";
2557 pr " PUSHs (sv_2mortal (newSViv (type)));\n";
2558 pr " PUSHs (sv_2mortal (newSVpvn (r, len)));\n";
2566 pr " r = hivex_%s (%s);\n"
2567 name (String.concat ", " c_params);
2569 pr " if (r == -1 && errno != 0)\n";
2570 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2572 pr " RETVAL = newSViv (r);\n";
2581 pr " r = hivex_%s (%s);\n"
2582 name (String.concat ", " c_params);
2584 pr " if (r == -1 && errno != 0)\n";
2585 pr " croak (\"%%s: %%s\", \"%s\", strerror (errno));\n"
2587 pr " RETVAL = my_newSVll (r);\n";
2595 and generate_python_c () =
2596 generate_header CStyle LGPLv2plus;
2599 #define PY_SSIZE_T_CLEAN 1
2602 #if PY_VERSION_HEX < 0x02050000
2603 typedef int Py_ssize_t;
2604 #define PY_SSIZE_T_MAX INT_MAX
2605 #define PY_SSIZE_T_MIN INT_MIN
2612 #include \"hivex.h\"
2614 #ifndef HAVE_PYCAPSULE_NEW
2622 get_handle (PyObject *obj)
2625 assert (obj != Py_None);
2626 #ifndef HAVE_PYCAPSULE_NEW
2627 return ((Pyhivex_Object *) obj)->h;
2629 return (hive_h *) PyCapsule_GetPointer(obj, \"hive_h\");
2634 put_handle (hive_h *h)
2637 #ifndef HAVE_PYCAPSULE_NEW
2639 PyCObject_FromVoidPtrAndDesc ((void *) h, (char *) \"hive_h\", NULL);
2641 return PyCapsule_New ((void *) h, \"hive_h\", NULL);
2645 /* This returns pointers into the Python objects, which should
2649 get_value (PyObject *v, hive_set_value *ret)
2653 obj = PyDict_GetItemString (v, \"key\");
2655 PyErr_SetString (PyExc_RuntimeError, \"no 'key' element in dictionary\");
2658 if (!PyString_Check (obj)) {
2659 PyErr_SetString (PyExc_RuntimeError, \"'key' element is not a string\");
2662 ret->key = PyString_AsString (obj);
2664 obj = PyDict_GetItemString (v, \"t\");
2666 PyErr_SetString (PyExc_RuntimeError, \"no 't' element in dictionary\");
2669 if (!PyInt_Check (obj)) {
2670 PyErr_SetString (PyExc_RuntimeError, \"'t' element is not an integer\");
2673 ret->t = PyInt_AsLong (obj);
2675 obj = PyDict_GetItemString (v, \"value\");
2677 PyErr_SetString (PyExc_RuntimeError, \"no 'value' element in dictionary\");
2680 if (!PyString_Check (obj)) {
2681 PyErr_SetString (PyExc_RuntimeError, \"'value' element is not a string\");
2684 ret->value = PyString_AsString (obj);
2685 ret->len = PyString_Size (obj);
2690 typedef struct py_set_values {
2692 hive_set_value *values;
2696 get_values (PyObject *v, py_set_values *ret)
2701 if (!PyList_Check (v)) {
2702 PyErr_SetString (PyExc_RuntimeError, \"expecting a list parameter\");
2706 slen = PyList_Size (v);
2708 PyErr_SetString (PyExc_RuntimeError, \"get_string_list: PyList_Size failure\");
2711 len = (size_t) slen;
2712 ret->nr_values = len;
2713 ret->values = malloc (len * sizeof (hive_set_value));
2715 PyErr_SetString (PyExc_RuntimeError, strerror (errno));
2719 for (i = 0; i < len; ++i) {
2720 if (get_value (PyList_GetItem (v, i), &(ret->values[i])) == -1) {
2730 put_string_list (char * const * const argv)
2735 for (argc = 0; argv[argc] != NULL; ++argc)
2738 list = PyList_New (argc);
2739 for (i = 0; i < argc; ++i)
2740 PyList_SetItem (list, i, PyString_FromString (argv[i]));
2746 free_strings (char **argv)
2750 for (argc = 0; argv[argc] != NULL; ++argc)
2755 /* Since hive_node_t is the same as hive_value_t this also works for values. */
2757 put_node_list (hive_node_h *nodes)
2762 for (argc = 0; nodes[argc] != 0; ++argc)
2765 list = PyList_New (argc);
2766 for (i = 0; i < argc; ++i)
2767 PyList_SetItem (list, i, PyLong_FromLongLong ((long) nodes[i]));
2773 put_len_type (size_t len, hive_type t)
2775 PyObject *r = PyTuple_New (2);
2776 PyTuple_SetItem (r, 0, PyInt_FromLong ((long) t));
2777 PyTuple_SetItem (r, 1, PyLong_FromLongLong ((long) len));
2782 put_val_type (char *val, size_t len, hive_type t)
2784 PyObject *r = PyTuple_New (2);
2785 PyTuple_SetItem (r, 0, PyInt_FromLong ((long) t));
2786 PyTuple_SetItem (r, 1, PyString_FromStringAndSize (val, len));
2792 (* Generate functions. *)
2794 fun (name, style, _, longdesc) ->
2795 pr "static PyObject *\n";
2796 pr "py_hivex_%s (PyObject *self, PyObject *args)\n" name;
2798 pr " PyObject *py_r;\n";
2801 match fst style with
2802 | RErr -> pr " int r;\n"; "-1"
2803 | RErrDispose -> pr " int r;\n"; "-1"
2804 | RHive -> pr " hive_h *r;\n"; "NULL"
2805 | RNode -> pr " hive_node_h r;\n"; "0"
2808 pr " hive_node_h r;\n";
2810 | RNodeList -> pr " hive_node_h *r;\n"; "NULL"
2811 | RValue -> pr " hive_value_h r;\n"; "0"
2812 | RValueList -> pr " hive_value_h *r;\n"; "NULL"
2813 | RString -> pr " char *r;\n"; "NULL"
2814 | RStringList -> pr " char **r;\n"; "NULL"
2817 pr " size_t len;\n";
2818 pr " hive_type t;\n";
2822 pr " size_t len;\n";
2823 pr " hive_type t;\n";
2832 "-1 && errno != 0" in
2834 (* Call and arguments. *)
2837 | AUnusedFlags -> "0"
2838 | ASetValues -> "values.nr_values, values.values"
2839 | ASetValue -> "&val"
2840 | arg -> name_of_argt arg) (snd style) in
2842 match fst style with
2843 | RLenType | RLenTypeVal -> c_params @ ["&t"; "&len"]
2850 pr " PyObject *py_h;\n"
2855 | AStringNullable n ->
2859 | AUnusedFlags -> ()
2861 pr " py_set_values values;\n";
2862 pr " PyObject *py_values;\n"
2864 pr " hive_set_value val;\n";
2865 pr " PyObject *py_val;\n"
2870 (* Convert the required parameters. *)
2871 pr " if (!PyArg_ParseTuple (args, (char *) \"";
2881 | AStringNullable n ->
2885 | AUnusedFlags -> ()
2891 pr ":hivex_%s\"" name;
2901 | AStringNullable n ->
2905 | AUnusedFlags -> ()
2913 pr " return NULL;\n";
2915 (* Convert some Python argument types to C. *)
2919 pr " h = get_handle (py_h);\n"
2925 | AUnusedFlags -> ()
2927 pr " if (get_values (py_values, &values) == -1)\n";
2928 pr " return NULL;\n"
2930 pr " if (get_value (py_val, &val) == -1)\n";
2931 pr " return NULL;\n"
2934 (* Call the C function. *)
2935 pr " r = hivex_%s (%s);\n" name (String.concat ", " c_params);
2937 (* Free up arguments. *)
2940 | AHive | ANode _ | AValue _
2941 | AString _ | AStringNullable _
2942 | AOpenFlags | AUnusedFlags -> ()
2944 pr " free (values.values);\n"
2948 (* Check for errors from C library. *)
2949 pr " if (r == %s) {\n" error_code;
2950 pr " PyErr_SetString (PyExc_RuntimeError,\n";
2951 pr " strerror (errno));\n";
2952 pr " return NULL;\n";
2956 (* Convert return value to Python. *)
2957 (match fst style with
2960 pr " Py_INCREF (Py_None);\n";
2961 pr " py_r = Py_None;\n"
2963 pr " py_r = put_handle (r);\n"
2965 pr " py_r = PyLong_FromLongLong (r);\n"
2968 pr " py_r = PyLong_FromLongLong (r);\n";
2970 pr " Py_INCREF (Py_None);\n";
2971 pr " py_r = Py_None;\n";
2975 pr " py_r = put_node_list (r);\n";
2978 pr " py_r = PyLong_FromLongLong (r);\n"
2980 pr " py_r = PyString_FromString (r);\n";
2983 pr " py_r = put_string_list (r);\n";
2984 pr " free_strings (r);\n"
2986 pr " py_r = put_len_type (len, t);\n"
2988 pr " py_r = put_val_type (r, len, t);\n";
2991 pr " py_r = PyInt_FromLong ((long) r);\n"
2993 pr " py_r = PyLong_FromLongLong (r);\n"
2995 pr " return py_r;\n";
3000 (* Table of functions. *)
3001 pr "static PyMethodDef methods[] = {\n";
3003 fun (name, _, _, _) ->
3004 pr " { (char *) \"%s\", py_hivex_%s, METH_VARARGS, NULL },\n"
3007 pr " { NULL, NULL, 0, NULL }\n";
3011 (* Init function. *)
3014 initlibhivexmod (void)
3016 static int initialized = 0;
3018 if (initialized) return;
3019 Py_InitModule ((char *) \"libhivexmod\", methods);
3024 and generate_python_py () =
3025 generate_header HashStyle LGPLv2plus;
3028 u\"\"\"Python bindings for hivex
3031 h = hivex.Hivex (filename)
3033 The hivex module provides Python bindings to the hivex API for
3034 examining and modifying Windows Registry 'hive' files.
3036 Read the hivex(3) man page to find out how to use the API.
3042 \"\"\"Instances of this class are hivex API handles.\"\"\"
3044 def __init__ (self, filename";
3047 fun (_, flag, _) -> pr ", %s = False" (String.lowercase flag)
3051 \"\"\"Create a new hivex handle.\"\"\"
3056 fun (n, flag, description) ->
3057 pr " # %s\n" description;
3058 pr " if %s: flags += %d\n" (String.lowercase flag) n
3061 pr " self._o = libhivexmod.open (filename, flags)
3064 libhivexmod.close (self._o)
3069 fun (name, style, shortdesc, _) ->
3070 (* The close and open calls are handled specially above. *)
3071 if fst style <> RErrDispose && List.hd (snd style) = AHive then (
3072 let args = List.tl (snd style) in
3073 let args = List.filter (
3074 function AOpenFlags | AUnusedFlags -> false
3078 pr " def %s (self" name;
3079 List.iter (fun arg -> pr ", %s" (name_of_argt arg)) args;
3081 pr " u\"\"\"%s\"\"\"\n" shortdesc;
3082 pr " return libhivexmod.%s (self._o" name;
3087 | AHive -> assert false
3088 | ANode n | AValue n
3089 | AString n | AStringNullable n -> pr "%s" n
3091 | AUnusedFlags -> assert false
3092 | ASetValues -> pr "values"
3093 | ASetValue -> pr "val"
3100 let output_to filename k =
3101 let filename_new = filename ^ ".new" in
3102 chan := open_out filename_new;
3105 chan := Pervasives.stdout;
3107 (* Is the new file different from the current file? *)
3108 if Sys.file_exists filename && files_equal filename filename_new then
3109 unlink filename_new (* same, so skip it *)
3111 (* different, overwrite old one *)
3112 (try chmod filename 0o644 with Unix_error _ -> ());
3113 rename filename_new filename;
3114 chmod filename 0o444;
3115 printf "written %s\n%!" filename;
3118 let perror msg = function
3119 | Unix_error (err, _, _) ->
3120 eprintf "%s: %s\n" msg (error_message err)
3122 eprintf "%s: %s\n" msg (Printexc.to_string exn)
3127 try openfile "configure.ac" [O_RDWR] 0
3129 | Unix_error (ENOENT, _, _) ->
3131 You are probably running this from the wrong directory.
3132 Run it from the top source directory using the command
3133 generator/generator.ml
3137 perror "open: configure.ac" exn;
3140 (* Acquire a lock so parallel builds won't try to run the generator
3141 * twice at the same time. Subsequent builds will wait for the first
3142 * one to finish. Note the lock is released implicitly when the
3145 (try lockf lock_fd F_LOCK 1
3147 perror "lock: configure.ac" exn;
3152 output_to "lib/hivex.h" generate_c_header;
3153 output_to "lib/hivex.pod" generate_c_pod;
3155 output_to "lib/hivex.syms" generate_linker_script;
3157 output_to "ocaml/hivex.mli" generate_ocaml_interface;
3158 output_to "ocaml/hivex.ml" generate_ocaml_implementation;
3159 output_to "ocaml/hivex_c.c" generate_ocaml_c;
3161 output_to "perl/lib/Win/Hivex.pm" generate_perl_pm;
3162 output_to "perl/Hivex.xs" generate_perl_xs;
3164 output_to "python/hivex.py" generate_python_py;
3165 output_to "python/hivex-py.c" generate_python_c;
3167 (* Always generate this file last, and unconditionally. It's used
3168 * by the Makefile to know when we must re-run the generator.
3170 let chan = open_out "generator/stamp-generator" in
3174 printf "generated %d lines of code\n" !lines