Commit | Line | Data |
---|---|---|
0f41c26f | 1 | ;;; GNU Guix --- Functional package management for GNU |
0d268c5d | 2 | ;;; Copyright © 2012, 2013, 2014, 2015, 2016, 2017 Ludovic Courtès <ludo@gnu.org> |
0f41c26f LC |
3 | ;;; |
4 | ;;; This file is part of GNU Guix. | |
5 | ;;; | |
6 | ;;; GNU Guix is free software; you can redistribute it and/or modify it | |
7 | ;;; under the terms of the GNU General Public License as published by | |
8 | ;;; the Free Software Foundation; either version 3 of the License, or (at | |
9 | ;;; your option) any later version. | |
10 | ;;; | |
11 | ;;; GNU Guix is distributed in the hope that it will be useful, but | |
12 | ;;; WITHOUT ANY WARRANTY; without even the implied warranty of | |
13 | ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
14 | ;;; GNU General Public License for more details. | |
15 | ;;; | |
16 | ;;; You should have received a copy of the GNU General Public License | |
17 | ;;; along with GNU Guix. If not, see <http://www.gnu.org/licenses/>. | |
18 | ||
19 | (define-module (guix serialization) | |
958dd3ce | 20 | #:use-module (guix combinators) |
0f41c26f | 21 | #:use-module (rnrs bytevectors) |
0f41c26f LC |
22 | #:use-module (srfi srfi-1) |
23 | #:use-module (srfi srfi-26) | |
0363991a LC |
24 | #:use-module (srfi srfi-34) |
25 | #:use-module (srfi srfi-35) | |
2535635f LC |
26 | #:use-module (ice-9 binary-ports) |
27 | #:use-module ((ice-9 rdelim) #:prefix rdelim:) | |
6c20d1d0 | 28 | #:use-module (ice-9 match) |
0363991a | 29 | #:use-module (ice-9 ftw) |
0f41c26f LC |
30 | #:export (write-int read-int |
31 | write-long-long read-long-long | |
32 | write-padding | |
0d268c5d | 33 | write-bytevector write-string |
ce72c780 | 34 | read-string read-latin1-string read-maybe-utf8-string |
0f41c26f | 35 | write-string-list read-string-list |
6c20d1d0 | 36 | write-string-pairs |
0f41c26f | 37 | write-store-path read-store-path |
0363991a LC |
38 | write-store-path-list read-store-path-list |
39 | ||
40 | &nar-error | |
41 | nar-error? | |
42 | nar-error-port | |
43 | nar-error-file | |
44 | ||
45 | &nar-read-error | |
46 | nar-read-error? | |
47 | nar-read-error-token | |
48 | ||
49 | write-file | |
50 | restore-file)) | |
0f41c26f LC |
51 | |
52 | ;;; Comment: | |
53 | ;;; | |
54 | ;;; Serialization procedures used by the RPCs and the Nar format. This module | |
55 | ;;; is for internal consumption. | |
56 | ;;; | |
57 | ;;; Code: | |
58 | ||
59 | ;; Similar to serialize.cc in Nix. | |
60 | ||
46b8aadb LC |
61 | (define-condition-type &nar-error &error ; XXX: inherit from &nix-error ? |
62 | nar-error? | |
63 | (file nar-error-file) ; file we were restoring, or #f | |
64 | (port nar-error-port)) ; port from which we read | |
65 | ||
66 | (define currently-restored-file | |
67 | ;; Name of the file being restored. Used internally for error reporting. | |
68 | (make-parameter #f)) | |
69 | ||
70 | ||
71 | (define (get-bytevector-n* port count) | |
72 | (let ((bv (get-bytevector-n port count))) | |
73 | (when (or (eof-object? bv) | |
74 | (< (bytevector-length bv) count)) | |
75 | (raise (condition (&nar-error | |
76 | (file (currently-restored-file)) | |
77 | (port port))))) | |
78 | bv)) | |
79 | ||
0f41c26f LC |
80 | (define (write-int n p) |
81 | (let ((b (make-bytevector 8 0))) | |
82 | (bytevector-u32-set! b 0 n (endianness little)) | |
83 | (put-bytevector p b))) | |
84 | ||
85 | (define (read-int p) | |
46b8aadb | 86 | (let ((b (get-bytevector-n* p 8))) |
0f41c26f LC |
87 | (bytevector-u32-ref b 0 (endianness little)))) |
88 | ||
89 | (define (write-long-long n p) | |
90 | (let ((b (make-bytevector 8 0))) | |
91 | (bytevector-u64-set! b 0 n (endianness little)) | |
92 | (put-bytevector p b))) | |
93 | ||
94 | (define (read-long-long p) | |
46b8aadb | 95 | (let ((b (get-bytevector-n* p 8))) |
0f41c26f LC |
96 | (bytevector-u64-ref b 0 (endianness little)))) |
97 | ||
98 | (define write-padding | |
99 | (let ((zero (make-bytevector 8 0))) | |
100 | (lambda (n p) | |
101 | (let ((m (modulo n 8))) | |
102 | (or (zero? m) | |
103 | (put-bytevector p zero 0 (- 8 m))))))) | |
104 | ||
0d268c5d LC |
105 | (define (write-bytevector s p) |
106 | (let* ((l (bytevector-length s)) | |
0f41c26f LC |
107 | (m (modulo l 8)) |
108 | (b (make-bytevector (+ 8 l (if (zero? m) 0 (- 8 m)))))) | |
109 | (bytevector-u32-set! b 0 l (endianness little)) | |
110 | (bytevector-copy! s 0 b 8 l) | |
111 | (put-bytevector p b))) | |
112 | ||
0d268c5d LC |
113 | (define (write-string s p) |
114 | (write-bytevector (string->utf8 s) p)) | |
115 | ||
472e4c43 | 116 | (define (read-byte-string p) |
0f41c26f LC |
117 | (let* ((len (read-int p)) |
118 | (m (modulo len 8)) | |
472e4c43 | 119 | (bv (get-bytevector-n* p len))) |
0f41c26f | 120 | (or (zero? m) |
46b8aadb | 121 | (get-bytevector-n* p (- 8 m))) |
472e4c43 | 122 | bv)) |
0f41c26f | 123 | |
472e4c43 LC |
124 | (define (read-string p) |
125 | (utf8->string (read-byte-string p))) | |
aa27b560 | 126 | |
472e4c43 LC |
127 | (define (read-latin1-string p) |
128 | "Read an ISO-8859-1 string from P." | |
129 | ;; Note: do not use 'get-string-n' to work around Guile bug | |
130 | ;; <http://bugs.gnu.org/19621>. See <http://bugs.gnu.org/19610> for | |
131 | ;; a discussion. | |
132 | (let ((bv (read-byte-string p))) | |
36626c55 | 133 | ;; XXX: Rewrite using (ice-9 iconv). |
472e4c43 | 134 | (list->string (map integer->char (bytevector->u8-list bv))))) |
0f41c26f | 135 | |
ce72c780 LC |
136 | (define (read-maybe-utf8-string p) |
137 | "Read a serialized string from port P. Attempt to decode it as UTF-8 and | |
138 | substitute invalid byte sequences with question marks. This is a | |
139 | \"permissive\" UTF-8 decoder." | |
140 | ;; XXX: We rely on the port's decoding mechanism to do permissive decoding | |
141 | ;; and substitute invalid byte sequences with question marks, but this is | |
142 | ;; not very efficient. Eventually Guile may provide a lightweight | |
143 | ;; permissive UTF-8 decoder. | |
144 | (let* ((bv (read-byte-string p)) | |
bc7d089a LC |
145 | (port (open-bytevector-input-port bv))) |
146 | (set-port-encoding! port "UTF-8") | |
147 | (set-port-conversion-strategy! port 'substitute) | |
2535635f | 148 | (rdelim:read-string port))) |
ce72c780 | 149 | |
0f41c26f LC |
150 | (define (write-string-list l p) |
151 | (write-int (length l) p) | |
152 | (for-each (cut write-string <> p) l)) | |
153 | ||
6c20d1d0 LC |
154 | (define (write-string-pairs l p) |
155 | (write-int (length l) p) | |
156 | (for-each (match-lambda | |
157 | ((first . second) | |
158 | (write-string first p) | |
159 | (write-string second p))) | |
160 | l)) | |
161 | ||
0f41c26f LC |
162 | (define (read-string-list p) |
163 | (let ((len (read-int p))) | |
164 | (unfold (cut >= <> len) | |
165 | (lambda (i) | |
166 | (read-string p)) | |
167 | 1+ | |
168 | 0))) | |
169 | ||
170 | (define (write-store-path f p) | |
171 | (write-string f p)) ; TODO: assert path | |
172 | ||
173 | (define (read-store-path p) | |
174 | (read-string p)) ; TODO: assert path | |
175 | ||
176 | (define write-store-path-list write-string-list) | |
177 | (define read-store-path-list read-string-list) | |
178 | ||
0363991a | 179 | \f |
0363991a LC |
180 | (define-condition-type &nar-read-error &nar-error |
181 | nar-read-error? | |
182 | (token nar-read-error-token)) ; faulty token, or #f | |
183 | ||
184 | ||
185 | (define (dump in out size) | |
186 | "Copy SIZE bytes from IN to OUT." | |
187 | (define buf-size 65536) | |
188 | (define buf (make-bytevector buf-size)) | |
189 | ||
190 | (let loop ((left size)) | |
191 | (if (<= left 0) | |
192 | 0 | |
193 | (let ((read (get-bytevector-n! in buf 0 (min left buf-size)))) | |
194 | (if (eof-object? read) | |
195 | left | |
196 | (begin | |
197 | (put-bytevector out buf 0 read) | |
198 | (loop (- left read)))))))) | |
199 | ||
200 | (define (write-contents file p size) | |
201 | "Write SIZE bytes from FILE to output port P." | |
202 | (define (call-with-binary-input-file file proc) | |
203 | ;; Open FILE as a binary file. This avoids scan-for-encoding, and thus | |
204 | ;; avoids any initial buffering. Disable file name canonicalization to | |
205 | ;; avoid stat'ing like crazy. | |
206 | (with-fluids ((%file-port-name-canonicalization #f)) | |
207 | (let ((port (open-file file "rb"))) | |
208 | (dynamic-wind | |
209 | (const #t) | |
210 | (cut proc port) | |
211 | (lambda () | |
212 | (close-port port)))))) | |
213 | ||
214 | (write-string "contents" p) | |
215 | (write-long-long size p) | |
216 | (call-with-binary-input-file file | |
217 | ;; Use `sendfile' when available (Guile 2.0.8+). | |
218 | (if (and (compile-time-value (defined? 'sendfile)) | |
219 | (file-port? p)) | |
220 | (cut sendfile p <> size 0) | |
221 | (cut dump <> p size))) | |
222 | (write-padding size p)) | |
223 | ||
224 | (define (read-contents in out) | |
225 | "Read the contents of a file from the Nar at IN, write it to OUT, and return | |
226 | the size in bytes." | |
227 | (define executable? | |
228 | (match (read-string in) | |
229 | ("contents" | |
230 | #f) | |
231 | ("executable" | |
232 | (match (list (read-string in) (read-string in)) | |
233 | (("" "contents") #t) | |
234 | (x (raise | |
235 | (condition (&message | |
236 | (message "unexpected executable file marker")) | |
237 | (&nar-read-error (port in) | |
238 | (file #f) | |
239 | (token x)))))) | |
240 | #t) | |
241 | (x | |
242 | (raise | |
243 | (condition (&message (message "unsupported nar file type")) | |
244 | (&nar-read-error (port in) (file #f) (token x))))))) | |
245 | ||
246 | (let ((size (read-long-long in))) | |
247 | ;; Note: `sendfile' cannot be used here because of port buffering on IN. | |
248 | (dump in out size) | |
249 | ||
250 | (when executable? | |
251 | (chmod out #o755)) | |
252 | (let ((m (modulo size 8))) | |
253 | (unless (zero? m) | |
46b8aadb | 254 | (get-bytevector-n* in (- 8 m)))) |
0363991a LC |
255 | size)) |
256 | ||
257 | (define %archive-version-1 | |
258 | ;; Magic cookie for Nix archives. | |
259 | "nix-archive-1") | |
260 | ||
fe585be9 LC |
261 | (define* (write-file file port |
262 | #:key (select? (const #t))) | |
0363991a | 263 | "Write the contents of FILE to PORT in Nar format, recursing into |
fe585be9 LC |
264 | sub-directories of FILE as needed. For each directory entry, call (SELECT? |
265 | FILE STAT), where FILE is the entry's absolute file name and STAT is the | |
266 | result of 'lstat'; exclude entries for which SELECT? does not return true." | |
0363991a LC |
267 | (define p port) |
268 | ||
269 | (write-string %archive-version-1 p) | |
270 | ||
fe585be9 LC |
271 | (let dump ((f file) (s (lstat file))) |
272 | (write-string "(" p) | |
273 | (case (stat:type s) | |
274 | ((regular) | |
275 | (write-string "type" p) | |
276 | (write-string "regular" p) | |
277 | (if (not (zero? (logand (stat:mode s) #o100))) | |
278 | (begin | |
279 | (write-string "executable" p) | |
280 | (write-string "" p))) | |
281 | (write-contents f p (stat:size s))) | |
282 | ((directory) | |
283 | (write-string "type" p) | |
284 | (write-string "directory" p) | |
285 | (let ((entries | |
286 | ;; 'scandir' defaults to 'string-locale<?' to sort files, but | |
287 | ;; this happens to be case-insensitive (at least in 'en_US' | |
288 | ;; locale on libc 2.18.) Conversely, we want files to be | |
289 | ;; sorted in a case-sensitive fashion. | |
290 | (scandir f (negate (cut member <> '("." ".."))) string<?))) | |
291 | (for-each (lambda (e) | |
292 | (let* ((f (string-append f "/" e)) | |
293 | (s (lstat f))) | |
294 | (when (select? f s) | |
0363991a LC |
295 | (write-string "entry" p) |
296 | (write-string "(" p) | |
297 | (write-string "name" p) | |
298 | (write-string e p) | |
299 | (write-string "node" p) | |
fe585be9 LC |
300 | (dump f s) |
301 | (write-string ")" p)))) | |
302 | entries))) | |
303 | ((symlink) | |
304 | (write-string "type" p) | |
305 | (write-string "symlink" p) | |
306 | (write-string "target" p) | |
307 | (write-string (readlink f) p)) | |
308 | (else | |
309 | (raise (condition (&message (message "unsupported file type")) | |
310 | (&nar-error (file f) (port port)))))) | |
311 | (write-string ")" p))) | |
0363991a LC |
312 | |
313 | (define (restore-file port file) | |
314 | "Read a file (possibly a directory structure) in Nar format from PORT. | |
315 | Restore it as FILE." | |
46b8aadb LC |
316 | (parameterize ((currently-restored-file file)) |
317 | (let ((signature (read-string port))) | |
318 | (unless (equal? signature %archive-version-1) | |
319 | (raise | |
320 | (condition (&message (message "invalid nar signature")) | |
321 | (&nar-read-error (port port) | |
322 | (token signature) | |
323 | (file #f)))))) | |
324 | ||
325 | (let restore ((file file)) | |
326 | (define (read-eof-marker) | |
327 | (match (read-string port) | |
328 | (")" #t) | |
329 | (x (raise | |
330 | (condition | |
331 | (&message (message "invalid nar end-of-file marker")) | |
332 | (&nar-read-error (port port) (file file) (token x))))))) | |
333 | ||
334 | (currently-restored-file file) | |
335 | ||
336 | (match (list (read-string port) (read-string port) (read-string port)) | |
337 | (("(" "type" "regular") | |
338 | (call-with-output-file file (cut read-contents port <>)) | |
339 | (read-eof-marker)) | |
340 | (("(" "type" "symlink") | |
341 | (match (list (read-string port) (read-string port)) | |
342 | (("target" target) | |
343 | (symlink target file) | |
344 | (read-eof-marker)) | |
345 | (x (raise | |
0363991a | 346 | (condition |
46b8aadb LC |
347 | (&message (message "invalid symlink tokens")) |
348 | (&nar-read-error (port port) (file file) (token x))))))) | |
349 | (("(" "type" "directory") | |
350 | (let ((dir file)) | |
351 | (mkdir dir) | |
352 | (let loop ((prefix (read-string port))) | |
353 | (match prefix | |
354 | ("entry" | |
355 | (match (list (read-string port) | |
356 | (read-string port) (read-string port) | |
357 | (read-string port)) | |
358 | (("(" "name" file "node") | |
359 | (restore (string-append dir "/" file)) | |
360 | (match (read-string port) | |
361 | (")" #t) | |
362 | (x | |
363 | (raise | |
364 | (condition | |
365 | (&message | |
366 | (message "unexpected directory entry termination")) | |
367 | (&nar-read-error (port port) | |
368 | (file file) | |
369 | (token x)))))) | |
370 | (loop (read-string port))))) | |
371 | (")" #t) ; done with DIR | |
372 | (x | |
373 | (raise | |
374 | (condition | |
375 | (&message (message "unexpected directory inter-entry marker")) | |
376 | (&nar-read-error (port port) (file file) (token x))))))))) | |
377 | (x | |
378 | (raise | |
379 | (condition | |
380 | (&message (message "unsupported nar entry type")) | |
381 | (&nar-read-error (port port) (file file) (token x))))))))) | |
0363991a | 382 | |
0f41c26f | 383 | ;;; serialization.scm ends here |