archive-zip: streaming for stored files
Write a data descriptor containing the CRC of the entry and its sizes after streaming it out. For simplicity, do that only if we're storing files (option -0) for now. t5000 verifies output. t1050 makes sure the command always respects core.bigfilethreshold Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
René Scharfe committed
May 3, 2012 at 08:51 UTC
2158f883d99a92f801534c91294305ccbe171f79
3 files changed
+88
-12
archive-zip.c
+78
-12
@@ -3,6 +3,7 @@
3
*/
4
#include "cache.h"
5
#include "archive.h"
6
+#include "streaming.h"
7
8
static int zip_date;
9
static int zip_time;
@@ -15,6 +16,7 @@ static unsigned int zip_dir_offset;
16
static unsigned int zip_dir_entries;
17
18
#define ZIP_DIRECTORY_MIN_SIZE (1024 * 1024)
19
+#define ZIP_STREAM (8)
20
21
struct zip_local_header {
22
unsigned char magic[4];
@@ -31,6 +33,14 @@ struct zip_local_header {
33
unsigned char _end[1];
34
};
35
36
+struct zip_data_desc {
37
+ unsigned char magic[4];
38
+ unsigned char crc32[4];
39
+ unsigned char compressed_size[4];
40
+ unsigned char size[4];
41
+ unsigned char _end[1];
42
+};
43
+
44
struct zip_dir_header {
45
unsigned char magic[4];
46
unsigned char creator_version[2];
@@ -70,6 +80,7 @@ struct zip_dir_trailer {
80
* we're interested in.
81
*/
82
#define ZIP_LOCAL_HEADER_SIZE offsetof(struct zip_local_header, _end)
83
+#define ZIP_DATA_DESC_SIZE offsetof(struct zip_data_desc, _end)
84
#define ZIP_DIR_HEADER_SIZE offsetof(struct zip_dir_header, _end)
85
#define ZIP_DIR_TRAILER_SIZE offsetof(struct zip_dir_trailer, _end)
86
@@ -120,6 +131,19 @@ static void *zlib_deflate(void *data, unsigned long size,
131
return buffer;
132
}
133
134
+static void write_zip_data_desc(unsigned long size,
135
+ unsigned long compressed_size,
136
+ unsigned long crc)
137
+{
138
+ struct zip_data_desc trailer;
139
+
140
+ copy_le32(trailer.magic, 0x08074b50);
141
+ copy_le32(trailer.crc32, crc);
142
+ copy_le32(trailer.compressed_size, compressed_size);
143
+ copy_le32(trailer.size, size);
144
+ write_or_die(1, &trailer, ZIP_DATA_DESC_SIZE);
145
+}
146
+
147
static void set_zip_dir_data_desc(struct zip_dir_header *header,
148
unsigned long size,
149
unsigned long compressed_size,
@@ -140,6 +164,8 @@ static void set_zip_header_data_desc(struct zip_local_header *header,
164
copy_le32(header->size, size);
165
}
166
167
+#define STREAM_BUFFER_SIZE (1024 * 16)
168
+
169
static int write_zip_entry(struct archiver_args *args,
170
const unsigned char *sha1,
171
const char *path, size_t pathlen,
@@ -155,6 +181,8 @@ static int write_zip_entry(struct archiver_args *args,
181
unsigned char *out;
182
void *deflated = NULL;
183
void *buffer;
184
+ struct git_istream *stream = NULL;
185
+ unsigned long flags = 0;
186
unsigned long size;
187
188
crc = crc32(0, NULL, 0);
@@ -173,25 +201,38 @@ static int write_zip_entry(struct archiver_args *args,
201
buffer = NULL;
202
size = 0;
203
} else if (S_ISREG(mode) || S_ISLNK(mode)) {
176
- enum object_type type;
177
- buffer = sha1_file_to_archive(args, path, sha1, mode, &type, &size);
178
- if (!buffer)
179
- return error("cannot read %s", sha1_to_hex(sha1));
204
+ enum object_type type = sha1_object_info(sha1, &size);
205
206
method = 0;
207
attr2 = S_ISLNK(mode) ? ((mode | 0777) << 16) :
208
(mode & 0111) ? ((mode) << 16) : 0;
184
- if (S_ISREG(mode) && args->compression_level != 0)
209
+ if (S_ISREG(mode) && args->compression_level != 0 && size > 0)
210
method = 8;
186
- crc = crc32(crc, buffer, size);
187
- out = buffer;
211
compressed_size = size;
212
+
213
+ if (S_ISREG(mode) && type == OBJ_BLOB && !args->convert &&
214
+ size > big_file_threshold && method == 0) {
215
+ stream = open_istream(sha1, &type, &size, NULL);
216
+ if (!stream)
217
+ return error("cannot stream blob %s",
218
+ sha1_to_hex(sha1));
219
+ flags |= ZIP_STREAM;
220
+ out = buffer = NULL;
221
+ } else {
222
+ buffer = sha1_file_to_archive(args, path, sha1, mode,
223
+ &type, &size);
224
+ if (!buffer)
225
+ return error("cannot read %s",
226
+ sha1_to_hex(sha1));
227
+ crc = crc32(crc, buffer, size);
228
+ out = buffer;
229
+ }
230
} else {
231
return error("unsupported file mode: 0%o (SHA1: %s)", mode,
232
sha1_to_hex(sha1));
233
}
234
194
- if (method == 8) {
235
+ if (buffer && method == 8) {
236
deflated = zlib_deflate(buffer, size, args->compression_level,
237
&compressed_size);
238
if (deflated && compressed_size - 6 < size) {
@@ -216,7 +257,7 @@ static int write_zip_entry(struct archiver_args *args,
257
copy_le16(dirent.creator_version,
258
S_ISLNK(mode) || (S_ISREG(mode) && (mode & 0111)) ? 0x0317 : 0);
259
copy_le16(dirent.version, 10);
219
- copy_le16(dirent.flags, 0);
260
+ copy_le16(dirent.flags, flags);
261
copy_le16(dirent.compression_method, method);
262
copy_le16(dirent.mtime, zip_time);
263
copy_le16(dirent.mdate, zip_date);
@@ -231,18 +272,43 @@ static int write_zip_entry(struct archiver_args *args,
272
273
copy_le32(header.magic, 0x04034b50);
274
copy_le16(header.version, 10);
234
- copy_le16(header.flags, 0);
275
+ copy_le16(header.flags, flags);
276
copy_le16(header.compression_method, method);
277
copy_le16(header.mtime, zip_time);
278
copy_le16(header.mdate, zip_date);
238
- set_zip_header_data_desc(&header, size, compressed_size, crc);
279
+ if (flags & ZIP_STREAM)
280
+ set_zip_header_data_desc(&header, 0, 0, 0);
281
+ else
282
+ set_zip_header_data_desc(&header, size, compressed_size, crc);
283
copy_le16(header.filename_length, pathlen);
284
copy_le16(header.extra_length, 0);
285
write_or_die(1, &header, ZIP_LOCAL_HEADER_SIZE);
286
zip_offset += ZIP_LOCAL_HEADER_SIZE;
287
write_or_die(1, path, pathlen);
288
zip_offset += pathlen;
245
- if (compressed_size > 0) {
289
+ if (stream && method == 0) {
290
+ unsigned char buf[STREAM_BUFFER_SIZE];
291
+ ssize_t readlen;
292
+
293
+ for (;;) {
294
+ readlen = read_istream(stream, buf, sizeof(buf));
295
+ if (readlen <= 0)
296
+ break;
297
+ crc = crc32(crc, buf, readlen);
298
+ write_or_die(1, buf, readlen);
299
+ }
300
+ close_istream(stream);
301
+ if (readlen)
302
+ return readlen;
303
+
304
+ compressed_size = size;
305
+ zip_offset += compressed_size;
306
+
307
+ write_zip_data_desc(size, compressed_size, crc);
308
+ zip_offset += ZIP_DATA_DESC_SIZE;
309
+
310
+ set_zip_dir_data_desc(&dirent, size, compressed_size, crc);
311
+ } else if (compressed_size > 0) {
312
write_or_die(1, out, compressed_size);
313
zip_offset += compressed_size;
314
}
t/t1050-large.sh
+4
@@ -138,4 +138,8 @@ test_expect_success 'tar achiving' '
138
git archive --format=tar HEAD >/dev/null
139
'
140
141
+test_expect_success 'zip achiving, store only' '
142
+ git archive --format=zip -0 HEAD >/dev/null
143
+'
144
+
145
test_done
t/t5000-tar-tree.sh
+6
@@ -244,6 +244,12 @@ test_expect_success UNZIP \
244
'validate file contents with prefix' \
245
'diff -r a e/prefix/a'
246
247
+test_expect_success UNZIP 'git archive -0 --format=zip on large files' '
248
+ test_config core.bigfilethreshold 1 &&
249
+ git archive -0 --format=zip HEAD >large.zip &&
250
+ (mkdir large && cd large && $UNZIP ../large.zip)
251
+'
252
+
253
test_expect_success \
254
'git archive --list outside of a git repo' \
255
'GIT_DIR=some/non-existing/directory git archive --list'