Merge branch 'db/length-as-hash' into svn-fe

* db/length-as-hash:
  vcs-svn: use strchr to find RFC822 delimiter
  vcs-svn: implement perfect hash for top-level keys
  vcs-svn: implement perfect hash for node-prop keys

Conflicts:
	vcs-svn/svndump.c
This commit is contained in:
Jonathan Nieder
2011-03-22 18:44:49 -05:00

View File

@ -14,6 +14,12 @@
#include "string_pool.h" #include "string_pool.h"
#include "strbuf.h" #include "strbuf.h"
/*
* Compare start of string to literal of equal length;
* must be guarded by length test.
*/
#define constcmp(s, ref) memcmp(s, ref, sizeof(ref) - 1)
#define NODEACT_REPLACE 4 #define NODEACT_REPLACE 4
#define NODEACT_DELETE 3 #define NODEACT_DELETE 3
#define NODEACT_ADD 2 #define NODEACT_ADD 2
@ -46,15 +52,6 @@ static struct {
struct strbuf uuid, url; struct strbuf uuid, url;
} dump_ctx; } dump_ctx;
static struct {
uint32_t svn_log, svn_author, svn_date, svn_executable, svn_special, uuid,
revision_number, node_path, node_kind, node_action,
node_copyfrom_path, node_copyfrom_rev, text_content_length,
prop_content_length, content_length, svn_fs_dump_format_version,
/* version 3 format */
text_delta, prop_delta;
} keys;
static void reset_node_ctx(char *fname) static void reset_node_ctx(char *fname)
{ {
node_ctx.type = 0; node_ctx.type = 0;
@ -85,47 +82,45 @@ static void reset_dump_ctx(const char *url)
strbuf_reset(&dump_ctx.uuid); strbuf_reset(&dump_ctx.uuid);
} }
static void init_keys(void) static void handle_property(const struct strbuf *key_buf,
{ const char *val, uint32_t len,
keys.svn_log = pool_intern("svn:log");
keys.svn_author = pool_intern("svn:author");
keys.svn_date = pool_intern("svn:date");
keys.svn_executable = pool_intern("svn:executable");
keys.svn_special = pool_intern("svn:special");
keys.uuid = pool_intern("UUID");
keys.revision_number = pool_intern("Revision-number");
keys.node_path = pool_intern("Node-path");
keys.node_kind = pool_intern("Node-kind");
keys.node_action = pool_intern("Node-action");
keys.node_copyfrom_path = pool_intern("Node-copyfrom-path");
keys.node_copyfrom_rev = pool_intern("Node-copyfrom-rev");
keys.text_content_length = pool_intern("Text-content-length");
keys.prop_content_length = pool_intern("Prop-content-length");
keys.content_length = pool_intern("Content-length");
keys.svn_fs_dump_format_version = pool_intern("SVN-fs-dump-format-version");
/* version 3 format (Subversion 1.1.0) */
keys.text_delta = pool_intern("Text-delta");
keys.prop_delta = pool_intern("Prop-delta");
}
static void handle_property(uint32_t key, const char *val, uint32_t len,
uint32_t *type_set) uint32_t *type_set)
{ {
if (key == keys.svn_log) { const char *key = key_buf->buf;
size_t keylen = key_buf->len;
switch (keylen + 1) {
case sizeof("svn:log"):
if (constcmp(key, "svn:log"))
break;
if (!val) if (!val)
die("invalid dump: unsets svn:log"); die("invalid dump: unsets svn:log");
strbuf_reset(&rev_ctx.log); strbuf_reset(&rev_ctx.log);
strbuf_add(&rev_ctx.log, val, len); strbuf_add(&rev_ctx.log, val, len);
} else if (key == keys.svn_author) { break;
case sizeof("svn:author"):
if (constcmp(key, "svn:author"))
break;
strbuf_reset(&rev_ctx.author); strbuf_reset(&rev_ctx.author);
if (val) if (val)
strbuf_add(&rev_ctx.author, val, len); strbuf_add(&rev_ctx.author, val, len);
} else if (key == keys.svn_date) { break;
case sizeof("svn:date"):
if (constcmp(key, "svn:date"))
break;
if (!val) if (!val)
die("invalid dump: unsets svn:date"); die("invalid dump: unsets svn:date");
if (parse_date_basic(val, &rev_ctx.timestamp, NULL)) if (parse_date_basic(val, &rev_ctx.timestamp, NULL))
warning("invalid timestamp: %s", val); warning("invalid timestamp: %s", val);
} else if (key == keys.svn_executable || key == keys.svn_special) { break;
case sizeof("svn:executable"):
case sizeof("svn:special"):
if (keylen == strlen("svn:executable") &&
constcmp(key, "svn:executable"))
break;
if (keylen == strlen("svn:special") &&
constcmp(key, "svn:special"))
break;
if (*type_set) { if (*type_set) {
if (!val) if (!val)
return; return;
@ -136,7 +131,7 @@ static void handle_property(uint32_t key, const char *val, uint32_t len,
return; return;
} }
*type_set = 1; *type_set = 1;
node_ctx.type = key == keys.svn_executable ? node_ctx.type = keylen == strlen("svn:executable") ?
REPO_MODE_EXE : REPO_MODE_EXE :
REPO_MODE_LNK; REPO_MODE_LNK;
} }
@ -151,7 +146,7 @@ static void die_short_read(void)
static void read_props(void) static void read_props(void)
{ {
uint32_t key = ~0; static struct strbuf key = STRBUF_INIT;
const char *t; const char *t;
/* /*
* NEEDSWORK: to support simple mode changes like * NEEDSWORK: to support simple mode changes like
@ -188,16 +183,19 @@ static void read_props(void)
switch (type) { switch (type) {
case 'K': case 'K':
key = pool_intern(val);
continue;
case 'D': case 'D':
key = pool_intern(val); strbuf_reset(&key);
if (val)
strbuf_add(&key, val, len);
if (type == 'K')
continue;
assert(type == 'D');
val = NULL; val = NULL;
len = 0; len = 0;
/* fall through */ /* fall through */
case 'V': case 'V':
handle_property(key, val, len, &type_set); handle_property(&key, val, len, &type_set);
key = ~0; strbuf_reset(&key);
continue; continue;
default: default:
die("invalid property line: %s\n", t); die("invalid property line: %s\n", t);
@ -290,45 +288,65 @@ void svndump_read(const char *url)
char *t; char *t;
uint32_t active_ctx = DUMP_CTX; uint32_t active_ctx = DUMP_CTX;
uint32_t len; uint32_t len;
uint32_t key;
reset_dump_ctx(url); reset_dump_ctx(url);
while ((t = buffer_read_line(&input))) { while ((t = buffer_read_line(&input))) {
val = strstr(t, ": "); val = strchr(t, ':');
if (!val) if (!val)
continue; continue;
*val++ = '\0'; val++;
*val++ = '\0'; if (*val != ' ')
key = pool_intern(t); continue;
val++;
if (key == keys.svn_fs_dump_format_version) { /* strlen(key) + 1 */
switch (val - t - 1) {
case sizeof("SVN-fs-dump-format-version"):
if (constcmp(t, "SVN-fs-dump-format-version"))
continue;
dump_ctx.version = atoi(val); dump_ctx.version = atoi(val);
if (dump_ctx.version > 3) if (dump_ctx.version > 3)
die("expected svn dump format version <= 3, found %"PRIu32, die("expected svn dump format version <= 3, found %"PRIu32,
dump_ctx.version); dump_ctx.version);
} else if (key == keys.uuid) { break;
case sizeof("UUID"):
if (constcmp(t, "UUID"))
continue;
strbuf_reset(&dump_ctx.uuid); strbuf_reset(&dump_ctx.uuid);
strbuf_addstr(&dump_ctx.uuid, val); strbuf_addstr(&dump_ctx.uuid, val);
} else if (key == keys.revision_number) { break;
case sizeof("Revision-number"):
if (constcmp(t, "Revision-number"))
continue;
if (active_ctx == NODE_CTX) if (active_ctx == NODE_CTX)
handle_node(); handle_node();
if (active_ctx != DUMP_CTX) if (active_ctx != DUMP_CTX)
handle_revision(); handle_revision();
active_ctx = REV_CTX; active_ctx = REV_CTX;
reset_rev_ctx(atoi(val)); reset_rev_ctx(atoi(val));
} else if (key == keys.node_path) { break;
if (active_ctx == NODE_CTX) case sizeof("Node-path"):
handle_node(); if (prefixcmp(t, "Node-"))
active_ctx = NODE_CTX; continue;
reset_node_ctx(val); if (!constcmp(t + strlen("Node-"), "path")) {
} else if (key == keys.node_kind) { if (active_ctx == NODE_CTX)
handle_node();
active_ctx = NODE_CTX;
reset_node_ctx(val);
break;
}
if (constcmp(t + strlen("Node-"), "kind"))
continue;
if (!strcmp(val, "dir")) if (!strcmp(val, "dir"))
node_ctx.type = REPO_MODE_DIR; node_ctx.type = REPO_MODE_DIR;
else if (!strcmp(val, "file")) else if (!strcmp(val, "file"))
node_ctx.type = REPO_MODE_BLB; node_ctx.type = REPO_MODE_BLB;
else else
fprintf(stderr, "Unknown node-kind: %s\n", val); fprintf(stderr, "Unknown node-kind: %s\n", val);
} else if (key == keys.node_action) { break;
case sizeof("Node-action"):
if (constcmp(t, "Node-action"))
continue;
if (!strcmp(val, "delete")) { if (!strcmp(val, "delete")) {
node_ctx.action = NODEACT_DELETE; node_ctx.action = NODEACT_DELETE;
} else if (!strcmp(val, "add")) { } else if (!strcmp(val, "add")) {
@ -341,19 +359,38 @@ void svndump_read(const char *url)
fprintf(stderr, "Unknown node-action: %s\n", val); fprintf(stderr, "Unknown node-action: %s\n", val);
node_ctx.action = NODEACT_UNKNOWN; node_ctx.action = NODEACT_UNKNOWN;
} }
} else if (key == keys.node_copyfrom_path) { break;
case sizeof("Node-copyfrom-path"):
if (constcmp(t, "Node-copyfrom-path"))
continue;
pool_tok_seq(REPO_MAX_PATH_DEPTH, node_ctx.src, "/", val); pool_tok_seq(REPO_MAX_PATH_DEPTH, node_ctx.src, "/", val);
} else if (key == keys.node_copyfrom_rev) { break;
case sizeof("Node-copyfrom-rev"):
if (constcmp(t, "Node-copyfrom-rev"))
continue;
node_ctx.srcRev = atoi(val); node_ctx.srcRev = atoi(val);
} else if (key == keys.text_content_length) { break;
node_ctx.textLength = atoi(val); case sizeof("Text-content-length"):
} else if (key == keys.prop_content_length) { if (!constcmp(t, "Text-content-length")) {
node_ctx.textLength = atoi(val);
break;
}
if (constcmp(t, "Prop-content-length"))
continue;
node_ctx.propLength = atoi(val); node_ctx.propLength = atoi(val);
} else if (key == keys.text_delta) { break;
node_ctx.text_delta = !strcmp(val, "true"); case sizeof("Text-delta"):
} else if (key == keys.prop_delta) { if (!constcmp(t, "Text-delta")) {
node_ctx.text_delta = !strcmp(val, "true");
break;
}
if (constcmp(t, "Prop-delta"))
continue;
node_ctx.prop_delta = !strcmp(val, "true"); node_ctx.prop_delta = !strcmp(val, "true");
} else if (key == keys.content_length) { break;
case sizeof("Content-length"):
if (constcmp(t, "Content-length"))
continue;
len = atoi(val); len = atoi(val);
t = buffer_read_line(&input); t = buffer_read_line(&input);
if (!t) if (!t)
@ -392,7 +429,6 @@ int svndump_init(const char *filename)
reset_dump_ctx(NULL); reset_dump_ctx(NULL);
reset_rev_ctx(0); reset_rev_ctx(0);
reset_node_ctx(NULL); reset_node_ctx(NULL);
init_keys();
return 0; return 0;
} }