-Wayne Davison wrote:
+Depends-On-Patch: g2r-basis-filename.diff
-I greatly simplified the changes to generator.c, making the patch
+The changes to generator.c were greatly simplified, making the patch
easier to maintain and fixing the failing test in the testsuite.
Very lightly tested.
---- generator.c 29 Jun 2004 19:19:00 -0000 1.92
-+++ generator.c 30 Jun 2004 07:35:56 -0000
-@@ -41,6 +41,7 @@ extern int ignore_times;
+Be sure to run "make proto" before "make".
+
+--- orig/generator.c 2004-09-20 19:57:58
++++ generator.c 2004-09-23 15:33:08
+@@ -43,6 +43,7 @@ extern int ignore_times;
extern int size_only;
extern int io_timeout;
extern int protocol_version;
-+extern int fuzzy;
++extern int fuzzy_basis;
extern int always_checksum;
+ extern char *partial_dir;
extern char *compare_dest;
- extern int link_dest;
-@@ -257,6 +258,94 @@ static void generate_and_send_sums(struc
+@@ -244,6 +245,92 @@ static void generate_and_send_sums(int f
}
+}
+
+
-+static int find_fuzzy(char **fname_ptr, char *buf, STRUCT_STAT *st_ptr)
++static int find_fuzzy(const char *fname, char *buf, STRUCT_STAT *st_ptr)
+{
+ DIR *d;
+ struct dirent *di;
+ unsigned int bestscore = 0;
+ const char *ext;
+
-+ strlcpy(mangled_name, *fname_ptr, sizeof mangled_name);
++ strlcpy(mangled_name, fname, sizeof mangled_name);
+
+ split_names(mangled_name, &dirname, &basename);
+ if (!(d = opendir(dirname))) {
+ }
+
+ /* Get final extension, eg. .gz; never full basename though. */
-+ ext = strrchr(basename + 1, '.');
-+ if (!ext)
++ if (!(ext = strrchr(basename + 1, '.')))
+ ext = basename + strlen(basename); /* ext = "" */
+
+ while ((di = readdir(d)) != NULL) {
+ pathjoin(buf, MAXPATHLEN, dirname, bestname);
+ if (verbose > 2) {
+ rprintf(FINFO, "[%s] fuzzy match %s->%s\n",
-+ who_am_i(), *fname_ptr, buf);
++ who_am_i(), fname, buf);
+ }
-+ *fname_ptr = buf;
+ return link_stat(buf, st_ptr, 0);
+ }
+ return -1;
/*
* Acts on file number @p i from @p flist, whose name is @p fname.
-@@ -267,12 +356,12 @@ static void generate_and_send_sums(struc
- * out. It might be wrong.
- */
+@@ -256,11 +343,11 @@ static void generate_and_send_sums(int f
static void recv_generator(char *fname, struct file_struct *file, int i,
-- int f_out)
-+ int f_out, int f_nameout)
+ int f_out, int f_out_name)
{
- int fd;
- STRUCT_STAT st;
- struct map_struct *mapbuf;
-- int statret;
-+ int statret, fuzzy_file = 0;
- char *fnamecmp;
+- int fd = -1, f_copy;
++ int fd = -1, f_copy = -1;
+ STRUCT_STAT st, partial_st;
+- struct file_struct *back_file;
++ struct file_struct *back_file = NULL;
+ int statret, stat_errno;
+- char *fnamecmp, *partialptr, *backupptr;
++ char *fnamecmp, *partialptr, *backupptr = NULL;
char fnamecmpbuf[MAXPATHLEN];
-
-@@ -431,8 +520,10 @@ static void recv_generator(char *fname,
- statret = link_stat(fnamecmpbuf, &st, 0);
- if (!S_ISREG(st.st_mode))
- statret = -1;
-- if (statret == -1)
-+ if (statret < 0) {
- errno = saveerrno;
-+ *fnamecmpbuf = '\0';
+ uchar fnamecmp_type;
+
+@@ -468,6 +555,15 @@ static void recv_generator(char *fname,
+ } else
+ partialptr = NULL;
+
++ if (statret == -1 && fuzzy_basis) {
++ if (find_fuzzy(fname, fnamecmpbuf, &st) == 0
++ && S_ISREG(st.st_mode)) {
++ statret = 0;
++ fnamecmp = fnamecmpbuf;
++ fnamecmp_type = G2R_FUZZY;
+ }
- #if HAVE_LINK
- else if (link_dest && !dry_run) {
- if (do_link(fnamecmpbuf, fname) != 0) {
-@@ -440,18 +531,30 @@ static void recv_generator(char *fname,
- rsyserr(FINFO, errno, "link %s => %s",
- fnamecmpbuf, fname);
- }
-- }
-- fnamecmp = fnamecmpbuf;
-+ fnamecmp = fnamecmpbuf;
-+ } else
-+ *fnamecmpbuf = '\0';
- }
- #endif
- else
- fnamecmp = fnamecmpbuf;
-+ } else
-+ *fnamecmpbuf = '\0';
++ }
+
-+ if (statret == -1 && fuzzy) {
-+ statret = find_fuzzy(&fnamecmp, fnamecmpbuf, &st);
-+ if (!S_ISREG(st.st_mode))
-+ statret = -1;
-+ else
-+ fuzzy_file = 1;
- }
-
if (statret == -1) {
if (preserve_hard_links && hard_link_check(file, HL_SKIP))
return;
- if (errno == ENOENT) {
-+ if (f_nameout >= 0)
-+ write(f_nameout, "", 1);
- write_int(f_out,i);
- if (!dry_run)
- write_sum_head(f_out, NULL);
-@@ -471,37 +574,43 @@ static void recv_generator(char *fname,
- /* now pretend the file didn't exist */
- if (preserve_hard_links && hard_link_check(file, HL_SKIP))
- return;
-+ if (f_nameout >= 0)
-+ write(f_nameout, "", 1);
- write_int(f_out,i);
- if (!dry_run)
- write_sum_head(f_out, NULL);
+@@ -494,7 +590,7 @@ static void recv_generator(char *fname,
return;
}
-- if (opt_ignore_existing && fnamecmp == fname) {
-+ if (opt_ignore_existing && !*fnamecmpbuf) {
- if (verbose > 1)
- rprintf(FINFO,"%s exists\n",fname);
- return;
- }
-
-- if (update_only && fnamecmp == fname
-+ if (update_only && !*fnamecmpbuf
- && cmp_modtime(st.st_mtime, file->modtime) > 0) {
- if (verbose > 1)
- rprintf(FINFO,"%s is newer\n",fname);
- return;
- }
-
-- if (skip_file(fname, file, &st)) {
-- if (fnamecmp == fname)
-+ if (!fuzzy_file && skip_file(fname, file, &st)) {
-+ if (!*fnamecmpbuf)
+- if (skip_file(fnamecmp, file, &st)) {
++ if (fnamecmp_type != G2R_FUZZY && skip_file(fnamecmp, file, &st)) {
+ if (fnamecmp_type == G2R_FNAME)
set_perms(fname, file, &st, PERMS_REPORT);
return;
- }
-
- if (dry_run) {
-+ if (f_nameout >= 0)
-+ write(f_nameout, "", 1);
- write_int(f_out,i);
- return;
- }
-
- if (disable_deltas_p()) {
-+ if (f_nameout >= 0)
-+ write(f_nameout, "", 1);
- write_int(f_out,i);
- write_sum_head(f_out, NULL);
- return;
-@@ -516,6 +625,8 @@ static void recv_generator(char *fname,
- /* pretend the file didn't exist */
- if (preserve_hard_links && hard_link_check(file, HL_SKIP))
+@@ -553,10 +649,6 @@ prepare_to_open:
return;
-+ if (f_nameout >= 0)
-+ write(f_nameout, "", 1);
- write_int(f_out,i);
- write_sum_head(f_out, NULL);
- return;
-@@ -534,6 +645,8 @@ static void recv_generator(char *fname,
- if (verbose > 2)
- rprintf(FINFO, "generating and sending sums for %d\n", i);
-
-+ if (f_nameout >= 0)
-+ write(f_nameout, fnamecmpbuf, strlen(fnamecmpbuf) + 1);
- write_int(f_out,i);
- generate_and_send_sums(mapbuf, st.st_size, f_out);
-
-@@ -543,10 +656,11 @@ static void recv_generator(char *fname,
- }
-
-
--void generate_files(int f, struct file_list *flist, char *local_name)
-+void generate_files(int f, struct file_list *flist, char *local_name,
-+ int f_nameout)
- {
- int i;
-- int phase=0;
-+ int phase = 0;
- char fbuf[MAXPATHLEN];
-
- if (verbose > 2) {
-@@ -584,7 +698,7 @@ void generate_files(int f, struct file_l
}
-
- recv_generator(local_name ? local_name : f_name_to(file, fbuf),
-- file, i, f);
-+ file, i, f, f_nameout);
+ fnamecmp_type = G2R_BACKUP;
+- } else {
+- backupptr = NULL;
+- back_file = NULL;
+- f_copy = -1;
}
- phase++;
-@@ -601,7 +715,7 @@ void generate_files(int f, struct file_l
- while ((i = get_redo_num()) != -1) {
- struct file_struct *file = flist->files[i];
- recv_generator(local_name ? local_name : f_name_to(file, fbuf),
-- file, i, f);
-+ file, i, f, f_nameout);
- }
-
- phase++;
-@@ -620,7 +734,7 @@ void generate_files(int f, struct file_l
- if (!file->basename || !S_ISDIR(file->mode))
- continue;
- recv_generator(local_name ? local_name : f_name(file),
-- file, i, -1);
-+ file, i, -1, -1);
+ if (verbose > 3) {
+@@ -570,6 +662,21 @@ prepare_to_open:
+ notify_others:
+ if (f_out_name >= 0) {
+ write_byte(f_out_name, fnamecmp_type);
++ if (fnamecmp_type == G2R_FUZZY) {
++ uchar lenbuf[3], *lb = lenbuf;
++ int len = strlen(fnamecmpbuf);
++ if (len > 0x7F) {
++#if MAXPATHLEN > 0x7FFF
++ *lb++ = len / 0x10000 + 0x80;
++ *lb++ = len / 0x100;
++#else
++ *lb++ = len / 0x100 + 0x80;
++#endif
++ }
++ *lb = len;
++ write_buf(f_out_name, lenbuf, lb - lenbuf + 1);
++ write_buf(f_out_name, fnamecmpbuf, len);
++ }
+ io_flush(NORMAL_FLUSH); /* XXX make this more efficient! */
}
- if (verbose > 2)
---- main.c 28 Jun 2004 17:45:40 -0000 1.201
-+++ main.c 30 Jun 2004 07:35:57 -0000
-@@ -428,8 +428,8 @@ static void do_server_sender(int f_in, i
- static int do_recv(int f_in,int f_out,struct file_list *flist,char *local_name)
- {
+--- orig/main.c 2004-07-22 00:10:43
++++ main.c 2004-07-22 00:32:31
+@@ -48,6 +48,7 @@ extern int keep_dirlinks;
+ extern int preserve_hard_links;
+ extern int protocol_version;
+ extern int recurse;
++extern int fuzzy_basis;
+ extern int relative_paths;
+ extern int rsync_port;
+ extern int whole_file;
+@@ -463,7 +464,7 @@ static int do_recv(int f_in,int f_out,st
int pid;
-- int status=0;
-- int error_pipe[2];
-+ int status = 0;
-+ int error_pipe[2], name_pipe[2];
-
- if (preserve_hard_links)
- init_hard_links(flist);
-@@ -441,17 +441,19 @@ static int do_recv(int f_in,int f_out,st
- }
- }
-
-- if (fd_pair(error_pipe) < 0) {
-- rprintf(FERROR,"error pipe failed in do_recv\n");
-+ if (fd_pair(error_pipe) < 0 || fd_pair(name_pipe) < 0) {
-+ rprintf(FERROR, "fd_pair() failed in do_recv\n");
- exit_cleanup(RERR_SOCKETIO);
- }
-
- io_flush(NORMAL_FLUSH);
-
-- if ((pid=do_fork()) == 0) {
-+ if ((pid = do_fork()) == 0) {
- close(error_pipe[0]);
-+ close(name_pipe[1]);
- if (f_in != f_out)
- close(f_out);
-+ set_blocking(name_pipe[0]);
-
- /* we can't let two processes write to the socket at one time */
- io_multiplexing_close();
-@@ -459,7 +461,7 @@ static int do_recv(int f_in,int f_out,st
- /* set place to send errors */
- set_msg_fd_out(error_pipe[1]);
-
-- recv_files(f_in,flist,local_name);
-+ recv_files(f_in, flist, local_name, name_pipe[0]);
- io_flush(FULL_FLUSH);
- report(f_in);
-
-@@ -475,14 +477,16 @@ static int do_recv(int f_in,int f_out,st
- am_generator = 1;
-
- close(error_pipe[1]);
-+ close(name_pipe[0]);
- if (f_in != f_out)
- close(f_in);
-+ set_blocking(name_pipe[1]);
-
- io_start_buffering_out(f_out);
-
- set_msg_fd_in(error_pipe[0]);
-
-- generate_files(f_out, flist, local_name);
-+ generate_files(f_out, flist, local_name, name_pipe[1]);
-
- get_redo_num(); /* Read final MSG_DONE and any prior messages. */
- report(-1);
---- options.c 20 Jun 2004 19:47:05 -0000 1.157
-+++ options.c 30 Jun 2004 07:35:57 -0000
-@@ -94,6 +94,7 @@ int ignore_errors = 0;
- int modify_window = 0;
- int blocking_io = -1;
- int checksum_seed = 0;
-+int fuzzy = 0;
- unsigned int block_size = 0;
-
-
-@@ -270,6 +271,7 @@ void usage(enum logcode F)
- rprintf(F," -T --temp-dir=DIR create temporary files in directory DIR\n");
+ int status = 0;
+ int error_pipe[2], name_pipe[2];
+- BOOL need_name_pipe = compare_dest && !dry_run;
++ BOOL need_name_pipe = (compare_dest || fuzzy_basis) && !dry_run;
+
+ /* The receiving side mustn't obey this, or an existing symlink that
+ * points to an identical file won't be replaced by the referent. */
+--- orig/options.c 2004-09-22 08:47:31
++++ options.c 2004-09-23 14:59:46
+@@ -85,6 +85,7 @@ int safe_symlinks = 0;
+ int copy_unsafe_links = 0;
+ int size_only = 0;
+ int bwlimit = 0;
++int fuzzy_basis = 0;
+ size_t bwlimit_writemax = 0;
+ int delete_after = 0;
+ int only_existing = 0;
+@@ -279,6 +280,7 @@ void usage(enum logcode F)
+ rprintf(F," -T, --temp-dir=DIR create temporary files in directory DIR\n");
rprintf(F," --compare-dest=DIR also compare destination files relative to DIR\n");
rprintf(F," --link-dest=DIR create hardlinks to DIR for unchanged files\n");
+ rprintf(F," --fuzzy use similar file as basis if basis doesn't exist\n");
rprintf(F," -P equivalent to --partial --progress\n");
rprintf(F," -z, --compress compress file data\n");
rprintf(F," -C, --cvs-exclude auto ignore files in the same way CVS does\n");
-@@ -368,6 +370,7 @@ static struct poptOption long_options[]
+@@ -378,6 +380,7 @@ static struct poptOption long_options[]
{"temp-dir", 'T', POPT_ARG_STRING, &tmpdir, 0, 0, 0 },
{"compare-dest", 0, POPT_ARG_STRING, &compare_dest, 0, 0, 0 },
{"link-dest", 0, POPT_ARG_STRING, &compare_dest, OPT_LINK_DEST, 0, 0 },
-+ {"fuzzy", 0, POPT_ARG_NONE, &fuzzy, 0, 0, 0 },
++ {"fuzzy", 0, POPT_ARG_NONE, &fuzzy_basis, 0, 0, 0 },
/* TODO: Should this take an optional int giving the compression level? */
{"compress", 'z', POPT_ARG_NONE, &do_compression, 0, 0, 0 },
{"daemon", 0, POPT_ARG_NONE, &daemon_opt, 0, 0, 0 },
-@@ -989,6 +992,9 @@ void server_options(char **args,int *arg
+@@ -833,6 +836,11 @@ int parse_arguments(int *argc, const cha
+ link_dest ? "--link-dest" : "--compare-dest");
+ return 0;
+ }
++ if (compare_dest) {
++ snprintf(err_buf, sizeof err_buf,
++ "--inplace does not yet work with --fuzzy\n");
++ return 0;
++ }
+ } else {
+ if (keep_partial && !partial_dir)
+ partial_dir = getenv("RSYNC_PARTIAL_DIR");
+@@ -1099,6 +1107,9 @@ void server_options(char **args,int *arg
}
}
-+ if (fuzzy && am_sender)
++ if (fuzzy_basis && am_sender)
+ args[ac++] = "--fuzzy";
+
*argc = ac;
return;
---- receiver.c 29 Jun 2004 15:12:01 -0000 1.83
-+++ receiver.c 30 Jun 2004 07:35:57 -0000
-@@ -36,7 +36,6 @@ extern int preserve_perms;
- extern int cvs_exclude;
- extern int io_error;
- extern char *tmpdir;
--extern char *compare_dest;
- extern int make_backups;
- extern int do_progress;
- extern char *backup_dir;
-@@ -293,14 +292,15 @@ static int receive_data(int f_in,struct
- * main routine for receiver process.
- *
- * Receiver process runs on the same host as the generator process. */
--int recv_files(int f_in,struct file_list *flist,char *local_name)
-+int recv_files(int f_in, struct file_list *flist, char *local_name,
-+ int f_name)
- {
- int fd1,fd2;
- STRUCT_STAT st;
- char *fname, fbuf[MAXPATHLEN];
- char template[MAXPATHLEN];
- char fnametmp[MAXPATHLEN];
-- char *fnamecmp;
-+ char *fnamecmp, *cp;
- char fnamecmpbuf[MAXPATHLEN];
- struct map_struct *mapbuf;
- struct file_struct *file;
-@@ -364,19 +364,19 @@ int recv_files(int f_in,struct file_list
- if (verbose > 2)
- rprintf(FINFO,"recv_files(%s)\n",fname);
-
-- fnamecmp = fname;
-+ for (cp = fnamecmpbuf; ; cp++) {
-+ if (read(f_name, cp, 1) <= 0) {
-+ rsyserr(FERROR, errno, "fname-pipe read failed");
-+ exit_cleanup(RERR_PROTOCOL);
-+ }
-+ if (!*cp)
-+ break;
-+ }
-+ fnamecmp = *fnamecmpbuf ? fnamecmpbuf : fname;
-
- /* open the file */
- fd1 = do_open(fnamecmp, O_RDONLY, 0);
+--- orig/receiver.c 2004-09-07 21:57:20
++++ receiver.c 2004-07-30 18:21:38
+@@ -319,6 +319,27 @@ static int receive_data(int f_in, char *
+ }
-- if (fd1 == -1 && compare_dest != NULL) {
-- /* try the file at compare_dest instead */
-- pathjoin(fnamecmpbuf, sizeof fnamecmpbuf,
-- compare_dest, fname);
-- fnamecmp = fnamecmpbuf;
-- fd1 = do_open(fnamecmp, O_RDONLY, 0);
-- }
--
- if (fd1 != -1 && do_fstat(fd1,&st) != 0) {
- rsyserr(FERROR, errno, "fstat %s failed",
- full_fname(fnamecmp));
-@@ -385,7 +385,7 @@ int recv_files(int f_in,struct file_list
- continue;
- }
-- if (fd1 != -1 && S_ISDIR(st.st_mode) && fnamecmp == fname) {
-+ if (fd1 != -1 && S_ISDIR(st.st_mode) && !*fnamecmpbuf) {
- /* this special handling for directories
- * wouldn't be necessary if robust_rename()
- * and the underlying robust_unlink could cope
---- rsync.yo 5 Jun 2004 16:16:30 -0000 1.171
-+++ rsync.yo 30 Jun 2004 07:35:58 -0000
-@@ -325,6 +325,7 @@ verb(
++static void read_gen_name(int fd, char *buf)
++{
++ int len = read_byte(fd);
++ if (len & 0x80) {
++#if MAXPATHLEN > 32767
++ uchar lenbuf[2];
++ read_buf(fd, (char *)lenbuf, 2);
++ len = (len & ~0x80) * 0x10000 + lenbuf[0] * 0x100 + lenbuf[1];
++#else
++ len = (len & ~0x80) * 0x100 + read_byte(fd);
++#endif
++ }
++ if (len >= MAXPATHLEN) {
++ rprintf(FERROR, "bogus data on generator name pipe\n");
++ exit_cleanup(RERR_PROTOCOL);
++ }
++
++ read_sbuf(fd, buf, len);
++}
++
++
+ static void discard_receive_data(int f_in, OFF_T length)
+ {
+ receive_data(f_in, NULL, -1, 0, NULL, -1, length);
+@@ -448,6 +469,10 @@ int recv_files(int f_in, struct file_lis
+ case G2R_BACKUP:
+ fnamecmp = get_backup_name(fname);
+ break;
++ case G2R_FUZZY:
++ read_gen_name(f_in_name, fnamecmpbuf);
++ fnamecmp = fnamecmpbuf;
++ break;
+ default:
+ pathjoin(fnamecmpbuf, sizeof fnamecmpbuf,
+ compare_dest, fname);
+--- orig/rsync.h 2004-09-07 21:52:22
++++ rsync.h 2004-09-07 22:02:37
+@@ -122,6 +122,7 @@
+ #define G2R_FNAME 0x80
+ #define G2R_PARTIAL_DIR 0x81
+ #define G2R_BACKUP 0x82
++#define G2R_FUZZY 0x83
+
+
+ /* Log-message categories. FLOG is only used on the daemon side to
+--- orig/rsync.yo 2004-09-22 08:47:31
++++ rsync.yo 2004-07-03 19:27:25
+@@ -356,6 +356,7 @@ verb(
-T --temp-dir=DIR create temporary files in directory DIR
--compare-dest=DIR also compare received files relative to DIR
--link-dest=DIR create hardlinks to DIR for unchanged files