Updated.

[rsync/rsync-patches.git] / g2r-basis-filename.diff

Added a pipe from the generator to the receiver that communicates
what basis file we used to generate the file data (if it was not
the default name).  This optimizes away the basis-file search in
the receiver and makes future options that do more basis-file
searching more efficient (such as the --fuzzy option and the
support for multiple --compare-dest options).

Also fixes a potential synchronization problem between the generator
and the receiver in read-batch mode.  Should consider making the
sending of the index value the default for this pipe (it's currently
only sent in batch mode due to the sender not listening to the
generator to determine what files get updated).

You must run "make proto" before compiling.

--- orig/generator.c	2004-07-20 21:36:07
+++ generator.c	2004-07-17 15:50:09
@@ -258,9 +258,9 @@ static void generate_and_send_sums(int f
  * out.  It might be wrong.
  */
 static void recv_generator(char *fname, struct file_struct *file, int i,
-			   int f_out)
+			   int f_out, int f_nameout)
 {
-	int fd;
+	int fd = -1;
 	STRUCT_STAT st;
 	int statret;
 	char *fnamecmp;
@@ -404,9 +404,6 @@ static void recv_generator(char *fname, 
 	}
 #endif
 
-	if (read_batch)
-		return;
-
 	if (preserve_hard_links && hard_link_check(file, HL_CHECK_MASTER))
 		return;
 
@@ -424,8 +421,10 @@ 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 == -1) {
 			errno = saveerrno;
+			*fnamecmpbuf = '\0';
+		}
 #if HAVE_LINK
 		else if (link_dest && !dry_run) {
 			if (do_link(fnamecmpbuf, fname) != 0) {
@@ -433,22 +432,22 @@ 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) {
 		if (preserve_hard_links && hard_link_check(file, HL_SKIP))
 			return;
-		if (errno == ENOENT) {
-			write_int(f_out,i);
-			if (!dry_run)
-				write_sum_head(f_out, NULL);
-		} else if (verbose > 1) {
+		if (errno == ENOENT)
+			goto notify_others;
+		if (verbose > 1) {
 			rsyserr(FERROR, errno,
 				"recv_generator: failed to open %s",
 				full_fname(fname));
@@ -457,26 +456,23 @@ static void recv_generator(char *fname, 
 	}
 
 	if (!S_ISREG(st.st_mode)) {
-		if (delete_file(fname) != 0) {
+		if (delete_file(fname) != 0)
 			return;
-		}
 
 		/* now pretend the file didn't exist */
 		if (preserve_hard_links && hard_link_check(file, HL_SKIP))
 			return;
-		write_int(f_out,i);
-		if (!dry_run)
-			write_sum_head(f_out, NULL);
-		return;
+		statret = -1;
+		goto notify_others;
 	}
 
-	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);
@@ -484,21 +480,17 @@ static void recv_generator(char *fname, 
 	}
 
 	if (skip_file(fname, file, &st)) {
-		if (fnamecmp == fname)
+		if (!*fnamecmpbuf)
 			set_perms(fname, file, &st, PERMS_REPORT);
 		return;
 	}
 
-	if (dry_run) {
-		write_int(f_out,i);
-		return;
-	}
-
-	if (whole_file > 0) {
-		write_int(f_out,i);
-		write_sum_head(f_out, NULL);
-		return;
+	if (dry_run || whole_file) {
+		statret = -1;
+		goto notify_others;
 	}
+	if (read_batch)
+		goto notify_others;
 
 	/* open the file */
 	fd = do_open(fnamecmp, O_RDONLY, 0);
@@ -509,9 +501,8 @@ static void recv_generator(char *fname, 
 		/* pretend the file didn't exist */
 		if (preserve_hard_links && hard_link_check(file, HL_SKIP))
 			return;
-		write_int(f_out,i);
-		write_sum_head(f_out, NULL);
-		return;
+		statret = -1;
+		goto notify_others;
 	}
 
 	if (verbose > 3) {
@@ -522,14 +513,41 @@ static void recv_generator(char *fname, 
 	if (verbose > 2)
 		rprintf(FINFO, "generating and sending sums for %d\n", i);
 
-	write_int(f_out,i);
-	generate_and_send_sums(fd, st.st_size, f_out);
+notify_others:
+	if (f_nameout >= 0) {
+		uchar lenbuf[3], *lb = lenbuf;
+		int len = statret == -1 ? 0 : strlen(fnamecmpbuf);
+		if (read_batch)
+			write_int(f_nameout, i);
+		if (len > 0x7F) {
+#if MAXPATHLEN > 0x7FFF
+			*lb++ = len / 0x10000 + 0x80;
+			*lb++ = len / 0x100;
+#else
+			*lb++ = len / 0x100 + 0x80;
+#endif
+		}
+		*lb = len;
+		write_buf(f_nameout, lenbuf, lb - lenbuf + 1);
+		if (len)
+			write_buf(f_nameout, fnamecmpbuf, len);
+	}
 
-	close(fd);
+	if (read_batch)
+		return;
+
+	write_int(f_out, i);
+	if (statret == 0) {
+		generate_and_send_sums(fd, st.st_size, f_out);
+
+		close(fd);
+	} else if (!dry_run)
+		write_sum_head(f_out, NULL);
 }
 
 
-void generate_files(int f_out, struct file_list *flist, char *local_name)
+void generate_files(int f_out, struct file_list *flist, char *local_name,
+		    int f_nameout)
 {
 	int i;
 	int phase = 0;
@@ -570,7 +588,7 @@ void generate_files(int f_out, struct fi
 		}
 
 		recv_generator(local_name ? local_name : f_name_to(file, fbuf),
-			       file, i, f_out);
+			       file, i, f_out, f_nameout);
 	}
 
 	phase++;
@@ -581,13 +599,15 @@ void generate_files(int f_out, struct fi
 		rprintf(FINFO,"generate_files phase=%d\n",phase);
 
 	write_int(f_out, -1);
+	if (read_batch)
+		write_int(f_nameout, flist->count);
 
 	/* files can cycle through the system more than once
 	 * to catch initial checksum errors */
 	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_out);
+			       file, i, f_out, f_nameout);
 	}
 
 	phase++;
@@ -595,6 +615,8 @@ void generate_files(int f_out, struct fi
 		rprintf(FINFO,"generate_files phase=%d\n",phase);
 
 	write_int(f_out, -1);
+	if (read_batch)
+		write_int(f_nameout, flist->count);
 
 	if (preserve_hard_links)
 		do_hard_links();
@@ -606,7 +628,7 @@ void generate_files(int f_out, struct fi
 		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 > 2)
--- orig/main.c	2004-07-19 17:14:44
+++ main.c	2004-07-17 15:58:11
@@ -57,6 +57,7 @@ extern int filesfrom_fd;
 extern pid_t cleanup_child_pid;
 extern char *files_from;
 extern char *remote_filesfrom_file;
+extern char *compare_dest;
 extern char *rsync_path;
 extern char *shell_cmd;
 extern char *batch_name;
@@ -444,20 +445,21 @@ static int do_recv(int f_in,int f_out,st
 {
 	int pid;
 	int status = 0;
-	int error_pipe[2];
+	int error_pipe[2], name_pipe[2];
+	int need_name_pipe = compare_dest || read_batch;
 
 	if (preserve_hard_links)
 		init_hard_links(flist);
 
 	if (!delete_after) {
 		/* I moved this here from recv_files() to prevent a race condition */
-		if (recurse && delete_mode && !local_name && flist->count>0) {
+		if (recurse && delete_mode && !local_name && flist->count > 0)
 			delete_files(flist);
-		}
 	}
 
-	if (fd_pair(error_pipe) < 0) {
-		rprintf(FERROR,"error pipe failed in do_recv\n");
+	if (fd_pair(error_pipe) < 0
+	    || (need_name_pipe && fd_pair(name_pipe) < 0)) {
+		rprintf(FERROR, "fd_pair() failed in do_recv\n");
 		exit_cleanup(RERR_SOCKETIO);
 	}
 
@@ -465,6 +467,11 @@ static int do_recv(int f_in,int f_out,st
 
 	if ((pid = do_fork()) == 0) {
 		close(error_pipe[0]);
+		if (need_name_pipe) {
+			close(name_pipe[1]);
+			set_blocking(name_pipe[0]);
+		} else
+			name_pipe[0] = -1;
 		if (f_in != f_out)
 			close(f_out);
 
@@ -474,7 +481,7 @@ 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);
 
@@ -492,6 +499,11 @@ static int do_recv(int f_in,int f_out,st
 		stop_write_batch();
 
 	close(error_pipe[1]);
+	if (need_name_pipe) {
+		close(name_pipe[0]);
+		set_nonblocking(name_pipe[1]);
+	} else
+		name_pipe[1] = -1;
 	if (f_in != f_out)
 		close(f_in);
 
@@ -499,7 +511,7 @@ static int do_recv(int f_in,int f_out,st
 
 	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);
--- orig/receiver.c	2004-07-20 21:36:07
+++ receiver.c	2004-07-19 16:44:39
@@ -28,6 +28,7 @@ extern int max_delete;
 extern int csum_length;
 extern struct stats stats;
 extern int dry_run;
+extern int read_batch;
 extern int am_server;
 extern int relative_paths;
 extern int keep_dirlinks;
@@ -318,6 +319,30 @@ static int receive_data(int f_in, char *
 }
 
 
+static char *read_gen_name(int fd, char *buf, char *realname)
+{
+	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) {
+		if (len >= MAXPATHLEN) {
+			rprintf(FERROR, "bogus data on generator name pipe\n");
+			exit_cleanup(RERR_PROTOCOL);
+		}
+		read_sbuf(fd, buf, len);
+		return buf;
+	}
+	return realname;
+}
+
+
 static void discard_receive_data(int f_in, OFF_T length)
 {
 	receive_data(f_in, NULL, -1, 0, NULL, -1, length);
@@ -328,8 +353,10 @@ static void discard_receive_data(int f_i
  * 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_in)
 {
+	int next_gen_i = -1;
 	int fd1,fd2;
 	STRUCT_STAT st;
 	char *fname, fbuf[MAXPATHLEN];
@@ -355,6 +382,17 @@ int recv_files(int f_in, struct file_lis
 
 		i = read_int(f_in);
 		if (i == -1) {
+			if (read_batch) {
+				if (next_gen_i < 0)
+					next_gen_i = read_int(f_name_in);
+				while (next_gen_i < flist->count) {
+					read_gen_name(f_name_in, fnamecmpbuf,
+						      NULL);
+					next_gen_i = read_int(f_name_in);
+				}
+				next_gen_i = -1;
+			}
+
 			if (phase)
 				break;
 
@@ -397,7 +435,26 @@ int recv_files(int f_in, struct file_lis
 		if (verbose > 2)
 			rprintf(FINFO,"recv_files(%s)\n",fname);
 
-		fnamecmp = fname;
+		if (read_batch) {
+			if (next_gen_i < 0)
+				next_gen_i = read_int(f_name_in);
+			while (i > next_gen_i) {
+				read_gen_name(f_name_in, fnamecmpbuf, NULL);
+				next_gen_i = read_int(f_name_in);
+			}
+			if (i < next_gen_i) {
+				rprintf(FINFO, "skipping update for \"%s\"\n",
+					fname);
+				discard_receive_data(f_in, file->length);
+				continue;
+			}
+			next_gen_i = -1;
+		}
+
+		if (f_name_in >= 0)
+			fnamecmp = read_gen_name(f_name_in, fnamecmpbuf, fname);
+		else
+			fnamecmp = fname;
 
 		if (server_exclude_list.head
 		    && check_exclude(&server_exclude_list, fname,
@@ -414,14 +471,6 @@ int recv_files(int f_in, struct file_lis
 		/* open the file */
 		fd1 = do_open(fnamecmp, O_RDONLY, 0);
 
-		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));