Fixed hard-linking when some of the files can get skipped. This adds

the FLAG_SKIP_HLINK flag, which gets set on any hard-linked file that
the user wants to skip (e.g. via --ignore-existing, --append, etc.).
The code in hlink.c now deals with the skipped files instead of
triggering an assert() error.

3 files changed, 129 insertions, 19 deletions

diff --git a/generator.c b/generator.c
index 10500a99..125b3132 100644
--- a/generator.c
+++ b/generator.c
@@ -135,8 +135,12 @@ enum delret {
     DR_SUCCESS = 0, DR_FAILURE, DR_AT_LIMIT, DR_NOT_EMPTY
 };
 
-/* Forward declaration for delete_item(). */
+/* Forward declarations. */
 static enum delret delete_dir_contents(char *fname, uint16 flags);
+#ifdef SUPPORT_HARD_LINKS
+static void handle_skipped_hlink(struct file_struct *file, int itemizing,
+				 enum logcode code, int f_out);
+#endif
 
 static int is_backup_file(char *fn)
 {
@@ -1215,7 +1219,8 @@ static int dflt_perms;
  * regular files that have changed, we try to find a basis file and then
  * start sending checksums.  The ndx is the file's unique index value.
  *
- * When fname is non-null, it must point to a MAXPATHLEN buffer!
+ * The fname parameter must point to a MAXPATHLEN buffer!  (e.g it gets
+ * passed to delete_item(), which can use it during a recursive delete.)
  *
  * Note that f_out is set to -1 when doing final directory-permission and
  * modification-time repair. */
@@ -1269,6 +1274,10 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 				excluded_dir = file;
 			}
 		  skipping:
+#ifdef SUPPORT_HARD_LINKS
+			if (F_IS_HLINKED(file))
+				handle_skipped_hlink(file, itemizing, code, f_out);
+#endif
 			rprintf(FERROR_XFER,
 				"skipping daemon-excluded file \"%s\"\n",
 				fname);
@@ -1285,6 +1294,10 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 		} else if (!dry_run) {
 			if (is_dir)
 				file->flags |= FLAG_MISSING_DIR;
+#ifdef SUPPORT_HARD_LINKS
+			if (F_IS_HLINKED(file))
+				handle_skipped_hlink(file, itemizing, code, f_out);
+#endif
 			return;
 		}
 	}
@@ -1347,6 +1360,10 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 			}
 			file->flags |= FLAG_MISSING_DIR;
 		}
+#ifdef SUPPORT_HARD_LINKS
+		else if (F_IS_HLINKED(file))
+			handle_skipped_hlink(file, itemizing, code, f_out);
+#endif
 		if (verbose > 1) {
 			rprintf(FINFO, "not creating new %s \"%s\"\n",
 				is_dir ? "directory" : "file", fname);
@@ -1362,6 +1379,10 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 	 && (!is_dir || !S_ISDIR(sx.st.st_mode))) {
 		if (verbose > 1 && is_dir >= 0)
 			rprintf(FINFO, "%s exists\n", fname);
+#ifdef SUPPORT_HARD_LINKS
+		if (F_IS_HLINKED(file))
+			handle_skipped_hlink(file, itemizing, code, f_out);
+#endif
 		goto cleanup;
 	}
 
@@ -1674,6 +1695,10 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 	    && cmp_time(sx.st.st_mtime, file->modtime) > 0) {
 		if (verbose > 1)
 			rprintf(FINFO, "%s is newer\n", fname);
+#ifdef SUPPORT_HARD_LINKS
+		if (F_IS_HLINKED(file))
+			handle_skipped_hlink(file, itemizing, code, f_out);
+#endif
 		goto cleanup;
 	}
 
@@ -1768,6 +1793,10 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 	}
 
 	if (append_mode > 0 && sx.st.st_size >= F_LENGTH(file)) {
+#ifdef SUPPORT_HARD_LINKS
+		if (F_IS_HLINKED(file))
+			handle_skipped_hlink(file, itemizing, code, f_out);
+#endif
 		goto cleanup;
 	}
 
@@ -1927,6 +1956,28 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
 	return;
 }
 
+#ifdef SUPPORT_HARD_LINKS
+static void handle_skipped_hlink(struct file_struct *file, int itemizing,
+				 enum logcode code, int f_out)
+{
+	char fbuf[MAXPATHLEN];
+	int new_last_ndx;
+	struct file_list *save_flist = cur_flist;
+
+	/* If we skip the last item in a chain of links and there was a
+	 * prior non-skipped hard-link waiting to finish, finish it now. */
+	if ((new_last_ndx = skip_hard_link(file, &cur_flist)) < 0)
+		return;
+
+	file = cur_flist->files[new_last_ndx - cur_flist->ndx_start];
+	cur_flist->in_progress--; /* undo prior increment */
+	f_name(file, fbuf);
+	recv_generator(fbuf, file, new_last_ndx, itemizing, code, f_out);
+
+	cur_flist = save_flist;
+}
+#endif
+
 static void touch_up_dirs(struct file_list *flist, int ndx)
 {
 	static int counter = 0;
diff --git a/hlink.c b/hlink.c
index 2a8b77eb..536b571f 100644
--- a/hlink.c
+++ b/hlink.c
@@ -125,12 +125,16 @@ static void match_gnums(int32 *ndx_list, int ndx_count)
 				prev = -1;
 			} else if (CVAL(node->data, 0) == 0) {
 				struct file_list *flist;
-				struct file_struct *fp;
 				prev = IVAL(node->data, 1);
 				flist = flist_for_ndx(prev);
-				assert(flist != NULL);
-				fp = flist->files[prev - flist->ndx_start];
-				fp->flags &= ~FLAG_HLINK_LAST;
+				if (flist)
+					flist->files[prev - flist->ndx_start]->flags &= ~FLAG_HLINK_LAST;
+				else {
+					/* We skipped all prior files in this
+					 * group, so mark this as a "first". */
+					file->flags |= FLAG_HLINK_FIRST;
+					prev = -1;
+				}
 			} else
 				prev = -1;
 		} else {
@@ -240,20 +244,45 @@ static int maybe_hard_link(struct file_struct *file, int ndx,
 }
 
 /* Figure out if a prior entry is still there or if we just have a
- * cached name for it.  Never called with a FLAG_HLINK_FIRST entry. */
-static char *check_prior(int prev_ndx, int gnum, struct file_list **flist_p)
+ * cached name for it. */
+static char *check_prior(struct file_struct *file, int gnum,
+			 int *prev_ndx_p, struct file_list **flist_p)
 {
-	struct file_list *flist = flist_for_ndx(prev_ndx);
+	struct file_struct *fp;
+	struct file_list *flist;
 	struct ht_int32_node *node;
+	int prev_ndx = F_HL_PREV(file);
 
-	if (flist) {
-		*flist_p = flist;
-		return NULL;
+	while (1) {
+		if (prev_ndx < 0) {
+			*prev_ndx_p = prev_ndx;
+			*flist_p = NULL;
+			return NULL;
+		}
+		if ((flist = flist_for_ndx(prev_ndx)) == NULL)
+			break;
+		fp = flist->files[prev_ndx - flist->ndx_start];
+		if (!(fp->flags & FLAG_SKIP_HLINK)) {
+			*prev_ndx_p = prev_ndx;
+			*flist_p = flist;
+			return NULL;
+		}
+		F_HL_PREV(file) = prev_ndx = F_HL_PREV(fp);
 	}
 
 	node = hashtable_find(prior_hlinks, gnum, 0);
 	assert(node != NULL && node->data);
-	assert(CVAL(node->data, 0) != 0);
+
+	if (CVAL(node->data, 0) == 0) {
+		/* The prior file must have been skipped. */
+		F_HL_PREV(file) = prev_ndx = -1;
+		*prev_ndx_p = prev_ndx;
+		*flist_p = NULL;
+		return NULL;
+	}
+
+	*prev_ndx_p = prev_ndx;
+	*flist_p = flist;
 	return node->data;
 }
 
@@ -268,12 +297,20 @@ int hard_link_check(struct file_struct *file, int ndx, const char *fname,
 	char *realname, *prev_name;
 	struct file_list *flist;
 	int gnum = inc_recurse ? F_HL_GNUM(file) : -1;
-	int prev_ndx = F_HL_PREV(file);
+	int prev_ndx;
 
-	prev_name = realname = check_prior(prev_ndx, gnum, &flist);
+	prev_name = realname = check_prior(file, gnum, &prev_ndx, &flist);
 
 	if (!prev_name) {
-		struct file_struct *prev_file = flist->files[prev_ndx - flist->ndx_start];
+		struct file_struct *prev_file;
+
+		if (!flist) {
+			/* The previous file was skipped, so this one is
+			 * treated as if it were the first in its group. */
+			return 0;
+		}
+
+		prev_file = flist->files[prev_ndx - flist->ndx_start];
 
 		/* Is the previous link not complete yet? */
 		if (!(prev_file->flags & FLAG_HLINK_DONE)) {
@@ -294,8 +331,8 @@ int hard_link_check(struct file_struct *file, int ndx, const char *fname,
 		/* There is a finished file to link with! */
 		if (!(prev_file->flags & FLAG_HLINK_FIRST)) {
 			/* The previous previous is FIRST when prev is not. */
-			prev_ndx = F_HL_PREV(prev_file);
-			prev_name = realname = check_prior(prev_ndx, gnum, &flist);
+			prev_name = realname = check_prior(prev_file, gnum, &prev_ndx, &flist);
+			assert(prev_name != NULL || flist != NULL);
 			/* Update our previous pointer to point to the FIRST. */
 			F_HL_PREV(file) = prev_ndx;
 		}
@@ -474,4 +511,25 @@ void finish_hard_link(struct file_struct *file, const char *fname, int fin_ndx,
 			out_of_memory("finish_hard_link");
 	}
 }
+
+int skip_hard_link(struct file_struct *file, struct file_list **flist_p)
+{
+	struct file_list *flist;
+	int prev_ndx;
+
+	file->flags |= FLAG_SKIP_HLINK;
+	if (!(file->flags & FLAG_HLINK_LAST))
+		return -1;
+
+	check_prior(file, F_HL_GNUM(file), &prev_ndx, &flist);
+	if (prev_ndx >= 0) {
+		file = flist->files[prev_ndx - flist->ndx_start];
+		if (file->flags & (FLAG_HLINK_DONE|FLAG_FILE_SENT))
+			return -1;
+		file->flags |= FLAG_HLINK_LAST;
+		*flist_p = flist;
+	}
+
+	return prev_ndx;
+}
 #endif
diff --git a/rsync.h b/rsync.h
index c1225563..c6293d0f 100644
--- a/rsync.h
+++ b/rsync.h
@@ -67,7 +67,8 @@
 #define FLAG_FILE_SENT (1<<1)	/* sender/receiver/generator */
 #define FLAG_DIR_CREATED (1<<1)	/* generator */
 #define FLAG_CONTENT_DIR (1<<2)	/* sender/receiver/generator */
-#define FLAG_MOUNT_DIR (1<<3)	/* sender/generator */
+#define FLAG_MOUNT_DIR (1<<3)	/* sender/generator (dirs only) */
+#define FLAG_SKIP_HLINK (1<<3)	/* receiver/generator (w/FLAG_HLINKED) */
 #define FLAG_DUPLICATE (1<<4)	/* sender */
 #define FLAG_MISSING_DIR (1<<4)	/* generator */
 #define FLAG_HLINKED (1<<5)	/* receiver/generator (checked on all types) */