From bc451aba36b4d64a1576cb4ae84f1ef475c01326 Mon Sep 17 00:00:00 2001
From: Moinak Ghosh <moinakg@gmail.com>
Date: Tue, 22 Oct 2013 23:41:51 +0530
Subject: [PATCH] Archiving support using Libarchive: Work in progress changes
 #2.

---
 README.md            |  2 +-
 archive/pc_archive.c | 90 ++++++++++++++++++++++++++++++++++++++++----
 archive/pc_archive.h |  1 +
 pcompress.c          |  9 +++++
 4 files changed, 94 insertions(+), 8 deletions(-)

diff --git a/README.md b/README.md
index 3c2713b..7a87c3f 100644
--- a/README.md
+++ b/README.md
@@ -26,7 +26,7 @@ repeated allocation of similar chunks. It can work in pipe mode, reading
 from stdin and writing to stdout. It also provides adaptive compression
 modes in which data analysis heuristics are used to identify near-optimal
 algorithms per chunk. Finally it supports 14 compression levels to allow
-for ultra compression modes in some algorithms.
+for ultra compression parameters in some algorithms.
 
 Pcompress also supports encryption via AES and uses Scrypt from Tarsnap
 for Password Based Key generation. A unique key is generated per session
diff --git a/archive/pc_archive.c b/archive/pc_archive.c
index e158775..e754b34 100644
--- a/archive/pc_archive.c
+++ b/archive/pc_archive.c
@@ -41,6 +41,16 @@
 #include <ftw.h>
 #include <stdint.h>
 
+/*
+AE_IFREG   Regular file
+AE_IFLNK   Symbolic link
+AE_IFSOCK  Socket
+AE_IFCHR   Character device
+AE_IFBLK   Block device
+AE_IFDIR   Directory
+AE_IFIFO   Named pipe (fifo)
+*/
+
 #define	ARC_ENTRY_OVRHEAD	500
 static struct arc_list_state {
 	uchar_t *pbuf;
@@ -58,6 +68,7 @@ add_pathname(const char *fpath, const struct stat *sb,
 {
 	short len;
 	uchar_t *buf;
+	struct hdr ehdr;
 
 	if (tflag == FTW_DP) return (0);
 	if (tflag == FTW_DNR || tflag == FTW_NS) {
@@ -78,11 +89,7 @@ add_pathname(const char *fpath, const struct stat *sb,
 	*((short *)buf) = len;
 	buf += 2;
 	memcpy(buf, fpath, len);
-	buf += len;
-	*((int *)buf) = tflag;
-	buf += 4;
-	*((uint64_t *)buf) = sb->st_size;
-	a_state.bufpos += (len + 14);
+	a_state.bufpos += (len + 2);
 	return (0);
 }
 
@@ -162,6 +169,7 @@ setup_archive(pc_ctx_t *pctx, struct stat *sbuf)
 	archive_write_set_format_pax_restricted(arc);
 	archive_write_open_fd(arc, pctx->archive_data_fd);
 	pctx->archive_ctx = arc;
+	pctx->archive_members_fd = fd;
 
 	return (0);
 }
@@ -170,8 +178,76 @@ setup_archive(pc_ctx_t *pctx, struct stat *sbuf)
  * Thread function. Archive members and write to pipe. The dispatcher thread
  * reads from the other end and compresses.
  */
-void *
-run_archiver(void *dat) {
+static void *
+archiver_thread(void *dat) {
+	pc_ctx_t *pctx = (pc_ctx_t *)dat;
+	char fpath[PATH_MAX], *name;
+	ssize_t rbytes;
+	short namelen;
+	int warn;
+	struct stat sbuf;
+	struct archive_entry *entry;
+	struct archive *arc, *ard;
+	struct archive_entry_linkresolver *resolver;
+
+	warn = 1;
+	entry = archive_entry_new();
+	arc = (struct archive *)(pctx->archive_ctx);
+
+	if ((resolver = archive_entry_linkresolver_new()) != NULL) {
+		archive_entry_linkresolver_set_strategy(resolver, archive_format(arc));
+	} else {
+		log_msg(LOG_WARN, 0, "Cannot create link resolver, hardlinks will be duplicated.");
+	}
+
+	ard = archive_read_disk_new();
+	archive_read_disk_set_standard_lookup(ard);
+	archive_read_disk_set_symlink_physical(ard);
+
+	/*
+	 * Read next path entry from list file.
+	 */
+	while ((rbytes = Read(pctx->archive_members_fd, &namelen, sizeof(namelen))) != 0) {
+		int fd;
+
+		if (rbytes < 2) break;
+		rbytes = Read(pctx->archive_members_fd, fpath, namelen);
+		if (rbytes < namelen) break;
+		archive_entry_copy_sourcepath(entry, fpath);
+		if (archive_read_disk_entry_from_file(ard, entry, 0, NULL) != ARCHIVE_OK) {
+			log_msg(LOG_WARN, 0, "%s", archive_error_string(ard);
+			archive_entry_clear(entry);
+			continue;
+		}
+
+		/*
+		 * Strip leading '/' or '../' or '/../' from member name.
+		 */
+		name = fpath;
+		while (name[0] == '/' || name[0] == '\\') {
+			if (warn) {
+				log_msg(LOG_WARN, 0, "Converting absolute paths.");
+				warn = 0;
+			}
+			if (name[1] == '.' && name[2] == '.' && (name[3] == '/' || name[3] == '\\')) {
+				name += 4; /* /.. is removed here and / is removed next. */
+			} else {
+				name += 1;
+			}
+		}
+		if (name != archive_entry_pathname(entry))
+			archive_entry_copy_pathname(entry, name);
+
+		if (archive_entry_filetype(entry) != AE_IFREG)
+			archive_entry_set_size(entry, 0);
+		archive_entry_linkify(bsdtar->resolver, &entry, &spare_entry);
+		archive_entry_write_header(arc, entry);
+		archive_entry_clear(entry);
+	}
 	return (NULL);
 }
 
+int
+start_archiver(pc_ctx_t *pctx) {
+	return (0);
+}
diff --git a/archive/pc_archive.h b/archive/pc_archive.h
index a83f12f..53e9c7d 100644
--- a/archive/pc_archive.h
+++ b/archive/pc_archive.h
@@ -42,6 +42,7 @@ typedef struct {
  * Archiving related functions.
  */
 int setup_archive(pc_ctx_t *pctx, struct stat *sbuf);
+int start_archiver(pc_ctx_t *pctx);
 
 #ifdef	__cplusplus
 }
diff --git a/pcompress.c b/pcompress.c
index 5962faa..3c64fe4 100644
--- a/pcompress.c
+++ b/pcompress.c
@@ -2055,6 +2055,15 @@ start_compress(pc_ctx_t *pctx, const char *filename, uint64_t chunksize, int lev
 	}
 	wthread = 1;
 
+	/*
+	 * Start the archiver thread if needed.
+	 */
+	if (pctx->archive_mode) {
+		if (start_archiver(pctx) != 0) {
+			COMP_BAIL;
+		}
+	}
+
 	/*
 	 * Write out file header. First insert hdr elements into mem buffer
 	 * then write out the full hdr in one shot.