From 1e18cbb71515a22b2a6f1eb4aaadea461929b834 Mon Sep 17 00:00:00 2001

From: Tim Kientzle <kientzle@acm.org>

Date: Sat, 21 Feb 2015 10:37:48 -0800

Subject: [PATCH] Issue 408: Fix escaped newline parsing

---

 libarchive/archive_read_support_format_mtree.c  | 56 ++++++++++++-------------

 libarchive/test/test_read_format_mtree.c        |  8 +++-

 libarchive/test/test_read_format_mtree.mtree.uu | 20 +++++----

 3 files changed, 45 insertions(+), 39 deletions(-)

diff --git a/libarchive/archive_read_support_format_mtree.c b/libarchive/archive_read_support_format_mtree.c

index c4e7021..397bfee 100644

--- a/libarchive/archive_read_support_format_mtree.c

+++ b/libarchive/archive_read_support_format_mtree.c

@@ -1661,6 +1661,10 @@ parse_escapes(char *src, struct mtree_entry *mentry)

 				c = '\v';

 				++src;

 				break;

+			case '\\':

+				c = '\\';

+				++src;

+				break;

 			}

 		}

 		*dest++ = c;

@@ -1804,8 +1808,7 @@ readline(struct archive_read *a, struct mtree *mtree, char **start, ssize_t limi

 	ssize_t total_size = 0;

 	ssize_t find_off = 0;

 	const void *t;

-	const char *s;

-	void *p;

+	void *nl;

 	char *u;

 	/* Accumulate line in a line buffer. */

@@ -1816,11 +1819,10 @@ readline(struct archive_read *a, struct mtree *mtree, char **start, ssize_t limi

 			return (0);

 		if (bytes_read < 0)

 			return (ARCHIVE_FATAL);

-		s = t;  /* Start of line? */

-		p = memchr(t, '\n', bytes_read);

-		/* If we found '\n', trim the read. */

-		if (p != NULL) {

-			bytes_read = 1 + ((const char *)p) - s;

+		nl = memchr(t, '\n', bytes_read);

+		/* If we found '\n', trim the read to end exactly there. */

+		if (nl != NULL) {

+			bytes_read = ((const char *)nl) - ((const char *)t) + 1;

 		}

 		if (total_size + bytes_read + 1 > limit) {

 			archive_set_error(&a->archive,

@@ -1834,38 +1836,34 @@ readline(struct archive_read *a, struct mtree *mtree, char **start, ssize_t limi

 			    "Can't allocate working buffer");

 			return (ARCHIVE_FATAL);

 		}

+		/* Append new bytes to string. */

 		memcpy(mtree->line.s + total_size, t, bytes_read);

 		__archive_read_consume(a, bytes_read);

 		total_size += bytes_read;

-		/* Null terminate. */

 		mtree->line.s[total_size] = '\0';

-		/* If we found an unescaped '\n', clean up and return. */

+

 		for (u = mtree->line.s + find_off; *u; ++u) {

 			if (u[0] == '\n') {

+				/* Ends with unescaped newline. */

 				*start = mtree->line.s;

 				return total_size;

-			}

-			if (u[0] == '#') {

-				if (p == NULL)

+			} else if (u[0] == '#') {

+				/* Ends with comment sequence #...\n */

+				if (nl == NULL) {

+					/* But we've not found the \n yet */

 					break;

-				*start = mtree->line.s;

-				return total_size;

-			}

-			if (u[0] != '\\')

-				continue;

-			if (u[1] == '\\') {

-				++u;

-				continue;

-			}

-			if (u[1] == '\n') {

-				memmove(u, u + 1,

-				    total_size - (u - mtree->line.s) + 1);

-				--total_size;

-				++u;

-				break;

+				}

+			} else if (u[0] == '\\') {

+				if (u[1] == '\n') {

+					/* Trim escaped newline. */

+					total_size -= 2;

+					mtree->line.s[total_size] = '\0';

+					break;

+				} else if (u[1] != '\0') {

+					/* Skip the two-char escape sequence */

+					++u;

+				}

 			}

-			if (u[1] == '\0')

-				break;

 		}

 		find_off = u - mtree->line.s;

 	}

diff --git a/libarchive/test/test_read_format_mtree.c b/libarchive/test/test_read_format_mtree.c

index efedc41..85a0e2e 100644

--- a/libarchive/test/test_read_format_mtree.c

+++ b/libarchive/test/test_read_format_mtree.c

@@ -110,6 +110,10 @@ test_read_format_mtree1(void)

 	assertEqualString(archive_entry_pathname(ae), "dir2/dir3b/indir3b");

 	assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae);;

+	assertEqualString(archive_entry_pathname(ae), "dir2/dir3b/filename\\with_esc\b\t\fapes");

+	assertEqualInt(archive_entry_filetype(ae), AE_IFREG);

+

+	assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae);;

 	assertEqualString(archive_entry_pathname(ae), "notindir");

 	assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae);;

@@ -151,7 +155,7 @@ test_read_format_mtree1(void)

 	assertEqualInt(archive_entry_mtime(ae), min_time);

 	assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae);;

-	assertEqualInt(19, archive_file_count(a));

+	assertEqualInt(20, archive_file_count(a));

 	assertEqualInt(ARCHIVE_OK, archive_read_close(a));

 	assertEqualInt(ARCHIVE_OK, archive_read_free(a));

 }

diff --git a/libarchive/test/test_read_format_mtree.mtree.uu b/libarchive/test/test_read_format_mtree.mtree.uu

index a0dff18..f1c9d60 100644

--- a/libarchive/test/test_read_format_mtree.mtree.uu

+++ b/libarchive/test/test_read_format_mtree.mtree.uu

@@ -5,14 +5,16 @@ M92!U:60],3@*("XN"F9I;&5<,#0P=VET:%PP-#!S<&%C92!T>7!E/69I;&4*

 M9&ER,B!T>7!E/61I<@H@9&ER,V$@='EP93UD:7(*("!I;F1I

 M9FEL90ID:7(R+V9U;&QI;F1I

 M"B!I;F1I<C(@='EP93UF:6QE"B!D:7(S8B!T>7!E/61I<@H@(&EN9&ER,V(@

-M='EP93UF:6QE"B`@+BX*("XN"FYO=&EN9&ER('1Y<&4]9FEL90ID:7(R+V9U

-M;&QI;F1I

-M:7IE/3!X,`ID:7(R+W-M86QL9FEL92!T>7!E/69I;&4@

-M,#`P,#`P,#`P,#`P,#`P,#`P,#`P,#$*9&ER,B]T;V]S;6%L;&9;;&4@='EP

-M93UF:6QE('-I>F4]+3$*9&ER,B]B:6=F:6QE('1Y<&4]9FEL92!S:7IE/3DR

-M,C,S-S(P,S8X-30W-S4X,#<*9&ER,B]T;V]B:6=F:6QE('1Y<&4]9FEL92!S

-M:7IE/3DR,C,S-S(P,S8X-30W-S4X,#@*9&ER,B]V97)Y;VQD9FEL92!T>7!E

-M/69I;&4@=&EM93TM.3(R,S,W,C`S-C@U-#

-H;&4@='EP93UF:6QE('1I;64]+3DR,C,S-S(P,S8X-30W-S4X,#D*"@``

+M='EP93UF:6QE"B`@9FEL96YA;65<7'=I=&A<"E]E

+M<&4]9FEL92!M;V1E/3`T-#0@

+M<&4]9FEL90ID:7(R+V9U;&QI;F1I

+M:6QE('1Y<&4]9FEL92!S:7IE/3!X,`ID:7(R+W-M86QL9FEL92!T>7!E/69I

+M;&4@

+M;V]S;6%L;&9;;&4@='EP93UF:6QE('-I>F4]+3$*9&ER,B]B:6=F:6QE('1Y

+M<&4]9FEL92!S:7IE/3DR,C,S-S(P,S8X-30W-S4X,#<*9&ER,B]T;V]B:6=F

+M:6QE('1Y<&4]9FEL92!S:7IE/3DR,C,S-S(P,S8X-30W-S4X,#@*9&ER,B]V

+M97)Y;VQD9FEL92!T>7!E/69I;&4@=&EM93TM.3(R,S,W,C`S-C@U-#

+M.`ID:7(R+W1O;V]L9&9;;&4@='EP93UF:6QE('1I;64]+3DR,C,S-S(P,S8X

+*-30W-S4X,#D*"@``

 `

 end

-- 

2.7.4