From: Giovanni Scafora <giovanni.archlinux.org>
Subject: unzip files encoded with non-latin, non-unicode file names
Last-Update: 2015-02-11
Updated 2015-02-11 by Marc Deslauriers <marc.deslauriers@canonical.com>
to fix buffer overflow in charset_to_intern()
Index: unzip-6.0/unix/unix.c
===================================================================
--- unzip-6.0.orig/unix/unix.c 2015-02-11 08:46:43.675324290 -0500
+++ unzip-6.0/unix/unix.c 2015-02-11 09:18:04.902081319 -0500
@@ -30,6 +30,9 @@
#define UNZIP_INTERNAL
#include "unzip.h"
+#include <iconv.h>
+#include <langinfo.h>
+
#ifdef SCO_XENIX
# define SYSNDIR
#else /* SCO Unix, AIX, DNIX, TI SysV, Coherent 4.x, ... */
@@ -1874,3 +1877,102 @@
}
}
#endif /* QLZIP */
+
+
+typedef struct {
+ char *local_charset;
+ char *archive_charset;
+} CHARSET_MAP;
+
+/* A mapping of local <-> archive charsets used by default to convert filenames
+ * of DOS/Windows Zip archives. Currently very basic. */
+static CHARSET_MAP dos_charset_map[] = {
+ { "ANSI_X3.4-1968", "CP850" },
+ { "ISO-8859-1", "CP850" },
+ { "CP1252", "CP850" },
+ { "UTF-8", "CP866" },
+ { "KOI8-R", "CP866" },
+ { "KOI8-U", "CP866" },
+ { "ISO-8859-5", "CP866" }
+};
+
+char OEM_CP[MAX_CP_NAME] = "";
+char ISO_CP[MAX_CP_NAME] = "";
+
+/* Try to guess the default value of OEM_CP based on the current locale.
+ * ISO_CP is left alone for now. */
+void init_conversion_charsets()
+{
+ const char *local_charset;
+ int i;
+
+ /* Make a guess only if OEM_CP not already set. */
+ if(*OEM_CP == '\0') {
+ local_charset = nl_langinfo(CODESET);
+ for(i = 0; i < sizeof(dos_charset_map)/sizeof(CHARSET_MAP); i++)
+ if(!strcasecmp(local_charset, dos_charset_map[i].local_charset)) {
+ strncpy(OEM_CP, dos_charset_map[i].archive_charset,
+ sizeof(OEM_CP));
+ break;
+ }
+ }
+}
+
+/* Convert a string from one encoding to the current locale using iconv().
+ * Be as non-intrusive as possible. If error is encountered during covertion
+ * just leave the string intact. */
+static void charset_to_intern(char *string, char *from_charset)
+{
+ iconv_t cd;
+ char *s,*d, *buf;
+ size_t slen, dlen, buflen;
+ const char *local_charset;
+
+ if(*from_charset == '\0')
+ return;
+
+ buf = NULL;
+ local_charset = nl_langinfo(CODESET);
+
+ if((cd = iconv_open(local_charset, from_charset)) == (iconv_t)-1)
+ return;
+
+ slen = strlen(string);
+ s = string;
+
+ /* Make sure OUTBUFSIZ + 1 never ends up smaller than FILNAMSIZ
+ * as this function also gets called with G.outbuf in fileio.c
+ */
+ buflen = FILNAMSIZ;
+ if (OUTBUFSIZ + 1 < FILNAMSIZ)
+ {
+ buflen = OUTBUFSIZ + 1;
+ }
+
+ d = buf = malloc(buflen);
+ if(!d)
+ goto cleanup;
+
+ bzero(buf,buflen);
+ dlen = buflen - 1;
+
+ if(iconv(cd, &s, &slen, &d, &dlen) == (size_t)-1)
+ goto cleanup;
+ strncpy(string, buf, buflen);
+
+ cleanup:
+ free(buf);
+ iconv_close(cd);
+}
+
+/* Convert a string from OEM_CP to the current locale charset. */
+inline void oem_intern(char *string)
+{
+ charset_to_intern(string, OEM_CP);
+}
+
+/* Convert a string from ISO_CP to the current locale charset. */
+inline void iso_intern(char *string)
+{
+ charset_to_intern(string, ISO_CP);
+}
Index: unzip-6.0/unix/unxcfg.h
===================================================================
--- unzip-6.0.orig/unix/unxcfg.h 2015-02-11 08:46:43.675324290 -0500
+++ unzip-6.0/unix/unxcfg.h 2015-02-11 08:46:43.671324260 -0500
@@ -228,4 +228,30 @@
/* wild_dir, dirname, wildname, matchname[], dirnamelen, have_dirname, */
/* and notfirstcall are used by do_wild(). */
+
+#define MAX_CP_NAME 25
+
+#ifdef SETLOCALE
+# undef SETLOCALE
+#endif
+#define SETLOCALE(category, locale) setlocale(category, locale)
+#include <locale.h>
+
+#ifdef _ISO_INTERN
+# undef _ISO_INTERN
+#endif
+#define _ISO_INTERN(str1) iso_intern(str1)
+
+#ifdef _OEM_INTERN
+# undef _OEM_INTERN
+#endif
+#ifndef IZ_OEM2ISO_ARRAY
+# define IZ_OEM2ISO_ARRAY
+#endif
+#define _OEM_INTERN(str1) oem_intern(str1)
+
+void iso_intern(char *);
+void oem_intern(char *);
+void init_conversion_charsets(void);
+
#endif /* !__unxcfg_h */
Index: unzip-6.0/unzip.c
===================================================================
--- unzip-6.0.orig/unzip.c 2015-02-11 08:46:43.675324290 -0500
+++ unzip-6.0/unzip.c 2015-02-11 08:46:43.675324290 -0500
@@ -327,12 +327,23 @@
-2 just filenames but allow -h/-t/-z -l long Unix \"ls -l\" format\n\
-v verbose, multi-page format\n";
+#ifndef UNIX
static ZCONST char Far ZipInfoUsageLine3[] = "miscellaneous options:\n\
-h print header line -t print totals for listed files or for all\n\
-z print zipfile comment -T print file times in sortable decimal format\
\n -C be case-insensitive %s\
-U use escapes for all non-ASCII Unicode\n\
-x exclude filenames that follow from listing\n";
+#else /* UNIX */
+static ZCONST char Far ZipInfoUsageLine3[] = "miscellaneous options:\n\
+ -h print header line -t print totals for listed files or for all\n\
+ -z print zipfile comment %c-T%c print file times in sortable decimal format\
+\n %c-C%c be case-insensitive %s\
+ -U use escapes for all non-ASCII Unicode\n\
+ -x exclude filenames that follow from listing\n\
+ -O CHARSET specify a character encoding for DOS, Windows and OS/2 archives\n\
+ -I CHARSET specify a character encoding for UNIX and other archives\n";
+#endif /* !UNIX */
#ifdef MORE
static ZCONST char Far ZipInfoUsageLine4[] =
" -M page output through built-in \"more\"\n";
@@ -664,6 +674,17 @@
-U use escapes for all non-ASCII Unicode -UU ignore any Unicode fields\n\
-C match filenames case-insensitively -L make (some) names \
lowercase\n %-42s -V retain VMS version numbers\n%s";
+#elif (defined UNIX)
+static ZCONST char Far UnzipUsageLine4[] = "\
+modifiers:\n\
+ -n never overwrite existing files -q quiet mode (-qq => quieter)\n\
+ -o overwrite files WITHOUT prompting -a auto-convert any text files\n\
+ -j junk paths (do not make directories) -aa treat ALL files as text\n\
+ -U use escapes for all non-ASCII Unicode -UU ignore any Unicode fields\n\
+ -C match filenames case-insensitively -L make (some) names \
+lowercase\n %-42s -V retain VMS version numbers\n%s\
+ -O CHARSET specify a character encoding for DOS, Windows and OS/2 archives\n\
+ -I CHARSET specify a character encoding for UNIX and other archives\n\n";
#else /* !VMS */
static ZCONST char Far UnzipUsageLine4[] = "\
modifiers:\n\
@@ -802,6 +823,10 @@
#endif /* UNICODE_SUPPORT */
+#ifdef UNIX
+ init_conversion_charsets();
+#endif
+
#if (defined(__IBMC__) && defined(__DEBUG_ALLOC__))
extern void DebugMalloc(void);
@@ -1335,6 +1360,11 @@
argc = *pargc;
argv = *pargv;
+#ifdef UNIX
+ extern char OEM_CP[MAX_CP_NAME];
+ extern char ISO_CP[MAX_CP_NAME];
+#endif
+
while (++argv, (--argc > 0 && *argv != NULL && **argv == '-')) {
s = *argv + 1;
while ((c = *s++) != 0) { /* "!= 0": prevent Turbo C warning */
@@ -1516,6 +1546,35 @@
}
break;
#endif /* MACOS */
+#ifdef UNIX
+ case ('I'):
+ if (negative) {
+ Info(slide, 0x401, ((char *)slide,
+ "error: encodings can't be negated"));
+ return(PK_PARAM);
+ } else {
+ if(*s) { /* Handle the -Icharset case */
+ /* Assume that charsets can't start with a dash to spot arguments misuse */
+ if(*s == '-') {
+ Info(slide, 0x401, ((char *)slide,
+ "error: a valid character encoding should follow the -I argument"));
+ return(PK_PARAM);
+ }
+ strncpy(ISO_CP, s, sizeof(ISO_CP));
+ } else { /* -I charset */
+ ++argv;
+ if(!(--argc > 0 && *argv != NULL && **argv != '-')) {
+ Info(slide, 0x401, ((char *)slide,
+ "error: a valid character encoding should follow the -I argument"));
+ return(PK_PARAM);
+ }
+ s = *argv;
+ strncpy(ISO_CP, s, sizeof(ISO_CP));
+ }
+ while(*(++s)); /* No params straight after charset name */
+ }
+ break;
+#endif /* ?UNIX */
case ('j'): /* junk pathnames/directory structure */
if (negative)
uO.jflag = FALSE, negative = 0;
@@ -1591,6 +1650,35 @@
} else
++uO.overwrite_all;
break;
+#ifdef UNIX
+ case ('O'):
+ if (negative) {
+ Info(slide, 0x401, ((char *)slide,
+ "error: encodings can't be negated"));
+ return(PK_PARAM);
+ } else {
+ if(*s) { /* Handle the -Ocharset case */
+ /* Assume that charsets can't start with a dash to spot arguments misuse */
+ if(*s == '-') {
+ Info(slide, 0x401, ((char *)slide,
+ "error: a valid character encoding should follow the -I argument"));
+ return(PK_PARAM);
+ }
+ strncpy(OEM_CP, s, sizeof(OEM_CP));
+ } else { /* -O charset */
+ ++argv;
+ if(!(--argc > 0 && *argv != NULL && **argv != '-')) {
+ Info(slide, 0x401, ((char *)slide,
+ "error: a valid character encoding should follow the -O argument"));
+ return(PK_PARAM);
+ }
+ s = *argv;
+ strncpy(OEM_CP, s, sizeof(OEM_CP));
+ }
+ while(*(++s)); /* No params straight after charset name */
+ }
+ break;
+#endif /* ?UNIX */
case ('p'): /* pipes: extract to stdout, no messages */
if (negative) {
uO.cflag = FALSE;
Index: unzip-6.0/unzpriv.h
===================================================================
--- unzip-6.0.orig/unzpriv.h 2015-02-11 08:46:43.675324290 -0500
+++ unzip-6.0/unzpriv.h 2015-02-11 08:46:43.675324290 -0500
@@ -3008,7 +3008,7 @@
!(((islochdr) || (isuxatt)) && \
((hostver) == 25 || (hostver) == 26 || (hostver) == 40))) || \
(hostnum) == FS_HPFS_ || \
- ((hostnum) == FS_NTFS_ && (hostver) == 50)) { \
+ ((hostnum) == FS_NTFS_ /* && (hostver) == 50 */ )) { \
_OEM_INTERN((string)); \
} else { \
_ISO_INTERN((string)); \
Index: unzip-6.0/zipinfo.c
===================================================================
--- unzip-6.0.orig/zipinfo.c 2015-02-11 08:46:43.675324290 -0500
+++ unzip-6.0/zipinfo.c 2015-02-11 08:46:43.675324290 -0500
@@ -457,6 +457,10 @@
int tflag_slm=TRUE, tflag_2v=FALSE;
int explicit_h=FALSE, explicit_t=FALSE;
+#ifdef UNIX
+ extern char OEM_CP[MAX_CP_NAME];
+ extern char ISO_CP[MAX_CP_NAME];
+#endif
#ifdef MACOS
uO.lflag = LFLAG; /* reset default on each call */
@@ -501,6 +505,35 @@
uO.lflag = 0;
}
break;
+#ifdef UNIX
+ case ('I'):
+ if (negative) {
+ Info(slide, 0x401, ((char *)slide,
+ "error: encodings can't be negated"));
+ return(PK_PARAM);
+ } else {
+ if(*s) { /* Handle the -Icharset case */
+ /* Assume that charsets can't start with a dash to spot arguments misuse */
+ if(*s == '-') {
+ Info(slide, 0x401, ((char *)slide,
+ "error: a valid character encoding should follow the -I argument"));
+ return(PK_PARAM);
+ }
+ strncpy(ISO_CP, s, sizeof(ISO_CP));
+ } else { /* -I charset */
+ ++argv;
+ if(!(--argc > 0 && *argv != NULL && **argv != '-')) {
+ Info(slide, 0x401, ((char *)slide,
+ "error: a valid character encoding should follow the -I argument"));
+ return(PK_PARAM);
+ }
+ s = *argv;
+ strncpy(ISO_CP, s, sizeof(ISO_CP));
+ }
+ while(*(++s)); /* No params straight after charset name */
+ }
+ break;
+#endif /* ?UNIX */
case 'l': /* longer form of "ls -l" type listing */
if (negative)
uO.lflag = -2, negative = 0;
@@ -521,6 +554,35 @@
G.M_flag = TRUE;
break;
#endif
+#ifdef UNIX
+ case ('O'):
+ if (negative) {
+ Info(slide, 0x401, ((char *)slide,
+ "error: encodings can't be negated"));
+ return(PK_PARAM);
+ } else {
+ if(*s) { /* Handle the -Ocharset case */
+ /* Assume that charsets can't start with a dash to spot arguments misuse */
+ if(*s == '-') {
+ Info(slide, 0x401, ((char *)slide,
+ "error: a valid character encoding should follow the -I argument"));
+ return(PK_PARAM);
+ }
+ strncpy(OEM_CP, s, sizeof(OEM_CP));
+ } else { /* -O charset */
+ ++argv;
+ if(!(--argc > 0 && *argv != NULL && **argv != '-')) {
+ Info(slide, 0x401, ((char *)slide,
+ "error: a valid character encoding should follow the -O argument"));
+ return(PK_PARAM);
+ }
+ s = *argv;
+ strncpy(OEM_CP, s, sizeof(OEM_CP));
+ }
+ while(*(++s)); /* No params straight after charset name */
+ }
+ break;
+#endif /* ?UNIX */
case 's': /* default: shorter "ls -l" type listing */
if (negative)
uO.lflag = -2, negative = 0;