--- Begin Message ---
Package: unzip
Version: 5.52-3
Severity: important
Tags: patch l10n
Problem: ZIP-files created on Windows have cyrillic file names encoded
in CP866 codepage. unzip incorrectly converts them as being CP850 to
CP1252 and renders cyrillic filenames to be unreadable.
Example:
$ unzip -l pikalov1.zip
Archive: pikalov1.zip
Length Date Time Name
-------- ---- ---- ----
3103866 11-09-05 19:43 05. ╩╚╪║╞ ╫╞М©╪А╢, ╩╚╪║╞.mp3
3193521 11-09-05 19:43 01. Х╫я║ ты ╪╚╝ ╚╩АСч©╝.mp3
3451644 11-09-05 19:43 02. ЪАЗА║щ ты ╪╚╢, ЪАЗА║щ.mp3
2819472 11-09-05 19:43 03.ьу Сы ъА║©, ъА║©.mp3
3375333 11-09-05 19:43 04. е©ъЩ╪╚ ╛ ╪Атяс©.mp3
-------- -------
15943836 5 files
I think this greately affects interoperability with other systems.
It seems it also affects graphical front archiver tools like
file-roller.
Known solution: There is a patch in russian distro Altlinux, that
adds `smart' filename encoding detection based on locale as well as
command line switches to specify file name encoding explicitely.
The details are available here (in Russian):
https://bugzilla.altlinux.ru/show_bug.cgi?id=4871
I attach their patch to this bug report (it is for 5.50, yet it
works for me with 5.52). Example is below:
$ ~/tmp/unzip-5.52/unzip -l pikalov1.zip
Archive: pikalov1.zip
Length Date Time Name
-------- ---- ---- ----
3103866 11-09-05 19:43 05. Я помню любимая, помню.mp3
3193521 11-09-05 19:43 01. Клен ты мой опавший.mp3
3451644 11-09-05 19:43 02. Шаганэ ты моя, Шаганэ.mp3
2819472 11-09-05 19:43 03.Эх вы сани, сани.mp3
3375333 11-09-05 19:43 04. Письмо к матери.mp3
-------- -------
15943836 5 files
-- System Information:
Debian Release: testing/unstable
APT prefers unstable
APT policy: (500, 'unstable'), (500, 'testing')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.8-2-686
Locale: LANG=ru_RU.KOI8-R, LC_CTYPE=ru_RU.KOI8-R (charmap=KOI8-R)
Versions of packages unzip depends on:
ii libc6 2.3.5-6 GNU C Library: Shared libraries an
unzip recommends no packages.
-- no debconf information
diff -Nur unzip-5.50.orig/unix/unix.c unzip-5.50/unix/unix.c
--- unzip-5.50.orig/unix/unix.c 2002-01-22 01:54:42 +0300
+++ unzip-5.50/unix/unix.c 2004-07-27 22:42:19 +0400
@@ -29,6 +29,9 @@
#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, ... */
@@ -1524,3 +1527,89 @@
}
}
#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" },
+ { "CP1251", "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 = dlen = buflen = strlen(string);
+ s = string;
+ d = buf = malloc(buflen + 1);
+ if(!d)
+ goto cleanup;
+
+ 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);
+}
diff -Nur unzip-5.50.orig/unix/unxcfg.h unzip-5.50/unix/unxcfg.h
--- unzip-5.50.orig/unix/unxcfg.h 2001-06-04 03:27:14 +0400
+++ unzip-5.50/unix/unxcfg.h 2004-07-27 22:42:19 +0400
@@ -123,4 +123,30 @@
/* 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 */
diff -Nur unzip-5.50.orig/unzip.c unzip-5.50/unzip.c
--- unzip-5.50.orig/unzip.c 2002-01-27 22:26:16 +0300
+++ unzip-5.50/unzip.c 2004-07-27 22:42:19 +0400
@@ -304,11 +304,21 @@
-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 %c-T%c print file times in sortable decimal format\
\n %c-C%c be case-insensitive %s\
-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\
+ -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
#ifdef VMS
static ZCONST char Far ZipInfoUsageLine4[] =
@@ -589,6 +599,7 @@
#endif /* ?VM_CMS */
#endif /* ?MACOS */
+#ifndef UNIX
static ZCONST char Far UnzipUsageLine4[] = "\
modifiers: -q quiet mode (-qq => quieter)\n\
-n never overwrite existing files -a auto-convert any text files\n\
@@ -596,6 +607,17 @@
-j junk paths (do not make directories) -v be verbose/print version info\n\
%c-C%c match filenames case-insensitively %c-L%c make (some) names \
lowercase\n %-42s %c-V%c retain VMS version numbers\n%s";
+#else /* UNIX */
+static ZCONST char Far UnzipUsageLine4[] = "\
+modifiers: -q quiet mode (-qq => quieter)\n\
+ -n never overwrite existing files -a auto-convert any text files\n\
+ -o overwrite files WITHOUT prompting -aa treat ALL files as text\n \
+ -j junk paths (do not make directories) -v be verbose/print version info\n\
+ %c-C%c match filenames case-insensitively %c-L%c make (some) names \
+lowercase\n %-42s %c-V%c retain VMS version numbers\n%s\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\n";
+#endif /* !UNIX */
static ZCONST char Far UnzipUsageLine5[] = "\
Examples (see unzip.txt for more info):\n\
@@ -656,6 +678,10 @@
SETLOCALE(LC_CTYPE,"");
+#ifdef UNIX
+ init_conversion_charsets();
+#endif
+
#if (defined(__IBMC__) && defined(__DEBUG_ALLOC__))
extern void DebugMalloc(void);
@@ -1070,6 +1096,11 @@
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 */
@@ -1233,6 +1264,35 @@
}
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;
@@ -1299,6 +1359,35 @@
} 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;
diff -Nur unzip-5.50.orig/unzpriv.h unzip-5.50/unzpriv.h
--- unzip-5.50.orig/unzpriv.h 2002-02-17 20:01:48 +0300
+++ unzip-5.50/unzpriv.h 2004-07-27 23:53:08 +0400
@@ -1103,7 +1103,9 @@
# define lastchar(ptr, len) (ptr[(len)-1])
# define MBSCHR(str, c) strchr(str, c)
# define MBSRCHR(str, c) strrchr(str, c)
-# define SETLOCALE(category, locale)
+# ifndef SETLOCALE
+# define SETLOCALE(category, locale)
+# endif
#endif /* ?_MBCS */
#define INCSTR(ptr) PREINCSTR(ptr)
@@ -2424,7 +2426,7 @@
!(((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)); \
diff -Nur unzip-5.50.orig/zipinfo.c unzip-5.50/zipinfo.c
--- unzip-5.50.orig/zipinfo.c 2001-12-26 01:56:40 +0300
+++ unzip-5.50/zipinfo.c 2004-07-27 05:54:16 +0400
@@ -440,6 +440,10 @@
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 */
@@ -484,6 +488,35 @@
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;
@@ -504,6 +537,35 @@
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;
--- End Message ---
