From 31d289a4759909d24dd309ac24d42902a8c20da0 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?L=C3=A9o=20Le=20Bouter?= Date: Wed, 10 Mar 2021 08:29:53 +0100 Subject: [PATCH] gnu: unzip: Add patches from Fedora [security fixes]. Non-exhaustively fixes CVE-2016-9844, CVE-2018-1000035, CVE-2018-18384, and CVE-2019-13232. * gnu/packages/patches/unzip-COVSCAN-fix-unterminated-string.patch, gnu/packages/patches/unzip-CVE-2016-9844.patch, gnu/packages/patches/unzip-CVE-2018-1000035.patch, gnu/packages/patches/unzip-CVE-2018-18384.patch, gnu/packages/patches/unzip-case-insensitive.patch, gnu/packages/patches/unzip-alt-iconv-utf8-print.patch, gnu/packages/patches/unzip-alt-iconv-utf8.patch, gnu/packages/patches/unzip-close.patch, gnu/packages/patches/unzip-exec-shield.patch, gnu/packages/patches/unzip-fix-recmatch.patch, gnu/packages/patches/unzip-manpage-fix.patch, gnu/packages/patches/unzip-overflow.patch, gnu/packages/patches/unzip-symlink.patch, gnu/packages/patches/unzip-timestamp.patch, gnu/packages/patches/unzip-valgrind.patch, gnu/packages/patches/unzip-x-option.patch, gnu/packages/patches/unzip-zipbomb-manpage.patch, gnu/packages/patches/unzip-zipbomb-part1.patch, gnu/packages/patches/unzip-zipbomb-part2.patch, gnu/packages/patches/unzip-zipbomb-part3.patch: New patches. * gnu/local.mk (dist_patch_DATA): Register them. * gnu/packages/compression.scm (unzip/fixed): New variable. Apply patches. (unzip)[replacement]: Graft. --- gnu/local.mk | 20 + gnu/packages/compression.scm | 49 ++ ...nzip-COVSCAN-fix-unterminated-string.patch | 131 +++++ .../patches/unzip-CVE-2016-9844.patch | 39 ++ .../patches/unzip-CVE-2018-1000035.patch | 34 ++ .../patches/unzip-CVE-2018-18384.patch | 35 ++ .../patches/unzip-alt-iconv-utf8-print.patch | 381 ++++++++++++++ .../patches/unzip-alt-iconv-utf8.patch | 398 +++++++++++++++ .../patches/unzip-caseinsensitive.patch | 131 +++++ gnu/packages/patches/unzip-close.patch | 176 +++++++ gnu/packages/patches/unzip-exec-shield.patch | 10 + gnu/packages/patches/unzip-fix-recmatch.patch | 477 ++++++++++++++++++ gnu/packages/patches/unzip-manpage-fix.patch | 11 + gnu/packages/patches/unzip-overflow.patch | 25 + gnu/packages/patches/unzip-symlink.patch | 17 + gnu/packages/patches/unzip-timestamp.patch | 41 ++ gnu/packages/patches/unzip-valgrind.patch | 26 + gnu/packages/patches/unzip-x-option.patch | 28 + .../patches/unzip-zipbomb-manpage.patch | 25 + .../patches/unzip-zipbomb-part1.patch | 25 + .../patches/unzip-zipbomb-part2.patch | 349 +++++++++++++ .../patches/unzip-zipbomb-part3.patch | 112 ++++ 22 files changed, 2540 insertions(+) create mode 100644 gnu/packages/patches/unzip-COVSCAN-fix-unterminated-string.patch create mode 100644 gnu/packages/patches/unzip-CVE-2016-9844.patch create mode 100644 gnu/packages/patches/unzip-CVE-2018-1000035.patch create mode 100644 gnu/packages/patches/unzip-CVE-2018-18384.patch create mode 100644 gnu/packages/patches/unzip-alt-iconv-utf8-print.patch create mode 100644 gnu/packages/patches/unzip-alt-iconv-utf8.patch create mode 100644 gnu/packages/patches/unzip-caseinsensitive.patch create mode 100644 gnu/packages/patches/unzip-close.patch create mode 100644 gnu/packages/patches/unzip-exec-shield.patch create mode 100644 gnu/packages/patches/unzip-fix-recmatch.patch create mode 100644 gnu/packages/patches/unzip-manpage-fix.patch create mode 100644 gnu/packages/patches/unzip-overflow.patch create mode 100644 gnu/packages/patches/unzip-symlink.patch create mode 100644 gnu/packages/patches/unzip-timestamp.patch create mode 100644 gnu/packages/patches/unzip-valgrind.patch create mode 100644 gnu/packages/patches/unzip-x-option.patch create mode 100644 gnu/packages/patches/unzip-zipbomb-manpage.patch create mode 100644 gnu/packages/patches/unzip-zipbomb-part1.patch create mode 100644 gnu/packages/patches/unzip-zipbomb-part2.patch create mode 100644 gnu/packages/patches/unzip-zipbomb-part3.patch diff --git a/gnu/local.mk b/gnu/local.mk index 551503e053..6500f5bda2 100644 --- a/gnu/local.mk +++ b/gnu/local.mk @@ -1740,6 +1740,26 @@ dist_patch_DATA = \ %D%/packages/patches/unzip-initialize-symlink-flag.patch \ %D%/packages/patches/unzip-overflow-long-fsize.patch \ %D%/packages/patches/unzip-remove-build-date.patch \ + %D%/packages/patches/unzip-case-insensitive.patch \ + %D%/packages/patches/unzip-COVSCAN-fix-unterminated-string.patch \ + %D%/packages/patches/unzip-CVE-2016-9844.patch \ + %D%/packages/patches/unzip-CVE-2018-1000035.patch \ + %D%/packages/patches/unzip-CVE-2018-18384.patch \ + %D%/packages/patches/unzip-alt-iconv-utf8-print.patch \ + %D%/packages/patches/unzip-alt-iconv-utf8.patch \ + %D%/packages/patches/unzip-close.patch \ + %D%/packages/patches/unzip-exec-shield.patch \ + %D%/packages/patches/unzip-fix-recmatch.patch \ + %D%/packages/patches/unzip-manpage-fix.patch \ + %D%/packages/patches/unzip-overflow.patch \ + %D%/packages/patches/unzip-symlink.patch \ + %D%/packages/patches/unzip-timestamp.patch \ + %D%/packages/patches/unzip-valgrind.patch \ + %D%/packages/patches/unzip-x-option.patch \ + %D%/packages/patches/unzip-zipbomb-manpage.patch \ + %D%/packages/patches/unzip-zipbomb-part1.patch \ + %D%/packages/patches/unzip-zipbomb-part2.patch \ + %D%/packages/patches/unzip-zipbomb-part3.patch \ %D%/packages/patches/ustr-fix-build-with-gcc-5.patch \ %D%/packages/patches/util-linux-tests.patch \ %D%/packages/patches/upower-builddir.patch \ diff --git a/gnu/packages/compression.scm b/gnu/packages/compression.scm index 99569f68c9..7a4f70f2b4 100644 --- a/gnu/packages/compression.scm +++ b/gnu/packages/compression.scm @@ -1594,6 +1594,7 @@ (define-public unzip "unzip-overflow-on-invalid-input.patch" "unzip-format-secure.patch" "unzip-overflow-long-fsize.patch")))) + (replacement unzip/fixed) (build-system gnu-build-system) ;; no inputs; bzip2 is not supported, since not compiled with BZ_NO_STDIO (arguments @@ -1629,6 +1630,54 @@ (define-public unzip (license (license:non-copyleft "file://LICENSE" "See LICENSE in the distribution.")))) +(define-public unzip/fixed + (package + (inherit unzip) + (version "6.0") + (source + (origin + (method url-fetch) + (uri (string-append "mirror://sourceforge/infozip" + "/UnZip%206.x%20%28latest%29/UnZip%206.0/unzip60.tar.gz")) + (sha256 + (base32 + "0dxx11knh3nk95p2gg2ak777dd11pr7jx5das2g49l262scrcv83")) + (patches (search-patches "unzip-CVE-2014-8139.patch" + "unzip-CVE-2014-8140.patch" + "unzip-CVE-2014-8141.patch" + "unzip-CVE-2014-9636.patch" + "unzip-CVE-2015-7696.patch" + "unzip-CVE-2015-7697.patch" + "unzip-allow-greater-hostver-values.patch" + "unzip-initialize-symlink-flag.patch" + "unzip-remove-build-date.patch" + "unzip-attribs-overflow.patch" + "unzip-overflow-on-invalid-input.patch" + "unzip-format-secure.patch" + "unzip-overflow-long-fsize.patch" + + ;; From Fedora + "unzip-alt-iconv-utf8.patch" + "unzip-alt-iconv-utf8-print.patch" + "unzip-fix-recmatch.patch" + "unzip-caseinsensitive.patch" + "unzip-close.patch" + "unzip-COVSCAN-fix-unterminated-string.patch" + "unzip-CVE-2016-9844.patch" + "unzip-CVE-2018-1000035.patch" + "unzip-CVE-2018-18384.patch" + "unzip-exec-shield.patch" + "unzip-manpage-fix.patch" + "unzip-overflow.patch" + "unzip-timestamp.patch" + "unzip-valgrind.patch" + "unzip-x-option.patch" + ;; CVE-2019-13232 + "unzip-zipbomb-manpage.patch" + "unzip-zipbomb-part1.patch" + "unzip-zipbomb-part2.patch" + "unzip-zipbomb-part3.patch")))))) + (define-public ziptime (let ((commit "2a5bc9dfbf7c6a80e5f7cb4dd05b4036741478bc") (revision "0")) diff --git a/gnu/packages/patches/unzip-COVSCAN-fix-unterminated-string.patch b/gnu/packages/patches/unzip-COVSCAN-fix-unterminated-string.patch new file mode 100644 index 0000000000..717377119b --- /dev/null +++ b/gnu/packages/patches/unzip-COVSCAN-fix-unterminated-string.patch @@ -0,0 +1,131 @@ +From 06d1b08aef94984256cad3c5a54cedb10295681f Mon Sep 17 00:00:00 2001 +From: Jakub Martisko +Date: Thu, 8 Nov 2018 09:31:18 +0100 +Subject: [PATCH] Possible unterminated string fix + +--- + unix/unix.c | 4 +++- + unix/unxcfg.h | 2 +- + unzip.c | 12 ++++++++---- + zipinfo.c | 12 ++++++++---- + 4 files changed, 20 insertions(+), 10 deletions(-) + +diff --git a/unix/unix.c b/unix/unix.c +index 59b622d..cd57f80 100644 +--- a/unix/unix.c ++++ b/unix/unix.c +@@ -1945,7 +1945,9 @@ void init_conversion_charsets() + 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)); ++ MAX_CP_NAME - 1); ++ ++ OEM_CP[MAX_CP_NAME - 1] = '\0'; + break; + } + } +diff --git a/unix/unxcfg.h b/unix/unxcfg.h +index 8729de2..9ee8cfe 100644 +--- a/unix/unxcfg.h ++++ b/unix/unxcfg.h +@@ -228,7 +228,7 @@ typedef struct stat z_stat; + /* and notfirstcall are used by do_wild(). */ + + +-#define MAX_CP_NAME 25 ++#define MAX_CP_NAME 25 + 1 + + #ifdef SETLOCALE + # undef SETLOCALE +diff --git a/unzip.c b/unzip.c +index 2d94a38..a485f2b 100644 +--- a/unzip.c ++++ b/unzip.c +@@ -1561,7 +1561,8 @@ int uz_opts(__G__ pargc, pargv) + "error: a valid character encoding should follow the -I argument")); + return(PK_PARAM); + } +- strncpy(ISO_CP, s, sizeof(ISO_CP)); ++ strncpy(ISO_CP, s, MAX_CP_NAME - 1); ++ ISO_CP[MAX_CP_NAME - 1] = '\0'; + } else { /* -I charset */ + ++argv; + if(!(--argc > 0 && *argv != NULL && **argv != '-')) { +@@ -1570,7 +1571,8 @@ int uz_opts(__G__ pargc, pargv) + return(PK_PARAM); + } + s = *argv; +- strncpy(ISO_CP, s, sizeof(ISO_CP)); ++ strncpy(ISO_CP, s, MAX_CP_NAME - 1); ++ ISO_CP[MAX_CP_NAME - 1] = '\0'; + } + while(*(++s)); /* No params straight after charset name */ + } +@@ -1665,7 +1667,8 @@ int uz_opts(__G__ pargc, pargv) + "error: a valid character encoding should follow the -I argument")); + return(PK_PARAM); + } +- strncpy(OEM_CP, s, sizeof(OEM_CP)); ++ strncpy(OEM_CP, s, MAX_CP_NAME - 1); ++ OEM_CP[MAX_CP_NAME - 1] = '\0'; + } else { /* -O charset */ + ++argv; + if(!(--argc > 0 && *argv != NULL && **argv != '-')) { +@@ -1674,7 +1677,8 @@ int uz_opts(__G__ pargc, pargv) + return(PK_PARAM); + } + s = *argv; +- strncpy(OEM_CP, s, sizeof(OEM_CP)); ++ strncpy(OEM_CP, s, MAX_CP_NAME - 1); ++ OEM_CP[MAX_CP_NAME - 1] = '\0'; + } + while(*(++s)); /* No params straight after charset name */ + } +diff --git a/zipinfo.c b/zipinfo.c +index accca2a..cb7e08d 100644 +--- a/zipinfo.c ++++ b/zipinfo.c +@@ -519,7 +519,8 @@ int zi_opts(__G__ pargc, pargv) + "error: a valid character encoding should follow the -I argument")); + return(PK_PARAM); + } +- strncpy(ISO_CP, s, sizeof(ISO_CP)); ++ strncpy(ISO_CP, s, MAX_CP_NAME - 1); ++ ISO_CP[MAX_CP_NAME - 1] = '\0'; + } else { /* -I charset */ + ++argv; + if(!(--argc > 0 && *argv != NULL && **argv != '-')) { +@@ -528,7 +529,8 @@ int zi_opts(__G__ pargc, pargv) + return(PK_PARAM); + } + s = *argv; +- strncpy(ISO_CP, s, sizeof(ISO_CP)); ++ strncpy(ISO_CP, s, MAX_CP_NAME - 1); ++ ISO_CP[MAX_CP_NAME - 1] = '\0'; + } + while(*(++s)); /* No params straight after charset name */ + } +@@ -568,7 +570,8 @@ int zi_opts(__G__ pargc, pargv) + "error: a valid character encoding should follow the -I argument")); + return(PK_PARAM); + } +- strncpy(OEM_CP, s, sizeof(OEM_CP)); ++ strncpy(OEM_CP, s, MAX_CP_NAME - 1); ++ OEM_CP[MAX_CP_NAME - 1] = '\0'; + } else { /* -O charset */ + ++argv; + if(!(--argc > 0 && *argv != NULL && **argv != '-')) { +@@ -577,7 +580,8 @@ int zi_opts(__G__ pargc, pargv) + return(PK_PARAM); + } + s = *argv; +- strncpy(OEM_CP, s, sizeof(OEM_CP)); ++ strncpy(OEM_CP, s, MAX_CP_NAME - 1); ++ OEM_CP[MAX_CP_NAME - 1] = '\0'; + } + while(*(++s)); /* No params straight after charset name */ + } +-- +2.14.5 + diff --git a/gnu/packages/patches/unzip-CVE-2016-9844.patch b/gnu/packages/patches/unzip-CVE-2016-9844.patch new file mode 100644 index 0000000000..0e4a173397 --- /dev/null +++ b/gnu/packages/patches/unzip-CVE-2016-9844.patch @@ -0,0 +1,39 @@ +From 754137e70cf58a64ad524b704a86b651ba0cde07 Mon Sep 17 00:00:00 2001 +From: Petr Stodulka +Date: Wed, 14 Dec 2016 16:30:36 +0100 +Subject: [PATCH] Fix CVE-2016-9844 (rhbz#1404283) + +Fixes buffer overflow in zipinfo in similar way like fix for +CVE-2014-9913 provided by upstream. +--- + zipinfo.c | 14 +++++++++++++- + 1 file changed, 13 insertions(+), 1 deletion(-) + +diff --git a/zipinfo.c b/zipinfo.c +index c03620e..accca2a 100644 +--- a/zipinfo.c ++++ b/zipinfo.c +@@ -1984,7 +1984,19 @@ static int zi_short(__G) /* return PK-type error code */ + ush dnum=(ush)((G.crec.general_purpose_bit_flag>>1) & 3); + methbuf[3] = dtype[dnum]; + } else if (methnum >= NUM_METHODS) { /* unknown */ +- sprintf(&methbuf[1], "%03u", G.crec.compression_method); ++ /* 2016-12-05 SMS. ++ * https://launchpad.net/bugs/1643750 ++ * Unexpectedly large compression methods overflow ++ * &methbuf[]. Use the old, three-digit decimal format ++ * for values which fit. Otherwise, sacrifice the "u", ++ * and use four-digit hexadecimal. ++ */ ++ if (G.crec.compression_method <= 999) { ++ sprintf( &methbuf[ 1], "%03u", G.crec.compression_method); ++ } else { ++ sprintf( &methbuf[ 0], "%04X", G.crec.compression_method); ++ } ++ + } + + for (k = 0; k < 15; ++k) +-- +2.5.5 + diff --git a/gnu/packages/patches/unzip-CVE-2018-1000035.patch b/gnu/packages/patches/unzip-CVE-2018-1000035.patch new file mode 100644 index 0000000000..8ca713865c --- /dev/null +++ b/gnu/packages/patches/unzip-CVE-2018-1000035.patch @@ -0,0 +1,34 @@ +--- a/fileio.c 2014-12-05 05:06:05 -0600 ++++ b/fileio.c 2017-11-14 01:06:28 -0600 +@@ -1,5 +1,5 @@ + /* +- Copyright (c) 1990-2009 Info-ZIP. All rights reserved. ++ Copyright (c) 1990-2017 Info-ZIP. All rights reserved. + + See the accompanying file LICENSE, version 2009-Jan-02 or later + (the contents of which are also included in unzip.h) for terms of use. +@@ -1582,6 +1582,8 @@ + int r = IZ_PW_ENTERED; + char *m; + char *prompt; ++ char *ep; ++ char *zp; + + #ifndef REENTRANT + /* tell picky compilers to shut up about "unused variable" warnings */ +@@ -1590,9 +1592,12 @@ + + if (*rcnt == 0) { /* First call for current entry */ + *rcnt = 2; +- if ((prompt = (char *)malloc(2*FILNAMSIZ + 15)) != (char *)NULL) { +- sprintf(prompt, LoadFarString(PasswPrompt), +- FnFilter1(zfn), FnFilter2(efn)); ++ zp = FnFilter1( zfn); ++ ep = FnFilter2( efn); ++ prompt = (char *)malloc( /* Slightly too long (2* "%s"). */ ++ sizeof( PasswPrompt)+ strlen( zp)+ strlen( ep)); ++ if (prompt != (char *)NULL) { ++ sprintf(prompt, LoadFarString(PasswPrompt), zp, ep); + m = prompt; + } else + m = (char *)LoadFarString(PasswPrompt2); diff --git a/gnu/packages/patches/unzip-CVE-2018-18384.patch b/gnu/packages/patches/unzip-CVE-2018-18384.patch new file mode 100644 index 0000000000..54d4b8cb64 --- /dev/null +++ b/gnu/packages/patches/unzip-CVE-2018-18384.patch @@ -0,0 +1,35 @@ +--- unzip60/list.c ++++ unzip60/list.c +@@ -97,7 +97,7 @@ int list_files(__G) /* return PK-type + { + int do_this_file=FALSE, cfactor, error, error_in_archive=PK_COOL; + #ifndef WINDLL +- char sgn, cfactorstr[13]; ++ char sgn, cfactorstr[1+10+1+1]; /* %NUL */ + int longhdr=(uO.vflag>1); + #endif + int date_format; +@@ -389,9 +389,9 @@ int list_files(__G) /* return PK-type + } + #else /* !WINDLL */ + if (cfactor == 100) +- sprintf(cfactorstr, LoadFarString(CompFactor100)); ++ snprintf(cfactorstr, sizeof(cfactorstr), LoadFarString(CompFactor100)); + else +- sprintf(cfactorstr, LoadFarString(CompFactorStr), sgn, cfactor); ++ snprintf(cfactorstr, sizeof(cfactorstr), LoadFarString(CompFactorStr), sgn, cfactor); + if (longhdr) + Info(slide, 0, ((char *)slide, LoadFarString(LongHdrStats), + FmZofft(G.crec.ucsize, "8", "u"), methbuf, +@@ -471,9 +471,9 @@ int list_files(__G) /* return PK-type + + #else /* !WINDLL */ + if (cfactor == 100) +- sprintf(cfactorstr, LoadFarString(CompFactor100)); ++ snprintf(cfactorstr, sizeof(cfactorstr), LoadFarString(CompFactor100)); + else +- sprintf(cfactorstr, LoadFarString(CompFactorStr), sgn, cfactor); ++ snprintf(cfactorstr, sizeof(cfactorstr), LoadFarString(CompFactorStr), sgn, cfactor); + if (longhdr) { + Info(slide, 0, ((char *)slide, LoadFarString(LongFileTrailer), + FmZofft(tot_ucsize, "8", "u"), FmZofft(tot_csize, "8", "u"), diff --git a/gnu/packages/patches/unzip-alt-iconv-utf8-print.patch b/gnu/packages/patches/unzip-alt-iconv-utf8-print.patch new file mode 100644 index 0000000000..0b0153ba54 --- /dev/null +++ b/gnu/packages/patches/unzip-alt-iconv-utf8-print.patch @@ -0,0 +1,381 @@ +From ca0212ba19b64488b9e8459a762c11ecd6e7d0bd Mon Sep 17 00:00:00 2001 +From: Petr Stodulka +Date: Tue, 24 Nov 2015 17:56:11 +0100 +Subject: [PATCH] print correctly non-ascii filenames + +--- + extract.c | 289 ++++++++++++++++++++++++++++++++++++++++++++++++-------------- + unzpriv.h | 7 ++ + 2 files changed, 233 insertions(+), 63 deletions(-) + +diff --git a/extract.c b/extract.c +index 0ee4e93..741b7e0 100644 +--- a/extract.c ++++ b/extract.c +@@ -2648,8 +2648,21 @@ static void set_deferred_symlink(__G__ slnk_entry) + } /* end function set_deferred_symlink() */ + #endif /* SYMLINKS */ + ++/* ++ * If Unicode is supported, assume we have what we need to do this ++ * check using wide characters, avoiding MBCS issues. ++ */ + +- ++#ifndef UZ_FNFILTER_REPLACECHAR ++ /* A convenient choice for the replacement of unprintable char codes is ++ * the "single char wildcard", as this character is quite unlikely to ++ * appear in filenames by itself. The following default definition ++ * sets the replacement char to a question mark as the most common ++ * "single char wildcard"; this setting should be overridden in the ++ * appropiate system-specific configuration header when needed. ++ */ ++# define UZ_FNFILTER_REPLACECHAR '?' ++#endif + + /*************************/ + /* Function fnfilter() */ /* here instead of in list.c for SFX */ +@@ -2661,48 +2674,168 @@ char *fnfilter(raw, space, size) /* convert name to safely printable form */ + extent size; + { + #ifndef NATIVE /* ASCII: filter ANSI escape codes, etc. */ +- ZCONST uch *r=(ZCONST uch *)raw; ++ ZCONST uch *r; // =(ZCONST uch *)raw; + uch *s=space; + uch *slim=NULL; + uch *se=NULL; + int have_overflow = FALSE; + +- if (size > 0) { +- slim = space + size +-#ifdef _MBCS +- - (MB_CUR_MAX - 1) +-#endif +- - 4; ++# if defined( UNICODE_SUPPORT) && defined( _MBCS) ++/* If Unicode support is enabled, and we have multi-byte characters, ++ * then do the isprint() checks by first converting to wide characters ++ * and checking those. This avoids our having to parse multi-byte ++ * characters for ourselves. After the wide-char replacements have been ++ * made, the wide string is converted back to the local character set. ++ */ ++ wchar_t *wstring; /* wchar_t version of raw */ ++ size_t wslen; /* length of wstring */ ++ wchar_t *wostring; /* wchar_t version of output string */ ++ size_t woslen; /* length of wostring */ ++ char *newraw; /* new raw */ ++ ++ /* 2012-11-06 SMS. ++ * Changed to check the value returned by mbstowcs(), and bypass the ++ * Unicode processing if it fails. This seems to fix a problem ++ * reported in the SourceForge forum, but it's not clear that we ++ * should be doing any Unicode processing without some evidence that ++ * the name actually is Unicode. (Check bit 11 in the flags before ++ * coming here?) ++ * http://sourceforge.net/p/infozip/bugs/40/ ++ */ ++ ++ if (MB_CUR_MAX <= 1) ++ { ++ /* There's no point to converting multi-byte chars if there are ++ * no multi-byte chars. ++ */ ++ wslen = (size_t)-1; + } +- while (*r) { +- if (size > 0 && s >= slim && se == NULL) { +- se = s; ++ else ++ { ++ /* Get Unicode wide character count (for storage allocation). */ ++ wslen = mbstowcs( NULL, raw, 0); ++ } ++ ++ if (wslen != (size_t)-1) ++ { ++ /* Apparently valid Unicode. Allocate wide-char storage. */ ++ wstring = (wchar_t *)malloc((wslen + 1) * sizeof(wchar_t)); ++ if (wstring == NULL) { ++ strcpy( (char *)space, raw); ++ return (char *)space; + } +-#ifdef QDOS +- if (qlflag & 2) { +- if (*r == '/' || *r == '.') { ++ wostring = (wchar_t *)malloc(2 * (wslen + 1) * sizeof(wchar_t)); ++ if (wostring == NULL) { ++ free(wstring); ++ strcpy( (char *)space, raw); ++ return (char *)space; ++ } ++ ++ /* Convert the multi-byte Unicode to wide chars. */ ++ wslen = mbstowcs(wstring, raw, wslen + 1); ++ ++ /* Filter the wide-character string. */ ++ fnfilterw( wstring, wostring, (2 * (wslen + 1) * sizeof(wchar_t))); ++ ++ /* Convert filtered wide chars back to multi-byte. */ ++ woslen = wcstombs( NULL, wostring, 0); ++ if ((newraw = malloc(woslen + 1)) == NULL) { ++ free(wstring); ++ free(wostring); ++ strcpy( (char *)space, raw); ++ return (char *)space; ++ } ++ woslen = wcstombs( newraw, wostring, (woslen * MB_CUR_MAX) + 1); ++ ++ if (size > 0) { ++ slim = space + size - 4; ++ } ++ r = (ZCONST uch *)newraw; ++ while (*r) { ++ if (size > 0 && s >= slim && se == NULL) { ++ se = s; ++ } ++# ifdef QDOS ++ if (qlflag & 2) { ++ if (*r == '/' || *r == '.') { ++ if (se != NULL && (s > (space + (size-3)))) { ++ have_overflow = TRUE; ++ break; ++ } ++ ++r; ++ *s++ = '_'; ++ continue; ++ } ++ } else ++# endif ++ { + if (se != NULL && (s > (space + (size-3)))) { + have_overflow = TRUE; + break; + } +- ++r; +- *s++ = '_'; +- continue; ++ *s++ = *r++; + } +- } else ++ } ++ if (have_overflow) { ++ strcpy((char *)se, "..."); ++ } else { ++ *s = '\0'; ++ } ++ ++ free(wstring); ++ free(wostring); ++ free(newraw); ++ } ++ else ++# endif /* defined( UNICODE_SUPPORT) && defined( _MBCS) */ ++ { ++ /* No Unicode support, or apparently invalid Unicode. */ ++ r = (ZCONST uch *)raw; ++ ++ if (size > 0) { ++ slim = space + size ++#ifdef _MBCS ++ - (MB_CUR_MAX - 1) ++#endif ++ - 4; ++ } ++ while (*r) { ++ if (size > 0 && s >= slim && se == NULL) { ++ se = s; ++ } ++#ifdef QDOS ++ if (qlflag & 2) { ++ if (*r == '/' || *r == '.') { ++ if (se != NULL && (s > (space + (size-3)))) { ++ have_overflow = TRUE; ++ break; ++ } ++ ++r; ++ *s++ = '_'; ++ continue; ++ } ++ } else + #endif + #ifdef HAVE_WORKING_ISPRINT +-# ifndef UZ_FNFILTER_REPLACECHAR +- /* A convenient choice for the replacement of unprintable char codes is +- * the "single char wildcard", as this character is quite unlikely to +- * appear in filenames by itself. The following default definition +- * sets the replacement char to a question mark as the most common +- * "single char wildcard"; this setting should be overridden in the +- * appropiate system-specific configuration header when needed. +- */ +-# define UZ_FNFILTER_REPLACECHAR '?' +-# endif +- if (!isprint(*r)) { ++ if (!isprint(*r)) { ++ if (*r < 32) { ++ /* ASCII control codes are escaped as "^{letter}". */ ++ if (se != NULL && (s > (space + (size-4)))) { ++ have_overflow = TRUE; ++ break; ++ } ++ *s++ = '^', *s++ = (uch)(64 + *r++); ++ } else { ++ /* Other unprintable codes are replaced by the ++ * placeholder character. */ ++ if (se != NULL && (s > (space + (size-3)))) { ++ have_overflow = TRUE; ++ break; ++ } ++ *s++ = UZ_FNFILTER_REPLACECHAR; ++ INCSTR(r); ++ } ++#else /* !HAVE_WORKING_ISPRINT */ + if (*r < 32) { + /* ASCII control codes are escaped as "^{letter}". */ + if (se != NULL && (s > (space + (size-4)))) { +@@ -2710,47 +2843,30 @@ char *fnfilter(raw, space, size) /* convert name to safely printable form */ + break; + } + *s++ = '^', *s++ = (uch)(64 + *r++); ++#endif /* ?HAVE_WORKING_ISPRINT */ + } else { +- /* Other unprintable codes are replaced by the +- * placeholder character. */ ++#ifdef _MBCS ++ unsigned i = CLEN(r); ++ if (se != NULL && (s > (space + (size-i-2)))) { ++ have_overflow = TRUE; ++ break; ++ } ++ for (; i > 0; i--) ++ *s++ = *r++; ++#else + if (se != NULL && (s > (space + (size-3)))) { + have_overflow = TRUE; + break; + } +- *s++ = UZ_FNFILTER_REPLACECHAR; +- INCSTR(r); +- } +-#else /* !HAVE_WORKING_ISPRINT */ +- if (*r < 32) { +- /* ASCII control codes are escaped as "^{letter}". */ +- if (se != NULL && (s > (space + (size-4)))) { +- have_overflow = TRUE; +- break; +- } +- *s++ = '^', *s++ = (uch)(64 + *r++); +-#endif /* ?HAVE_WORKING_ISPRINT */ +- } else { +-#ifdef _MBCS +- unsigned i = CLEN(r); +- if (se != NULL && (s > (space + (size-i-2)))) { +- have_overflow = TRUE; +- break; +- } +- for (; i > 0; i--) + *s++ = *r++; +-#else +- if (se != NULL && (s > (space + (size-3)))) { +- have_overflow = TRUE; +- break; +- } +- *s++ = *r++; + #endif +- } +- } +- if (have_overflow) { +- strcpy((char *)se, "..."); +- } else { +- *s = '\0'; ++ } ++ } ++ if (have_overflow) { ++ strcpy((char *)se, "..."); ++ } else { ++ *s = '\0'; ++ } + } + + #ifdef WINDLL +@@ -2772,6 +2888,53 @@ char *fnfilter(raw, space, size) /* convert name to safely printable form */ + } /* end function fnfilter() */ + + ++#if defined( UNICODE_SUPPORT) && defined( _MBCS) ++ ++/****************************/ ++/* Function fnfilter[w]() */ /* (Here instead of in list.c for SFX.) */ ++/****************************/ ++ ++/* fnfilterw() - Convert wide name to safely printable form. */ ++ ++/* fnfilterw() - Convert wide-character name to safely printable form. */ ++ ++wchar_t *fnfilterw( src, dst, siz) ++ ZCONST wchar_t *src; /* Pointer to source char (string). */ ++ wchar_t *dst; /* Pointer to destination char (string). */ ++ extent siz; /* Not used (!). */ ++{ ++ wchar_t *dsx = dst; ++ ++ /* Filter the wide chars. */ ++ while (*src) ++ { ++ if (iswprint( *src)) ++ { ++ /* Printable code. Copy it. */ ++ *dst++ = *src; ++ } ++ else ++ { ++ /* Unprintable code. Substitute something printable for it. */ ++ if (*src < 32) ++ { ++ /* Replace ASCII control code with "^{letter}". */ ++ *dst++ = (wchar_t)'^'; ++ *dst++ = (wchar_t)(64 + *src); ++ } ++ else ++ { ++ /* Replace other unprintable code with the placeholder. */ ++ *dst++ = (wchar_t)UZ_FNFILTER_REPLACECHAR; ++ } ++ } ++ src++; ++ } ++ *dst = (wchar_t)0; /* NUL-terminate the destination string. */ ++ return dsx; ++} /* fnfilterw(). */ ++ ++#endif /* defined( UNICODE_SUPPORT) && defined( _MBCS) */ + + + #ifdef SET_DIR_ATTRIB +diff --git a/unzpriv.h b/unzpriv.h +index 22d3923..e48a652 100644 +--- a/unzpriv.h ++++ b/unzpriv.h +@@ -1212,6 +1212,7 @@ + # ifdef UNICODE_WCHAR + # if !(defined(_WIN32_WCE) || defined(POCKET_UNZIP)) + # include ++# include + # endif + # endif + # ifndef _MBCS /* no need to include twice, see below */ +@@ -2410,6 +2411,12 @@ int memflush OF((__GPRO__ ZCONST uch *rawbuf, ulg size)); + char *fnfilter OF((ZCONST char *raw, uch *space, + extent size)); + ++# if defined( UNICODE_SUPPORT) && defined( _MBCS) ++wchar_t *fnfilterw OF((ZCONST wchar_t *src, wchar_t *dst, ++ extent siz)); ++#endif ++ ++ + /*--------------------------------------------------------------------------- + Decompression functions: + ---------------------------------------------------------------------------*/ +-- +2.4.3 + diff --git a/gnu/packages/patches/unzip-alt-iconv-utf8.patch b/gnu/packages/patches/unzip-alt-iconv-utf8.patch new file mode 100644 index 0000000000..b9e37774e2 --- /dev/null +++ b/gnu/packages/patches/unzip-alt-iconv-utf8.patch @@ -0,0 +1,398 @@ +From: Giovanni Scafora +Subject: unzip files encoded with non-latin, non-unicode file names +Last-Update: 2015-02-11 + +Updated 2015-02-11 by Marc Deslauriers +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 ++#include ++ + #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 ++ ++#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 +@@ -327,11 +327,21 @@ + -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 -T print file times in sortable decimal format\ + \n -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 + 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; diff --git a/gnu/packages/patches/unzip-caseinsensitive.patch b/gnu/packages/patches/unzip-caseinsensitive.patch new file mode 100644 index 0000000000..3cb68450b9 --- /dev/null +++ b/gnu/packages/patches/unzip-caseinsensitive.patch @@ -0,0 +1,131 @@ +diff --git a/match.c b/match.c +index 6cd656f..4e569f5 100644 +--- a/match.c ++++ b/match.c +@@ -190,10 +190,10 @@ char *___tmp_ptr; + + #endif + +-static int recmatch(p, s, cs) ++static int recmatch(p, s, ci) + ZCONST char *p; /* sh pattern to match */ + ZCONST char *s; /* string to match it to */ +-int cs; /* flag: force case-sensitive matching */ ++int ci; /* flag: force case-insensitive matching */ + /* Recursively compare the sh pattern p with the string s and return 1 if + they match, and 0 or 2 if they don't or if there is a syntax error in the + pattern. This routine recurses on itself no deeper than the number of +@@ -214,7 +214,7 @@ int cs; /* flag: force case-sensitive matching */ + if (CLEN(p) == 2) { + if (CLEN(s) == 2) { + return (*p == *s && *(p+1) == *(s+1)) ? +- recmatch(p + 2, s + 2, cs) : 0; ++ recmatch(p + 2, s + 2, ci) : 0; + } else { + return 0; + } +@@ -230,9 +230,9 @@ int cs; /* flag: force case-sensitive matching */ + /* '?' (or '%' or '#') matches any character (but not an empty string) */ + if (c == WILDCHR_SINGLE) { + if (wild_stop_at_dir) +- return (*s && *s != DIRSEP_CHR) ? recmatch(p, s + CLEN(s), cs) : 0; ++ return (*s && *s != DIRSEP_CHR) ? recmatch(p, s + CLEN(s), ci) : 0; + else +- return *s ? recmatch(p, s + CLEN(s), cs) : 0; ++ return *s ? recmatch(p, s + CLEN(s), ci) : 0; + } + + /* WILDCHR_MULTI ('*') matches any number of characters, including zero */ +@@ -253,14 +253,14 @@ int cs; /* flag: force case-sensitive matching */ + # endif /* ?AMIGA */ + /* Single WILDCHR_MULTI ('*'): this doesn't match slashes */ + for (; *s && *s != DIRSEP_CHR; INCSTR(s)) +- if ((c = recmatch(p, s, cs)) != 0) ++ if ((c = recmatch(p, s, ci)) != 0) + return c; + /* end of pattern: matched if at end of string, else continue */ + if (*p == 0) + return (*s == 0); + /* continue to match if at DIRSEP_CHR in pattern, else give up */ + return (*p == DIRSEP_CHR || (*p == '\\' && p[1] == DIRSEP_CHR)) +- ? recmatch(p, s, cs) : 2; ++ ? recmatch(p, s, ci) : 2; + } + /* Two consecutive WILDCHR_MULTI ("**"): this matches DIRSEP_CHR ('/') */ + p++; /* move p past the second WILDCHR_MULTI */ +@@ -308,17 +308,17 @@ int cs; /* flag: force case-sensitive matching */ + */ + if (q != srest) + return 0; +- return ((cs ? strcmp(p, q) : namecmp(p, q)) == 0); ++ return ((!ci ? strcmp(p, q) : namecmp(p, q)) == 0); + } + #else /* !_MBCS */ +- return ((cs ? strcmp(p, srest) : namecmp(p, srest)) == 0); ++ return ((!ci ? strcmp(p, srest) : namecmp(p, srest)) == 0); + #endif /* ?_MBCS */ + } + else + { + /* pattern contains more wildcards, continue with recursion... */ + for (; *s; INCSTR(s)) +- if ((c = recmatch(p, s, cs)) != 0) ++ if ((c = recmatch(p, s, ci)) != 0) + return c; + return 2; /* 2 means give up--shmatch will return false */ + } +@@ -353,17 +353,17 @@ int cs; /* flag: force case-sensitive matching */ + c = *(p-1); + else + { +- uch cc = (cs ? (uch)*s : case_map((uch)*s)); ++ uch cc = (!ci ? (uch)*s : to_up((uch)*s)); + uch uc = (uch) c; + if (*(p+1) != '-') + for (uc = uc ? uc : (uch)*p; uc <= (uch)*p; uc++) + /* compare range */ +- if ((cs ? uc : case_map(uc)) == cc) +- return r ? 0 : recmatch(q + CLEN(q), s + CLEN(s), cs); ++ if ((!ci ? uc : to_up(uc)) == cc) ++ return r ? 0 : recmatch(q + CLEN(q), s + CLEN(s), ci); + c = e = 0; /* clear range, escape flags */ + } + } +- return r ? recmatch(q + CLEN(q), s + CLEN(s), cs) : 0; ++ return r ? recmatch(q + CLEN(q), s + CLEN(s), ci) : 0; + /* bracket match failed */ + } + #endif /* !VMS */ +@@ -382,18 +382,18 @@ int cs; /* flag: force case-sensitive matching */ + { + /* Match "...]" with "]". Continue after "]" in both. */ + if ((*(p+ 2* CLEN( p)) == ']') && (*s == ']')) +- return recmatch( (p+ 3* CLEN( p)), (s+ CLEN( s)), cs); ++ return recmatch( (p+ 3* CLEN( p)), (s+ CLEN( s)), ci); + + /* Else, look for a reduced match in s, until "]" in or end of s. */ + for (; *s && (*s != ']'); INCSTR(s)) + if (*s == '.') + /* If reduced match, then continue after "..." in p, "." in s. */ +- if ((c = recmatch( (p+ CLEN( p)), s, cs)) != 0) ++ if ((c = recmatch( (p+ CLEN( p)), s, ci)) != 0) + return (int)c; + + /* Match "...]" with "]". Continue after "]" in both. */ + if ((*(p+ 2* CLEN( p)) == ']') && (*s == ']')) +- return recmatch( (p+ 3* CLEN( p)), (s+ CLEN( s)), cs); ++ return recmatch( (p+ 3* CLEN( p)), (s+ CLEN( s)), ci); + + /* No reduced match. Quit. */ + return 2; +@@ -402,8 +402,8 @@ int cs; /* flag: force case-sensitive matching */ + #endif /* def VMS */ + + /* Just a character--compare it */ +- return (cs ? c == *s : case_map((uch)c) == case_map((uch)*s)) ? +- recmatch(p, s + CLEN(s), cs) : 0; ++ return (!ci ? c == *s : to_up((uch)c) == to_up((uch)*s)) ? ++ recmatch(p, s + CLEN(s), ci) : 0; + } + + diff --git a/gnu/packages/patches/unzip-close.patch b/gnu/packages/patches/unzip-close.patch new file mode 100644 index 0000000000..9a238a9006 --- /dev/null +++ b/gnu/packages/patches/unzip-close.patch @@ -0,0 +1,176 @@ +diff -up unzip60/extract.c.close unzip60/extract.c +--- unzip60/extract.c.close 2009-03-14 02:32:52.000000000 +0100 ++++ unzip60/extract.c 2009-11-19 08:17:23.481263496 +0100 +@@ -1924,24 +1924,21 @@ static int extract_or_test_member(__G) + + #ifdef VMS /* VMS: required even for stdout! (final flush) */ + if (!uO.tflag) /* don't close NULL file */ +- close_outfile(__G); ++ error = close_outfile(__G); + #else + #ifdef DLL + if (!uO.tflag && (!uO.cflag || G.redirect_data)) { + if (G.redirect_data) + FINISH_REDIRECT(); + else +- close_outfile(__G); ++ error = close_outfile(__G); + } + #else + if (!uO.tflag && !uO.cflag) /* don't close NULL file or stdout */ +- close_outfile(__G); ++ error = close_outfile(__G); + #endif + #endif /* VMS */ + +- /* GRR: CONVERT close_outfile() TO NON-VOID: CHECK FOR ERRORS! */ +- +- + if (G.disk_full) { /* set by flush() */ + if (G.disk_full > 1) { + #if (defined(DELETE_IF_FULL) && defined(HAVE_UNLINK)) +diff -up unzip60/unix/unix.c.close unzip60/unix/unix.c +--- unzip60/unix/unix.c.close 2009-01-24 00:31:26.000000000 +0100 ++++ unzip60/unix/unix.c 2009-11-19 08:33:25.568389171 +0100 +@@ -1096,10 +1096,41 @@ static int get_extattribs(__G__ pzt, z_u + #ifndef MTS + + /****************************/ ++/* Function CloseError() */ ++/***************************/ ++ ++int CloseError(__G) ++ __GDEF ++{ ++ int errval = PK_OK; ++ ++ if (fclose(G.outfile) < 0) { ++ switch (errno) { ++ case ENOSPC: ++ /* Do we need this on fileio.c? */ ++ Info(slide, 0x4a1, ((char *)slide, "%s: write error (disk full?). Continue? (y/n/^C) ", ++ FnFilter1(G.filename))); ++ fgets(G.answerbuf, 9, stdin); ++ if (*G.answerbuf == 'y') /* stop writing to this file */ ++ G.disk_full = 1; /* pass to next */ ++ else ++ G.disk_full = 2; /* no: exit program */ ++ ++ errval = PK_DISK; ++ break; ++ ++ default: ++ errval = PK_WARN; ++ } ++ } ++ return errval; ++} /* End of CloseError() */ ++ ++/****************************/ + /* Function close_outfile() */ + /****************************/ + +-void close_outfile(__G) /* GRR: change to return PK-style warning level */ ++int close_outfile(__G) + __GDEF + { + union { +@@ -1108,6 +1139,7 @@ void close_outfile(__G) /* GRR: chang + } zt; + ulg z_uidgid[2]; + int have_uidgid_flg; ++ int errval = PK_OK; + + have_uidgid_flg = get_extattribs(__G__ &(zt.t3), z_uidgid); + +@@ -1141,16 +1173,16 @@ void close_outfile(__G) /* GRR: chang + Info(slide, 0x201, ((char *)slide, + "warning: symbolic link (%s) failed: mem alloc overflow\n", + FnFilter1(G.filename))); +- fclose(G.outfile); +- return; ++ errval = CloseError(G.outfile, G.filename); ++ return errval ? errval : PK_WARN; + } + + if ((slnk_entry = (slinkentry *)malloc(slnk_entrysize)) == NULL) { + Info(slide, 0x201, ((char *)slide, + "warning: symbolic link (%s) failed: no mem\n", + FnFilter1(G.filename))); +- fclose(G.outfile); +- return; ++ errval = CloseError(G.outfile, G.filename); ++ return errval ? errval : PK_WARN; + } + slnk_entry->next = NULL; + slnk_entry->targetlen = ucsize; +@@ -1174,10 +1206,10 @@ void close_outfile(__G) /* GRR: chang + "warning: symbolic link (%s) failed\n", + FnFilter1(G.filename))); + free(slnk_entry); +- fclose(G.outfile); +- return; ++ errval = CloseError(G.outfile, G.filename); ++ return errval ? errval : PK_WARN; + } +- fclose(G.outfile); /* close "link" file for good... */ ++ errval = CloseError(G.outfile, G.filename); /* close "link" file for good... */ + slnk_entry->target[ucsize] = '\0'; + if (QCOND2) + Info(slide, 0, ((char *)slide, "-> %s ", +@@ -1188,7 +1220,7 @@ void close_outfile(__G) /* GRR: chang + else + G.slink_head = slnk_entry; + G.slink_last = slnk_entry; +- return; ++ return errval; + } + #endif /* SYMLINKS */ + +@@ -1201,7 +1233,7 @@ void close_outfile(__G) /* GRR: chang + #endif + + #if (defined(NO_FCHOWN)) +- fclose(G.outfile); ++ errval = CloseError(G.outfile, G.filename); + #endif + + /* if -X option was specified and we have UID/GID info, restore it */ +@@ -1227,7 +1259,7 @@ void close_outfile(__G) /* GRR: chang + } + + #if (!defined(NO_FCHOWN) && defined(NO_FCHMOD)) +- fclose(G.outfile); ++ errval = CloseError(G.outfile, G.filename); + #endif + + #if (!defined(NO_FCHOWN) && !defined(NO_FCHMOD)) +@@ -1239,7 +1271,7 @@ void close_outfile(__G) /* GRR: chang + if (fchmod(fileno(G.outfile), filtattr(__G__ G.pInfo->file_attr))) + perror("fchmod (file attributes) error"); + +- fclose(G.outfile); ++ errval = CloseError(G.outfile, G.filename); + #endif /* !NO_FCHOWN && !NO_FCHMOD */ + + /* skip restoring time stamps on user's request */ +@@ -1267,6 +1299,7 @@ void close_outfile(__G) /* GRR: chang + #endif + #endif /* NO_FCHOWN || NO_FCHMOD */ + ++ return errval; + } /* end function close_outfile() */ + + #endif /* !MTS */ +diff -up unzip60/unzpriv.h.close unzip60/unzpriv.h +--- unzip60/unzpriv.h.close 2009-04-20 01:59:26.000000000 +0200 ++++ unzip60/unzpriv.h 2009-11-19 08:19:08.610388618 +0100 +@@ -2604,7 +2604,7 @@ char *GetLoadPath OF((__GPRO)); + int SetFileSize OF((FILE *file, zusz_t filesize)); /* local */ + #endif + #ifndef MTS /* macro in MTS */ +- void close_outfile OF((__GPRO)); /* local */ ++ int close_outfile OF((__GPRO)); /* local */ + #endif + #ifdef SET_SYMLINK_ATTRIBS + int set_symlnk_attribs OF((__GPRO__ slinkentry *slnk_entry)); /* local */ diff --git a/gnu/packages/patches/unzip-exec-shield.patch b/gnu/packages/patches/unzip-exec-shield.patch new file mode 100644 index 0000000000..74500aa5d4 --- /dev/null +++ b/gnu/packages/patches/unzip-exec-shield.patch @@ -0,0 +1,10 @@ +diff -up unzip60/crc_i386.S.exec-shield unzip60/crc_i386.S +--- unzip60/crc_i386.S.exec-shield 2007-01-07 06:02:58.000000000 +0100 ++++ unzip60/crc_i386.S 2009-11-18 11:16:39.630389312 +0100 +@@ -302,3 +302,6 @@ _crc32: /* ulg c + #endif /* i386 || _i386 || _I386 || __i386 */ + + #endif /* !USE_ZLIB && !CRC_TABLE_ONLY */ ++ ++.section .note.GNU-stack, "", @progbits ++.previous diff --git a/gnu/packages/patches/unzip-fix-recmatch.patch b/gnu/packages/patches/unzip-fix-recmatch.patch new file mode 100644 index 0000000000..2a8583c8ca --- /dev/null +++ b/gnu/packages/patches/unzip-fix-recmatch.patch @@ -0,0 +1,477 @@ +diff -up unzip60/match.c.recmatch unzip60/match.c +--- unzip60/match.c.recmatch 2005-08-14 13:00:36.000000000 -0400 ++++ unzip60/match.c 2013-05-28 10:29:57.949077543 -0400 +@@ -27,16 +27,14 @@ + + --------------------------------------------------------------------------- + +- Copyright on recmatch() from Zip's util.c (although recmatch() was almost +- certainly written by Mark Adler...ask me how I can tell :-) ): ++ Copyright on recmatch() from Zip's util.c ++ Copyright (c) 1990-2005 Info-ZIP. All rights reserved. + +- Copyright (C) 1990-1992 Mark Adler, Richard B. Wales, Jean-loup Gailly, +- Kai Uwe Rommel and Igor Mandrichenko. ++ See the accompanying file LICENSE, version 2004-May-22 or later ++ for terms of use. ++ If, for some reason, both of these files are missing, the Info-ZIP license ++ also may be found at: ftp://ftp.info-zip.org/pub/infozip/license.html + +- Permission is granted to any individual or institution to use, copy, +- or redistribute this software so long as all of the original files are +- included unmodified, that it is not sold for profit, and that this copy- +- right notice is retained. + + --------------------------------------------------------------------------- + +@@ -53,7 +51,7 @@ + + A set is composed of characters or ranges; a range looks like ``character + hyphen character'' (as in 0-9 or A-Z). [0-9a-zA-Z_] is the minimal set of +- characters allowed in the [..] pattern construct. Other characters are ++ characters ALlowed in the [..] pattern construct. Other characters are + allowed (i.e., 8-bit characters) if your system will support them. + + To suppress the special syntactic significance of any of ``[]*?!^-\'', in- +@@ -101,8 +99,32 @@ + # define WILDCHAR '?' + # define BEG_RANGE '[' + # define END_RANGE ']' ++# define WILDCHR_SINGLE '?' ++# define DIRSEP_CHR '/' ++# define WILDCHR_MULTI '*' + #endif + ++#ifdef WILD_STOP_AT_DIR ++ int wild_stop_at_dir = 1; /* default wildcards do not include / in matches */ ++#else ++ int wild_stop_at_dir = 0; /* default wildcards do include / in matches */ ++#endif ++ ++ ++ ++/* ++ * case mapping functions. case_map is used to ignore case in comparisons, ++ * to_up is used to force upper case even on Unix (for dosify option). ++ */ ++#ifdef USE_CASE_MAP ++# define case_map(c) upper[(c) & 0xff] ++# define to_up(c) upper[(c) & 0xff] ++#else ++# define case_map(c) (c) ++# define to_up(c) ((c) >= 'a' && (c) <= 'z' ? (c)-'a'+'A' : (c)) ++#endif /* USE_CASE_MAP */ ++ ++ + #if 0 /* GRR: add this to unzip.h someday... */ + #if !(defined(MSDOS) && defined(DOSWILD)) + #ifdef WILD_STOP_AT_DIR +@@ -114,8 +136,8 @@ int recmatch OF((ZCONST uch *pattern, ZC + int ignore_case __WDLPRO)); + #endif + #endif /* 0 */ +-static int recmatch OF((ZCONST uch *pattern, ZCONST uch *string, +- int ignore_case __WDLPRO)); ++static int recmatch OF((ZCONST char *, ZCONST char *, ++ int)); + static char *isshexp OF((ZCONST char *p)); + static int namecmp OF((ZCONST char *s1, ZCONST char *s2)); + +@@ -154,192 +176,240 @@ int match(string, pattern, ignore_case _ + } + dospattern[j-1] = '\0'; /* nuke the end "." */ + } +- j = recmatch((uch *)dospattern, (uch *)string, ignore_case __WDL); ++ j = recmatch(dospattern, string, ignore_case); + free(dospattern); + return j == 1; + } else + #endif /* MSDOS && DOSWILD */ +- return recmatch((uch *)pattern, (uch *)string, ignore_case __WDL) == 1; ++ return recmatch(pattern, string, ignore_case) == 1; + } + ++#ifdef _MBCS ++ ++char *___tmp_ptr; + ++#endif + +-static int recmatch(p, s, ic __WDL) +- ZCONST uch *p; /* sh pattern to match */ +- ZCONST uch *s; /* string to which to match it */ +- int ic; /* true for case insensitivity */ +- __WDLDEF /* directory sepchar for WildStopAtDir mode, or 0 */ ++static int recmatch(p, s, cs) ++ZCONST char *p; /* sh pattern to match */ ++ZCONST char *s; /* string to match it to */ ++int cs; /* flag: force case-sensitive matching */ + /* Recursively compare the sh pattern p with the string s and return 1 if +- * they match, and 0 or 2 if they don't or if there is a syntax error in the +- * pattern. This routine recurses on itself no more deeply than the number +- * of characters in the pattern. */ ++ they match, and 0 or 2 if they don't or if there is a syntax error in the ++ pattern. This routine recurses on itself no deeper than the number of ++ characters in the pattern. */ + { +- unsigned int c; /* pattern char or start of range in [-] loop */ ++ int c; /* pattern char or start of range in [-] loop */ ++ /* Get first character, the pattern for new recmatch calls follows */ ++ /* borrowed from Zip's global.c */ ++ int no_wild = 0; ++ int allow_regex=1; ++ /* This fix provided by akt@m5.dion.ne.jp for Japanese. ++ See 21 July 2006 mail. ++ It only applies when p is pointing to a doublebyte character and ++ things like / and wildcards are not doublebyte. This probably ++ should not be needed. */ + +- /* Get first character, the pattern for new recmatch calls follows */ +- c = *p; INCSTR(p); ++#ifdef _MBCS ++ if (CLEN(p) == 2) { ++ if (CLEN(s) == 2) { ++ return (*p == *s && *(p+1) == *(s+1)) ? ++ recmatch(p + 2, s + 2, cs) : 0; ++ } else { ++ return 0; ++ } ++ } ++#endif /* ?_MBCS */ + +- /* If that was the end of the pattern, match if string empty too */ +- if (c == 0) +- return *s == 0; ++ c = *POSTINCSTR(p); + +- /* '?' (or '%') matches any character (but not an empty string). */ +- if (c == WILDCHAR) +-#ifdef WILD_STOP_AT_DIR +- /* If uO.W_flag is non-zero, it won't match '/' */ +- return (*s && (!sepc || *s != (uch)sepc)) +- ? recmatch(p, s + CLEN(s), ic, sepc) : 0; +-#else +- return *s ? recmatch(p, s + CLEN(s), ic) : 0; +-#endif ++ /* If that was the end of the pattern, match if string empty too */ ++ if (c == 0) ++ return *s == 0; ++ ++ /* '?' (or '%' or '#') matches any character (but not an empty string) */ ++ if (c == WILDCHR_SINGLE) { ++ if (wild_stop_at_dir) ++ return (*s && *s != DIRSEP_CHR) ? recmatch(p, s + CLEN(s), cs) : 0; ++ else ++ return *s ? recmatch(p, s + CLEN(s), cs) : 0; ++ } + +- /* '*' matches any number of characters, including zero */ ++ /* WILDCHR_MULTI ('*') matches any number of characters, including zero */ + #ifdef AMIGA +- if (c == '#' && *p == '?') /* "#?" is Amiga-ese for "*" */ +- c = '*', p++; ++ if (!no_wild && c == '#' && *p == '?') /* "#?" is Amiga-ese for "*" */ ++ c = WILDCHR_MULTI, p++; + #endif /* AMIGA */ +- if (c == '*') { +-#ifdef WILD_STOP_AT_DIR +- if (sepc) { +- /* check for single "*" or double "**" */ +-# ifdef AMIGA +- if ((c = p[0]) == '#' && p[1] == '?') /* "#?" is Amiga-ese for "*" */ +- c = '*', p++; +- if (c != '*') { +-# else /* !AMIGA */ +- if (*p != '*') { +-# endif /* ?AMIGA */ +- /* single "*": this doesn't match the dirsep character */ +- for (; *s && *s != (uch)sepc; INCSTR(s)) +- if ((c = recmatch(p, s, ic, sepc)) != 0) +- return (int)c; +- /* end of pattern: matched if at end of string, else continue */ +- if (*p == '\0') +- return (*s == 0); +- /* continue to match if at sepc in pattern, else give up */ +- return (*p == (uch)sepc || (*p == '\\' && p[1] == (uch)sepc)) +- ? recmatch(p, s, ic, sepc) : 2; +- } +- /* "**": this matches slashes */ +- ++p; /* move p behind the second '*' */ +- /* and continue with the non-W_flag code variant */ +- } +-#endif /* WILD_STOP_AT_DIR */ ++ if (!no_wild && c == WILDCHR_MULTI) ++ { ++ if (wild_stop_at_dir) { ++ /* Check for an immediately following WILDCHR_MULTI */ ++# ifdef AMIGA ++ if ((c = p[0]) == '#' && p[1] == '?') /* "#?" is Amiga-ese for "*" */ ++ c = WILDCHR_MULTI, p++; ++ if (c != WILDCHR_MULTI) { ++# else /* !AMIGA */ ++ if (*p != WILDCHR_MULTI) { ++# endif /* ?AMIGA */ ++ /* Single WILDCHR_MULTI ('*'): this doesn't match slashes */ ++ for (; *s && *s != DIRSEP_CHR; INCSTR(s)) ++ if ((c = recmatch(p, s, cs)) != 0) ++ return c; ++ /* end of pattern: matched if at end of string, else continue */ + if (*p == 0) +- return 1; +- if (isshexp((ZCONST char *)p) == NULL) { +- /* Optimization for rest of pattern being a literal string: +- * If there are no other shell expression chars in the rest +- * of the pattern behind the multi-char wildcard, then just +- * compare the literal string tail. +- */ +- ZCONST uch *srest; +- +- srest = s + (strlen((ZCONST char *)s) - strlen((ZCONST char *)p)); +- if (srest - s < 0) +- /* remaining literal string from pattern is longer than rest +- * of test string, there can't be a match +- */ +- return 0; +- else +- /* compare the remaining literal pattern string with the last +- * bytes of the test string to check for a match +- */ ++ return (*s == 0); ++ /* continue to match if at DIRSEP_CHR in pattern, else give up */ ++ return (*p == DIRSEP_CHR || (*p == '\\' && p[1] == DIRSEP_CHR)) ++ ? recmatch(p, s, cs) : 2; ++ } ++ /* Two consecutive WILDCHR_MULTI ("**"): this matches DIRSEP_CHR ('/') */ ++ p++; /* move p past the second WILDCHR_MULTI */ ++ /* continue with the normal non-WILD_STOP_AT_DIR code */ ++ } /* wild_stop_at_dir */ ++ ++ /* Not wild_stop_at_dir */ ++ if (*p == 0) ++ return 1; ++ if (!isshexp((char *)p)) ++ { ++ /* optimization for rest of pattern being a literal string */ ++ ++ /* optimization to handle patterns like *.txt */ ++ /* if the first char in the pattern is '*' and there */ ++ /* are no other shell expression chars, i.e. a literal string */ ++ /* then just compare the literal string at the end */ ++ ++ ZCONST char *srest; ++ ++ srest = s + (strlen(s) - strlen(p)); ++ if (srest - s < 0) ++ /* remaining literal string from pattern is longer than rest of ++ test string, there can't be a match ++ */ ++ return 0; ++ else ++ /* compare the remaining literal pattern string with the last bytes ++ of the test string to check for a match */ + #ifdef _MBCS +- { +- ZCONST uch *q = s; ++ { ++ ZCONST char *q = s; + +- /* MBCS-aware code must not scan backwards into a string from +- * the end. +- * So, we have to move forward by character from our well-known +- * character position s in the test string until we have +- * advanced to the srest position. +- */ +- while (q < srest) +- INCSTR(q); +- /* In case the byte *srest is a trailing byte of a multibyte +- * character in the test string s, we have actually advanced +- * past the position (srest). +- * For this case, the match has failed! +- */ +- if (q != srest) +- return 0; +- return ((ic +- ? namecmp((ZCONST char *)p, (ZCONST char *)q) +- : strcmp((ZCONST char *)p, (ZCONST char *)q) +- ) == 0); +- } ++ /* MBCS-aware code must not scan backwards into a string from ++ * the end. ++ * So, we have to move forward by character from our well-known ++ * character position s in the test string until we have advanced ++ * to the srest position. ++ */ ++ while (q < srest) ++ INCSTR(q); ++ /* In case the byte *srest is a trailing byte of a multibyte ++ * character, we have actually advanced past the position (srest). ++ * For this case, the match has failed! ++ */ ++ if (q != srest) ++ return 0; ++ return ((cs ? strcmp(p, q) : namecmp(p, q)) == 0); ++ } + #else /* !_MBCS */ +- return ((ic +- ? namecmp((ZCONST char *)p, (ZCONST char *)srest) +- : strcmp((ZCONST char *)p, (ZCONST char *)srest) +- ) == 0); ++ return ((cs ? strcmp(p, srest) : namecmp(p, srest)) == 0); + #endif /* ?_MBCS */ +- } else { +- /* pattern contains more wildcards, continue with recursion... */ +- for (; *s; INCSTR(s)) +- if ((c = recmatch(p, s, ic __WDL)) != 0) +- return (int)c; +- return 2; /* 2 means give up--match will return false */ +- } + } +- +- /* Parse and process the list of characters and ranges in brackets */ +- if (c == BEG_RANGE) { +- int e; /* flag true if next char to be taken literally */ +- ZCONST uch *q; /* pointer to end of [-] group */ +- int r; /* flag true to match anything but the range */ +- +- if (*s == 0) /* need a character to match */ +- return 0; +- p += (r = (*p == '!' || *p == '^')); /* see if reverse */ +- for (q = p, e = 0; *q; INCSTR(q)) /* find closing bracket */ +- if (e) +- e = 0; +- else +- if (*q == '\\') /* GRR: change to ^ for MS-DOS, OS/2? */ +- e = 1; +- else if (*q == END_RANGE) +- break; +- if (*q != END_RANGE) /* nothing matches if bad syntax */ +- return 0; +- for (c = 0, e = (*p == '-'); p < q; INCSTR(p)) { +- /* go through the list */ +- if (!e && *p == '\\') /* set escape flag if \ */ +- e = 1; +- else if (!e && *p == '-') /* set start of range if - */ +- c = *(p-1); +- else { +- unsigned int cc = Case(*s); +- +- if (*(p+1) != '-') +- for (c = c ? c : *p; c <= *p; c++) /* compare range */ +- if ((unsigned)Case(c) == cc) /* typecast for MSC bug */ +- return r ? 0 : recmatch(q + 1, s + 1, ic __WDL); +- c = e = 0; /* clear range, escape flags */ +- } +- } +- return r ? recmatch(q + CLEN(q), s + CLEN(s), ic __WDL) : 0; +- /* bracket match failed */ ++ else ++ { ++ /* pattern contains more wildcards, continue with recursion... */ ++ for (; *s; INCSTR(s)) ++ if ((c = recmatch(p, s, cs)) != 0) ++ return c; ++ return 2; /* 2 means give up--shmatch will return false */ + } ++ } + +- /* if escape ('\\'), just compare next character */ +- if (c == '\\' && (c = *p++) == 0) /* if \ at end, then syntax error */ +- return 0; ++#ifndef VMS /* No bracket matching in VMS */ ++ /* Parse and process the list of characters and ranges in brackets */ ++ if (!no_wild && allow_regex && c == '[') ++ { ++ int e; /* flag true if next char to be taken literally */ ++ ZCONST char *q; /* pointer to end of [-] group */ ++ int r; /* flag true to match anything but the range */ ++ ++ if (*s == 0) /* need a character to match */ ++ return 0; ++ p += (r = (*p == '!' || *p == '^')); /* see if reverse */ ++ for (q = p, e = 0; *q; q++) /* find closing bracket */ ++ if (e) ++ e = 0; ++ else ++ if (*q == '\\') ++ e = 1; ++ else if (*q == ']') ++ break; ++ if (*q != ']') /* nothing matches if bad syntax */ ++ return 0; ++ for (c = 0, e = *p == '-'; p < q; p++) /* go through the list */ ++ { ++ if (e == 0 && *p == '\\') /* set escape flag if \ */ ++ e = 1; ++ else if (e == 0 && *p == '-') /* set start of range if - */ ++ c = *(p-1); ++ else ++ { ++ uch cc = (cs ? (uch)*s : case_map((uch)*s)); ++ uch uc = (uch) c; ++ if (*(p+1) != '-') ++ for (uc = uc ? uc : (uch)*p; uc <= (uch)*p; uc++) ++ /* compare range */ ++ if ((cs ? uc : case_map(uc)) == cc) ++ return r ? 0 : recmatch(q + CLEN(q), s + CLEN(s), cs); ++ c = e = 0; /* clear range, escape flags */ ++ } ++ } ++ return r ? recmatch(q + CLEN(q), s + CLEN(s), cs) : 0; ++ /* bracket match failed */ ++ } ++#endif /* !VMS */ + +- /* just a character--compare it */ +-#ifdef QDOS +- return QMatch(Case((uch)c), Case(*s)) ? +- recmatch(p, s + CLEN(s), ic __WDL) : 0; +-#else +- return Case((uch)c) == Case(*s) ? +- recmatch(p, s + CLEN(s), ic __WDL) : 0; +-#endif ++ /* If escape ('\'), just compare next character */ ++ if (!no_wild && c == '\\') ++ if ((c = *p++) == '\0') /* if \ at end, then syntax error */ ++ return 0; ++ ++#ifdef VMS ++ /* 2005-11-06 SMS. ++ Handle "..." wildcard in p with "." or "]" in s. ++ */ ++ if ((c == '.') && (*p == '.') && (*(p+ CLEN( p)) == '.') && ++ ((*s == '.') || (*s == ']'))) ++ { ++ /* Match "...]" with "]". Continue after "]" in both. */ ++ if ((*(p+ 2* CLEN( p)) == ']') && (*s == ']')) ++ return recmatch( (p+ 3* CLEN( p)), (s+ CLEN( s)), cs); ++ ++ /* Else, look for a reduced match in s, until "]" in or end of s. */ ++ for (; *s && (*s != ']'); INCSTR(s)) ++ if (*s == '.') ++ /* If reduced match, then continue after "..." in p, "." in s. */ ++ if ((c = recmatch( (p+ CLEN( p)), s, cs)) != 0) ++ return (int)c; ++ ++ /* Match "...]" with "]". Continue after "]" in both. */ ++ if ((*(p+ 2* CLEN( p)) == ']') && (*s == ']')) ++ return recmatch( (p+ 3* CLEN( p)), (s+ CLEN( s)), cs); ++ ++ /* No reduced match. Quit. */ ++ return 2; ++ } ++ ++#endif /* def VMS */ ++ ++ /* Just a character--compare it */ ++ return (cs ? c == *s : case_map((uch)c) == case_map((uch)*s)) ? ++ recmatch(p, s + CLEN(s), cs) : 0; ++} + +-} /* end function recmatch() */ + + + ++/*************************************************************************************************/ + static char *isshexp(p) + ZCONST char *p; + /* If p is a sh expression, a pointer to the first special character is diff --git a/gnu/packages/patches/unzip-manpage-fix.patch b/gnu/packages/patches/unzip-manpage-fix.patch new file mode 100644 index 0000000000..ec06516bbb --- /dev/null +++ b/gnu/packages/patches/unzip-manpage-fix.patch @@ -0,0 +1,11 @@ +--- unzip60/man/unzip.1 2011-01-11 11:59:59.000000000 +0000 ++++ unzip60/man/unzip_new 2011-02-05 18:45:55.000000000 +0000 +@@ -424,7 +424,7 @@ + .\" Amiga support possible eventually, but not yet + [MS-DOS, OS/2, NT] restore the volume label if the extraction medium is + removable (e.g., a diskette). Doubling the option (\fB\-$$\fP) allows fixed +-media (hard disks) to be labelled as well. By default, volume labels are ++media (hard disks) to be labeled as well. By default, volume labels are + ignored. + .IP \fB\-/\fP\ \fIextensions\fP + [Acorn only] overrides the extension list supplied by Unzip$Ext environment diff --git a/gnu/packages/patches/unzip-overflow.patch b/gnu/packages/patches/unzip-overflow.patch new file mode 100644 index 0000000000..228c28377f --- /dev/null +++ b/gnu/packages/patches/unzip-overflow.patch @@ -0,0 +1,25 @@ +diff --git a/extract.c b/extract.c +index a0a4929..9ef80b3 100644 +--- a/extract.c ++++ b/extract.c +@@ -2214,6 +2214,7 @@ static int test_compr_eb(__G__ eb, eb_size, compr_offset, test_uc_ebdata) + ulg eb_ucsize; + uch *eb_ucptr; + int r; ++ ush method; + + if (compr_offset < 4) /* field is not compressed: */ + return PK_OK; /* do nothing and signal OK */ +@@ -2223,6 +2224,12 @@ static int test_compr_eb(__G__ eb, eb_size, compr_offset, test_uc_ebdata) + eb_size <= (compr_offset + EB_CMPRHEADLEN))) + return IZ_EF_TRUNC; /* no compressed data! */ + ++ method = makeword(eb + (EB_HEADSIZE + compr_offset)); ++ if ((method == STORED) && (eb_size != compr_offset + EB_CMPRHEADLEN + eb_ucsize)) ++ return PK_ERR; /* compressed & uncompressed ++ * should match in STORED ++ * method */ ++ + if ( + #ifdef INT_16BIT + (((ulg)(extent)eb_ucsize) != eb_ucsize) || diff --git a/gnu/packages/patches/unzip-symlink.patch b/gnu/packages/patches/unzip-symlink.patch new file mode 100644 index 0000000000..d7eaf160b4 --- /dev/null +++ b/gnu/packages/patches/unzip-symlink.patch @@ -0,0 +1,17 @@ +diff --git a/process.c b/process.c +index 1e9a1e1..905732b 100644 +--- a/process.c ++++ b/process.c +@@ -1751,6 +1751,12 @@ int process_cdir_file_hdr(__G) /* return PK-type error code */ + = (G.crec.general_purpose_bit_flag & (1 << 11)) == (1 << 11); + #endif + ++#ifdef SYMLINKS ++ /* Initialize the symlink flag, may be set by the platform-specific ++ mapattr function. */ ++ G.pInfo->symlink = 0; ++#endif ++ + return PK_COOL; + + } /* end function process_cdir_file_hdr() */ diff --git a/gnu/packages/patches/unzip-timestamp.patch b/gnu/packages/patches/unzip-timestamp.patch new file mode 100644 index 0000000000..2aa9424eb8 --- /dev/null +++ b/gnu/packages/patches/unzip-timestamp.patch @@ -0,0 +1,41 @@ +From: "Steven M. Schweda" +Subject: Do not ignore extra fields containing Unix Timestamps +Bug-Debian: https://bugs.debian.org/842993 +X-Debian-version: 6.0-21 + +--- a/process.c ++++ b/process.c +@@ -2914,10 +2914,13 @@ + break; + + case EF_IZUNIX2: +- if (have_new_type_eb == 0) { +- flags &= ~0x0ff; /* ignore any previous IZUNIX field */ ++ if (have_new_type_eb == 0) { /* (< 1) */ + have_new_type_eb = 1; + } ++ if (have_new_type_eb <= 1) { ++ /* Ignore any prior (EF_IZUNIX/EF_PKUNIX) UID/GID. */ ++ flags &= 0x0ff; ++ } + #ifdef IZ_HAVE_UXUIDGID + if (have_new_type_eb > 1) + break; /* IZUNIX3 overrides IZUNIX2 e.f. block ! */ +@@ -2933,6 +2936,8 @@ + /* new 3rd generation Unix ef */ + have_new_type_eb = 2; + ++ /* Ignore any prior EF_IZUNIX/EF_PKUNIX/EF_IZUNIX2 UID/GID. */ ++ flags &= 0x0ff; + /* + Version 1 byte version of this extra field, currently 1 + UIDSize 1 byte Size of UID field +@@ -2953,8 +2958,6 @@ + uid_size = *((EB_HEADSIZE + 1) + ef_buf); + gid_size = *((EB_HEADSIZE + uid_size + 2) + ef_buf); + +- flags &= ~0x0ff; /* ignore any previous UNIX field */ +- + if ( read_ux3_value((EB_HEADSIZE + 2) + ef_buf, + uid_size, &z_uidgid[0]) + && diff --git a/gnu/packages/patches/unzip-valgrind.patch b/gnu/packages/patches/unzip-valgrind.patch new file mode 100644 index 0000000000..41a752023b --- /dev/null +++ b/gnu/packages/patches/unzip-valgrind.patch @@ -0,0 +1,26 @@ +diff --git a/fileio.c b/fileio.c +index ba0a1d0..03fc4be 100644 +--- a/fileio.c ++++ b/fileio.c +@@ -2006,6 +2006,7 @@ int do_string(__G__ length, option) /* return PK-type error code */ + unsigned comment_bytes_left; + unsigned int block_len; + int error=PK_OK; ++ unsigned int length2; + #ifdef AMIGA + char tmp_fnote[2 * AMIGA_FILENOTELEN]; /* extra room for squozen chars */ + #endif +@@ -2292,8 +2293,12 @@ int do_string(__G__ length, option) /* return PK-type error code */ + seek_zipf(__G__ G.cur_zipfile_bufstart - G.extra_bytes + + (G.inptr-G.inbuf) + length); + } else { +- if (readbuf(__G__ (char *)G.extra_field, length) == 0) ++ if ((length2 = readbuf(__G__ (char *)G.extra_field, length)) == 0) + return PK_EOF; ++ if(length2 < length) { ++ memset (__G__ (char *)G.extra_field+length2, 0 , length-length2); ++ length = length2; ++ } + /* Looks like here is where extra fields are read */ + getZip64Data(__G__ G.extra_field, length); + #ifdef UNICODE_SUPPORT diff --git a/gnu/packages/patches/unzip-x-option.patch b/gnu/packages/patches/unzip-x-option.patch new file mode 100644 index 0000000000..72c77aeb78 --- /dev/null +++ b/gnu/packages/patches/unzip-x-option.patch @@ -0,0 +1,28 @@ +--- ./process.c.orig 2009-03-06 02:25:10.000000000 +0100 ++++ ./process.c 2013-09-12 10:51:16.000000000 +0200 +@@ -2901,9 +2901,9 @@ + */ + + #ifdef IZ_HAVE_UXUIDGID +- if (eb_len >= EB_UX3_MINLEN +- && z_uidgid != NULL +- && (*((EB_HEADSIZE + 0) + ef_buf) == 1) ++ if ((eb_len >= EB_UX3_MINLEN) ++ && (z_uidgid != NULL) ++ && ((*((EB_HEADSIZE + 0) + ef_buf) == 1))) + /* only know about version 1 */ + { + uch uid_size; +@@ -2915,10 +2915,10 @@ + flags &= ~0x0ff; /* ignore any previous UNIX field */ + + if ( read_ux3_value((EB_HEADSIZE + 2) + ef_buf, +- uid_size, z_uidgid[0]) ++ uid_size, &z_uidgid[0]) + && + read_ux3_value((EB_HEADSIZE + uid_size + 3) + ef_buf, +- gid_size, z_uidgid[1]) ) ++ gid_size, &z_uidgid[1]) ) + { + flags |= EB_UX2_VALID; /* signal success */ + } diff --git a/gnu/packages/patches/unzip-zipbomb-manpage.patch b/gnu/packages/patches/unzip-zipbomb-manpage.patch new file mode 100644 index 0000000000..cdeeea5c93 --- /dev/null +++ b/gnu/packages/patches/unzip-zipbomb-manpage.patch @@ -0,0 +1,25 @@ +From 6fe72291a5563cdbcd2bdd87e36528537b7cdcfb Mon Sep 17 00:00:00 2001 +From: Jakub Martisko +Date: Mon, 18 Nov 2019 14:17:46 +0100 +Subject: [PATCH] update the man page + +--- + man/unzip.1 | 2 ++ + 1 file changed, 2 insertions(+) + +diff --git a/man/unzip.1 b/man/unzip.1 +index 21816d1..4d66073 100644 +--- a/man/unzip.1 ++++ b/man/unzip.1 +@@ -850,6 +850,8 @@ the specified zipfiles were not found. + invalid options were specified on the command line. + .IP 11 + no matching files were found. ++.IP 12 ++invalid zip file with overlapped components (possible zip bomb). + .IP 50 + the disk is (or was) full during extraction. + .IP 51 +-- +2.23.0 + diff --git a/gnu/packages/patches/unzip-zipbomb-part1.patch b/gnu/packages/patches/unzip-zipbomb-part1.patch new file mode 100644 index 0000000000..35cf856522 --- /dev/null +++ b/gnu/packages/patches/unzip-zipbomb-part1.patch @@ -0,0 +1,25 @@ +From 41beb477c5744bc396fa1162ee0c14218ec12213 Mon Sep 17 00:00:00 2001 +From: Mark Adler +Date: Mon, 27 May 2019 08:20:32 -0700 +Subject: [PATCH] Fix bug in undefer_input() that misplaced the input state. + +--- + fileio.c | 4 +++- + 1 file changed, 3 insertions(+), 1 deletion(-) + +diff --git a/fileio.c b/fileio.c +index c042987..bc00d74 100644 +--- a/fileio.c ++++ b/fileio.c +@@ -530,8 +530,10 @@ void undefer_input(__G) + * This condition was checked when G.incnt_leftover was set > 0 in + * defer_leftover_input(), and it is NOT allowed to touch G.csize + * before calling undefer_input() when (G.incnt_leftover > 0) +- * (single exception: see read_byte()'s "G.csize <= 0" handling) !! ++ * (single exception: see readbyte()'s "G.csize <= 0" handling) !! + */ ++ if (G.csize < 0L) ++ G.csize = 0L; + G.incnt = G.incnt_leftover + (int)G.csize; + G.inptr = G.inptr_leftover - (int)G.csize; + G.incnt_leftover = 0; diff --git a/gnu/packages/patches/unzip-zipbomb-part2.patch b/gnu/packages/patches/unzip-zipbomb-part2.patch new file mode 100644 index 0000000000..903c845763 --- /dev/null +++ b/gnu/packages/patches/unzip-zipbomb-part2.patch @@ -0,0 +1,349 @@ +From 47b3ceae397d21bf822bc2ac73052a4b1daf8e1c Mon Sep 17 00:00:00 2001 +From: Mark Adler +Date: Tue, 11 Jun 2019 22:01:18 -0700 +Subject: [PATCH] Detect and reject a zip bomb using overlapped entries. + +This detects an invalid zip file that has at least one entry that +overlaps with another entry or with the central directory to the +end of the file. A Fifield zip bomb uses overlapped local entries +to vastly increase the potential inflation ratio. Such an invalid +zip file is rejected. + +See https://www.bamsoftware.com/hacks/zipbomb/ for David Fifield's +analysis, construction, and examples of such zip bombs. + +The detection maintains a list of covered spans of the zip files +so far, where the central directory to the end of the file and any +bytes preceding the first entry at zip file offset zero are +considered covered initially. Then as each entry is decompressed +or tested, it is considered covered. When a new entry is about to +be processed, its initial offset is checked to see if it is +contained by a covered span. If so, the zip file is rejected as +invalid. + +This commit depends on a preceding commit: "Fix bug in +undefer_input() that misplaced the input state." +--- + extract.c | 190 +++++++++++++++++++++++++++++++++++++++++++++++++++++- + globals.c | 1 + + globals.h | 3 + + process.c | 11 ++++ + unzip.h | 1 + + 5 files changed, 205 insertions(+), 1 deletion(-) + +diff --git a/extract.c b/extract.c +index 1acd769..0973a33 100644 +--- a/extract.c ++++ b/extract.c +@@ -319,6 +319,125 @@ static ZCONST char Far UnsupportedExtraField[] = + "\nerror: unsupported extra-field compression type (%u)--skipping\n"; + static ZCONST char Far BadExtraFieldCRC[] = + "error [%s]: bad extra-field CRC %08lx (should be %08lx)\n"; ++static ZCONST char Far NotEnoughMemCover[] = ++ "error: not enough memory for bomb detection\n"; ++static ZCONST char Far OverlappedComponents[] = ++ "error: invalid zip file with overlapped components (possible zip bomb)\n"; ++ ++ ++ ++ ++ ++/* A growable list of spans. */ ++typedef zoff_t bound_t; ++typedef struct { ++ bound_t beg; /* start of the span */ ++ bound_t end; /* one past the end of the span */ ++} span_t; ++typedef struct { ++ span_t *span; /* allocated, distinct, and sorted list of spans */ ++ size_t num; /* number of spans in the list */ ++ size_t max; /* allocated number of spans (num <= max) */ ++} cover_t; ++ ++/* ++ * Return the index of the first span in cover whose beg is greater than val. ++ * If there is no such span, then cover->num is returned. ++ */ ++static size_t cover_find(cover, val) ++ cover_t *cover; ++ bound_t val; ++{ ++ size_t lo = 0, hi = cover->num; ++ while (lo < hi) { ++ size_t mid = (lo + hi) >> 1; ++ if (val < cover->span[mid].beg) ++ hi = mid; ++ else ++ lo = mid + 1; ++ } ++ return hi; ++} ++ ++/* Return true if val lies within any one of the spans in cover. */ ++static int cover_within(cover, val) ++ cover_t *cover; ++ bound_t val; ++{ ++ size_t pos = cover_find(cover, val); ++ return pos > 0 && val < cover->span[pos - 1].end; ++} ++ ++/* ++ * Add a new span to the list, but only if the new span does not overlap any ++ * spans already in the list. The new span covers the values beg..end-1. beg ++ * must be less than end. ++ * ++ * Keep the list sorted and merge adjacent spans. Grow the allocated space for ++ * the list as needed. On success, 0 is returned. If the new span overlaps any ++ * existing spans, then 1 is returned and the new span is not added to the ++ * list. If the new span is invalid because beg is greater than or equal to ++ * end, then -1 is returned. If the list needs to be grown but the memory ++ * allocation fails, then -2 is returned. ++ */ ++static int cover_add(cover, beg, end) ++ cover_t *cover; ++ bound_t beg; ++ bound_t end; ++{ ++ size_t pos; ++ int prec, foll; ++ ++ if (beg >= end) ++ /* The new span is invalid. */ ++ return -1; ++ ++ /* Find where the new span should go, and make sure that it does not ++ overlap with any existing spans. */ ++ pos = cover_find(cover, beg); ++ if ((pos > 0 && beg < cover->span[pos - 1].end) || ++ (pos < cover->num && end > cover->span[pos].beg)) ++ return 1; ++ ++ /* Check for adjacencies. */ ++ prec = pos > 0 && beg == cover->span[pos - 1].end; ++ foll = pos < cover->num && end == cover->span[pos].beg; ++ if (prec && foll) { ++ /* The new span connects the preceding and following spans. Merge the ++ following span into the preceding span, and delete the following ++ span. */ ++ cover->span[pos - 1].end = cover->span[pos].end; ++ cover->num--; ++ memmove(cover->span + pos, cover->span + pos + 1, ++ (cover->num - pos) * sizeof(span_t)); ++ } ++ else if (prec) ++ /* The new span is adjacent only to the preceding span. Extend the end ++ of the preceding span. */ ++ cover->span[pos - 1].end = end; ++ else if (foll) ++ /* The new span is adjacent only to the following span. Extend the ++ beginning of the following span. */ ++ cover->span[pos].beg = beg; ++ else { ++ /* The new span has gaps between both the preceding and the following ++ spans. Assure that there is room and insert the span. */ ++ if (cover->num == cover->max) { ++ size_t max = cover->max == 0 ? 16 : cover->max << 1; ++ span_t *span = realloc(cover->span, max * sizeof(span_t)); ++ if (span == NULL) ++ return -2; ++ cover->span = span; ++ cover->max = max; ++ } ++ memmove(cover->span + pos + 1, cover->span + pos, ++ (cover->num - pos) * sizeof(span_t)); ++ cover->num++; ++ cover->span[pos].beg = beg; ++ cover->span[pos].end = end; ++ } ++ return 0; ++} + + + +@@ -374,6 +493,29 @@ int extract_or_test_files(__G) /* return PK-type error code */ + } + #endif /* !SFX || SFX_EXDIR */ + ++ /* One more: initialize cover structure for bomb detection. Start with a ++ span that covers the central directory though the end of the file. */ ++ if (G.cover == NULL) { ++ G.cover = malloc(sizeof(cover_t)); ++ if (G.cover == NULL) { ++ Info(slide, 0x401, ((char *)slide, ++ LoadFarString(NotEnoughMemCover))); ++ return PK_MEM; ++ } ++ ((cover_t *)G.cover)->span = NULL; ++ ((cover_t *)G.cover)->max = 0; ++ } ++ ((cover_t *)G.cover)->num = 0; ++ if ((G.extra_bytes != 0 && ++ cover_add((cover_t *)G.cover, 0, G.extra_bytes) != 0) || ++ cover_add((cover_t *)G.cover, ++ G.extra_bytes + G.ecrec.offset_start_central_directory, ++ G.ziplen) != 0) { ++ Info(slide, 0x401, ((char *)slide, ++ LoadFarString(NotEnoughMemCover))); ++ return PK_MEM; ++ } ++ + /*--------------------------------------------------------------------------- + The basic idea of this function is as follows. Since the central di- + rectory lies at the end of the zipfile and the member files lie at the +@@ -591,7 +733,8 @@ int extract_or_test_files(__G) /* return PK-type error code */ + if (error > error_in_archive) + error_in_archive = error; + /* ...and keep going (unless disk full or user break) */ +- if (G.disk_full > 1 || error_in_archive == IZ_CTRLC) { ++ if (G.disk_full > 1 || error_in_archive == IZ_CTRLC || ++ error == PK_BOMB) { + /* clear reached_end to signal premature stop ... */ + reached_end = FALSE; + /* ... and cancel scanning the central directory */ +@@ -1060,6 +1203,11 @@ static int extract_or_test_entrylist(__G__ numchunk, + + /* seek_zipf(__G__ pInfo->offset); */ + request = G.pInfo->offset + G.extra_bytes; ++ if (cover_within((cover_t *)G.cover, request)) { ++ Info(slide, 0x401, ((char *)slide, ++ LoadFarString(OverlappedComponents))); ++ return PK_BOMB; ++ } + inbuf_offset = request % INBUFSIZ; + bufstart = request - inbuf_offset; + +@@ -1591,6 +1739,18 @@ static int extract_or_test_entrylist(__G__ numchunk, + return IZ_CTRLC; /* cancel operation by user request */ + } + #endif ++ error = cover_add((cover_t *)G.cover, request, ++ G.cur_zipfile_bufstart + (G.inptr - G.inbuf)); ++ if (error < 0) { ++ Info(slide, 0x401, ((char *)slide, ++ LoadFarString(NotEnoughMemCover))); ++ return PK_MEM; ++ } ++ if (error != 0) { ++ Info(slide, 0x401, ((char *)slide, ++ LoadFarString(OverlappedComponents))); ++ return PK_BOMB; ++ } + #ifdef MACOS /* MacOS is no preemptive OS, thus call event-handling by hand */ + UserStop(); + #endif +@@ -1992,6 +2152,34 @@ static int extract_or_test_member(__G) /* return PK-type error code */ + } + + undefer_input(__G); ++ ++ if ((G.lrec.general_purpose_bit_flag & 8) != 0) { ++ /* skip over data descriptor (harder than it sounds, due to signature ++ * ambiguity) ++ */ ++# define SIG 0x08074b50 ++# define LOW 0xffffffff ++ uch buf[12]; ++ unsigned shy = 12 - readbuf((char *)buf, 12); ++ ulg crc = shy ? 0 : makelong(buf); ++ ulg clen = shy ? 0 : makelong(buf + 4); ++ ulg ulen = shy ? 0 : makelong(buf + 8); /* or high clen if ZIP64 */ ++ if (crc == SIG && /* if not SIG, no signature */ ++ (G.lrec.crc32 != SIG || /* if not SIG, have signature */ ++ (clen == SIG && /* if not SIG, no signature */ ++ ((G.lrec.csize & LOW) != SIG || /* if not SIG, have signature */ ++ (ulen == SIG && /* if not SIG, no signature */ ++ (G.zip64 ? G.lrec.csize >> 32 : G.lrec.ucsize) != SIG ++ /* if not SIG, have signature */ ++ ))))) ++ /* skip four more bytes to account for signature */ ++ shy += 4 - readbuf((char *)buf, 4); ++ if (G.zip64) ++ shy += 8 - readbuf((char *)buf, 8); /* skip eight more for ZIP64 */ ++ if (shy) ++ error = PK_ERR; ++ } ++ + return error; + + } /* end function extract_or_test_member() */ +diff --git a/globals.c b/globals.c +index fa8cca5..1e0f608 100644 +--- a/globals.c ++++ b/globals.c +@@ -181,6 +181,7 @@ Uz_Globs *globalsCtor() + # if (!defined(NO_TIMESTAMPS)) + uO.D_flag=1; /* default to '-D', no restoration of dir timestamps */ + # endif ++ G.cover = NULL; /* not allocated yet */ + #endif + + uO.lflag=(-1); +diff --git a/globals.h b/globals.h +index 11b7215..2bdcdeb 100644 +--- a/globals.h ++++ b/globals.h +@@ -260,12 +260,15 @@ typedef struct Globals { + ecdir_rec ecrec; /* used in unzip.c, extract.c */ + z_stat statbuf; /* used by main, mapname, check_for_newer */ + ++ int zip64; /* true if Zip64 info in extra field */ ++ + int mem_mode; + uch *outbufptr; /* extract.c static */ + ulg outsize; /* extract.c static */ + int reported_backslash; /* extract.c static */ + int disk_full; + int newfile; ++ void **cover; /* used in extract.c for bomb detection */ + + int didCRlast; /* fileio static */ + ulg numlines; /* fileio static: number of lines printed */ +diff --git a/process.c b/process.c +index 1e9a1e1..d2e4dc3 100644 +--- a/process.c ++++ b/process.c +@@ -637,6 +637,13 @@ void free_G_buffers(__G) /* releases all memory allocated in global vars */ + } + #endif + ++ /* Free the cover span list and the cover structure. */ ++ if (G.cover != NULL) { ++ free(*(G.cover)); ++ free(G.cover); ++ G.cover = NULL; ++ } ++ + } /* end function free_G_buffers() */ + + +@@ -1890,6 +1897,8 @@ int getZip64Data(__G__ ef_buf, ef_len) + #define Z64FLGS 0xffff + #define Z64FLGL 0xffffffff + ++ G.zip64 = FALSE; ++ + if (ef_len == 0 || ef_buf == NULL) + return PK_COOL; + +@@ -1927,6 +1936,8 @@ int getZip64Data(__G__ ef_buf, ef_len) + #if 0 + break; /* Expect only one EF_PKSZ64 block. */ + #endif /* 0 */ ++ ++ G.zip64 = TRUE; + } + + /* Skip this extra field block. */ +diff --git a/unzip.h b/unzip.h +index 5b2a326..ed24a5b 100644 +--- a/unzip.h ++++ b/unzip.h +@@ -645,6 +645,7 @@ typedef struct _Uzp_cdir_Rec { + #define PK_NOZIP 9 /* zipfile not found */ + #define PK_PARAM 10 /* bad or illegal parameters specified */ + #define PK_FIND 11 /* no files found */ ++#define PK_BOMB 12 /* likely zip bomb */ + #define PK_DISK 50 /* disk full */ + #define PK_EOF 51 /* unexpected EOF */ + diff --git a/gnu/packages/patches/unzip-zipbomb-part3.patch b/gnu/packages/patches/unzip-zipbomb-part3.patch new file mode 100644 index 0000000000..3b8d67b773 --- /dev/null +++ b/gnu/packages/patches/unzip-zipbomb-part3.patch @@ -0,0 +1,112 @@ +From 6d351831be705cc26d897db44f878a978f4138fc Mon Sep 17 00:00:00 2001 +From: Mark Adler +Date: Thu, 25 Jul 2019 20:43:17 -0700 +Subject: [PATCH] Do not raise a zip bomb alert for a misplaced central + directory. + +There is a zip-like file in the Firefox distribution, omni.ja, +which is a zip container with the central directory placed at the +start of the file instead of after the local entries as required +by the zip standard. This commit marks the actual location of the +central directory, as well as the end of central directory records, +as disallowed locations. This now permits such containers to not +raise a zip bomb alert, where in fact there are no overlaps. +--- + extract.c | 25 +++++++++++++++++++------ + process.c | 6 ++++++ + unzpriv.h | 10 ++++++++++ + 3 files changed, 35 insertions(+), 6 deletions(-) + +diff --git a/extract.c b/extract.c +index 0973a33..1b73cb0 100644 +--- a/extract.c ++++ b/extract.c +@@ -493,8 +493,11 @@ int extract_or_test_files(__G) /* return PK-type error code */ + } + #endif /* !SFX || SFX_EXDIR */ + +- /* One more: initialize cover structure for bomb detection. Start with a +- span that covers the central directory though the end of the file. */ ++ /* One more: initialize cover structure for bomb detection. Start with ++ spans that cover any extra bytes at the start, the central directory, ++ the end of central directory record (including the Zip64 end of central ++ directory locator, if present), and the Zip64 end of central directory ++ record, if present. */ + if (G.cover == NULL) { + G.cover = malloc(sizeof(cover_t)); + if (G.cover == NULL) { +@@ -506,15 +509,25 @@ int extract_or_test_files(__G) /* return PK-type error code */ + ((cover_t *)G.cover)->max = 0; + } + ((cover_t *)G.cover)->num = 0; +- if ((G.extra_bytes != 0 && +- cover_add((cover_t *)G.cover, 0, G.extra_bytes) != 0) || +- cover_add((cover_t *)G.cover, ++ if (cover_add((cover_t *)G.cover, + G.extra_bytes + G.ecrec.offset_start_central_directory, +- G.ziplen) != 0) { ++ G.extra_bytes + G.ecrec.offset_start_central_directory + ++ G.ecrec.size_central_directory) != 0) { + Info(slide, 0x401, ((char *)slide, + LoadFarString(NotEnoughMemCover))); + return PK_MEM; + } ++ if ((G.extra_bytes != 0 && ++ cover_add((cover_t *)G.cover, 0, G.extra_bytes) != 0) || ++ (G.ecrec.have_ecr64 && ++ cover_add((cover_t *)G.cover, G.ecrec.ec64_start, ++ G.ecrec.ec64_end) != 0) || ++ cover_add((cover_t *)G.cover, G.ecrec.ec_start, ++ G.ecrec.ec_end) != 0) { ++ Info(slide, 0x401, ((char *)slide, ++ LoadFarString(OverlappedComponents))); ++ return PK_BOMB; ++ } + + /*--------------------------------------------------------------------------- + The basic idea of this function is as follows. Since the central di- +diff --git a/process.c b/process.c +index d2e4dc3..d75d405 100644 +--- a/process.c ++++ b/process.c +@@ -1408,6 +1408,10 @@ static int find_ecrec64(__G__ searchlen) /* return PK-class error */ + + /* Now, we are (almost) sure that we have a Zip64 archive. */ + G.ecrec.have_ecr64 = 1; ++ G.ecrec.ec_start -= ECLOC64_SIZE+4; ++ G.ecrec.ec64_start = ecrec64_start_offset; ++ G.ecrec.ec64_end = ecrec64_start_offset + ++ 12 + makeint64(&byterec[ECREC64_LENGTH]); + + /* Update the "end-of-central-dir offset" for later checks. */ + G.real_ecrec_offset = ecrec64_start_offset; +@@ -1542,6 +1546,8 @@ static int find_ecrec(__G__ searchlen) /* return PK-class error */ + makelong(&byterec[OFFSET_START_CENTRAL_DIRECTORY]); + G.ecrec.zipfile_comment_length = + makeword(&byterec[ZIPFILE_COMMENT_LENGTH]); ++ G.ecrec.ec_start = G.real_ecrec_offset; ++ G.ecrec.ec_end = G.ecrec.ec_start + 22 + G.ecrec.zipfile_comment_length; + + /* Now, we have to read the archive comment, BEFORE the file pointer + is moved away backwards to seek for a Zip64 ECLOC64 structure. +diff --git a/unzpriv.h b/unzpriv.h +index dc9eff5..297b3c7 100644 +--- a/unzpriv.h ++++ b/unzpriv.h +@@ -2185,6 +2185,16 @@ typedef struct VMStimbuf { + int have_ecr64; /* valid Zip64 ecdir-record exists */ + int is_zip64_archive; /* Zip64 ecdir-record is mandatory */ + ush zipfile_comment_length; ++ zusz_t ec_start, ec_end; /* offsets of start and end of the ++ end of central directory record, ++ including if present the Zip64 ++ end of central directory locator, ++ which immediately precedes the ++ end of central directory record */ ++ zusz_t ec64_start, ec64_end; /* if have_ecr64 is true, then these ++ are the offsets of the start and ++ end of the Zip64 end of central ++ directory record */ + } ecdir_rec; + +