Attachment 1402 Details for Bug 4871 – Патч для перекодировки с поддержкой UTF-8

[patch] Патч для перекодировки с поддержкой UTF-8

unzip-5.50-alt-iconv-v1.2-utf8.patch (text/plain), 14.87 KB, created by m0sia on 2006-02-23 21:42:36 MSK

(hide)

Description:

Filename:

MIME Type:

Creator: m0sia

Created: 2006-02-23 21:42:36 MSK

Size: 14.87 KB

patch

obsolete

>*** unzip-5.50-real/unix/unix.c	2002-01-22 03:54:42.000000000 +0500
>--- unzip-5.50/unix/unix.c	2006-02-20 14:42:21.607294176 +0500
>***************
>*** 29,34 ****
>--- 29,37 ----
>  #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, ... */
>***************
>*** 1524,1526 ****
>--- 1527,1616 ----
>      }
>  }
>  #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;
>+     dlen = buflen = 2*slen;
>+     d = buf = malloc(buflen + 1);
>+     if(!d)
>+     	goto cleanup;
>+     bzero(buf,buflen);
>+     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);
>+ }
>*** unzip-5.50-real/unix/unxcfg.h	2001-06-04 05:27:14.000000000 +0600
>--- unzip-5.50/unix/unxcfg.h	2006-02-20 14:16:27.194600824 +0500
>***************
>*** 123,126 ****
>--- 123,152 ----
>  /* 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 */
>*** unzip-5.50-real/unzip.c	2002-01-28 00:26:16.000000000 +0500
>--- unzip-5.50/unzip.c	2006-02-20 14:16:27.196600520 +0500
>***************
>*** 304,314 ****
>--- 304,324 ----
>    -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[] =
>***************
>*** 589,594 ****
>--- 599,605 ----
>  #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\
>***************
>*** 596,601 ****
>--- 607,623 ----
>   -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\
>***************
>*** 656,661 ****
>--- 678,687 ----
>  
>      SETLOCALE(LC_CTYPE,"");
>  
>+ #ifdef UNIX
>+     init_conversion_charsets();
>+ #endif
>+ 
>  #if (defined(__IBMC__) && defined(__DEBUG_ALLOC__))
>      extern void DebugMalloc(void);
>  
>***************
>*** 1070,1075 ****
>--- 1096,1106 ----
>      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 */
>***************
>*** 1233,1238 ****
>--- 1264,1298 ----
>                      }
>                      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;
>***************
>*** 1299,1304 ****
>--- 1359,1393 ----
>                      } 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;
>*** unzip-5.50-real/unzpriv.h	2002-02-17 22:01:48.000000000 +0500
>--- unzip-5.50/unzpriv.h	2006-02-20 14:16:27.198600216 +0500
>***************
>*** 1103,1109 ****
>  #  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)
>  #endif /* ?_MBCS */
>  #define INCSTR(ptr) PREINCSTR(ptr)
>  
>--- 1103,1111 ----
>  #  define lastchar(ptr, len) (ptr[(len)-1])
>  #  define MBSCHR(str, c) strchr(str, c)
>  #  define MBSRCHR(str, c) strrchr(str, c)
>! #  ifndef SETLOCALE
>! #    define SETLOCALE(category, locale)
>! #  endif
>  #endif /* ?_MBCS */
>  #define INCSTR(ptr) PREINCSTR(ptr)
>  
>***************
>*** 2424,2430 ****
>           !(((islochdr) || (isuxatt)) && \
>             ((hostver) == 25 || (hostver) == 26 || (hostver) == 40))) || \
>          (hostnum) == FS_HPFS_ || \
>!         ((hostnum) == FS_NTFS_ && (hostver) == 50)) { \
>          _OEM_INTERN((string)); \
>      } else { \
>          _ISO_INTERN((string)); \
>--- 2426,2432 ----
>           !(((islochdr) || (isuxatt)) && \
>             ((hostver) == 25 || (hostver) == 26 || (hostver) == 40))) || \
>          (hostnum) == FS_HPFS_ || \
>!         ((hostnum) == FS_NTFS_/* && (hostver) == 50*/)) { \
>          _OEM_INTERN((string)); \
>      } else { \
>          _ISO_INTERN((string)); \
>*** unzip-5.50-real/zipinfo.c	2001-12-26 03:56:40.000000000 +0500
>--- unzip-5.50/zipinfo.c	2006-02-20 14:16:27.200599912 +0500
>***************
>*** 440,445 ****
>--- 440,449 ----
>      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 */
>***************
>*** 484,489 ****
>--- 488,522 ----
>                              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;
>***************
>*** 504,509 ****
>--- 537,571 ----
>                          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;

Actions: View | Diff

Attachments on bug 4871: 502 | 531 | 532 | 1402