--- parser3/src/main/pa_common.C 2006/12/02 12:35:50 1.221 +++ parser3/src/main/pa_common.C 2008/06/07 11:01:09 1.233 @@ -26,7 +26,7 @@ * */ -static const char * const IDENT_COMMON_C="$Date: 2006/12/02 12:35:50 $"; +static const char * const IDENT_COMMON_C="$Date: 2008/06/07 11:01:09 $"; #include "pa_common.h" #include "pa_exception.h" @@ -34,6 +34,7 @@ static const char * const IDENT_COMMON_C #include "pa_globals.h" #include "pa_charsets.h" #include "pa_http.h" +#include "pa_request_charsets.h" // some maybe-undefined constants @@ -91,9 +92,9 @@ void fix_line_breaks(char *str, size_t& } char* file_read_text(Request_charsets& charsets, - const String& file_spec, - bool fail_on_read_problem, - HashStringValue* params/*, HashStringValue* * out_fields*/) { + const String& file_spec, + bool fail_on_read_problem, + HashStringValue* params/*, HashStringValue* * out_fields*/) { File_read_result file= file_read(charsets, file_spec, true, params, fail_on_read_problem); return file.success?file.str:0; @@ -111,6 +112,8 @@ int pa_get_valid_file_options_count(Hash result++; if(options.get(PA_COLUMN_ENCLOSER_NAME)) result++; + if(options.get(PA_CHARSET_NAME)) + result++; return result; } @@ -121,10 +124,10 @@ struct File_read_action_info { }; #endif static void file_read_action( - struct stat& finfo, - int f, - const String& file_spec, const char* /*fname*/, bool as_text, - void *context) { + struct stat& finfo, + int f, + const String& file_spec, const char* /*fname*/, bool as_text, + void *context) { File_read_action_info& info=*static_cast(context); size_t to_read_size=info.count; if(!to_read_size) @@ -152,13 +155,13 @@ static void file_read_action( } } File_read_result file_read(Request_charsets& charsets, const String& file_spec, - bool as_text, HashStringValue *params, - bool fail_on_read_problem, - char* buf, size_t offset, size_t count) { + bool as_text, HashStringValue *params, + bool fail_on_read_problem, + char* buf, size_t offset, size_t count) { File_read_result result={false, 0, 0, 0}; if(file_spec.starts_with("http://")) { if(offset || count) - throw Exception("parser.runtime", + throw Exception(PARSER_RUNTIME, 0, "offset and load options are not supported for HTTP:// file load"); @@ -172,7 +175,7 @@ File_read_result file_read(Request_chars if(params) { int valid_options=pa_get_valid_file_options_count(*params); if(valid_options!=params->count()) - throw Exception("parser.runtime", + throw Exception(PARSER_RUNTIME, 0, "invalid option passed"); } @@ -182,6 +185,19 @@ File_read_result file_read(Request_chars result.success=file_read_action_under_lock(file_spec, "read", file_read_action, &info, as_text, fail_on_read_problem); + + if(result.length && as_text && params) { + if( Value* vcharset_name=params->get(PA_CHARSET_NAME) ) { + Charset asked_charset=::charsets.get(vcharset_name->as_string(). + change_case(charsets.source(), String::CC_UPPER)); + + String::C body=String::C(result.str, result.length); + body=Charset::transcode(body, asked_charset, charsets.source()); + + result.str=const_cast(body.str); // hacking a little + result.length=body.length; + } + } } if(result.success && as_text) { @@ -203,7 +219,7 @@ File_read_result file_read(Request_chars void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) { if(finfo.st_uid/*foreign?*/!=geteuid() && finfo.st_gid/*foreign?*/!=getegid()) - throw Exception("parser.runtime", + throw Exception(PARSER_RUNTIME, &file_spec, "parser is in safe mode: " "reading files of foreign group and user disabled " @@ -232,7 +248,7 @@ bool file_read_action_under_lock(const S // they delay update till open, so we would receive "!^test[" string // if would do stat, next open. // later: it seems, even this does not help sometimes - if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) { + if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) { try { if(pa_lock_shared_blocking(f)!=0) throw Exception("file.lock", @@ -261,7 +277,7 @@ bool file_read_action_under_lock(const S pa_unlock(f);close(f); return true; - } else { + } else { if(fail_on_read_problem) throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, &file_spec, @@ -282,7 +298,9 @@ void create_dir_for_file(const String& f bool file_write_action_under_lock( const String& file_spec, - const char* action_name, File_write_action action, void *context, + const char* action_name, + File_write_action action, + void *context, bool as_text, bool do_append, bool do_block, @@ -353,9 +371,12 @@ void file_write( bool as_text, bool do_append) { File_write_action_info info={data, size}; + file_write_action_under_lock( file_spec, - "write", file_write_action, &info, + "write", + file_write_action, + &info, as_text, do_append); } @@ -447,11 +468,11 @@ bool file_executable(const String& file_ } bool file_stat(const String& file_spec, - size_t& rsize, - time_t& ratime, - time_t& rmtime, - time_t& rctime, - bool fail_on_read_problem) { + size_t& rsize, + time_t& ratime, + time_t& rmtime, + time_t& rctime, + bool fail_on_read_problem) { const char* fname=file_spec.cstr(String::L_FILE_SPEC); struct stat finfo; if(stat(fname, &finfo)!=0) @@ -470,62 +491,140 @@ bool file_stat(const String& file_spec, } char* getrow(char* *row_ref, char delim) { - char* result=*row_ref; - if(result) { + char* result=*row_ref; + if(result) { *row_ref=strchr(result, delim); if(*row_ref) *((*row_ref)++)=0; else if(!*result) return 0; - } - return result; + } + return result; } char* lsplit(char* string, char delim) { - if(string) { + if(string) { char* v=strchr(string, delim); if(v) { *v=0; return v+1; } - } - return 0; + } + return 0; } char* lsplit(char* *string_ref, char delim) { - char* result=*string_ref; + char* result=*string_ref; char* next=lsplit(*string_ref, delim); - *string_ref=next; - return result; + *string_ref=next; + return result; } char* rsplit(char* string, char delim) { - if(string) { + if(string) { char* v=strrchr(string, delim); if(v) { *v=0; return v+1; } - } - return NULL; + } + return NULL; } -/// @todo less stupid type detection + +// format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx +// flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx +// width, precision: non negative decimal number +enum FormatType { + FormatInvalid, + FormatInt, + FormatUInt, + FormatDouble +}; +FormatType format_type(char* fmt){ + enum FormatState { + Percent, + Flags, + Width, + Precision, + Done + } state=Percent; + + FormatType result=FormatInvalid; + + char* pos=fmt; + while(char c=*(pos++)){ + switch(state){ + case Percent: + if(c=='%'){ + state=Flags; + } else { + return FormatInvalid; // 1st char must be '%' only + } + break; + case Flags: + if(strchr("-+ #0", c)!=0){ + break; + } + // go to the next step + case Width: + if(c=='.'){ + state=Precision; + break; + } + // go to the next step + case Precision: + if(c>='0' && c<='9'){ + if(state == Flags) state=Width; // no more flags + break; + } else if(c=='d' || c=='i'){ + result=FormatInt; + } else if(strchr("feEgG", c)!=0){ + result=FormatDouble; + } else if(strchr("uoxX", c)!=0){ + result=FormatUInt; + } else { + return FormatInvalid; // invalid char + } + state=Done; + break; + case Done: + return FormatInvalid; // no chars allowed after 'type' + } + } + return result; +} + + const char* format(double value, char* fmt) { - char local_buf[MAX_NUMBER]; + char local_buf[MAX_NUMBER]; size_t size; - - if(fmt) - if(strpbrk(fmt, "diouxX")) - if(strpbrk(fmt, "ouxX")) - size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value); - else + + if(fmt && strlen(fmt)){ + switch(format_type(fmt)){ + case FormatDouble: + size=snprintf(local_buf, sizeof(local_buf), fmt, value); + break; + case FormatInt: size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value); - else - size=snprintf(local_buf, sizeof(local_buf), fmt, value); - else - size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value); - + break; + case FormatUInt: + size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value); + break; + case FormatInvalid: + throw Exception(PARSER_RUNTIME, + 0, + "Incorrect format string '%s' was specified.", fmt); + } + } else + size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value); + + if(size < 0 || size >= MAX_NUMBER-1){ // on win32 we manually reduce max size while printing + throw Exception(PARSER_RUNTIME, + 0, + "Error occure white executing snprintf with format string '%s'.", fmt); + } + return pa_strdup(local_buf, size); } @@ -546,13 +645,16 @@ size_t stdout_write(const void *buf, siz #endif } +enum EscapeState { + EscapeRest, + EscapeFirst, + EscapeSecond, + EscapeUnicode +}; +/* char* unescape_chars(const char* cp, int len) { char* s=new(PointerFreeGC) char[len + 1]; - enum EscapeState { - EscapeRest, - EscapeFirst, - EscapeSecond - } escapeState=EscapeRest; + EscapeState escapeState=EscapeRest; uint escapedValue=0; int srcPos=0; int dstPos=0; @@ -560,29 +662,96 @@ char* unescape_chars(const char* cp, int uchar ch=(uchar)cp[srcPos]; switch(escapeState) { case EscapeRest: - if(ch=='%') { - escapeState=EscapeFirst; - } else if(ch=='+') { - s[dstPos++]=' '; - } else { - s[dstPos++]=ch; - } - break; + if(ch=='%') { + escapeState=EscapeFirst; + } else if(ch=='+'){ + s[dstPos++]=' '; + } else { + s[dstPos++]=ch; + } + break; case EscapeFirst: - escapedValue=hex_value[ch] << 4; - escapeState=EscapeSecond; - break; + escapedValue=hex_value[ch] << 4; + escapeState=EscapeSecond; + break; case EscapeSecond: - escapedValue +=hex_value[ch]; - s[dstPos++]=(char)escapedValue; - escapeState=EscapeRest; - break; + escapedValue +=hex_value[ch]; + s[dstPos++]=(char)escapedValue; + escapeState=EscapeRest; + break; } srcPos++; } s[dstPos]=0; return s; } +*/ + +char* unescape_chars(const char* cp, int len, Charset* charset){ + char* s=new(PointerFreeGC) char[len + 1]; // enough (%uXXXX==6 bytes, max utf-8 char length==6 bytes) + char* dst=s; + EscapeState escapeState=EscapeRest; + uint escapedValue=0; + int srcPos=0; + short int jsCnt=0; + while(srcPos < len) { + uchar c=(uchar)cp[srcPos]; + if(c=='%'){ + escapeState=EscapeFirst; + } else { + switch(escapeState) { + case EscapeRest: + if(c=='+'){ + *dst++=' '; + } else { + *dst++=c; + } + break; + case EscapeFirst: + if(charset && c=='u'){ + // escaped unicode value: %u0430 + jsCnt=0; + escapedValue=0; + escapeState=EscapeUnicode; + } else { + if(isxdigit(c)){ + escapedValue=hex_value[c] << 4; + escapeState=EscapeSecond; + } else { + *dst++=c; + escapeState=EscapeRest; + } + } + break; + case EscapeSecond: + if(isxdigit(c)){ + escapedValue+=hex_value[c]; + *dst++=(char)escapedValue; + } + escapeState=EscapeRest; + break; + case EscapeUnicode: + if(isxdigit(c)){ + escapedValue=(escapedValue << 4) + hex_value[c]; + if(++jsCnt==4){ + // transcode utf8 char to client charset (we can lost some chars here) + charset->store_Char((XMLByte*&)dst, (XMLCh)escapedValue, '?'); + escapeState=EscapeRest; + } + } else { + // not full unicode value + escapeState=EscapeRest; + } + break; + } + } + + srcPos++; + } + + *dst=0; // zero-termination + return s; +} #ifdef WIN32 void back_slashes_to_slashes(char* s) { @@ -601,55 +770,145 @@ void slashes_to_back_slashes(char* s) { */ #endif -bool StrEqNc(const char* s1, const char* s2, bool strict) { +bool StrStartFromNC(const char* str, const char* substr, bool equal){ while(true) { - if(!(*s1)) { - if(!(*s2)) + if(!(*substr)){ + if(!(*str)) return true; else - return !strict; - } else if(!(*s2)) - return !strict; - if(isalpha((unsigned char)*s1)) { - if(tolower((unsigned char)*s1) !=tolower((unsigned char)*s2)) + return !equal; + } + if(!(*str)) + return false; + if(isalpha((unsigned char)*str)) { + if(tolower((unsigned char)*str)!=tolower((unsigned char)*substr)) return false; - } else if((*s1) !=(*s2)) + } else if((*str) != (*substr)) return false; - s1++; - s2++; + str++; + substr++; } } +size_t strpos(const char *str, const char *substr) { + const char *p = strstr(str, substr); + return (p==0)?STRING_NOT_FOUND:p-str; +} + +// content-type: xxx; charset=WE-NEED-THIS +// content-type: xxx; charset="WE-NEED-THIS" +// content-type: xxx; charset="WE-NEED-THIS"; +Charset* detect_charset(const char* content_type){ + if(content_type){ + size_t len=strlen(content_type); + char* CONTENT_TYPE=new(PointerFreeGC) char[len+1]; + memcpy(CONTENT_TYPE, content_type, len); + for(char *p=CONTENT_TYPE; *p; p++) + *p=(char)toupper((unsigned char)*p); + + if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){ + begin+=8; // skip "CHARSET=" + char* end=0; + if(*begin && (*begin=='"' || *begin =='\'')){ + char quote=*begin; + begin++; + end=(char*)strchr(begin, quote); + } + if(!end) + end=(char*)strchr(begin, ';'); + + size_t len; + if(end){ + *end=0; // terminator + len=end-begin; + } else { + len=strlen(begin); + } + + String::Body NAME=String::Body(begin, len); + return &charsets.get(NAME); + } + } + return 0; +} + +Charset* detect_charset(Charset& source_charset, const String& content_type){ + const String& CONTENT_TYPE=content_type.change_case(source_charset, String::CC_UPPER); + size_t begin=CONTENT_TYPE.pos("CHARSET="); + if(begin!=STRING_NOT_FOUND) { + begin+=8; // skip "CHARSET=" + size_t end=STRING_NOT_FOUND; + + if(CONTENT_TYPE.pos('"', begin)==begin){ + begin++; + end=CONTENT_TYPE.pos('"', begin); + } else if(CONTENT_TYPE.pos('\'', begin)==begin){ + begin++; + end=CONTENT_TYPE.pos('\'', begin); + } + + if(end==STRING_NOT_FOUND) + end=CONTENT_TYPE.pos(';', begin); + + if(end==STRING_NOT_FOUND) + end=CONTENT_TYPE.length(); + + const String::Body NAME=CONTENT_TYPE.mid(begin, end); + return &charsets.get(NAME); + } + return 0; +} + + static bool isLeap(int year) { - return !( - (year % 4) || ((year % 400) && !(year % 100)) - ); + return !( + (year % 4) || ((year % 400) && !(year % 100)) + ); } int getMonthDays(int year, int month) { static int monthDays[]={ - 31, - 28, - 31, - 30, - 31, - 30, - 31, - 31, - 30, - 31, - 30, - 31 - }; - return (month == 2 && isLeap(year)) ? 29 : monthDays[month]; -} - -void remove_crlf(char* start, char* end) { - for(char* p=start; p(context); + *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6]; + unsigned char* base64 = *info.base64; + int state=0; + int save=0; + int nCount; + do { + unsigned char buffer[FILE_BUFFER_SIZE]; + nCount = file_block_read(f, buffer, sizeof(buffer)); + if( nCount ){ + size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save); + base64+=filled; + } + } while(nCount > 0); + g_mime_utils_base64_encode_close (0, 0, base64, &state, &save); + } +} + void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size) { /* wont go to more than had (overly conservative) */ @@ -1007,10 +1304,10 @@ const unsigned long pa_crc32(const Strin } static void file_crc32_file_action( - struct stat& finfo, - int f, - const String&, const char* /*fname*/, bool, - void *context) + struct stat& finfo, + int f, + const String&, const char* /*fname*/, bool, + void *context) { unsigned long& crc32=*static_cast(context); if(finfo.st_size) {