--- parser3/src/main/pa_common.C 2012/03/02 21:57:28 1.264 +++ parser3/src/main/pa_common.C 2013/07/23 12:10:56 1.276 @@ -1,7 +1,7 @@ /** @file Parser: commonly functions. - Copyright(c) 2001-2012 ArtLebedev Group (http://www.artlebedev.com) + Copyright (c) 2000-2012 Art. Lebedev Studio (http://www.artlebedev.com) Author: Alexandr Petrosian (http://paf.design.ru) * BASE64 part @@ -26,8 +26,6 @@ * */ -static const char * const IDENT_COMMON_C="$Date: 2012/03/02 21:57:28 $"; - #include "pa_common.h" #include "pa_exception.h" #include "pa_hash.h" @@ -38,6 +36,13 @@ static const char * const IDENT_COMMON_C #include "pcre.h" #include "pa_request.h" +#ifdef _MSC_VER +#include +#include +#endif + +volatile const char * IDENT_PA_COMMON_C="$Id: pa_common.C,v 1.276 2013/07/23 12:10:56 moko Exp $" IDENT_PA_COMMON_H IDENT_PA_HASH_H IDENT_PA_ARRAY_H IDENT_PA_STACK_H; + // some maybe-undefined constants #ifndef _O_TEXT @@ -118,23 +123,13 @@ void fix_line_breaks(char *str, size_t& str[length]=0; // terminating } -char* file_read_text(Request_charsets& charsets, - const String& file_spec, - bool fail_on_read_problem, - HashStringValue* params, - bool transcode_result) { - File_read_result file= - file_read(charsets, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); +char* file_read_text(Request_charsets& charsets, const String& file_spec, bool fail_on_read_problem, HashStringValue* params, bool transcode_result) { + File_read_result file=file_read(charsets, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); return file.success?file.str:0; } -char* file_load_text(Request& r, - const String& file_spec, - bool fail_on_read_problem, - HashStringValue* params, - bool transcode_result) { - File_read_result file= - file_load(r, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); +char* file_load_text(Request& r, const String& file_spec, bool fail_on_read_problem, HashStringValue* params, bool transcode_result) { + File_read_result file=file_load(r, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); return file.success?file.str:0; } @@ -160,29 +155,21 @@ struct File_read_action_info { char* buf; size_t offset; size_t count; }; #endif -static void file_read_action( - struct stat& finfo, - int f, - const String& file_spec, const char* /*fname*/, bool as_text, - void *context) { + +static void file_read_action(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) to_read_size=(size_t)finfo.st_size; assert( !(info.buf && as_text) ); - if(to_read_size) { + if(to_read_size) { if(info.offset) lseek(f, info.offset, SEEK_SET); - *info.data=info.buf - ? info.buf - : (char *)pa_malloc_atomic(to_read_size+1); - *info.data_size=(size_t)read(f, *info.data, to_read_size); - - if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size) - throw Exception("file.read", - &file_spec, - "read failed: actually read %u bytes count not in [0..%u] valid range", - *info.data_size, to_read_size); + *info.data=info.buf ? info.buf : (char *)pa_malloc_atomic(to_read_size+1); + ssize_t result=read(f, *info.data, to_read_size); + if(result<0) + throw Exception("file.read", &file_spec, "read failed: %s (%d)", strerror(errno), errno); + *info.data_size=result; } else { // empty file // for both, text and binary: for text we need that terminator, for binary we need nonzero pointer to be able to save such files *info.data=(char *)pa_malloc_atomic(1); @@ -414,16 +401,18 @@ struct File_write_action_info { size_t length; }; #endif + static void file_write_action(int f, void *context) { File_write_action_info& info=*static_cast(context); if(info.length) { - int written=write(f, info.str, info.length); + ssize_t written=write(f, info.str, info.length); if(written<0) - throw Exception("file.access", - 0, - "write failed: %s (%d)", strerror(errno), errno); + throw Exception("file.write", 0, "write failed: %s (%d)", strerror(errno), errno); + if((size_t)written!=info.length) + throw Exception("file.write", 0, "write failed: %u of %u bytes written", written, info.length); } } + void file_write( Request_charsets& charsets, const String& file_spec, @@ -493,7 +482,7 @@ static void rmdir(const String& file_spe char* dir_spec=file_spec.taint_cstrm(String::L_FILE_SPEC); size_t length=strlen(dir_spec); while( (length=get_dir(dir_spec, length)) && (length > pos_after) ){ -#ifdef WIN32 +#ifdef _MSC_VER if(!entry_readable(dir_spec, true)) break; DWORD attrs=GetFileAttributes(dir_spec); @@ -509,7 +498,7 @@ static void rmdir(const String& file_spe }; } -bool file_delete(const String& file_spec, bool fail_on_problem) { +bool file_delete(const String& file_spec, bool fail_on_problem, bool keep_empty_dirs) { const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); if(unlink(fname)!=0) if(fail_on_problem) @@ -520,11 +509,13 @@ bool file_delete(const String& file_spec else return false; - rmdir(file_spec, 1); + if(!keep_empty_dirs) + rmdir(file_spec, 1); + return true; } -void file_move(const String& old_spec, const String& new_spec) { +void file_move(const String& old_spec, const String& new_spec, bool keep_empty_dirs) { const char* old_spec_cstr=old_spec.taint_cstr(String::L_FILE_SPEC); const char* new_spec_cstr=new_spec.taint_cstr(String::L_FILE_SPEC); @@ -536,7 +527,8 @@ void file_move(const String& old_spec, c "rename failed: %s (%d), actual filename '%s' to '%s'", strerror(errno), errno, old_spec_cstr, new_spec_cstr); - rmdir(old_spec, 1); + if(!keep_empty_dirs) + rmdir(old_spec, 1); } @@ -563,7 +555,8 @@ bool dir_exists(const String& file_spec) const String* file_exist(const String& path, const String& name) { String& result=*new String(path); - result << "/"; + if(path.last_char() != '/') + result << "/"; result << name; return file_exist(result)?&result:0; } @@ -646,7 +639,7 @@ enum FormatType { FormatUInt, FormatDouble }; -FormatType format_type(char* fmt){ +FormatType format_type(const char* fmt){ enum FormatState { Percent, Flags, @@ -657,7 +650,7 @@ FormatType format_type(char* fmt){ FormatType result=FormatInvalid; - char* pos=fmt; + const char* pos=fmt; while(char c=*(pos++)){ switch(state){ case Percent: @@ -701,7 +694,7 @@ FormatType format_type(char* fmt){ } -const char* format(double value, char* fmt) { +const char* format(double value, const char* fmt) { char local_buf[MAX_NUMBER]; int size=-1; @@ -824,6 +817,26 @@ char* unescape_chars(const char* cp, int return s; } +char *search_stop(char*& current, char cstop_at) { + // sanity check + if(!current) + return 0; + + // skip leading WS + while(*current==' ' || *current=='\t') + current++; + if(!*current) + return current=0; + + char *result=current; + if(char *pstop_at=strchr(current, cstop_at)) { + *pstop_at=0; + current=pstop_at+1; + } else + current=0; + return result; +} + #ifdef WIN32 void back_slashes_to_slashes(char* s) { if(s) @@ -831,14 +844,6 @@ void back_slashes_to_slashes(char* s) { if(*s=='\\') *s='/'; } -/* -void slashes_to_back_slashes(char* s) { - if(s) - for(; *s; s++) - if(*s=='/') - *s='\\'; -} -*/ #endif bool StrStartFromNC(const char* str, const char* substr, bool equal){ @@ -964,7 +969,7 @@ int __vsnprintf(char* b, size_t s, const if((ssize_t)s<0) return 0; -#if _MSC_VER +#ifdef _MSC_VER /* win32: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm @@ -1026,7 +1031,7 @@ int __snprintf(char* b, size_t s, const * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA. * */ -static char *base64_alphabet = +static const char *base64_alphabet = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; /** @@ -1168,9 +1173,9 @@ g_mime_utils_base64_encode_close (const } static unsigned char gmime_base64_rank[256] = { + 255,255,255,255,255,255,255,255,255,254,254,255,255,254,255,255, 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, - 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, - 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63, + 254,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255, 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255, @@ -1193,13 +1198,14 @@ static unsigned char gmime_base64_rank[2 * @out: output stream * @state: holds the number of bits that are stored in @save * @save: leftover bits that have not yet been decoded + * @strict: only base64 and whitespace chars are allowed * * Decodes a chunk of base64 encoded data. * * Returns the number of bytes decoded (which have been dumped in @out). **/ size_t -g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) +g_mime_utils_base64_decode_step(const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save, bool strict=false) { const unsigned char *inptr; unsigned char *outptr; @@ -1217,15 +1223,21 @@ g_mime_utils_base64_decode_step (const u inptr = in; while (inptr < inend) { c = gmime_base64_rank[*inptr++]; - if (c != 0xff) { - saved = (saved << 6) | c; - i++; - if (i == 4) { - *outptr++ = (unsigned char)(saved >> 16); - *outptr++ = (unsigned char)(saved >> 8); - *outptr++ = (unsigned char)(saved); - i = 0; - } + switch(c) { + case 0xff: // non-base64 and non-whitespace chars. not allowed in strict mode + if(strict) + throw Exception(BASE64_FORMAT, 0, "Invalid base64 char on position %d is detected", inptr-in-1); + case 0xfe: // whitespace chars 0x09, 0x0A, 0x0D, 0x20 are allowed in any mode + break; + default: + saved = (saved << 6) | c; + i++; + if (i == 4) { + *outptr++ = (unsigned char)(saved >> 16); + *outptr++ = (unsigned char)(saved >> 8); + *outptr++ = (unsigned char)(saved); + i = 0; + } } } @@ -1237,7 +1249,7 @@ g_mime_utils_base64_decode_step (const u i = 2; while (inptr > in && i) { inptr--; - if (gmime_base64_rank[*inptr] != 0xff) { + if (gmime_base64_rank[*inptr] <= 0xfe) { if (*inptr == '=' && outptr > out) outptr--; i--; @@ -1303,7 +1315,7 @@ static void file_base64_file_action( } } -void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size) { +void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size, bool strict) { // every 4 base64 bytes are converted into 3 normal bytes // not full set (tail) of 4-bytes set is ignored size_t new_size=in_size/4*3; @@ -1312,10 +1324,13 @@ void pa_base64_decode(const char *in, si int state=0; int save=0; result_size= - g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size, - (unsigned char*)result, &state, &save); + g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size, + (unsigned char*)result, &state, &save, strict); assert(result_size <= new_size); result[result_size]=0; // for text files + + if(strict && state!=0) + throw Exception(BASE64_FORMAT, 0, "Unexpected end of chars"); }