--- parser3/src/main/pa_common.C 2016/12/28 22:50:07 1.298 +++ parser3/src/main/pa_common.C 2019/11/12 21:13:31 1.306 @@ -1,30 +1,9 @@ /** @file Parser: commonly functions. - Copyright (c) 2000-2015 Art. Lebedev Studio (http://www.artlebedev.com) + Copyright (c) 2000-2017 Art. Lebedev Studio (http://www.artlebedev.com) Author: Alexandr Petrosian (http://paf.design.ru) - - * BASE64 part - * Authors: Michael Zucchi - * Jeffrey Stedfast - * - * Copyright 2000-2004 Ximian, Inc. (www.ximian.com) - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA. - * - */ +*/ #include "pa_common.h" #include "pa_exception.h" @@ -50,7 +29,7 @@ #define pa_mkdir(path, mode) mkdir(path, mode) #endif -volatile const char * IDENT_PA_COMMON_C="$Id: pa_common.C,v 1.298 2016/12/28 22:50:07 moko Exp $" IDENT_PA_COMMON_H IDENT_PA_HASH_H IDENT_PA_ARRAY_H IDENT_PA_STACK_H; +volatile const char * IDENT_PA_COMMON_C="$Id: pa_common.C,v 1.306 2019/11/12 21:13:31 moko Exp $" IDENT_PA_COMMON_H IDENT_PA_HASH_H IDENT_PA_ARRAY_H IDENT_PA_STACK_H; // some maybe-undefined constants @@ -79,8 +58,39 @@ volatile const char * IDENT_PA_COMMON_C= const String file_status_name(FILE_STATUS_NAME); +String sql_bind_name(SQL_BIND_NAME); +String sql_limit_name(PA_SQL_LIMIT_NAME); +String sql_offset_name(PA_SQL_OFFSET_NAME); +String sql_default_name(SQL_DEFAULT_NAME); +String sql_distinct_name(SQL_DISTINCT_NAME); +String sql_value_type_name(SQL_VALUE_TYPE_NAME); + +// forwards + +const UTF16* pa_utf16_encode(const char* in, Charset& source_charset); + // functions +#ifdef _MSC_VER + +int pa_stat(const char *pathname, struct stat *buffer){ + const UTF16* utf16name=pa_utf16_encode(pathname, pa_thread_request().charsets.source()); + return _wstat64((const wchar_t *)utf16name, buffer); +} + +int pa_open(const char *pathname, int flags, int mode){ + const UTF16* utf16name=pa_utf16_encode(pathname, pa_thread_request().charsets.source()); + return _wopen((const wchar_t *)utf16name, flags, mode); +} + +FILE *pa_fopen(const char *pathname, const char *mode){ + const UTF16* utf16name=pa_utf16_encode(pathname, pa_thread_request().charsets.source()); + const UTF16* utf16mode=pa_utf16_encode(mode, pa_thread_request().charsets.source()); + return _wfopen((const wchar_t *)utf16name, (const wchar_t *)utf16mode); +} + +#endif + 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; @@ -118,7 +128,7 @@ static void file_read_action(struct stat 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; + to_read_size=check_file_size(finfo.st_size, file_spec); if(to_read_size) { if(info.offset) lseek(f, info.offset, SEEK_SET); @@ -234,7 +244,7 @@ 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=pa_open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) { try { if(pa_lock_shared_blocking(f)!=0) throw Exception("file.lock", &file_spec, "shared lock failed: %s (%d), actual filename '%s'", strerror(errno), errno, fname); @@ -287,7 +297,7 @@ bool file_write_action_under_lock( if(access(fname, W_OK)!=0) // no create_dir_for_file(file_spec); - if((f=open(fname, + if((f=pa_open(fname, O_CREAT|O_RDWR |(as_text?_O_TEXT:_O_BINARY) |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) { @@ -503,12 +513,18 @@ bool file_stat(const String& file_spec, return false; } rsize=finfo.st_size; - ratime=finfo.st_atime; - rmtime=finfo.st_mtime; - rctime=finfo.st_ctime; + ratime=(time_t)finfo.st_atime; + rmtime=(time_t)finfo.st_mtime; + rctime=(time_t)finfo.st_ctime; return true; } +size_t check_file_size(uint64_t size, const String& file_spec){ + if(size > pa_file_size_limit) + throw Exception(PARSER_RUNTIME, &file_spec, "content size of %.15g bytes exceeds the limit (%.15g bytes)", (double)size, (double)pa_file_size_limit); + return (size_t)size; +} + /** String related functions */ @@ -713,7 +729,7 @@ const char* format(double value, const c 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); + throw Exception(PARSER_RUNTIME, 0, "Error occurred white executing snprintf with format string '%s'.", fmt); } return pa_strdup(local_buf, (size_t)size); @@ -890,326 +906,6 @@ const char* hex_string(unsigned char* by return bytes_hex; } -/* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */ -/* - * Authors: Michael Zucchi - * Jeffrey Stedfast - * - * Copyright 2000 Helix Code, Inc. (www.helixcode.com) - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA. - * - */ -static const char *base64_alphabet = - "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; - -/** - * g_mime_utils_base64_encode_step: - * @in: input stream - * @inlen: length of the input - * @out: output string - * @state: holds the number of bits that are stored in @save - * @save: leftover bits that have not yet been encoded - * - * Base64 encodes a chunk of data. Performs an 'encode step', only - * encodes blocks of 3 characters to the output at a time, saves - * left-over state in state and save (initialise to 0 on first - * invocation). - * - * Returns the number of bytes encoded. - **/ - -#define BASE64_GROUPS_IN_LINE 19 - -static size_t -g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) -{ - register const unsigned char *inptr; - register unsigned char *outptr; - - if (inlen <= 0) - return 0; - - inptr = in; - outptr = out; - - if (inlen + ((unsigned char *)save)[0] > 2) { - const unsigned char *inend = in + inlen - 2; - register int c1 = 0, c2 = 0, c3 = 0; - register int already; - - already = *state; - - switch (((char *)save)[0]) { - case 1: c1 = ((unsigned char *)save)[1]; goto skip1; - case 2: c1 = ((unsigned char *)save)[1]; - c2 = ((unsigned char *)save)[2]; goto skip2; - } - - /* yes, we jump into the loop, no i'm not going to change it, its beautiful! */ - while (inptr < inend) { - c1 = *inptr++; - skip1: - c2 = *inptr++; - skip2: - c3 = *inptr++; - *outptr++ = base64_alphabet [c1 >> 2]; - *outptr++ = base64_alphabet [(c2 >> 4) | ((c1 & 0x3) << 4)]; - *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)]; - *outptr++ = base64_alphabet [c3 & 0x3f]; - /* this is a bit ugly ... */ - if ((++already) >= BASE64_GROUPS_IN_LINE) { - *outptr++ = '\n'; - already = 0; - } - } - - ((unsigned char *)save)[0] = 0; - inlen = 2 - (inptr - inend); - *state = already; - } - - //d(printf ("state = %d, inlen = %d\n", (int)((char *)save)[0], inlen)); - - if (inlen > 0) { - register char *saveout; - - /* points to the slot for the next char to save */ - saveout = & (((char *)save)[1]) + ((char *)save)[0]; - - /* inlen can only be 0 1 or 2 */ - switch (inlen) { - case 2: *saveout++ = *inptr++; - case 1: *saveout++ = *inptr++; - } - *(char *)save = *(char *)save+(char)inlen; - } - - /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n", - (int)((char *)save)[0], - (int)((char *)save)[1], - (int)((char *)save)[2]));*/ - - return (outptr - out); -} - -/** - * g_mime_utils_base64_encode_close: - * @in: input stream - * @inlen: length of the input - * @out: output string - * @state: holds the number of bits that are stored in @save - * @save: leftover bits that have not yet been encoded - * - * Base64 encodes the input stream to the output stream. Call this - * when finished encoding data with g_mime_utils_base64_encode_step to - * flush off the last little bit. - * - * Returns the number of bytes encoded. - **/ -static size_t -g_mime_utils_base64_encode_close (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) -{ - unsigned char *outptr = out; - int c1, c2; - - if (inlen > 0) - outptr += g_mime_utils_base64_encode_step (in, inlen, outptr, state, save); - - c1 = ((unsigned char *)save)[1]; - c2 = ((unsigned char *)save)[2]; - - switch (((unsigned char *)save)[0]) { - case 2: - outptr[2] = base64_alphabet [(c2 & 0x0f) << 2]; - goto skip; - case 1: - outptr[2] = '='; - skip: - outptr[0] = base64_alphabet [c1 >> 2]; - outptr[1] = base64_alphabet [c2 >> 4 | ((c1 & 0x3) << 4)]; - outptr[3] = '='; - outptr += 4; - break; - } - - *outptr++ = 0; - - *save = 0; - *state = 0; - - return (outptr - out); -} - -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, - 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, - 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, - 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255, - 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, - 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, - 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, - 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, - 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, -}; - -/** - * g_mime_utils_base64_decode_step: - * @in: input stream - * @inlen: max length of data to decode - * @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, bool strict=false) -{ - const unsigned char *inptr; - unsigned char *outptr; - const unsigned char *inend; - int saved; - unsigned char c; - int i; - - inend = in + inlen; - outptr = out; - - /* convert 4 base64 bytes to 3 normal bytes */ - saved = *save; - i = *state; - inptr = in; - while (inptr < inend) { - c = gmime_base64_rank[*inptr++]; - 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; - } - } - } - - *save = saved; - *state = i; - - /* quick scan back for '=' on the end somewhere */ - /* fortunately we can drop 1 output char for each trailing = (upto 2) */ - i = 2; - while (inptr > in && i) { - inptr--; - if (gmime_base64_rank[*inptr] <= 0xfe) { - if (*inptr == '=' && outptr > out) - outptr--; - i--; - } - } - - /* if i != 0 then there is a truncation error! */ - return (outptr - out); -} - - -char* pa_base64_encode(const char *in, size_t in_size){ - size_t new_size = ((in_size / 3 + 1) * 4); - new_size += new_size / (BASE64_GROUPS_IN_LINE * 4)/*new lines*/ + 1/*zero terminator*/; - char* result = new(PointerFreeGC) char[new_size]; - int state=0; - int save=0; -#ifndef NDEBUG - size_t filled= -#endif - g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size, (unsigned char*)result, &state, &save); - - //throw Exception(PARSER_RUNTIME, 0, "%d %d %d", in_size, new_size, filled); - assert(filled <= new_size); - - return result; -} - -struct File_base64_action_info { - unsigned char** base64; -}; - -static void file_base64_file_action(struct stat& finfo, int f, const String&, void *context) { - - if(finfo.st_size) { - File_base64_action_info& info=*static_cast(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); - } -} - -char* pa_base64_encode(const String& file_spec){ - unsigned char* base64=0; - File_base64_action_info info={&base64}; - - file_read_action_under_lock(file_spec, "pa_base64_encode", file_base64_file_action, &info); - - return (char*)base64; -} - -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; - result=new(PointerFreeGC) char[new_size+1/*terminator*/]; - - 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, 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"); -} - - int file_block_read(const int f, unsigned char* buffer, const size_t size){ int nCount = read(f, buffer, size); if (nCount < 0) @@ -1302,6 +998,51 @@ Charset* detect_charset(const char* cont return 0; } +const UTF16* pa_utf16_encode(const char* in, Charset& source_charset){ + if(!in) + return 0; + + String::C sIn(in,strlen(in)); + + UTF16* utf16=(UTF16*)pa_malloc_atomic(sIn.length*2+2); + UTF16* utf16_end=utf16; + + if(!source_charset.isUTF8()) + sIn=Charset::transcode(sIn, source_charset, pa_UTF8_charset); + + int status=pa_convertUTF8toUTF16((const UTF8**)&sIn.str, (const UTF8*)(sIn.str+sIn.length), &utf16_end, utf16+sIn.length, strictConversion); + if(status != conversionOK) + throw Exception("utf-16 encode", new String(in), "utf-16 conversion failed (%d)", status); + + *utf16_end=0; + + return utf16; +} + +const char* pa_utf16_decode(const UTF16* in, Charset& asked_charset){ + if(!in) + return 0; + + const UTF16* utf16_start=in; + const UTF16* utf16_end; + + for(utf16_end=in; *utf16_end; utf16_end++); + + char *result = (char *)pa_malloc_atomic((utf16_end-in)*6+1); + char *result_end = result; + + int status=pa_convertUTF16toUTF8(&utf16_start, utf16_end, (UTF8**)&result_end, (UTF8*)(result+(utf16_end-in)*6), strictConversion); + + if(status != conversionOK) + throw Exception("utf-16 decode", 0, "utf conversion failed (%d)", status); + + *result_end='\0'; + + if(asked_charset.isUTF8()) + return result; + + return Charset::transcode(result, pa_UTF8_charset, asked_charset).cstr(); +} static bool is_latin(const char *in){ for(; *in; in++){ @@ -1313,7 +1054,7 @@ static bool is_latin(const char *in){ #define MAX_IDNA_LENGTH 256 -const char *pa_idna_encode(const char *in, Charset &source_charset){ +const char *pa_idna_encode(const char *in, Charset& source_charset){ if(!in || is_latin(in)) return in;