|
|
| version 1.78, 2001/10/29 13:04:46 | version 1.284, 2015/08/20 22:06:49 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: commonly functions. | Parser: commonly functions. |
| Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com) | Copyright (c) 2000-2012 Art. Lebedev Studio (http://www.artlebedev.com) |
| Author: Alexander Petrosyan <paf@design.ru>(http://design.ru/paf) | Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) |
| $Id$ | * BASE64 part |
| */ | * Authors: Michael Zucchi <notzed@ximian.com> |
| * Jeffrey Stedfast <fejj@ximian.com> | |
| * | |
| * 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_common.h" |
| #include "pa_types.h" | |
| #include "pa_exception.h" | #include "pa_exception.h" |
| #include "pa_pool.h" | |
| #include "pa_globals.h" | |
| #include "pa_value.h" | |
| #include "pa_hash.h" | #include "pa_hash.h" |
| #include "pa_string.h" | #include "pa_globals.h" |
| #include "pa_charsets.h" | |
| #include "pa_http.h" | |
| #include "pa_request_charsets.h" | |
| #include "pcre.h" | |
| #include "pa_request.h" | |
| #include "pa_idna.h" | |
| #include "pa_convert_utf.h" | |
| #ifdef _MSC_VER | |
| #include <windows.h> | |
| #include <direct.h> | |
| #endif | |
| #ifndef WIN32 | #ifdef _MSC_VER |
| # ifndef _O_TEXT | #define pa_mkdir(path, mode) _mkdir(path) |
| # define _O_TEXT 0 | #else |
| # endif | #define pa_mkdir(path, mode) mkdir(path, mode) |
| # ifndef _O_BINARY | #endif |
| # define _O_BINARY 0 | |
| volatile const char * IDENT_PA_COMMON_C="$Id$" IDENT_PA_COMMON_H IDENT_PA_HASH_H IDENT_PA_ARRAY_H IDENT_PA_STACK_H; | |
| // some maybe-undefined constants | |
| #ifndef _O_TEXT | |
| # define _O_TEXT 0 | |
| #endif | |
| #ifndef _O_BINARY | |
| # define _O_BINARY 0 | |
| #endif | |
| #ifdef HAVE_FTRUNCATE | |
| # define PA_O_TRUNC 0 | |
| #else | |
| # ifdef _O_TRUNC | |
| # define PA_O_TRUNC _O_TRUNC | |
| # else | |
| # error you must have either ftruncate function or _O_TRUNC bit declared | |
| # endif | # endif |
| #endif | #endif |
| #if _MSC_VER | // defines for globals |
| int __vsnprintf(char *b, size_t s, const char *f, va_list l) { | #define FILE_STATUS_NAME "status" |
| int r=_vsnprintf(b, --s, f, l); | |
| b[s]=0; | // globals |
| return r; | |
| const String file_status_name(FILE_STATUS_NAME); | |
| // functions | |
| 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; | |
| } | } |
| int __snprintf(char *b, size_t s, const char *f, ...) { | |
| va_list l; | char* file_load_text(Request& r, const String& file_spec, bool fail_on_read_problem, HashStringValue* params, bool transcode_result) { |
| va_start(l, f); | File_read_result file=file_load(r, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); |
| int r=__vsnprintf(b, s, f, l); | return file.success?file.str:0; |
| va_end(l); | } |
| return r; | |
| /// these options were handled but not checked elsewhere, now check them | |
| int pa_get_valid_file_options_count(HashStringValue& options) { | |
| int result=0; | |
| if(options.get(PA_SQL_LIMIT_NAME)) | |
| result++; | |
| if(options.get(PA_SQL_OFFSET_NAME)) | |
| result++; | |
| if(options.get(PA_COLUMN_SEPARATOR_NAME)) | |
| result++; | |
| if(options.get(PA_COLUMN_ENCLOSER_NAME)) | |
| result++; | |
| if(options.get(PA_CHARSET_NAME)) | |
| result++; | |
| return result; | |
| } | } |
| #ifndef DOXYGEN | |
| struct File_read_action_info { | |
| char **data; size_t *data_size; | |
| char* buf; size_t offset; size_t count; | |
| }; | |
| #endif | #endif |
| void fix_line_breaks(char *src, size_t& size) { | static void file_read_action(struct stat& finfo, int f, const String& file_spec, const char* /*fname*/, bool as_text, void *context) { |
| char *dest=src; | File_read_action_info& info=*static_cast<File_read_action_info *>(context); |
| // fix DOS: \r\n -> \n | size_t to_read_size=info.count; |
| // fix Macintosh: \r -> \n | if(!to_read_size) |
| char *bol=src; | to_read_size=(size_t)finfo.st_size; |
| while(char *eol=strchr(bol, '\r')) { | assert( !(info.buf && as_text) ); |
| size_t len=eol-bol; | if(to_read_size) { |
| if(dest!=bol) | if(info.offset) |
| memcpy(dest, bol, len); | lseek(f, info.offset, SEEK_SET); |
| dest+=len; | *info.data=info.buf ? info.buf : (char *)pa_malloc_atomic(to_read_size+1); |
| *dest++='\n'; | 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); | |
| *(char*)(*info.data)=0; | |
| *info.data_size=0; | |
| return; | |
| } | |
| } | |
| if(eol[1]=='\n') { // \r,\n = DOS | File_read_result file_read(Request_charsets& charsets, const String& file_spec, |
| bol=eol+2; | bool as_text, HashStringValue *params, |
| size--; | bool fail_on_read_problem, |
| } else // \r,not \n = Macintosh | char* buf, size_t offset, size_t count, bool transcode_text_result) { |
| bol=eol+1; | File_read_result result={false, 0, 0, 0}; |
| if(params){ | |
| int valid_options=pa_get_valid_file_options_count(*params); | |
| if(valid_options!=params->count()) | |
| throw Exception(PARSER_RUNTIME, 0, CALLED_WITH_INVALID_OPTION); | |
| } | } |
| // last piece without \r, including terminating 0 | |
| if(dest!=bol) | File_read_action_info info={&result.str, &result.length, buf, offset, count}; |
| strcpy(dest, bol); | |
| result.success=file_read_action_under_lock(file_spec, | |
| "read", file_read_action, &info, | |
| as_text, fail_on_read_problem); | |
| if(as_text){ | |
| if(result.success){ | |
| Charset* asked_charset=0; | |
| if(result.length>=3 && strncmp(result.str, "\xEF\xBB\xBF", 3)==0){ | |
| // skip UTF-8 signature (BOM code) | |
| result.str+=3; | |
| result.length-=3; | |
| asked_charset=&UTF8_charset; | |
| } | |
| if(params) | |
| if(Value* vcharset_name=params->get(PA_CHARSET_NAME)) | |
| asked_charset=&::charsets.get(vcharset_name->as_string().change_case(charsets.source(), String::CC_UPPER)); | |
| if(result.length && transcode_text_result && asked_charset){ // length must be checked because transcode returns CONST string in case length==0, which contradicts hacking few lines below | |
| String::C body=String::C(result.str, result.length); | |
| body=Charset::transcode(body, *asked_charset, charsets.source()); | |
| result.str=const_cast<char*>(body.str); // hacking a little | |
| result.length=body.length; | |
| } | |
| } | |
| if(result.length) | |
| fix_line_breaks(result.str, result.length); | |
| } | |
| return result; | |
| } | |
| File_read_result file_load(Request& r, const String& file_spec, | |
| bool as_text, HashStringValue *params, | |
| bool fail_on_read_problem, | |
| char* buf, size_t offset, size_t count, bool transcode_text_result) { | |
| File_read_result result={false, 0, 0, 0}; | |
| if(file_spec.starts_with("http://")) { | |
| if(offset || count) | |
| throw Exception(PARSER_RUNTIME, | |
| 0, | |
| "offset and load options are not supported for HTTP:// file load"); | |
| // fail on read problem | |
| File_read_http_result http=pa_internal_file_read_http(r, file_spec, as_text, params, transcode_text_result); | |
| result.success=true; | |
| result.str=http.str; | |
| result.length=http.length; | |
| result.headers=http.headers; | |
| } else | |
| result= | |
| file_read(r.charsets, file_spec, as_text, params, fail_on_read_problem, buf, offset, count, transcode_text_result); | |
| return result; | |
| } | |
| #ifdef PA_SAFE_MODE | |
| 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, | |
| &file_spec, | |
| "parser is in safe mode: " | |
| "reading files of foreign group and user disabled " | |
| "[recompile parser with --disable-safe-mode configure option], " | |
| "actual filename '%s', " | |
| "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)", | |
| fname, | |
| finfo.st_uid, geteuid(), | |
| finfo.st_gid, getegid()); | |
| } | |
| #else | |
| void check_safe_mode(struct stat, const String&, const char*) { | |
| } | } |
| #endif | |
| char *file_read_text(Pool& pool, const String& file_spec, bool fail_on_read_problem) { | |
| void *result; size_t size; | bool file_read_action_under_lock(const String& file_spec, |
| return file_read(pool, file_spec, result, size, true, fail_on_read_problem)?(char *)result:0; | const char* action_name, File_read_action action, void *context, |
| } | bool as_text, |
| bool file_read(Pool& pool, const String& file_spec, | bool fail_on_read_problem) { |
| void*& data, size_t& read_size, bool as_text, | const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); |
| bool fail_on_read_problem, | |
| size_t offset, size_t limit) { | |
| const char *fname=file_spec.cstr(String::UL_FILE_SPEC); | |
| //printf("file_read(%s)\n", fname); | |
| int f; | int f; |
| struct stat finfo; | |
| // first open, next stat: | // first open, next stat: |
| // directory update of NTFS hard links performed on open. | // directory update of NTFS hard links performed on open. |
| // ex: | // ex: |
| // a.html:^test[] and b.html hardlink to a.html | // a.html:^test[] and b.html hardlink to a.html |
| // user inserts ! before ^test in a.html | // user inserts ! before ^test in a.html |
| // directory entry of b.html in NTFS not updated at once, | // directory entry of b.html in NTFS not updated at once, |
| // they delay update till open, so we would receive "!^test[" string | // they delay update till open, so we would receive "!^test[" string |
| // if would do stat, next open. | // if would do stat, next open. |
| if( | // later: it seems, even this does not help sometimes |
| (f=open(fname, O_RDONLY|(as_text?_O_BINARY/*_O_TEXT*/:_O_BINARY)))>=0 && | if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) { |
| stat(fname, &finfo)==0) { | try { |
| /*if(exclusive) | if(pa_lock_shared_blocking(f)!=0) |
| flock(f, LOCK_EX);*/ | throw Exception("file.lock", |
| size_t max_size=limit?min(offset+limit, finfo.st_size)-offset:finfo.st_size; | &file_spec, |
| if(!max_size) { // eof | "shared lock failed: %s (%d), actual filename '%s'", |
| data=0; | strerror(errno), errno, fname); |
| read_size=0; | |
| } else { | struct stat finfo; |
| data=pool.malloc(max_size+(as_text?1:0), 3); | if(fstat(f, &finfo)!=0) |
| if(offset) | throw Exception("file.missing", // hardly possible: we just opened it OK |
| lseek(f, offset, SEEK_SET); | &file_spec, |
| read_size=read(f, data, max_size); | "stat failed: %s (%d), actual filename '%s'", |
| } | strerror(errno), errno, fname); |
| /*if(exclusive) | |
| flock(f, LOCK_UN);*/ | check_safe_mode(finfo, file_spec, fname); |
| close(f); | |
| if(!max_size) // eof | action(finfo, f, file_spec, fname, as_text, context); |
| return true; | } catch(...) { |
| pa_unlock(f);close(f); | |
| if(read_size>=0 && read_size<=max_size) { | if(fail_on_read_problem) |
| if(as_text) | rethrow; |
| ((char*&)data)[read_size]=0; | return false; |
| } else | } |
| throw Exception(0, 0, | |
| &file_spec, | pa_unlock(f);close(f); |
| "read failed: actually read %d bytes count not in [0..%lu] valid range", | |
| read_size, (unsigned long)max_size); //never | |
| if(as_text) | |
| fix_line_breaks((char *)data, read_size); | |
| return true; | return true; |
| } | } else { |
| if(fail_on_read_problem) | if(fail_on_read_problem) |
| throw Exception(0, 0, | throw Exception(errno==EACCES?"file.access" |
| &file_spec, | :(errno==ENOENT || errno==ENOTDIR || errno==ENODEV)?"file.missing":0, |
| "read failed: %s (%d), actual filename '%s'", | &file_spec, |
| strerror(errno), errno, fname); | "%s failed: %s (%d), actual filename '%s'", |
| return false; | action_name, strerror(errno), errno, fname); |
| return false; | |
| } | |
| } | } |
| static void create_dir_for_file(const String& file_spec) { | void create_dir_for_file(const String& file_spec) { |
| size_t pos_after=1; | size_t pos_after=1; |
| int pos_before; | size_t pos_before; |
| while((pos_before=file_spec.pos("/", 1, pos_after))>=0) { | while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) { |
| mkdir(file_spec.mid(0, pos_before).cstr(String::UL_FILE_SPEC), 0775); | pa_mkdir(file_spec.mid(0, pos_before).taint_cstr(String::L_FILE_SPEC), 0775); |
| pos_after=pos_before+1; | pos_after=pos_before+1; |
| } | } |
| } | } |
| void file_write(Pool& pool, | bool file_write_action_under_lock( |
| const String& file_spec, | const String& file_spec, |
| const void *data, size_t size, | const char* action_name, |
| bool as_text/*, | File_write_action action, |
| bool exclusive*/) { | void *context, |
| const char *fname=file_spec.cstr(String::UL_FILE_SPEC); | bool as_text, |
| bool do_append, | |
| bool do_block, | |
| bool fail_on_lock_problem) { | |
| const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); | |
| int f; | int f; |
| if(access(fname, W_OK)!=0) {/*no*/ | if(access(fname, W_OK)!=0) // no |
| create_dir_for_file(file_spec); | create_dir_for_file(file_spec); |
| if((f=open(fname, O_WRONLY|O_CREAT|_O_BINARY, 0666))>0) | if((f=open(fname, |
| close(f); | O_CREAT|O_RDWR |
| } | |(as_text?_O_TEXT:_O_BINARY) |
| if(access(fname, R_OK|W_OK)==0) { | |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) { |
| int mode=O_RDWR|(as_text?_O_TEXT:_O_BINARY) | if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) { |
| #ifdef WIN32 | Exception e("file.lock", |
| |O_TRUNC | &file_spec, |
| "shared lock failed: %s (%d), actual filename '%s'", | |
| strerror(errno), errno, fname); | |
| close(f); | |
| if(fail_on_lock_problem) | |
| throw e; | |
| return false; | |
| } | |
| try { | |
| #if (defined(HAVE_FCHMOD) && defined(PA_SAFE_MODE)) | |
| struct stat finfo; | |
| if(fstat(f, &finfo)==0 && finfo.st_mode & 0111) | |
| fchmod(f, finfo.st_mode & 0666/*clear executable bits*/); // backward: ignore errors if any | |
| #endif | #endif |
| ; | action(f, context); |
| if((f=open(fname, mode, 0666))>=0) { | } catch(...) { |
| /*if(exclusive) | #ifdef HAVE_FTRUNCATE |
| flock(f, LOCK_EX);*/ | if(!do_append) |
| ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower | |
| if(size) write(f, data, size); | #endif |
| #ifndef WIN32 | pa_unlock(f);close(f); |
| ftruncate(f, size); | rethrow; |
| #endif | } |
| /*if(exclusive) | |
| flock(f, LOCK_UN);*/ | #ifdef HAVE_FTRUNCATE |
| close(f); | if(!do_append) |
| return; | ftruncate(f, lseek(f, 0, SEEK_CUR)); // O_TRUNC truncates even exclusevely write-locked file [thanks to Igor Milyakov <virtan@rotabanner.com> for discovering] |
| #endif | |
| pa_unlock(f);close(f); | |
| return true; | |
| } else | |
| throw Exception(errno==EACCES?"file.access":0, | |
| &file_spec, | |
| "%s failed: %s (%d), actual filename '%s'", | |
| action_name, strerror(errno), errno, fname); | |
| // here should be nothing, see rethrow above | |
| } | |
| #ifndef DOXYGEN | |
| struct File_write_action_info { | |
| const char* str; | |
| size_t length; | |
| }; | |
| #endif | |
| static void file_write_action(int f, void *context) { | |
| File_write_action_info& info=*static_cast<File_write_action_info *>(context); | |
| if(info.length) { | |
| ssize_t written=write(f, info.str, info.length); | |
| if(written<0) | |
| 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, | |
| const char* data, | |
| size_t size, | |
| bool as_text, | |
| bool do_append, | |
| Charset* asked_charset) { | |
| if(as_text && asked_charset){ | |
| String::C body=String::C(data, size); | |
| body=Charset::transcode(body, charsets.source(), *asked_charset); | |
| data=body.str; | |
| size=body.length; | |
| }; | |
| File_write_action_info info={data, size}; | |
| file_write_action_under_lock( | |
| file_spec, | |
| "write", | |
| file_write_action, | |
| &info, | |
| as_text, | |
| do_append); | |
| } | |
| static size_t get_dir(char* fname, size_t helper_length){ | |
| bool dir=false; | |
| size_t pos=0; | |
| for(pos=helper_length; pos; pos--){ | |
| char c=fname[pos-1]; | |
| if(c=='/' || c=='\\'){ | |
| fname[pos-1]=0; | |
| dir=true; | |
| } else if(dir) break; | |
| } | |
| return pos; | |
| } | |
| static bool entry_readable(char* fname, bool need_dir) { | |
| if(need_dir){ | |
| size_t size=strlen(fname); | |
| while(size) { | |
| char c=fname[size-1]; | |
| if(c=='/' || c=='\\') | |
| fname[--size]=0; | |
| else | |
| break; | |
| } | } |
| } | } |
| throw Exception(0, 0, | |
| &file_spec, | struct stat finfo; |
| "write failed: %s (%d), actual filename '%s'", | if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) { |
| strerror(errno), errno, fname); | bool is_dir=(finfo.st_mode&S_IFDIR) != 0; |
| return is_dir==need_dir; | |
| } | |
| return false; | |
| } | |
| static bool entry_readable(const String& file_spec, bool need_dir) { | |
| return entry_readable(file_spec.taint_cstrm(String::L_FILE_SPEC), need_dir); | |
| } | } |
| // throws nothing! [this is required in file_move & file_delete] | // throws nothing! [this is required in file_move & file_delete] |
| static void rmdir(const String& file_spec, size_t pos_after) { | static void rmdir(const String& file_spec, size_t pos_after) { |
| int pos_before; | char* dir_spec=file_spec.taint_cstrm(String::L_FILE_SPEC); |
| if((pos_before=file_spec.pos("/", 1, pos_after))>=0) | size_t length=strlen(dir_spec); |
| rmdir(file_spec, pos_before+1); | while( (length=get_dir(dir_spec, length)) && (length > pos_after) ){ |
| #ifdef _MSC_VER | |
| rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::UL_FILE_SPEC)); | if(!entry_readable(dir_spec, true)) |
| } | break; |
| void file_delete(Pool& pool, const String& file_spec) { | DWORD attrs=GetFileAttributes(dir_spec); |
| const char *fname=file_spec.cstr(String::UL_FILE_SPEC); | if( |
| if(unlink(fname)!=0) | (attrs==INVALID_FILE_ATTRIBUTES) |
| throw Exception(0, 0, | || !(attrs & FILE_ATTRIBUTE_DIRECTORY) |
| &file_spec, | || (attrs & FILE_ATTRIBUTE_REPARSE_POINT) |
| "unlink failed: %s (%d), actual filename '%s'", | ) |
| strerror(errno), errno, fname); | break; |
| #endif | |
| if( rmdir(dir_spec) ) | |
| break; | |
| }; | |
| } | |
| rmdir(file_spec, 1); | 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) | |
| throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, | |
| &file_spec, | |
| "unlink failed: %s (%d), actual filename '%s'", | |
| strerror(errno), errno, fname); | |
| else | |
| return false; | |
| } | |
| if(!keep_empty_dirs) | |
| rmdir(file_spec, 1); | |
| return true; | |
| } | } |
| void file_move(Pool& pool, const String& old_spec, const String& new_spec) { | |
| const char *old_spec_cstr=old_spec.cstr(String::UL_FILE_SPEC); | void file_move(const String& old_spec, const String& new_spec, bool keep_empty_dirs) { |
| const char *new_spec_cstr=new_spec.cstr(String::UL_FILE_SPEC); | 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); | |
| create_dir_for_file(new_spec); | create_dir_for_file(new_spec); |
| if(rename(old_spec_cstr, new_spec_cstr)!=0) | if(rename(old_spec_cstr, new_spec_cstr)!=0) |
| throw Exception(0, 0, | throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, |
| &old_spec, | &old_spec, |
| "rename failed: %s (%d), actual filename '%s' to '%s'", | "rename failed: %s (%d), actual filename '%s' to '%s'", |
| strerror(errno), errno, old_spec_cstr, new_spec_cstr); | strerror(errno), errno, old_spec_cstr, new_spec_cstr); |
| rmdir(old_spec, 1); | if(!keep_empty_dirs) |
| rmdir(old_spec, 1); | |
| } | } |
| static bool entry_readable(const String& file_spec, bool need_dir) { | bool entry_exists(const char* fname, struct stat *afinfo) { |
| const char *fname=file_spec.cstr(String::UL_FILE_SPEC); | struct stat lfinfo; |
| struct stat finfo; | bool result=stat(fname, &lfinfo)==0; |
| if(access(fname, R_OK)==0 && stat(fname, &finfo)==0) { | if(afinfo) |
| bool is_dir=finfo.st_mode&S_IFDIR != 0; | *afinfo=lfinfo; |
| return is_dir==need_dir; | return result; |
| } | |
| return false; | |
| } | } |
| bool file_readable(const String& file_spec) { | |
| return entry_readable(file_spec, false); | bool entry_exists(const String& file_spec) { |
| const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); | |
| return entry_exists(fname, 0); | |
| } | } |
| bool dir_readable(const String& file_spec) { | |
| return entry_readable(file_spec, true); | bool file_exist(const String& file_spec) { |
| return entry_readable(file_spec, false); | |
| } | } |
| String *file_readable(const String& path, const String& name) { | |
| String *result=new(path.pool()) String(path); | bool dir_exists(const String& file_spec) { |
| *result << "/"; | return entry_readable(file_spec, true); |
| *result << name; | } |
| return file_readable(*result)?result:0; | |
| const String* file_exist(const String& path, const String& name) { | |
| String& result=*new String(path); | |
| if(path.last_char() != '/') | |
| result << "/"; | |
| result << name; | |
| return file_exist(result)?&result:0; | |
| } | } |
| bool file_executable(const String& file_spec) { | bool file_executable(const String& file_spec) { |
| return access(file_spec.cstr(String::UL_FILE_SPEC), X_OK)==0; | return access(file_spec.taint_cstr(String::L_FILE_SPEC), X_OK)==0; |
| } | } |
| bool file_stat(const String& file_spec, | bool file_stat(const String& file_spec, |
| size_t& rsize, | size_t& rsize, |
| time_t& ratime, | time_t& ratime, |
| time_t& rmtime, | time_t& rmtime, |
| time_t& rctime, | time_t& rctime, |
| bool fail_on_read_problem) { | bool fail_on_read_problem) { |
| Pool& pool=file_spec.pool(); | const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); |
| const char *fname=file_spec.cstr(String::UL_FILE_SPEC); | struct stat finfo; |
| struct stat finfo; | if(stat(fname, &finfo)!=0) { |
| if(stat(fname, &finfo)!=0) | |
| if(fail_on_read_problem) | if(fail_on_read_problem) |
| throw Exception(0, 0, | throw Exception("file.missing", |
| &file_spec, | &file_spec, |
| "getting file size failed: %s (%d), real filename '%s'", | "getting file size failed: %s (%d), real filename '%s'", |
| strerror(errno), errno, fname); | strerror(errno), errno, fname); |
| else | else |
| return false; | return false; |
| } | |
| rsize=finfo.st_size; | rsize=finfo.st_size; |
| ratime=finfo.st_atime; | ratime=finfo.st_atime; |
| rmtime=finfo.st_mtime; | rmtime=finfo.st_mtime; |
| Line 259 bool file_stat(const String& file_spec, | Line 548 bool file_stat(const String& file_spec, |
| return true; | return true; |
| } | } |
| char *getrow(char **row_ref, char delim) { | /** |
| char *result=*row_ref; | String related functions |
| if(result) { | */ |
| *row_ref=strchr(result, delim); | |
| bool capitalized(const char* s){ | |
| bool upper=true; | |
| for(const char* c=s; *c; c++){ | |
| if(*c != (upper ? toupper((unsigned char)*c) : tolower((unsigned char)*c))) | |
| return false; | |
| upper=strchr("-_ ", *c) != 0; | |
| } | |
| return true; | |
| } | |
| const char* capitalize(const char* s){ | |
| if(!s || capitalized(s)) | |
| return s; | |
| char* result=pa_strdup(s); | |
| if(result){ | |
| bool upper=true; | |
| for(char* c=result; *c; c++){ | |
| *c=upper ? (char)toupper((unsigned char)*c) : (char)tolower((unsigned char)*c); | |
| upper=strchr("-_ ", *c) != 0; | |
| } | |
| } | |
| return (const char*)result; | |
| } | |
| void fix_line_breaks(char *str, size_t& length) { | |
| //_asm int 3; | |
| const char* const eob=str+length; | |
| char* dest=str; | |
| // fix DOS: \r\n -> \n | |
| // fix Macintosh: \r -> \n | |
| char* bol=str; | |
| while(char* eol=(char*)memchr(bol, '\r', eob -bol)) { | |
| size_t len=eol-bol; | |
| if(dest!=bol) | |
| memmove(dest, bol, len); | |
| dest+=len; | |
| *dest++='\n'; | |
| if(&eol[1]<eob && eol[1]=='\n') { // \r, \n = DOS | |
| bol=eol+2; | |
| length--; | |
| } else // \r, not \n = Macintosh | |
| bol=eol+1; | |
| } | |
| // last piece without \r | |
| if(dest!=bol) | |
| memmove(dest, bol, eob-bol); | |
| str[length]=0; // terminating | |
| } | |
| /** | |
| scans for @a delim[default \n] in @a *row_ref, | |
| @return piece of line before it or end of string, if no @a delim found | |
| assigns @a *row_ref to point right after delimiter if there were one | |
| or to zero if no @a delim were found. | |
| */ | |
| char* getrow(char* *row_ref, char delim) { | |
| char* result=*row_ref; | |
| if(result) { | |
| *row_ref=strchr(result, delim); | |
| if(*row_ref) | if(*row_ref) |
| *((*row_ref)++)=0; | *((*row_ref)++)=0; |
| else if(!*result) | else if(!*result) |
| return 0; | return 0; |
| } | } |
| return result; | return result; |
| } | } |
| char *lsplit(char *string, char delim) { | char* lsplit(char* string, char delim) { |
| if(string) { | if(string) { |
| char *v=strchr(string, delim); | char* v=strchr(string, delim); |
| if(v) { | if(v) { |
| *v=0; | *v=0; |
| return v+1; | return v+1; |
| } | } |
| } | } |
| return 0; | return 0; |
| } | } |
| char *lsplit(char **string_ref, char delim) { | char* lsplit(char* *string_ref, char delim) { |
| char *result=*string_ref; | char* result=*string_ref; |
| char *next=lsplit(*string_ref, delim); | char* next=lsplit(*string_ref, delim); |
| *string_ref=next; | *string_ref=next; |
| return result; | return result; |
| } | } |
| char *rsplit(char *string, char delim) { | char* rsplit(char* string, char delim) { |
| if(string) { | if(string) { |
| char *v=strrchr(string, delim); | char* v=strrchr(string, delim); |
| if(v) { | if(v) { |
| *v=0; | *v=0; |
| return v+1; | return v+1; |
| } | } |
| } | } |
| return NULL; | return NULL; |
| } | } |
| /// @todo less stupid type detection | |
| char *format(Pool& pool, double value, char *fmt) { | // format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx |
| char *result=(char *)pool.malloc(MAX_NUMBER, 4); | // flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx |
| if(fmt) | // width, precision: non negative decimal number |
| if(strpbrk(fmt, "diouxX")) | enum FormatType { |
| if(strpbrk(fmt, "ouxX")) | FormatInvalid, |
| snprintf(result, MAX_NUMBER, fmt, (uint)value ); | FormatInt, |
| else | FormatUInt, |
| snprintf(result, MAX_NUMBER, fmt, (int)value ); | FormatDouble |
| else | }; |
| snprintf(result, MAX_NUMBER, fmt, value); | FormatType format_type(const char* fmt){ |
| else | enum FormatState { |
| snprintf(result, MAX_NUMBER, "%d", (int)value); | Percent, |
| Flags, | |
| Width, | |
| Precision, | |
| Done | |
| } state=Percent; | |
| FormatType result=FormatInvalid; | |
| const 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; | return result; |
| } | } |
| const char* format(double value, const char* fmt) { | |
| char local_buf[MAX_NUMBER]; | |
| int size=-1; | |
| 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); | |
| 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_t)size); | |
| } | |
| size_t stdout_write(const void *buf, size_t size) { | size_t stdout_write(const void *buf, size_t size) { |
| #ifdef WIN32 | #ifdef WIN32 |
| size_t to_write = size; | |
| do{ | do{ |
| int chunk_written=fwrite(buf, 1, min(8*0x400, size), stdout); | int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout); |
| if(chunk_written<=0) | if(chunk_written<=0) |
| break; | break; |
| size-=chunk_written; | size-=chunk_written; |
| buf=((const char*)buf)+chunk_written; | buf=((const char*)buf)+chunk_written; |
| } while(size>0); | } while(size>0); |
| return size; | return to_write-size; |
| #else | #else |
| return fwrite(buf, 1, size, stdout); | return fwrite(buf, 1, size, stdout); |
| #endif | #endif |
| } | } |
| char *unescape_chars(Pool& pool, const char *cp, int len) { | enum EscapeState { |
| char *s=(char *)pool.malloc(len + 1, 5); | EscapeRest, |
| enum EscapeState { | EscapeFirst, |
| EscapeRest, | EscapeSecond, |
| EscapeFirst, | EscapeUnicode |
| EscapeSecond | }; |
| } escapeState=EscapeRest; | |
| int escapedValue=0; | // @todo prescan for reduce required size (unescaped sting in 1 byte charset requires less memory usually) |
| char* unescape_chars(const char* cp, int len, Charset* charset, bool js){ | |
| char* s=new(PointerFreeGC) char[len+1]; // must be enough (%uXXXX==6 bytes, max utf-8 char length==6 bytes) | |
| char* dst=s; | |
| EscapeState escapeState=EscapeRest; | |
| uint escapedValue=0; | |
| int srcPos=0; | int srcPos=0; |
| int dstPos=0; | short int jsCnt=0; |
| while(srcPos < len) { | while(srcPos<len){ |
| int ch=cp[srcPos]; | uchar c=(uchar)cp[srcPos]; |
| switch(escapeState) { | if(c=='%' || (c=='\\' && js)){ |
| case EscapeRest: | escapeState=EscapeFirst; |
| if(ch=='%') { | } else { |
| escapeState=EscapeFirst; | switch(escapeState) { |
| } else if(ch=='+') { | case EscapeRest: |
| s[dstPos++]=' '; | if(!js && c=='+'){ |
| } else { | *dst++=' '; |
| s[dstPos++]=ch; | } 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; | |
| } | } |
| break; | |
| case EscapeFirst: | |
| escapedValue=hex_value[ch] << 4; | |
| escapeState=EscapeSecond; | |
| break; | |
| case EscapeSecond: | |
| escapedValue +=hex_value[ch]; | |
| s[dstPos++]=escapedValue; | |
| escapeState=EscapeRest; | |
| break; | |
| } | } |
| srcPos++; | srcPos++; |
| } | } |
| s[dstPos]=0; | |
| *dst=0; // zero-termination | |
| return s; | return s; |
| } | } |
| /// used by attributed_meaning_to_string / append_attribute_subattribute | char *search_stop(char*& current, char cstop_at) { |
| struct Attributed_meaning_info { | // sanity check |
| String *header; // header line being constructed | if(!current) |
| String::Untaint_lang lang; // language in which to append to that line | return 0; |
| }; | |
| static void append_attribute_subattribute(const Hash::Key& akey, Hash::Val *avalue, | // skip leading WS |
| void *info) { | while(*current==' ' || *current=='\t') |
| if(akey==VALUE_NAME) | current++; |
| return; | if(!*current) |
| return current=0; | |
| Attributed_meaning_info& ami=*static_cast<Attributed_meaning_info *>(info); | |
| char *result=current; | |
| // ...; charset=windows1251 | if(char *pstop_at=strchr(current, cstop_at)) { |
| *ami.header << "; "; | *pstop_at=0; |
| ami.header->append(akey, ami.lang); | current=pstop_at+1; |
| *ami.header << "="; | } else |
| ami.header->append(static_cast<Value *>(avalue)->as_string(), ami.lang); | current=0; |
| } | |
| const String& attributed_meaning_to_string(Value& meaning, | |
| String::Untaint_lang lang) { | |
| String &result=*new(meaning.pool()) String(meaning.pool()); | |
| if(Hash *hash=meaning.get_hash(0)) { | |
| // $value(value) $subattribute(subattribute value) | |
| if(Value *value=static_cast<Value *>(hash->get(*value_name))) | |
| result.append(value->as_string(), lang, true); | |
| Attributed_meaning_info attributed_meaning_info={ | |
| &result, | |
| lang | |
| }; | |
| hash->for_each(append_attribute_subattribute, &attributed_meaning_info); | |
| } else // result value | |
| result.append(meaning.as_string(), lang, true); | |
| return result; | return result; |
| } | } |
| #ifdef WIN32 | #ifdef WIN32 |
| void back_slashes_to_slashes(char *s) { | void back_slashes_to_slashes(char* s) { |
| if(s) | if(s) |
| for(; *s; s++) | for(; *s; s++) |
| if(*s=='\\') | if(*s=='\\') |
| *s='/'; | *s='/'; |
| } | |
| #endif | |
| size_t strpos(const char *str, const char *substr) { | |
| const char *p = strstr(str, substr); | |
| return (p==0)?STRING_NOT_FOUND:p-str; | |
| } | |
| int remove_crlf(char* start, char* end) { | |
| char* from=start; | |
| char* to=start; | |
| bool skip=false; | |
| while(from < end){ | |
| switch(*from){ | |
| case '\n': | |
| case '\r': | |
| case '\t': | |
| case ' ': | |
| if(!skip){ | |
| *to=' '; | |
| to++; | |
| skip=true; | |
| } | |
| break; | |
| default: | |
| if(from != to) | |
| *to=*from; | |
| to++; | |
| skip=false; | |
| } | |
| from++; | |
| } | |
| return to-start; | |
| } | |
| const char* hex_digits="0123456789ABCDEF"; | |
| const char* hex_string(unsigned char* bytes, size_t size, bool upcase) { | |
| char *bytes_hex=new(PointerFreeGC) char [size*2/*byte->hh*/+1/*for zero-teminator*/]; | |
| unsigned char *src=bytes; | |
| unsigned char *end=bytes+size; | |
| char *dest=bytes_hex; | |
| const char *hex=upcase? hex_digits : "0123456789abcdef"; | |
| for(; src<end; src++) { | |
| *dest++=hex[*src/0x10]; | |
| *dest++=hex[*src%0x10]; | |
| } | |
| *dest=0; | |
| return bytes_hex; | |
| } | |
| /// must be last in this file | |
| #undef vsnprintf | |
| int __vsnprintf(char* b, size_t s, const char* f, va_list l) { | |
| if(!s) | |
| return 0; | |
| int r; | |
| // note: on win32 & maybe somewhere else | |
| // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing | |
| // http://stackoverflow.com/questions/2915672/snprintf-and-visual-studio-2010 | |
| --s; | |
| // clients do not check for negative 's', feature: ignore such prints | |
| if((ssize_t)s<0) | |
| return 0; | |
| #ifdef _MSC_VER | |
| // win32: if the number of bytes to write exceeds buffer, then count bytes are written and -1 is returned | |
| r=_vsnprintf(b, s, f, l); | |
| if(r<0) | |
| r=s; | |
| #else | |
| r=vsnprintf(b, s, f, l); | |
| /* | |
| solaris: man vsnprintf | |
| The snprintf() function returns the number of characters | |
| formatted, that is, the number of characters that would have | |
| been written to the buffer if it were large enough. If the | |
| value of n is 0 on a call to snprintf(), an unspecified | |
| value less than 1 is returned. | |
| */ | |
| if(r<0) | |
| r=0; | |
| else if((size_t)r>s) | |
| r=s; | |
| #endif | |
| b[r]=0; | |
| return r; | |
| } | |
| int __snprintf(char* b, size_t s, const char* f, ...) { | |
| va_list l; | |
| va_start(l, f); | |
| int r=__vsnprintf(b, s, f, l); | |
| va_end(l); | |
| return r; | |
| } | } |
| /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */ | |
| /* | /* |
| void slashes_to_back_slashes(char *s) { | * Authors: Michael Zucchi <notzed@helixcode.com> |
| if(s) | * Jeffrey Stedfast <fejj@helixcode.com> |
| for(; *s; s++) | * |
| if(*s=='/') | * Copyright 2000 Helix Code, Inc. (www.helixcode.com) |
| *s='\\'; | * |
| * 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 | #endif |
| g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size, (unsigned char*)result, &state, &save); | |
| bool StrEqNc(const char *s1, const char *s2, bool strict) { | //throw Exception(PARSER_RUNTIME, 0, "%d %d %d", in_size, new_size, filled); |
| while(true) { | assert(filled <= new_size); |
| if(!(*s1)) { | |
| if(!(*s2)) | return result; |
| return true; | } |
| else | |
| return !strict; | struct File_base64_action_info { |
| } else if(!(*s2)) | unsigned char** base64; |
| return !strict; | }; |
| if(isalpha(*s1)) { | |
| if(tolower(*s1) !=tolower(*s2)) | static void file_base64_file_action( |
| return false; | struct stat& finfo, |
| } else if((*s1) !=(*s2)) | int f, |
| const String&, const char* /*fname*/, bool, | |
| void *context) { | |
| if(finfo.st_size) { | |
| File_base64_action_info& info=*static_cast<File_base64_action_info *>(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) | |
| throw Exception("file.read", | |
| 0, | |
| "read failed: %s (%d)", strerror(errno), errno); | |
| return nCount; | |
| } | |
| static unsigned long crc32Table[256]; | |
| static void InitCrc32Table() | |
| { | |
| if(crc32Table[1] == 0){ | |
| // This is the official polynomial used by CRC32 in PKZip. | |
| // Often times the polynomial shown reversed as 0x04C11DB7. | |
| static const unsigned long dwPolynomial = 0xEDB88320; | |
| for(int i = 0; i < 256; i++) | |
| { | |
| unsigned long dwCrc = i; | |
| for(int j = 8; j > 0; j--) | |
| { | |
| if(dwCrc & 1) | |
| dwCrc = (dwCrc >> 1) ^ dwPolynomial; | |
| else | |
| dwCrc >>= 1; | |
| } | |
| crc32Table[i] = dwCrc; | |
| } | |
| } | |
| } | |
| inline void CalcCrc32(const unsigned char byte, unsigned long &crc32) | |
| { | |
| crc32 = ((crc32) >> 8) ^ crc32Table[(byte) ^ ((crc32) & 0x000000FF)]; | |
| } | |
| const unsigned long pa_crc32(const char *in, size_t in_size){ | |
| unsigned long crc32=0xFFFFFFFF; | |
| InitCrc32Table(); | |
| for(size_t i = 0; i<in_size; i++) | |
| CalcCrc32(in[i], crc32); | |
| return ~crc32; | |
| } | |
| static void file_crc32_file_action( | |
| struct stat& finfo, | |
| int f, | |
| const String&, const char* /*fname*/, bool, | |
| void *context) { | |
| unsigned long& crc32=*static_cast<unsigned long *>(context); | |
| if(finfo.st_size) { | |
| InitCrc32Table(); | |
| int nCount=0; | |
| do { | |
| unsigned char buffer[FILE_BUFFER_SIZE]; | |
| nCount = file_block_read(f, buffer, sizeof(buffer)); | |
| for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32); | |
| } while(nCount > 0); | |
| } | |
| } | |
| const unsigned long pa_crc32(const String& file_spec){ | |
| unsigned long crc32=0xFFFFFFFF; | |
| file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32); | |
| return ~crc32; | |
| } | |
| // 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){ | |
| char* CONTENT_TYPE=pa_strdup(content_type); | |
| 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, ';'); | |
| if(end) | |
| *end=0; // terminator | |
| return *begin?&charsets.get(begin):0; | |
| } | |
| } | |
| return 0; | |
| } | |
| static bool is_latin(const char *in){ | |
| for(; *in; in++){ | |
| if ((unsigned char)(*in) > 0x7F) | |
| return false; | return false; |
| s1++; | |
| s2++; | |
| } | } |
| return true; | |
| } | } |
| static int isLeap(int year) { | #define MAX_IDNA_LENGTH 256 |
| return !( | |
| (year % 4) || ((year % 400) && !(year % 100)) | const char *pa_idna_encode(const char *in, Charset &source_charset){ |
| ); | if(!in || is_latin(in)) |
| return in; | |
| uint32_t utf32[MAX_IDNA_LENGTH]; | |
| uint32_t *utf32_end=utf32; | |
| String::C sIn(in,strlen(in)); | |
| if(!source_charset.isUTF8()) | |
| sIn=Charset::transcode(sIn, source_charset, UTF8_charset); | |
| int status=pa_convertUTF8toUTF32((const UTF8**)&sIn.str, (const UTF8*)(sIn.str+sIn.length), &utf32_end, utf32+MAX_IDNA_LENGTH-1, strictConversion); | |
| if(status != conversionOK) | |
| throw Exception("idna encode", new String(in), "utf conversion failed (%d)", status); | |
| *utf32_end=0; | |
| char *result = (char *)pa_malloc(MAX_IDNA_LENGTH); | |
| status=pa_idna_to_ascii_4z(utf32, result, MAX_IDNA_LENGTH, 0); | |
| if(status != IDNA_SUCCESS) | |
| throw Exception("idna encode", new String(in), "encode failed: %s", pa_idna_strerror(status)); | |
| return result; | |
| } | |
| const char *pa_idna_decode(const char *in, Charset &asked_charset){ | |
| if(!in || !(*in)) | |
| return in; | |
| uint32_t utf32[MAX_IDNA_LENGTH]; | |
| const uint32_t *utf32_start=utf32; | |
| uint32_t *utf32_end; | |
| int status=pa_idna_to_unicode_4z(in, utf32, MAX_IDNA_LENGTH, 0); | |
| if(status != IDNA_SUCCESS) | |
| throw Exception("idna decode", new String(in), "decode failed: %s", pa_idna_strerror(status)); | |
| for(utf32_end=utf32; *utf32_end; utf32_end++); | |
| char *result = (char *)pa_malloc(MAX_IDNA_LENGTH); | |
| char *result_end = result; | |
| status=pa_convertUTF32toUTF8(&utf32_start, utf32_end, (UTF8**)&result_end, (UTF8*)(result+MAX_IDNA_LENGTH-1), strictConversion); | |
| if(status != conversionOK) | |
| throw Exception("idna decode", new String(in), "utf conversion failed (%d)", status); | |
| *result_end='\0'; | |
| if(!asked_charset.isUTF8()) | |
| result = (char *)Charset::transcode(result, UTF8_charset, asked_charset).cstr(); | |
| return result; | |
| } | |
| static bool isLeap(int year) { | |
| return !( | |
| (year % 4) || ((year % 400) && !(year % 100)) | |
| ); | |
| } | } |
| int getMonthDays(int year, int month) { | int getMonthDays(int year, int month) { |
| int monthDays[]={ | static int monthDays[]={ |
| 31, | 31, |
| isLeap(year) ? 29 : 28, | 28, |
| 31, | 31, |
| 30, | 30, |
| 31, | 31, |
| 30, | 30, |
| 31, | 31, |
| 31, | 31, |
| 30, | 30, |
| 31, | 31, |
| 30, | 30, |
| 31 | 31 |
| }; | }; |
| return monthDays[month]; | return (month == 1 /* january -- 0 */ && isLeap(year)) ? 29 : monthDays[month]; |
| } | } |
| void remove_crlf(char *start, char *end) { | String::C date_gmt_string(tm* tms) { |
| for(char *p=start; p<end; p++) | /// http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.3 |
| switch(*p) { | static const char month_names[12][4]={ |
| case '\n': *p='|'; break; | "Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec"}; |
| case '\r': *p=' '; break; | static const char days[7][4]={ |
| } | "Sun","Mon","Tue","Wed","Thu","Fri","Sat"}; |
| char *buf=new(PointerFreeGC) char[MAX_STRING]; | |
| return String::C(buf, | |
| snprintf(buf, MAX_STRING, "%s, %.2d %s %.4d %.2d:%.2d:%.2d GMT", | |
| days[tms->tm_wday], | |
| tms->tm_mday,month_names[tms->tm_mon],tms->tm_year+1900, | |
| tms->tm_hour,tms->tm_min,tms->tm_sec)); | |
| } | } |