|
|
| version 1.281, 2015/04/02 22:04:41 | version 1.306, 2019/11/12 21:13:31 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: commonly functions. | Parser: commonly functions. |
| Copyright (c) 2000-2012 Art. Lebedev Studio (http://www.artlebedev.com) | Copyright (c) 2000-2017 Art. Lebedev Studio (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) |
| */ | |
| * 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_exception.h" | #include "pa_exception.h" |
| Line 36 | Line 15 |
| #include "pcre.h" | #include "pcre.h" |
| #include "pa_request.h" | #include "pa_request.h" |
| #include "pa_idna.h" | |
| #include "pa_convert_utf.h" | |
| #ifdef _MSC_VER | #ifdef _MSC_VER |
| #include <windows.h> | #include <windows.h> |
| #include <direct.h> | #include <direct.h> |
| Line 76 volatile const char * IDENT_PA_COMMON_C= | Line 58 volatile const char * IDENT_PA_COMMON_C= |
| const String file_status_name(FILE_STATUS_NAME); | 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 | // 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) { | 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); | 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; | return file.success?file.str:0; |
| Line 111 struct File_read_action_info { | Line 124 struct File_read_action_info { |
| }; | }; |
| #endif | #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, void *context) { |
| File_read_action_info& info=*static_cast<File_read_action_info *>(context); | File_read_action_info& info=*static_cast<File_read_action_info *>(context); |
| size_t to_read_size=info.count; | size_t to_read_size=info.count; |
| if(!to_read_size) | if(!to_read_size) |
| to_read_size=(size_t)finfo.st_size; | to_read_size=check_file_size(finfo.st_size, file_spec); |
| assert( !(info.buf && as_text) ); | |
| if(to_read_size) { | if(to_read_size) { |
| if(info.offset) | if(info.offset) |
| lseek(f, info.offset, SEEK_SET); | lseek(f, info.offset, SEEK_SET); |
| Line 147 File_read_result file_read(Request_chars | Line 159 File_read_result file_read(Request_chars |
| File_read_action_info info={&result.str, &result.length, buf, offset, count}; | File_read_action_info info={&result.str, &result.length, buf, offset, count}; |
| result.success=file_read_action_under_lock(file_spec, | result.success=file_read_action_under_lock(file_spec, "read", file_read_action, &info, as_text, fail_on_read_problem); |
| "read", file_read_action, &info, | |
| as_text, fail_on_read_problem); | |
| if(as_text){ | if(as_text){ |
| if(result.success){ | if(result.success){ |
| Charset* asked_charset=0; | 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(params) |
| if(Value* vcharset_name=params->get(PA_CHARSET_NAME)) | if(Value* vcharset_name=params->get(PA_CHARSET_NAME)) |
| asked_charset=&::charsets.get(vcharset_name->as_string().change_case(charsets.source(), String::CC_UPPER)); | asked_charset=&pa_charsets.get(vcharset_name->as_string()); |
| asked_charset=pa_charsets.checkBOM(result.str, result.length, asked_charset); | |
| 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 | 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); | String::C body=String::C(result.str, result.length); |
| Line 188 File_read_result file_load(Request& r, c | Line 193 File_read_result file_load(Request& r, c |
| File_read_result result={false, 0, 0, 0}; | File_read_result result={false, 0, 0, 0}; |
| if(file_spec.starts_with("http://")) { | if(file_spec.starts_with("http://")) { |
| if(offset || count) | if(offset || count) |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "offset and load options are not supported for HTTP:// file load"); |
| 0, | |
| "offset and load options are not supported for HTTP:// file load"); | |
| // fail on read problem | // fail on read problem |
| File_read_http_result http=pa_internal_file_read_http(r, file_spec, as_text, params, transcode_text_result); | File_read_http_result http=pa_internal_file_read_http(r, file_spec, as_text, params, transcode_text_result); |
| Line 210 File_read_result file_load(Request& r, c | Line 213 File_read_result file_load(Request& r, c |
| void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) { | void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) { |
| if(finfo.st_uid/*foreign?*/!=geteuid() | if(finfo.st_uid/*foreign?*/!=geteuid() |
| && finfo.st_gid/*foreign?*/!=getegid()) | && finfo.st_gid/*foreign?*/!=getegid()) |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, |
| &file_spec, | &file_spec, |
| "parser is in safe mode: " | "parser is in safe mode: reading files of foreign group and user disabled " |
| "reading files of foreign group and user disabled " | "[recompile parser with --disable-safe-mode configure option], " |
| "[recompile parser with --disable-safe-mode configure option], " | "actual filename '%s', fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)", |
| "actual filename '%s', " | fname, finfo.st_uid, geteuid(), finfo.st_gid, getegid() |
| "fuid(%d)!=euid(%d) or fgid(%d)!=egid(%d)", | ); |
| fname, | |
| finfo.st_uid, geteuid(), | |
| finfo.st_gid, getegid()); | |
| } | } |
| #else | #else |
| void check_safe_mode(struct stat, const String&, const char*) { | void check_safe_mode(struct stat, const String&, const char*) { |
| Line 244 bool file_read_action_under_lock(const S | Line 244 bool file_read_action_under_lock(const S |
| // 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. |
| // later: it seems, even this does not help sometimes | // 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 { | try { |
| if(pa_lock_shared_blocking(f)!=0) | if(pa_lock_shared_blocking(f)!=0) |
| throw Exception("file.lock", | throw Exception("file.lock", &file_spec, "shared lock failed: %s (%d), actual filename '%s'", strerror(errno), errno, fname); |
| &file_spec, | |
| "shared lock failed: %s (%d), actual filename '%s'", | |
| strerror(errno), errno, fname); | |
| struct stat finfo; | struct stat finfo; |
| if(fstat(f, &finfo)!=0) | if(pa_fstat(f, &finfo)!=0) |
| throw Exception("file.missing", // hardly possible: we just opened it OK | throw Exception("file.missing", // hardly possible: we just opened it OK |
| &file_spec, | &file_spec, "stat failed: %s (%d), actual filename '%s'", strerror(errno), errno, fname); |
| "stat failed: %s (%d), actual filename '%s'", | |
| strerror(errno), errno, fname); | |
| check_safe_mode(finfo, file_spec, fname); | check_safe_mode(finfo, file_spec, fname); |
| action(finfo, f, file_spec, fname, as_text, context); | action(finfo, f, file_spec, context); |
| } catch(...) { | } catch(...) { |
| pa_unlock(f);close(f); | pa_unlock(f);close(f); |
| if(fail_on_read_problem) | if(fail_on_read_problem) |
| rethrow; | rethrow; |
| return false; | return false; |
| } | } |
| pa_unlock(f);close(f); | pa_unlock(f);close(f); |
| return true; | return true; |
| } else { | } else { |
| if(fail_on_read_problem) | if(fail_on_read_problem) |
| throw Exception(errno==EACCES?"file.access" | throw Exception(errno==EACCES ? "file.access" : (errno==ENOENT || errno==ENOTDIR || errno==ENODEV) ? "file.missing" : 0, |
| :(errno==ENOENT || errno==ENOTDIR || errno==ENODEV)?"file.missing":0, | &file_spec, "%s failed: %s (%d), actual filename '%s'", action_name, strerror(errno), errno, fname); |
| &file_spec, | |
| "%s failed: %s (%d), actual filename '%s'", | |
| action_name, strerror(errno), errno, fname); | |
| return false; | return false; |
| } | } |
| } | } |
| Line 305 bool file_write_action_under_lock( | Line 297 bool file_write_action_under_lock( |
| 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, | if((f=pa_open(fname, |
| O_CREAT|O_RDWR | O_CREAT|O_RDWR |
| |(as_text?_O_TEXT:_O_BINARY) | |(as_text?_O_TEXT:_O_BINARY) |
| |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) { | |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) { |
| if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) { | if((do_block?pa_lock_exclusive_blocking(f):pa_lock_exclusive_nonblocking(f))!=0) { |
| Exception e("file.lock", | Exception e("file.lock", &file_spec, "shared lock failed: %s (%d), actual filename '%s'", strerror(errno), errno, fname); |
| &file_spec, | |
| "shared lock failed: %s (%d), actual filename '%s'", | |
| strerror(errno), errno, fname); | |
| close(f); | close(f); |
| if(fail_on_lock_problem) | if(fail_on_lock_problem) |
| throw e; | throw e; |
| Line 323 bool file_write_action_under_lock( | Line 312 bool file_write_action_under_lock( |
| try { | try { |
| #if (defined(HAVE_FCHMOD) && defined(PA_SAFE_MODE)) | #if (defined(HAVE_FCHMOD) && defined(PA_SAFE_MODE)) |
| struct stat finfo; | struct stat finfo; |
| if(fstat(f, &finfo)==0 && finfo.st_mode & 0111) | if(pa_fstat(f, &finfo)==0 && finfo.st_mode & 0111) |
| fchmod(f, finfo.st_mode & 0666/*clear executable bits*/); // backward: ignore errors if any | fchmod(f, finfo.st_mode & 0666/*clear executable bits*/); // backward: ignore errors if any |
| #endif | #endif |
| action(f, context); | action(f, context); |
| Line 343 bool file_write_action_under_lock( | Line 332 bool file_write_action_under_lock( |
| pa_unlock(f);close(f); | pa_unlock(f);close(f); |
| return true; | return true; |
| } else | } else |
| throw Exception(errno==EACCES?"file.access":0, | throw Exception(errno==EACCES ? "file.access" : 0, &file_spec, "%s failed: %s (%d), actual filename '%s'", action_name, strerror(errno), errno, fname); |
| &file_spec, | |
| "%s failed: %s (%d), actual filename '%s'", | |
| action_name, strerror(errno), errno, fname); | |
| // here should be nothing, see rethrow above | // here should be nothing, see rethrow above |
| } | } |
| Line 455 static void rmdir(const String& file_spe | Line 441 static void rmdir(const String& file_spe |
| bool file_delete(const String& file_spec, bool fail_on_problem, bool keep_empty_dirs) { | 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); | const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); |
| if(unlink(fname)!=0) | if(unlink(fname)!=0) { |
| if(fail_on_problem) | if(fail_on_problem) |
| throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, | throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, |
| &file_spec, | &file_spec, "unlink failed: %s (%d), actual filename '%s'", strerror(errno), errno, fname); |
| "unlink failed: %s (%d), actual filename '%s'", | |
| strerror(errno), errno, fname); | |
| else | else |
| return false; | return false; |
| } | |
| if(!keep_empty_dirs) | if(!keep_empty_dirs) |
| rmdir(file_spec, 1); | rmdir(file_spec, 1); |
| Line 477 void file_move(const String& old_spec, c | Line 462 void file_move(const String& old_spec, c |
| 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(errno==EACCES?"file.access":errno==ENOENT?"file.missing":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'", strerror(errno), errno, old_spec_cstr, new_spec_cstr); |
| "rename failed: %s (%d), actual filename '%s' to '%s'", | |
| strerror(errno), errno, old_spec_cstr, new_spec_cstr); | |
| if(!keep_empty_dirs) | if(!keep_empty_dirs) |
| rmdir(old_spec, 1); | rmdir(old_spec, 1); |
| Line 489 void file_move(const String& old_spec, c | Line 472 void file_move(const String& old_spec, c |
| bool entry_exists(const char* fname, struct stat *afinfo) { | bool entry_exists(const char* fname, struct stat *afinfo) { |
| struct stat lfinfo; | struct stat lfinfo; |
| bool result=stat(fname, &lfinfo)==0; | bool result=pa_stat(fname, &lfinfo)==0; |
| if(afinfo) | if(afinfo) |
| *afinfo=lfinfo; | *afinfo=lfinfo; |
| return result; | return result; |
| Line 520 bool file_executable(const String& file_ | Line 503 bool file_executable(const String& file_ |
| return access(file_spec.taint_cstr(String::L_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, uint64_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.taint_cstr(String::L_FILE_SPEC); | const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); |
| struct stat finfo; | struct stat finfo; |
| if(stat(fname, &finfo)!=0) | if(pa_stat(fname, &finfo)!=0) { |
| if(fail_on_read_problem) | if(fail_on_read_problem) |
| throw Exception("file.missing", | throw Exception("file.missing", &file_spec, "getting file size failed: %s (%d), real filename '%s'", strerror(errno), errno, fname); |
| &file_spec, | |
| "getting file size failed: %s (%d), real filename '%s'", | |
| strerror(errno), errno, fname); | |
| else | else |
| return false; | return false; |
| } | |
| rsize=finfo.st_size; | rsize=finfo.st_size; |
| ratime=finfo.st_atime; | ratime=(time_t)finfo.st_atime; |
| rmtime=finfo.st_mtime; | rmtime=(time_t)finfo.st_mtime; |
| rctime=finfo.st_ctime; | rctime=(time_t)finfo.st_ctime; |
| return true; | 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 | String related functions |
| */ | */ |
| Line 572 const char* capitalize(const char* s){ | Line 554 const char* capitalize(const char* s){ |
| return (const char*)result; | return (const char*)result; |
| } | } |
| char *str_lower(const char *s, size_t helper_length){ | |
| char *result=pa_strdup(s, helper_length); | |
| for(char* c=result; *c; c++) | |
| *c=(char)tolower((unsigned char)*c); | |
| return result; | |
| } | |
| char *str_upper(const char *s, size_t helper_length){ | |
| char *result=pa_strdup(s, helper_length); | |
| for(char* c=result; *c; c++) | |
| *c=(char)toupper((unsigned char)*c); | |
| return result; | |
| } | |
| void fix_line_breaks(char *str, size_t& length) { | void fix_line_breaks(char *str, size_t& length) { |
| //_asm int 3; | //_asm int 3; |
| const char* const eob=str+length; | const char* const eob=str+length; |
| Line 727 const char* format(double value, const c | Line 723 const char* format(double value, const c |
| size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value); | size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value); |
| break; | break; |
| case FormatInvalid: | case FormatInvalid: |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "Incorrect format string '%s' was specified.", fmt); |
| 0, | |
| "Incorrect format string '%s' was specified.", fmt); | |
| } | } |
| } else | } else |
| size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value); | 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 | if(size < 0 || size >= MAX_NUMBER-1){ // on win32 we manually reduce max size while printing |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "Error occurred white executing snprintf with format string '%s'.", fmt); |
| 0, | |
| "Error occure white executing snprintf with format string '%s'.", fmt); | |
| } | } |
| return pa_strdup(local_buf, (size_t)size); | return pa_strdup(local_buf, (size_t)size); |
| Line 782 char* unescape_chars(const char* cp, int | Line 774 char* unescape_chars(const char* cp, int |
| } else { | } else { |
| switch(escapeState) { | switch(escapeState) { |
| case EscapeRest: | case EscapeRest: |
| if(!js && c=='+'){ | if(c=='+' && !js){ |
| *dst++=' '; | *dst++=' '; |
| } else { | } else { |
| *dst++=c; | *dst++=c; |
| Line 914 const char* hex_string(unsigned char* by | Line 906 const char* hex_string(unsigned char* by |
| return bytes_hex; | 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 | |
| --s; | |
| // clients do not check for negative 's', feature: ignore such prints | |
| if((ssize_t)s<0) | |
| return 0; | |
| #ifdef _MSC_VER | |
| /* | |
| win32: | |
| mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm | |
| if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 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 */ | |
| /* | |
| * Authors: Michael Zucchi <notzed@helixcode.com> | |
| * Jeffrey Stedfast <fejj@helixcode.com> | |
| * | |
| * 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&, 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 file_block_read(const int f, unsigned char* buffer, const size_t size){ |
| int nCount = read(f, buffer, size); | int nCount = read(f, buffer, size); |
| if (nCount < 0) | if (nCount < 0) |
| throw Exception("file.read", | throw Exception("file.read", 0, "read failed: %s (%d)", strerror(errno), errno); |
| 0, | |
| "read failed: %s (%d)", strerror(errno), errno); | |
| return nCount; | return nCount; |
| } | } |
| Line 1334 inline void CalcCrc32(const unsigned cha | Line 942 inline void CalcCrc32(const unsigned cha |
| } | } |
| const unsigned long pa_crc32(const char *in, size_t in_size){ | unsigned long pa_crc32(const char *in, size_t in_size){ |
| unsigned long crc32=0xFFFFFFFF; | unsigned long crc32=0xFFFFFFFF; |
| InitCrc32Table(); | InitCrc32Table(); |
| Line 1344 const unsigned long pa_crc32(const char | Line 952 const unsigned long pa_crc32(const char |
| return ~crc32; | return ~crc32; |
| } | } |
| static void file_crc32_file_action( | static void file_crc32_file_action(struct stat& finfo, int f, const String&, void *context) { |
| struct stat& finfo, | |
| int f, | |
| const String&, const char* /*fname*/, bool, | |
| void *context) { | |
| unsigned long& crc32=*static_cast<unsigned long *>(context); | unsigned long& crc32=*static_cast<unsigned long *>(context); |
| if(finfo.st_size) { | if(finfo.st_size) { |
| InitCrc32Table(); | InitCrc32Table(); |
| Line 1361 static void file_crc32_file_action( | Line 965 static void file_crc32_file_action( |
| } | } |
| } | } |
| const unsigned long pa_crc32(const String& file_spec){ | unsigned long pa_crc32(const String& file_spec){ |
| unsigned long crc32=0xFFFFFFFF; | unsigned long crc32=0xFFFFFFFF; |
| file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32); | file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32); |
| return ~crc32; | return ~crc32; |
| Line 1372 const unsigned long pa_crc32(const Strin | Line 976 const unsigned long pa_crc32(const Strin |
| // content-type: xxx; charset="WE-NEED-THIS"; | // content-type: xxx; charset="WE-NEED-THIS"; |
| Charset* detect_charset(const char* content_type){ | Charset* detect_charset(const char* content_type){ |
| if(content_type){ | if(content_type){ |
| char* CONTENT_TYPE=pa_strdup(content_type); | char* CONTENT_TYPE=str_upper(content_type); |
| for(char *p=CONTENT_TYPE; *p; p++) | |
| *p=(char)toupper((unsigned char)*p); | |
| if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){ | if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){ |
| begin+=8; // skip "CHARSET=" | begin+=8; // skip "CHARSET=" |
| Line 1391 Charset* detect_charset(const char* cont | Line 992 Charset* detect_charset(const char* cont |
| if(end) | if(end) |
| *end=0; // terminator | *end=0; // terminator |
| return *begin?&charsets.get(begin):0; | return *begin ? &pa_charsets.get_direct(begin) : 0; |
| } | } |
| } | } |
| return 0; | 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++){ | |
| if ((unsigned char)(*in) > 0x7F) | |
| return false; | |
| } | |
| return true; | |
| } | |
| #define MAX_IDNA_LENGTH 256 | |
| 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, pa_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); | |
| static bool isLeap(int year) { | *result_end='\0'; |
| return !( | |
| (year % 4) || ((year % 400) && !(year % 100)) | if(!asked_charset.isUTF8()) |
| ); | result = (char *)Charset::transcode(result, pa_UTF8_charset, asked_charset).cstr(); |
| } | |
| return result; | |
| int getMonthDays(int year, int month) { | |
| static int monthDays[]={ | |
| 31, | |
| 28, | |
| 31, | |
| 30, | |
| 31, | |
| 30, | |
| 31, | |
| 31, | |
| 30, | |
| 31, | |
| 30, | |
| 31 | |
| }; | |
| return (month == 1 /* january -- 0 */ && isLeap(year)) ? 29 : monthDays[month]; | |
| } | |
| String::C date_gmt_string(tm* tms) { | |
| /// http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.3 | |
| static const char month_names[12][4]={ | |
| "Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec"}; | |
| 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)); | |
| } | } |
| /// must be last in this file | |
| #undef vsnprintf | |
| int pa_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 pa_snprintf(char* b, size_t s, const char* f, ...) { | |
| va_list l; | |
| va_start(l, f); | |
| int r=pa_vsnprintf(b, s, f, l); | |
| va_end(l); | |
| return r; | |
| } | |