|
|
| version 1.219, 2006/11/14 17:25:21 | version 1.303, 2017/02/07 22:00:41 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: commonly functions. | Parser: commonly functions. |
| Copyright(c) 2001-2005 ArtLebedev Group (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 | * BASE64 part |
| Line 26 | Line 26 |
| * | * |
| */ | */ |
| static const char * const IDENT_COMMON_C="$Date$"; | |
| #include "pa_common.h" | #include "pa_common.h" |
| #include "pa_exception.h" | #include "pa_exception.h" |
| #include "pa_hash.h" | #include "pa_hash.h" |
| #include "pa_globals.h" | #include "pa_globals.h" |
| #include "pa_charsets.h" | #include "pa_charsets.h" |
| #include "pa_http.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 | |
| #ifdef _MSC_VER | |
| #define pa_mkdir(path, mode) _mkdir(path) | |
| #else | |
| #define pa_mkdir(path, mode) mkdir(path, mode) | |
| #endif | |
| 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 | // some maybe-undefined constants |
| Line 62 static const char * const IDENT_COMMON_C | Line 79 static const char * const IDENT_COMMON_C |
| const String file_status_name(FILE_STATUS_NAME); | const String file_status_name(FILE_STATUS_NAME); |
| // forwards | |
| const UTF16* pa_utf16_encode(const char* in, Charset& source_charset); | |
| // functions | // functions |
| void fix_line_breaks(char *str, size_t& length) { | #ifdef _MSC_VER |
| //_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) | |
| memcpy(dest, bol, len); | |
| dest+=len; | |
| *dest++='\n'; | |
| if(&eol[1]<eob && eol[1]=='\n') { // \r, \n = DOS | int pa_stat(const char *pathname, struct stat *buffer){ |
| bol=eol+2; | const UTF16* utf16name=pa_utf16_encode(pathname, pa_thread_request().charsets.source()); |
| length--; | return _wstat64((const wchar_t *)utf16name, buffer); |
| } else // \r, not \n = Macintosh | } |
| bol=eol+1; | |
| } | int pa_open(const char *pathname, int flags, int mode){ |
| // last piece without \r | const UTF16* utf16name=pa_utf16_encode(pathname, pa_thread_request().charsets.source()); |
| if(dest!=bol) | return _wopen((const wchar_t *)utf16name, flags, mode); |
| memcpy(dest, bol, eob-bol); | } |
| str[length]=0; // terminating | |
| 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); | |
| } | } |
| char* file_read_text(Request_charsets& charsets, | #endif |
| const String& file_spec, | |
| bool fail_on_read_problem, | char* file_read_text(Request_charsets& charsets, const String& file_spec, bool fail_on_read_problem, HashStringValue* params, bool transcode_result) { |
| HashStringValue* params/*, HashStringValue* * out_fields*/) { | File_read_result file=file_read(charsets, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); |
| File_read_result file= | return file.success?file.str:0; |
| file_read(charsets, file_spec, true, params, fail_on_read_problem); | } |
| char* file_load_text(Request& r, const String& file_spec, bool fail_on_read_problem, HashStringValue* params, bool transcode_result) { | |
| File_read_result file=file_load(r, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); | |
| return file.success?file.str:0; | return file.success?file.str:0; |
| } | } |
| /// these options were handled but not checked elsewhere, now check them | /// these options were handled but not checked elsewhere, now check them |
| int pa_get_valid_file_options_count(HashStringValue& options) | int pa_get_valid_file_options_count(HashStringValue& options) { |
| { | |
| int result=0; | int result=0; |
| if(options.get(PA_SQL_LIMIT_NAME)) | if(options.get(PA_SQL_LIMIT_NAME)) |
| result++; | result++; |
| Line 111 int pa_get_valid_file_options_count(Hash | Line 126 int pa_get_valid_file_options_count(Hash |
| result++; | result++; |
| if(options.get(PA_COLUMN_ENCLOSER_NAME)) | if(options.get(PA_COLUMN_ENCLOSER_NAME)) |
| result++; | result++; |
| if(options.get(PA_CHARSET_NAME)) | |
| result++; | |
| return result; | return result; |
| } | } |
| Line 120 struct File_read_action_info { | Line 137 struct File_read_action_info { |
| char* buf; size_t offset; size_t count; | char* buf; size_t offset; size_t count; |
| }; | }; |
| #endif | #endif |
| static void file_read_action( | |
| struct stat& finfo, | static void file_read_action(struct stat& finfo, int f, const String& file_spec, void *context) { |
| int f, | |
| const String& file_spec, const char* /*fname*/, bool as_text, | |
| 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); |
| *info.data=info.buf | *info.data=info.buf ? info.buf : (char *)pa_malloc_atomic(to_read_size+1); |
| ? info.buf | ssize_t result=read(f, *info.data, to_read_size); |
| : new(PointerFreeGC) char[to_read_size+(as_text?1:0)]; | if(result<0) |
| *info.data_size=(size_t)read(f, *info.data, to_read_size); | throw Exception("file.read", &file_spec, "read failed: %s (%d)", strerror(errno), errno); |
| *info.data_size=result; | |
| if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size) | |
| throw Exception(0, | |
| &file_spec, | |
| "read failed: actually read %u bytes count not in [0..%u] valid range", | |
| *info.data_size, to_read_size); | |
| } else { // empty file | } 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 | // 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=new(PointerFreeGC) char[1]; | *info.data=(char *)pa_malloc_atomic(1); |
| *(char*)(*info.data)=0; | *(char*)(*info.data)=0; |
| *info.data_size=0; | *info.data_size=0; |
| return; | return; |
| } | } |
| } | } |
| File_read_result file_read(Request_charsets& charsets, const String& file_spec, | File_read_result file_read(Request_charsets& charsets, const String& file_spec, |
| bool as_text, HashStringValue *params, | bool as_text, HashStringValue *params, |
| bool fail_on_read_problem, | bool fail_on_read_problem, |
| char* buf, size_t offset, size_t count) { | char* buf, size_t offset, size_t count, bool transcode_text_result) { |
| 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); | |
| } | |
| File_read_action_info info={&result.str, &result.length, buf, offset, count}; | |
| 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(params) | |
| if(Value* vcharset_name=params->get(PA_CHARSET_NAME)) | |
| 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 | |
| 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}; | 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(charsets, file_spec, as_text, params); | File_read_http_result http=pa_internal_file_read_http(r, file_spec, as_text, params, transcode_text_result); |
| result.success=true; | result.success=true; |
| result.str=http.str; | result.str=http.str; |
| result.length=http.length; | result.length=http.length; |
| result.headers=http.headers; | result.headers=http.headers; |
| } else { | } else |
| if(params) { | result= |
| int valid_options=pa_get_valid_file_options_count(*params); | file_read(r.charsets, file_spec, as_text, params, fail_on_read_problem, buf, offset, count, transcode_text_result); |
| if(valid_options!=params->count()) | |
| throw Exception("parser.runtime", | |
| 0, | |
| "invalid option passed"); | |
| } | |
| File_read_action_info info={&result.str, &result.length, | |
| buf, offset, count}; | |
| result.success=file_read_action_under_lock(file_spec, | |
| "read", file_read_action, &info, | |
| as_text, fail_on_read_problem); | |
| } | |
| if(result.success && as_text) { | |
| // UTF-8 signature: EF BB BF | |
| if(result.length>=3) { | |
| char *in=(char *)result.str; | |
| if(strncmp(in, "\xEF\xBB\xBF", 3)==0) { | |
| result.str=in+3; result.length-=3;// skip prefix | |
| } | |
| } | |
| fix_line_breaks((char *)(result.str), result.length); | |
| } | |
| return result; | return result; |
| } | } |
| #ifdef PA_SAFE_MODE | #ifdef PA_SAFE_MODE |
| 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(), | #else |
| finfo.st_gid, getegid()); | void check_safe_mode(struct stat, const String&, const char*) { |
| } | } |
| #endif | #endif |
| bool file_read_action_under_lock(const String& file_spec, | bool file_read_action_under_lock(const String& file_spec, |
| const char* action_name, File_read_action action, void *context, | const char* action_name, File_read_action action, void *context, |
| bool as_text, | bool as_text, |
| bool fail_on_read_problem) { | bool fail_on_read_problem) { |
| const char* fname=file_spec.cstr(String::L_FILE_SPEC); | const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); |
| int f; | int f; |
| // first open, next stat: | // first open, next stat: |
| Line 232 bool file_read_action_under_lock(const S | Line 258 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(stat(fname, &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); | |
| #ifdef PA_SAFE_MODE | |
| check_safe_mode(finfo, file_spec, fname); | check_safe_mode(finfo, file_spec, fname); |
| #endif | |
| 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":errno==ENOENT?"file.missing":0, | throw Exception(errno==EACCES ? "file.access" : (errno==ENOENT || errno==ENOTDIR || errno==ENODEV) ? "file.missing" : 0, |
| &file_spec, | &file_spec, "%s failed: %s (%d), actual filename '%s'", action_name, strerror(errno), errno, fname); |
| "%s failed: %s (%d), actual filename '%s'", | |
| action_name, strerror(errno), errno, fname); | |
| return false; | return false; |
| } | } |
| } | } |
| Line 275 void create_dir_for_file(const String& f | Line 292 void create_dir_for_file(const String& f |
| size_t pos_after=1; | size_t pos_after=1; |
| size_t pos_before; | size_t pos_before; |
| while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) { | while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) { |
| mkdir(file_spec.mid(0, pos_before).cstr(String::L_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; |
| } | } |
| } | } |
| bool file_write_action_under_lock( | bool file_write_action_under_lock( |
| const String& file_spec, | const String& file_spec, |
| const char* action_name, File_write_action action, void *context, | const char* action_name, |
| File_write_action action, | |
| void *context, | |
| bool as_text, | bool as_text, |
| bool do_append, | bool do_append, |
| bool do_block, | bool do_block, |
| bool fail_on_lock_problem) { | bool fail_on_lock_problem) { |
| const char* fname=file_spec.cstr(String::L_FILE_SPEC); | 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, | 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 308 bool file_write_action_under_lock( | Line 324 bool file_write_action_under_lock( |
| } | } |
| try { | try { |
| action(f, context); | #if (defined(HAVE_FCHMOD) && defined(PA_SAFE_MODE)) |
| struct stat finfo; | |
| if(pa_fstat(f, &finfo)==0 && finfo.st_mode & 0111) | |
| fchmod(f, finfo.st_mode & 0666/*clear executable bits*/); // backward: ignore errors if any | |
| #endif | |
| action(f, context); | |
| } catch(...) { | } catch(...) { |
| #ifdef HAVE_FTRUNCATE | #ifdef HAVE_FTRUNCATE |
| if(!do_append) | if(!do_append) |
| Line 325 bool file_write_action_under_lock( | Line 346 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 |
| } | } |
| #ifndef DOXYGEN | #ifndef DOXYGEN |
| struct File_write_action_info { | struct File_write_action_info { |
| const char* str; size_t length; | const char* str; |
| size_t length; | |
| }; | }; |
| #endif | #endif |
| static void file_write_action(int f, void *context) { | static void file_write_action(int f, void *context) { |
| File_write_action_info& info=*static_cast<File_write_action_info *>(context); | File_write_action_info& info=*static_cast<File_write_action_info *>(context); |
| if(info.length) { | if(info.length) { |
| int written=write(f, info.str, info.length); | ssize_t written=write(f, info.str, info.length); |
| if(written<0) | if(written<0) |
| throw Exception(0, | throw Exception("file.write", 0, "write failed: %s (%d)", strerror(errno), errno); |
| 0, | if((size_t)written!=info.length) |
| "write failed: %s (%d)", strerror(errno), errno); | throw Exception("file.write", 0, "write failed: %u of %u bytes written", written, info.length); |
| } | } |
| } | } |
| void file_write( | void file_write( |
| const String& file_spec, | Request_charsets& charsets, |
| const char* data, size_t size, | const String& file_spec, |
| const char* data, | |
| size_t size, | |
| bool as_text, | bool as_text, |
| bool do_append) { | 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_info info={data, size}; |
| file_write_action_under_lock( | file_write_action_under_lock( |
| file_spec, | file_spec, |
| "write", file_write_action, &info, | "write", |
| file_write_action, | |
| &info, | |
| as_text, | as_text, |
| do_append); | 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; | |
| } | |
| } | |
| struct stat finfo; | |
| if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) { | |
| 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) { |
| size_t pos_before; | char* dir_spec=file_spec.taint_cstrm(String::L_FILE_SPEC); |
| if((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) | 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::L_FILE_SPEC)); | if(!entry_readable(dir_spec, true)) |
| break; | |
| DWORD attrs=GetFileAttributes(dir_spec); | |
| if( | |
| (attrs==INVALID_FILE_ATTRIBUTES) | |
| || !(attrs & FILE_ATTRIBUTE_DIRECTORY) | |
| || (attrs & FILE_ATTRIBUTE_REPARSE_POINT) | |
| ) | |
| break; | |
| #endif | |
| if( rmdir(dir_spec) ) | |
| break; | |
| }; | |
| } | } |
| bool file_delete(const String& file_spec, bool fail_on_problem) { | |
| const char* fname=file_spec.cstr(String::L_FILE_SPEC); | bool file_delete(const String& file_spec, bool fail_on_problem, bool keep_empty_dirs) { |
| if(unlink(fname)!=0) | const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); |
| 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) | |
| rmdir(file_spec, 1); | |
| rmdir(file_spec, 1); | |
| return true; | return true; |
| } | } |
| void file_move(const String& old_spec, const String& new_spec) { | |
| const char* old_spec_cstr=old_spec.cstr(String::L_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::L_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(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); | |
| rmdir(old_spec, 1); | if(!keep_empty_dirs) |
| rmdir(old_spec, 1); | |
| } | } |
| 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; |
| } | } |
| bool entry_exists(const String& file_spec) { | bool entry_exists(const String& file_spec) { |
| const char* fname=file_spec.cstr(String::L_FILE_SPEC); | const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); |
| return entry_exists(fname, 0); | return entry_exists(fname, 0); |
| } | } |
| static bool entry_readable(const String& file_spec, bool need_dir) { | |
| char* fname=file_spec.cstrm(String::L_FILE_SPEC); | |
| if(need_dir) { | |
| size_t size=strlen(fname); | |
| while(size) { | |
| char c=fname[size-1]; | |
| if(c=='/' || c=='\\') | |
| fname[--size]=0; | |
| else | |
| break; | |
| } | |
| } | |
| struct stat finfo; | |
| if(access(fname, R_OK)==0 && entry_exists(fname, &finfo)) { | |
| bool is_dir=(finfo.st_mode&S_IFDIR) != 0; | |
| return is_dir==need_dir; | |
| } | |
| return false; | |
| } | |
| bool file_exist(const String& file_spec) { | bool file_exist(const String& file_spec) { |
| return entry_readable(file_spec, false); | return entry_readable(file_spec, false); |
| } | } |
| bool dir_exists(const String& file_spec) { | bool dir_exists(const String& file_spec) { |
| return entry_readable(file_spec, true); | return entry_readable(file_spec, true); |
| } | } |
| const String* file_exist(const String& path, const String& name) { | const String* file_exist(const String& path, const String& name) { |
| String& result=*new String(path); | String& result=*new String(path); |
| result << "/"; | if(path.last_char() != '/') |
| result << "/"; | |
| result << name; | result << name; |
| return file_exist(result)?&result:0; | 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::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, | const char* fname=file_spec.taint_cstr(String::L_FILE_SPEC); |
| time_t& ratime, | |
| time_t& rmtime, | |
| time_t& rctime, | |
| bool fail_on_read_problem) { | |
| const char* fname=file_spec.cstr(String::L_FILE_SPEC); | |
| struct stat finfo; | 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 | |
| */ | |
| 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; | |
| } | |
| 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) { | |
| //_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* getrow(char* *row_ref, char delim) { |
| char* result=*row_ref; | char* result=*row_ref; |
| if(result) { | if(result) { |
| *row_ref=strchr(result, delim); | *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 | |
| const char* format(double value, char* fmt) { | // format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx |
| char local_buf[MAX_NUMBER]; | // flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx |
| size_t size; | // width, precision: non negative decimal number |
| enum FormatType { | |
| if(fmt) | FormatInvalid, |
| if(strpbrk(fmt, "diouxX")) | FormatInt, |
| if(strpbrk(fmt, "ouxX")) | FormatUInt, |
| size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value); | FormatDouble |
| else | }; |
| FormatType format_type(const char* fmt){ | |
| enum FormatState { | |
| 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; | |
| } | |
| 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); | size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value); |
| else | break; |
| size=snprintf(local_buf, sizeof(local_buf), fmt, value); | case FormatUInt: |
| else | size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value); |
| size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value); | break; |
| case FormatInvalid: | |
| return pa_strdup(local_buf, size); | 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) { |
| Line 546 size_t stdout_write(const void *buf, siz | Line 766 size_t stdout_write(const void *buf, siz |
| #endif | #endif |
| } | } |
| char* unescape_chars(const char* cp, int len) { | enum EscapeState { |
| char* s=new(PointerFreeGC) char[len + 1]; | EscapeRest, |
| enum EscapeState { | EscapeFirst, |
| EscapeRest, | EscapeSecond, |
| EscapeFirst, | EscapeUnicode |
| EscapeSecond | }; |
| } escapeState=EscapeRest; | |
| // @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; | uint escapedValue=0; |
| int srcPos=0; | int srcPos=0; |
| int dstPos=0; | short int jsCnt=0; |
| while(srcPos < len) { | while(srcPos<len){ |
| uchar ch=(uchar)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(c=='+' && !js){ |
| } 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++]=(char)escapedValue; | |
| escapeState=EscapeRest; | |
| break; | |
| } | } |
| srcPos++; | |
| srcPos++; | |
| } | } |
| s[dstPos]=0; | |
| *dst=0; // zero-termination | |
| return s; | return s; |
| } | } |
| char *search_stop(char*& current, char cstop_at) { | |
| // sanity check | |
| if(!current) | |
| return 0; | |
| // skip leading WS | |
| while(*current==' ' || *current=='\t') | |
| current++; | |
| if(!*current) | |
| return current=0; | |
| char *result=current; | |
| if(char *pstop_at=strchr(current, cstop_at)) { | |
| *pstop_at=0; | |
| current=pstop_at+1; | |
| } else | |
| current=0; | |
| return result; | |
| } | |
| #ifdef WIN32 | #ifdef WIN32 |
| void back_slashes_to_slashes(char* s) { | void back_slashes_to_slashes(char* s) { |
| if(s) | if(s) |
| Line 591 void back_slashes_to_slashes(char* s) { | Line 867 void back_slashes_to_slashes(char* s) { |
| if(*s=='\\') | if(*s=='\\') |
| *s='/'; | *s='/'; |
| } | } |
| /* | |
| void slashes_to_back_slashes(char* s) { | |
| if(s) | |
| for(; *s; s++) | |
| if(*s=='/') | |
| *s='\\'; | |
| } | |
| */ | |
| #endif | #endif |
| bool StrEqNc(const char* s1, const char* s2, bool strict) { | size_t strpos(const char *str, const char *substr) { |
| while(true) { | const char *p = strstr(str, substr); |
| if(!(*s1)) { | return (p==0)?STRING_NOT_FOUND:p-str; |
| if(!(*s2)) | } |
| return true; | |
| else | int remove_crlf(char* start, char* end) { |
| return !strict; | char* from=start; |
| } else if(!(*s2)) | char* to=start; |
| return !strict; | bool skip=false; |
| if(isalpha((unsigned char)*s1)) { | while(from < end){ |
| if(tolower((unsigned char)*s1) !=tolower((unsigned char)*s2)) | switch(*from){ |
| return false; | case '\n': |
| } else if((*s1) !=(*s2)) | case '\r': |
| return false; | case '\t': |
| s1++; | case ' ': |
| s2++; | if(!skip){ |
| } | *to=' '; |
| } | to++; |
| skip=true; | |
| static bool isLeap(int year) { | } |
| return !( | break; |
| (year % 4) || ((year % 400) && !(year % 100)) | default: |
| ); | if(from != to) |
| } | *to=*from; |
| to++; | |
| int getMonthDays(int year, int month) { | skip=false; |
| int monthDays[]={ | |
| 31, | |
| isLeap(year) ? 29 : 28, | |
| 31, | |
| 30, | |
| 31, | |
| 30, | |
| 31, | |
| 31, | |
| 30, | |
| 31, | |
| 30, | |
| 31 | |
| }; | |
| return monthDays[month]; | |
| } | |
| void remove_crlf(char* start, char* end) { | |
| for(char* p=start; p<end; p++) | |
| switch(*p) { | |
| case '\n': *p='|'; break; | |
| case '\r': *p=' '; break; | |
| } | } |
| from++; | |
| } | |
| return to-start; | |
| } | } |
| const char* hex_digits="0123456789ABCDEF"; | |
| /// must be last in this file | const char* hex_string(unsigned char* bytes, size_t size, bool upcase) { |
| #undef vsnprintf | char *bytes_hex=new(PointerFreeGC) char [size*2/*byte->hh*/+1/*for zero-teminator*/]; |
| int __vsnprintf(char* b, size_t s, const char* f, va_list l) { | unsigned char *src=bytes; |
| if(!s) | unsigned char *end=bytes+size; |
| return 0; | char *dest=bytes_hex; |
| 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 | const char *hex=upcase? hex_digits : "0123456789abcdef"; |
| if((ssize_t)s<0) | |
| return 0; | |
| #if _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 | for(; src<end; src++) { |
| formatted, that is, the number of characters that would have | *dest++=hex[*src/0x10]; |
| been written to the buffer if it were large enough. If the | *dest++=hex[*src%0x10]; |
| value of n is 0 on a call to snprintf(), an unspecified | } |
| value less than 1 is returned. | *dest=0; |
| */ | |
| 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, ...) { | return bytes_hex; |
| 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 */ | /* mime64 functions are from libgmime[http://spruce.sourceforge.net/gmime/] lib */ |
| Line 730 int __snprintf(char* b, size_t s, const | Line 942 int __snprintf(char* b, size_t s, const |
| * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA. | * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA. |
| * | * |
| */ | */ |
| static char *base64_alphabet = | static const char *base64_alphabet = |
| "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; | "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; |
| /** | /** |
| Line 748 static char *base64_alphabet = | Line 960 static char *base64_alphabet = |
| * | * |
| * Returns the number of bytes encoded. | * Returns the number of bytes encoded. |
| **/ | **/ |
| #define BASE64_GROUPS_IN_LINE 19 | |
| static size_t | static size_t |
| g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) | g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) |
| { | { |
| Line 785 g_mime_utils_base64_encode_step (const u | Line 1000 g_mime_utils_base64_encode_step (const u |
| *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)]; | *outptr++ = base64_alphabet [((c2 & 0x0f) << 2) | (c3 >> 6)]; |
| *outptr++ = base64_alphabet [c3 & 0x3f]; | *outptr++ = base64_alphabet [c3 & 0x3f]; |
| /* this is a bit ugly ... */ | /* this is a bit ugly ... */ |
| if ((++already) >= 19) { | if ((++already) >= BASE64_GROUPS_IN_LINE) { |
| *outptr++ = '\n'; | *outptr++ = '\n'; |
| already = 0; | already = 0; |
| } | } |
| Line 869 g_mime_utils_base64_encode_close (const | Line 1084 g_mime_utils_base64_encode_close (const |
| } | } |
| static unsigned char gmime_base64_rank[256] = { | static unsigned char gmime_base64_rank[256] = { |
| 255,255,255,255,255,255,255,255,255,254,254,255,255,254,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | 254,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63, |
| 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63, | |
| 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255, | 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, | 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, | 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255, |
| Line 894 static unsigned char gmime_base64_rank[2 | Line 1109 static unsigned char gmime_base64_rank[2 |
| * @out: output stream | * @out: output stream |
| * @state: holds the number of bits that are stored in @save | * @state: holds the number of bits that are stored in @save |
| * @save: leftover bits that have not yet been decoded | * @save: leftover bits that have not yet been decoded |
| * @strict: only base64 and whitespace chars are allowed | |
| * | * |
| * Decodes a chunk of base64 encoded data. | * Decodes a chunk of base64 encoded data. |
| * | * |
| * Returns the number of bytes decoded (which have been dumped in @out). | * Returns the number of bytes decoded (which have been dumped in @out). |
| **/ | **/ |
| size_t | size_t |
| g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) | g_mime_utils_base64_decode_step(const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save, bool strict=false) |
| { | { |
| const unsigned char *inptr; | const unsigned char *inptr; |
| unsigned char *outptr; | unsigned char *outptr; |
| Line 918 g_mime_utils_base64_decode_step (const u | Line 1134 g_mime_utils_base64_decode_step (const u |
| inptr = in; | inptr = in; |
| while (inptr < inend) { | while (inptr < inend) { |
| c = gmime_base64_rank[*inptr++]; | c = gmime_base64_rank[*inptr++]; |
| if (c != 0xff) { | switch(c) { |
| saved = (saved << 6) | c; | case 0xff: // non-base64 and non-whitespace chars. not allowed in strict mode |
| i++; | if(strict) |
| if (i == 4) { | throw Exception(BASE64_FORMAT, 0, "Invalid base64 char on position %d is detected", inptr-in-1); |
| *outptr++ = (unsigned char)(saved >> 16); | case 0xfe: // whitespace chars 0x09, 0x0A, 0x0D, 0x20 are allowed in any mode |
| *outptr++ = (unsigned char)(saved >> 8); | break; |
| *outptr++ = (unsigned char)(saved); | default: |
| i = 0; | saved = (saved << 6) | c; |
| } | i++; |
| if (i == 4) { | |
| *outptr++ = (unsigned char)(saved >> 16); | |
| *outptr++ = (unsigned char)(saved >> 8); | |
| *outptr++ = (unsigned char)(saved); | |
| i = 0; | |
| } | |
| } | } |
| } | } |
| Line 938 g_mime_utils_base64_decode_step (const u | Line 1160 g_mime_utils_base64_decode_step (const u |
| i = 2; | i = 2; |
| while (inptr > in && i) { | while (inptr > in && i) { |
| inptr--; | inptr--; |
| if (gmime_base64_rank[*inptr] != 0xff) { | if (gmime_base64_rank[*inptr] <= 0xfe) { |
| if (*inptr == '=' && outptr > out) | if (*inptr == '=' && outptr > out) |
| outptr--; | outptr--; |
| i--; | i--; |
| Line 950 g_mime_utils_base64_decode_step (const u | Line 1172 g_mime_utils_base64_decode_step (const u |
| } | } |
| char* pa_base64_encode(const char *in, size_t in_size) | char* pa_base64_encode(const char *in, size_t in_size){ |
| { | size_t new_size = ((in_size / 3 + 1) * 4); |
| /* wont go to more than 2x size (overly conservative) */ | new_size += new_size / (BASE64_GROUPS_IN_LINE * 4)/*new lines*/ + 1/*zero terminator*/; |
| char* result=new(PointerFreeGC) char[in_size * 2 + 6]; | char* result = new(PointerFreeGC) char[new_size]; |
| int state=0; | int state=0; |
| int save=0; | int save=0; |
| #ifndef NDEBUG | #ifndef NDEBUG |
| size_t filled= | size_t filled= |
| #endif | #endif |
| g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size, | g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size, (unsigned char*)result, &state, &save); |
| (unsigned char*)result, &state, &save); | |
| assert(filled <= in_size * 2 + 6); | //throw Exception(PARSER_RUNTIME, 0, "%d %d %d", in_size, new_size, filled); |
| assert(filled <= new_size); | |
| return result; | return result; |
| } | } |
| void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size) | struct File_base64_action_info { |
| { | unsigned char** base64; |
| /* wont go to more than had (overly conservative) */ | }; |
| result=new(PointerFreeGC) char[in_size+1/*terminator*/]; | |
| static void file_base64_file_action(struct stat& finfo, int f, const String& file_spec, void *context) { | |
| if(finfo.st_size) { | |
| File_base64_action_info& info=*static_cast<File_base64_action_info *>(context); | |
| *info.base64=new(PointerFreeGC) unsigned char[check_file_size(finfo.st_size, file_spec) * 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 state=0; |
| int save=0; | int save=0; |
| result_size= | result_size=g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size, (unsigned char*)result, &state, &save, strict); |
| g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size, | assert(result_size <= new_size); |
| (unsigned char*)result, &state, &save); | |
| assert(result_size <= in_size); | |
| result[result_size]=0; // for text files | result[result_size]=0; // for text files |
| if(strict && state!=0) | |
| throw Exception(BASE64_FORMAT, 0, "Unexpected end of chars"); | |
| } | } |
| const unsigned long pa_crc32(const char *in, size_t in_size) | 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() | |
| { | { |
| unsigned long crc32=0xFFFFFFFF; | if(crc32Table[1] == 0){ |
| if(in_size){ | // This is the official polynomial used by CRC32 in PKZip. |
| InitCrc32Table(); | // Often times the polynomial shown reversed as 0x04C11DB7. |
| for(size_t i = 0; i < in_size; i++) CalcCrc32(in[i], crc32); | 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; | |
| } | |
| } | } |
| return ~crc32; | |
| } | } |
| const unsigned long pa_crc32(const String& file_spec) | inline void CalcCrc32(const unsigned char byte, unsigned long &crc32) |
| { | { |
| crc32 = ((crc32) >> 8) ^ crc32Table[(byte) ^ ((crc32) & 0x000000FF)]; | |
| } | |
| unsigned long pa_crc32(const char *in, size_t in_size){ | |
| unsigned long crc32=0xFFFFFFFF; | unsigned long crc32=0xFFFFFFFF; |
| file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32); | |
| InitCrc32Table(); | |
| for(size_t i = 0; i<in_size; i++) | |
| CalcCrc32(in[i], crc32); | |
| 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(); |
| size_t nCount=0; | int nCount=0; |
| do { | do { |
| char buffer[FILE_BUFFER_SIZE]; | unsigned char buffer[FILE_BUFFER_SIZE]; |
| nCount = read(f, buffer, sizeof(buffer)); | nCount = file_block_read(f, buffer, sizeof(buffer)); |
| for(size_t i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32); | for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32); |
| } while(nCount); | } while(nCount > 0); |
| } | |
| } | |
| 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=str_upper(content_type); | |
| 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 ? &pa_charsets.get_direct(begin) : 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); | |
| *result_end='\0'; | |
| if(!asked_charset.isUTF8()) | |
| result = (char *)Charset::transcode(result, pa_UTF8_charset, asked_charset).cstr(); | |
| return result; | |
| } | |
| /// 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; | |
| } | |