|
|
| version 1.271, 2013/07/04 10:27:49 | version 1.301, 2017/02/06 16:17:12 |
|---|---|
| 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) 2000-2015 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 36 | Line 36 |
| #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 | |
| #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; | 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 65 volatile const char * IDENT_PA_COMMON_C= | Line 79 volatile const char * IDENT_PA_COMMON_C= |
| const String file_status_name(FILE_STATUS_NAME); | const String file_status_name(FILE_STATUS_NAME); |
| // functions | // forwards |
| bool capitalized(const char* s){ | const UTF16* pa_utf16_encode(const char* in, Charset& source_charset); |
| 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){ | // functions |
| if(!s || capitalized(s)) | |
| return s; | |
| char* result=pa_strdup(s); | #ifdef _MSC_VER |
| if(result){ | |
| bool upper=true; | int pa_stat(const char *pathname, struct stat *buffer){ |
| for(char* c=result; *c; c++){ | const UTF16* utf16name=pa_utf16_encode(pathname, pa_thread_request().charsets.source()); |
| *c=upper ? (char)toupper((unsigned char)*c) : (char)tolower((unsigned char)*c); | return _wstat64((const wchar_t *)utf16name, buffer); |
| upper=strchr("-_ ", *c) != 0; | |
| } | |
| } | |
| return (const char*)result; | |
| } | } |
| void fix_line_breaks(char *str, size_t& length) { | int pa_open(const char *pathname, int flags, int mode){ |
| //_asm int 3; | const UTF16* utf16name=pa_utf16_encode(pathname, pa_thread_request().charsets.source()); |
| const char* const eob=str+length; | return _wopen((const wchar_t *)utf16name, flags, mode); |
| 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 | FILE *pa_fopen(const char *pathname, const char *mode){ |
| bol=eol+2; | const UTF16* utf16name=pa_utf16_encode(pathname, pa_thread_request().charsets.source()); |
| length--; | const UTF16* utf16mode=pa_utf16_encode(mode, pa_thread_request().charsets.source()); |
| } else // \r, not \n = Macintosh | return _wfopen((const wchar_t *)utf16name, (const wchar_t *)utf16mode); |
| bol=eol+1; | |
| } | |
| // last piece without \r | |
| if(dest!=bol) | |
| memmove(dest, bol, eob-bol); | |
| str[length]=0; // terminating | |
| } | } |
| #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 151 struct File_read_action_info { | Line 138 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 187 File_read_result file_read(Request_chars | Line 173 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 228 File_read_result file_load(Request& r, c | Line 207 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 250 File_read_result file_load(Request& r, c | Line 227 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 284 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(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 326 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).taint_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; |
| } | } |
| } | } |
| Line 345 bool file_write_action_under_lock( | Line 311 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 363 bool file_write_action_under_lock( | Line 326 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 383 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 |
| } | } |
| Line 403 static void file_write_action(int f, voi | Line 363 static void file_write_action(int f, voi |
| ssize_t written=write(f, info.str, info.length); | ssize_t written=write(f, info.str, info.length); |
| if(written<0) | if(written<0) |
| throw Exception("file.write", 0, "write failed: %s (%d)", strerror(errno), errno); | throw Exception("file.write", 0, "write failed: %s (%d)", strerror(errno), errno); |
| if(written!=info.length) | if((size_t)written!=info.length) |
| throw Exception("file.write", 0, "write failed: %u of %u bytes written", written, info.length); | throw Exception("file.write", 0, "write failed: %u of %u bytes written", written, info.length); |
| } | } |
| } | } |
| Line 473 static bool entry_readable(const String& | Line 433 static bool entry_readable(const String& |
| } | } |
| // 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=0) { | static void rmdir(const String& file_spec, size_t pos_after) { |
| char* dir_spec=file_spec.taint_cstrm(String::L_FILE_SPEC); | char* dir_spec=file_spec.taint_cstrm(String::L_FILE_SPEC); |
| size_t length=strlen(dir_spec); | size_t length=strlen(dir_spec); |
| while( (length=get_dir(dir_spec, length)) && (length > pos_after) ){ | while( (length=get_dir(dir_spec, length)) && (length > pos_after) ){ |
| #ifdef WIN32 | #ifdef _MSC_VER |
| if(!entry_readable(dir_spec, true)) | if(!entry_readable(dir_spec, true)) |
| break; | break; |
| DWORD attrs=GetFileAttributes(dir_spec); | DWORD attrs=GetFileAttributes(dir_spec); |
| Line 495 static void rmdir(const String& file_spe | Line 455 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 517 void file_move(const String& old_spec, c | Line 476 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 529 void file_move(const String& old_spec, c | Line 486 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 560 bool file_executable(const String& file_ | Line 517 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 | |
| */ | |
| 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) { |
| Line 634 enum FormatType { | Line 666 enum FormatType { |
| FormatUInt, | FormatUInt, |
| FormatDouble | FormatDouble |
| }; | }; |
| FormatType format_type(char* fmt){ | FormatType format_type(const char* fmt){ |
| enum FormatState { | enum FormatState { |
| Percent, | Percent, |
| Flags, | Flags, |
| Line 645 FormatType format_type(char* fmt){ | Line 677 FormatType format_type(char* fmt){ |
| FormatType result=FormatInvalid; | FormatType result=FormatInvalid; |
| char* pos=fmt; | const char* pos=fmt; |
| while(char c=*(pos++)){ | while(char c=*(pos++)){ |
| switch(state){ | switch(state){ |
| case Percent: | case Percent: |
| Line 689 FormatType format_type(char* fmt){ | Line 721 FormatType format_type(char* fmt){ |
| } | } |
| const char* format(double value, char* fmt) { | const char* format(double value, const char* fmt) { |
| char local_buf[MAX_NUMBER]; | char local_buf[MAX_NUMBER]; |
| int size=-1; | int size=-1; |
| Line 705 const char* format(double value, char* f | Line 737 const char* format(double value, char* f |
| 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 occure 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 760 char* unescape_chars(const char* cp, int | Line 788 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 839 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 StrStartFromNC(const char* str, const char* substr, bool equal){ | |
| while(true) { | |
| if(!(*substr)){ | |
| if(!(*str)) | |
| return true; | |
| else | |
| return !equal; | |
| } | |
| if(!(*str)) | |
| return false; | |
| if(isalpha((unsigned char)*str)) { | |
| if(tolower((unsigned char)*str)!=tolower((unsigned char)*substr)) | |
| return false; | |
| } else if((*str) != (*substr)) | |
| return false; | |
| str++; | |
| substr++; | |
| } | |
| } | |
| size_t strpos(const char *str, const char *substr) { | size_t strpos(const char *str, const char *substr) { |
| const char *p = strstr(str, substr); | const char *p = strstr(str, substr); |
| return (p==0)?STRING_NOT_FOUND:p-str; | return (p==0)?STRING_NOT_FOUND:p-str; |
| } | } |
| // content-type: xxx; charset=WE-NEED-THIS | |
| // content-type: xxx; charset="WE-NEED-THIS" | |
| // content-type: xxx; charset="WE-NEED-THIS"; | |
| Charset* detect_charset(const char* content_type){ | |
| if(content_type){ | |
| char* CONTENT_TYPE=pa_strdup(content_type); | |
| for(char *p=CONTENT_TYPE; *p; p++) | |
| *p=(char)toupper((unsigned char)*p); | |
| if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){ | |
| begin+=8; // skip "CHARSET=" | |
| char* end=0; | |
| if(*begin && (*begin=='"' || *begin =='\'')){ | |
| char quote=*begin; | |
| begin++; | |
| end=(char*)strchr(begin, quote); | |
| } | |
| if(!end) | |
| end=(char*)strchr(begin, ';'); | |
| if(end) | |
| *end=0; // terminator | |
| return *begin?&charsets.get(begin):0; | |
| } | |
| } | |
| return 0; | |
| } | |
| static bool isLeap(int year) { | |
| return !( | |
| (year % 4) || ((year % 400) && !(year % 100)) | |
| ); | |
| } | |
| 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]; | |
| } | |
| int remove_crlf(char* start, char* end) { | int remove_crlf(char* start, char* end) { |
| char* from=start; | char* from=start; |
| char* to=start; | char* to=start; |
| Line 956 int remove_crlf(char* start, char* end) | Line 901 int remove_crlf(char* start, char* end) |
| return to-start; | 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 | |
| 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 | const char *hex=upcase? hex_digits : "0123456789abcdef"; |
| 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) | for(; src<end; src++) { |
| r=0; | *dest++=hex[*src/0x10]; |
| else if((size_t)r>s) | *dest++=hex[*src%0x10]; |
| r=s; | } |
| #endif | *dest=0; |
| 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 1281 char* pa_base64_encode(const char *in, s | Line 1189 char* pa_base64_encode(const char *in, s |
| return result; | return result; |
| } | } |
| struct File_base64_action_info { | |
| unsigned char** base64; | |
| }; | |
| char* pa_base64_encode(const String& file_spec){ | static void file_base64_file_action(struct stat& finfo, int f, const String& file_spec, void *context) { |
| 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; | |
| } | |
| static void file_base64_file_action( | |
| struct stat& finfo, | |
| int f, | |
| const String&, const char* /*fname*/, bool, | |
| void *context) { | |
| if(finfo.st_size) { | if(finfo.st_size) { |
| File_base64_action_info& info=*static_cast<File_base64_action_info *>(context); | File_base64_action_info& info=*static_cast<File_base64_action_info *>(context); |
| *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6]; | *info.base64=new(PointerFreeGC) unsigned char[check_file_size(finfo.st_size, file_spec) * 2 + 6]; |
| unsigned char* base64 = *info.base64; | unsigned char* base64 = *info.base64; |
| int state=0; | int state=0; |
| int save=0; | int save=0; |
| Line 1318 static void file_base64_file_action( | Line 1214 static void file_base64_file_action( |
| } | } |
| } | } |
| 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) { | 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 | // every 4 base64 bytes are converted into 3 normal bytes |
| // not full set (tail) of 4-bytes set is ignored | // not full set (tail) of 4-bytes set is ignored |
| Line 1326 void pa_base64_decode(const char *in, si | Line 1231 void pa_base64_decode(const char *in, si |
| 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, | |
| (unsigned char*)result, &state, &save, strict); | |
| assert(result_size <= new_size); | assert(result_size <= new_size); |
| result[result_size]=0; // for text files | result[result_size]=0; // for text files |
| Line 1340 void pa_base64_decode(const char *in, si | Line 1243 void pa_base64_decode(const char *in, si |
| 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; |
| } | } |
| const unsigned long pa_crc32(const char *in, size_t in_size){ | static unsigned long crc32Table[256]; |
| static void InitCrc32Table() | |
| { | |
| if(crc32Table[1] == 0){ | |
| // This is the official polynomial used by CRC32 in PKZip. | |
| // Often times the polynomial shown reversed as 0x04C11DB7. | |
| static const unsigned long dwPolynomial = 0xEDB88320; | |
| for(int i = 0; i < 256; i++) | |
| { | |
| unsigned long dwCrc = i; | |
| for(int j = 8; j > 0; j--) | |
| { | |
| if(dwCrc & 1) | |
| dwCrc = (dwCrc >> 1) ^ dwPolynomial; | |
| else | |
| dwCrc >>= 1; | |
| } | |
| crc32Table[i] = dwCrc; | |
| } | |
| } | |
| } | |
| inline void CalcCrc32(const unsigned char byte, unsigned long &crc32) | |
| { | |
| crc32 = ((crc32) >> 8) ^ crc32Table[(byte) ^ ((crc32) & 0x000000FF)]; | |
| } | |
| unsigned long pa_crc32(const char *in, size_t in_size){ | |
| unsigned long crc32=0xFFFFFFFF; | unsigned long crc32=0xFFFFFFFF; |
| InitCrc32Table(); | InitCrc32Table(); |
| Line 1356 const unsigned long pa_crc32(const char | Line 1286 const unsigned long pa_crc32(const char |
| return ~crc32; | return ~crc32; |
| } | } |
| const unsigned long pa_crc32(const String& file_spec){ | static void file_crc32_file_action(struct stat& finfo, int f, const String&, void *context) { |
| unsigned long crc32=0xFFFFFFFF; | |
| file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32); | |
| return ~crc32; | |
| } | |
| static void file_crc32_file_action( | |
| struct stat& finfo, | |
| int f, | |
| const String&, const char* /*fname*/, bool, | |
| void *context) { | |
| unsigned long& crc32=*static_cast<unsigned long *>(context); | unsigned long& crc32=*static_cast<unsigned long *>(context); |
| if(finfo.st_size) { | if(finfo.st_size) { |
| InitCrc32Table(); | InitCrc32Table(); |
| Line 1379 static void file_crc32_file_action( | Line 1299 static void file_crc32_file_action( |
| } | } |
| } | } |
| 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+1, 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()) | |
| result = (char *)Charset::transcode(result, pa_UTF8_charset, asked_charset).cstr(); | |
| return result; | |
| } | |
| 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; | |
| } | |