--- parser3/src/main/pa_common.C 2001/03/22 12:09:59 1.21 +++ parser3/src/main/pa_common.C 2001/10/08 09:04:08 1.69 @@ -2,29 +2,11 @@ Parser: commonly functions. Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com) - Author: Alexander Petrosyan (http://design.ru/paf) - $Id: pa_common.C,v 1.21 2001/03/22 12:09:59 paf Exp $ + $Id: pa_common.C,v 1.69 2001/10/08 09:04:08 parser Exp $ */ -#ifdef HAVE_CONFIG_H -# include "pa_config.h" -#endif - -#include -#include -#include -#include -#include -#include -#include - -#ifdef WIN32 -#else -# include -#endif - #include "pa_common.h" #include "pa_types.h" #include "pa_exception.h" @@ -34,7 +16,16 @@ #include "pa_hash.h" #include "pa_string.h" -#ifdef WIN32 +#ifndef WIN32 +# ifndef _O_TEXT +# define _O_TEXT 0 +# endif +# ifndef _O_BINARY +# define _O_BINARY 0 +# endif +#endif + +#if _MSC_VER int __vsnprintf(char *b, size_t s, const char *f, va_list l) { int r=_vsnprintf(b, --s, f, l); @@ -52,75 +43,198 @@ int __snprintf(char *b, size_t s, const #endif -char *file_read_text(Pool& pool, const char *fname, bool fail_on_read_problem) { - int f; +char *file_read_text(Pool& pool, const String& file_spec, bool fail_on_read_problem) { + void *result; + size_t size; + return file_read(pool, file_spec, result, size, true, + fail_on_read_problem)?(char *)result:0; +} +bool file_read(Pool& pool, const String& file_spec, + void*& data, size_t& read_size, bool as_text, + bool fail_on_read_problem, + size_t offset, size_t limit) { + const char *fname=file_spec.cstr(String::UL_FILE_SPEC); +//printf("file_read(%s)\n", fname); + int f; struct stat finfo; - if(fname && !stat(fname,&finfo) &&(f=open(fname,O_RDONLY -#ifdef WIN32 - |O_TEXT -#endif - ))>=0) { + + // first open, next stat: + // directory update of NTFS hard links performed on open. + // ex: + // a.html:^test[] and b.html hardlink to a.html + // user inserts ! before ^test in a.html + // directory entry of b.html in NTFS not updated at once, + // they delay update till open, so we would receive "!^test[" string + // if would do stat, next open. + if( + (f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0 && + stat(fname, &finfo)==0) { /*if(exclusive) flock(f, LOCK_EX);*/ - - char *result=(char *)pool.malloc(finfo.st_size+1); - int read_size=read(f,result,finfo.st_size); - if(read_size>=0 && read_size<=finfo.st_size) - result[read_size]='\0'; + size_t max_size=limit?min(offset+limit, finfo.st_size)-offset:finfo.st_size; + if(!max_size) { // eof + data=0; + read_size=0; + } else { + data=pool.malloc(max_size+(as_text?1:0)); + if(offset) + lseek(f, offset, SEEK_SET); + read_size=read(f, data, max_size); + } /*if(exclusive) flock(f, LOCK_UN);*/ close(f); - return result;//prepare_config(result, remove_empty_lines); + if(!max_size) // eof + return true; + + if(read_size>=0 && read_size<=max_size) { + if(as_text) + ((char*&)data)[read_size]=0; + } else + PTHROW(0, 0, + &file_spec, + "read failed: actually read %d bytes count not in [0..%lu] valid range", + read_size, (unsigned long)max_size); //never + + return true;//prepare_config(result, remove_empty_lines); } if(fail_on_read_problem) - PTHROW(0,0, - 0, - "can not read '%s' file", fname); - return 0; + PTHROW(0, 0, + &file_spec, + "read failed: %s (%d), actual filename '%s'", + strerror(errno), errno, fname); + return false; +} + +static void create_dir_for_file(const String& file_spec) { + size_t pos_after=1; + int pos_before; + while((pos_before=file_spec.pos("/", 1, pos_after))>=0) { + mkdir(file_spec.mid(0, pos_before).cstr(String::UL_FILE_SPEC), 0775); + pos_after=pos_before+1; + } } void file_write(Pool& pool, - const char *fname, - const char *data, size_t size, - bool as_text/*, + const String& file_spec, + const void *data, size_t size, + bool as_text/*, bool exclusive*/) { - if(fname) { - int f; - if(access(fname, F_OK)!=0) {/*no*/ - if((f=open(fname,O_WRONLY|O_CREAT|_O_BINARY,0666))>0) - close(f); - } - if(access(fname, R_OK|W_OK)==0) { - int mode=O_RDWR + const char *fname=file_spec.cstr(String::UL_FILE_SPEC); + int f; + if(access(fname, W_OK)!=0) {/*no*/ + create_dir_for_file(file_spec); + + if((f=open(fname, O_WRONLY|O_CREAT|_O_BINARY, 0666))>0) + close(f); + } + if(access(fname, R_OK|W_OK)==0) { + int mode=O_RDWR|(as_text?_O_TEXT:_O_BINARY) #ifdef WIN32 - |O_TRUNC + |O_TRUNC #endif - ; - mode|=as_text?_O_TEXT:_O_BINARY; - if((f=open(fname,mode,0666))>=0) { - /*if(exclusive) - flock(f, LOCK_EX);*/ - - if(size) write(f,data,size); + ; + if((f=open(fname, mode, 0666))>=0) { + /*if(exclusive) + flock(f, LOCK_EX);*/ + + if(size) write(f, data, size); #ifndef WIN32 - ftruncate(f,size); + ftruncate(f, size); #endif - /*if(exclusive) - flock(f, LOCK_UN);*/ - close(f); - return; - } + /*if(exclusive) + flock(f, LOCK_UN);*/ + close(f); + return; } } - if(fname) - PTHROW(0, 0, - 0, - "file_write('%s'): %s (#%d)", - fname, strerror(errno), errno); - else - PTHROW(0, 0, - 0, - "file_write: no filename specified"); + PTHROW(0, 0, + &file_spec, + "write failed: %s (%d), actual filename '%s'", + strerror(errno), errno, fname); +} + +// throws nothing! [this is required in file_move & file_delete] +static void rmdir(const String& file_spec, size_t pos_after) { + int pos_before; + if((pos_before=file_spec.pos("/", 1, pos_after))>=0) + rmdir(file_spec, pos_before+1); + + rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::UL_FILE_SPEC)); +} +void file_delete(Pool& pool, const String& file_spec) { + const char *fname=file_spec.cstr(String::UL_FILE_SPEC); + if(unlink(fname)!=0) + PTHROW(0, 0, + &file_spec, + "unlink failed: %s (%d), actual filename '%s'", + strerror(errno), errno, fname); + + rmdir(file_spec, 1); +} +void file_move(Pool& pool, const String& old_spec, const String& new_spec) { + const char *old_spec_cstr=old_spec.cstr(String::UL_FILE_SPEC); + const char *new_spec_cstr=new_spec.cstr(String::UL_FILE_SPEC); + + create_dir_for_file(new_spec); + + if(rename(old_spec_cstr, new_spec_cstr)!=0) + PTHROW(0, 0, + &old_spec, + "rename failed: %s (%d), actual filename '%s' to '%s'", + strerror(errno), errno, old_spec_cstr, new_spec_cstr); + + rmdir(old_spec, 1); +} + + +static bool entry_readable(const String& file_spec, bool need_dir) { + const char *fname=file_spec.cstr(String::UL_FILE_SPEC); + struct stat finfo; + if(access(fname, R_OK)==0 && stat(fname, &finfo)==0) { + bool is_dir=(bool)(finfo.st_mode&S_IFDIR); + return is_dir==need_dir; + } + return false; +} +bool file_readable(const String& file_spec) { + return entry_readable(file_spec, false); +} +bool dir_readable(const String& file_spec) { + return entry_readable(file_spec, true); +} +String *file_readable(const String& path, const String& name) { + String *result=new(path.pool()) String(path); + *result << "/"; + *result << name; + return file_readable(*result)?result:0; +} +bool file_executable(const String& file_spec) { + return access(file_spec.cstr(String::UL_FILE_SPEC), X_OK)==0; +} + +bool file_stat(const String& file_spec, + size_t& rsize, + time_t& ratime, + time_t& rmtime, + time_t& rctime, + bool fail_on_read_problem) { + Pool& pool=file_spec.pool(); + const char *fname=file_spec.cstr(String::UL_FILE_SPEC); + struct stat finfo; + if(stat(fname, &finfo)!=0) + if(fail_on_read_problem) + PTHROW(0, 0, + &file_spec, + "getting file size failed: %s (%d), real filename '%s'", + strerror(errno), errno, fname); + else + return false; + rsize=finfo.st_size; + ratime=finfo.st_atime; + rmtime=finfo.st_mtime; + rctime=finfo.st_ctime; + return true; } char *getrow(char **row_ref, char delim) { @@ -135,9 +249,9 @@ char *getrow(char **row_ref, char delim) return result; } -char *lsplit(char *string_ref, char delim) { - if(string_ref) { - char *v=strchr(string_ref, delim); +char *lsplit(char *string, char delim) { + if(string) { + char *v=strchr(string, delim); if(v) { *v=0; return v+1; @@ -164,30 +278,31 @@ char *rsplit(char *string, char delim) { return NULL; } +/// @todo less stupid type detection char *format(Pool& pool, double value, char *fmt) { char *result=(char *)pool.malloc(MAX_NUMBER); if(fmt) if(strpbrk(fmt, "diouxX")) if(strpbrk(fmt, "ouxX")) - snprintf(result, MAX_NUMBER, fmt,(uint)value ); + snprintf(result, MAX_NUMBER, fmt, (uint)value ); else - snprintf(result, MAX_NUMBER, fmt,(int)value ); + snprintf(result, MAX_NUMBER, fmt, (int)value ); else snprintf(result, MAX_NUMBER, fmt, value); else - snprintf(result, MAX_NUMBER, "%d",(int)value); + snprintf(result, MAX_NUMBER, "%d", (int)value); return result; } -size_t stdout_write(const char *buf, size_t size) { +size_t stdout_write(const void *buf, size_t size) { #ifdef WIN32 do{ int chunk_written=fwrite(buf, 1, min(8*0x400, size), stdout); if(chunk_written<=0) break; size-=chunk_written; - buf+=chunk_written; + buf=((const char*)buf)+chunk_written; } while(size>0); return size; @@ -199,8 +314,8 @@ size_t stdout_write(const char *buf, siz const char *unescape_chars(Pool& pool, const char *cp, int len) { char *s=(char *)pool.malloc(len + 1); enum EscapeState { - EscapeRest, - EscapeFirst, + EscapeRest, + EscapeFirst, EscapeSecond } escapeState=EscapeRest; int escapedValue=0; @@ -234,29 +349,144 @@ const char *unescape_chars(Pool& pool, c return s; } +/// used by attributed_meaning_to_string / append_attribute_subattribute +struct Attributed_meaning_info { + String *header; // header line being constructed + String::Untaint_lang lang; // language in which to append to that line +}; static void append_attribute_subattribute(const Hash::Key& akey, Hash::Val *avalue, void *info) { if(akey==VALUE_NAME) return; + Attributed_meaning_info& ami=*static_cast(info); + // ...; charset=windows1251 - String *string=static_cast(info); - string->APPEND_CONST("; "); - string->append(akey, String::UL_HEADER, true); - string->APPEND_CONST("="); - string->append(static_cast(avalue)->as_string(), - String::UL_HEADER, true); + *ami.header << "; "; + ami.header->append(akey, ami.lang, true); + *ami.header << "="; + ami.header->append(static_cast(avalue)->as_string(), ami.lang, true); } -const String& attributed_meaning_to_string(Value& meaning) { +const String& attributed_meaning_to_string(Value& meaning, + String::Untaint_lang lang) { String &result=*new(meaning.pool()) String(meaning.pool()); if(Hash *hash=meaning.get_hash()) { // $value(value) $subattribute(subattribute value) if(Value *value=static_cast(hash->get(*value_name))) - result.append(value->as_string(), String::UL_HEADER, true); + result.append(value->as_string(), lang, true); - hash->foreach(append_attribute_subattribute, &result); + Attributed_meaning_info attributed_meaning_info={ + &result, + lang + }; + hash->for_each(append_attribute_subattribute, &attributed_meaning_info); } else // result value - result.append(meaning.as_string(), String::UL_HEADER, true); + result.append(meaning.as_string(), lang, true); + + return result; +} +#ifdef WIN32 +void back_slashes_to_slashes(char *s) { + if(s) + for(; *s; s++) + if(*s=='\\') + *s='/'; +} +/* +void slashes_to_back_slashes(char *s) { + if(s) + for(; *s; s++) + if(*s=='/') + *s='\\'; +} +*/ +#endif + +bool StrEqNc(const char *s1, const char *s2, bool strict) { + while(true) { + if(!(*s1)) { + if(!(*s2)) + return true; + else + return !strict; + } else if(!(*s2)) + return !strict; + if(isalpha(*s1)) { + if(tolower(*s1) !=tolower(*s2)) + return false; + } else if((*s1) !=(*s2)) + return false; + s1++; + s2++; + } +} + +char *unquote(char*& current, char stop_at) { + char *result=current; + char *dest=current; + // skip leading WS + while(*current==' ' || *current=='\t') + current++; + if(!*current) + return current=0; + + bool quoted=*current=='"'; + if(quoted) + current++; + + for(; *current; ) { + if(quoted) + switch(*current) { + case '\\': // "...\c + if(current[1]) + current++; + break; + case '"': + current++; + goto break2; + } + else + if(*current==stop_at) + break; + + *dest++=*current++; + } +break2: + if(*current) + current++; // skip 'stop_at' + *dest=0; return result; +} + +static int isLeap(int year) { + return !( + (year % 4) || ((year % 400) && !(year % 100)) + ); +} + +int getMonthDays(int year, int month) { + 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