/*! @decl array(array(string)|string) tokenize(string code) *! *! Tokenize a string of Pike tokens. *! *! @returns *! Returns an array with Pike-level tokens and the remainder (a *! partial token), if any. */ static void f_tokenize( INT32 args ) { struct array *res; struct pike_string *left_s = NULL; /* Make gcc happy. */ struct pike_string *data; int left; ONERROR tmp; get_all_args("tokenize", args, "%W", &data); if(!data->len) { pop_n_elems(args); push_empty_array(); push_empty_string(); f_aggregate(2); return; } res = allocate_array_no_init( 0, 128 ); SET_ONERROR(tmp, do_free_arrayptr, &res); switch(data->size_shift) { case 0: left = tokenize0(&res, STR0(data), data->len); left_s = make_shared_binary_string0(STR0(data)+left, data->len-left); break; case 1: left = tokenize1(&res, STR1(data), data->len); left_s = make_shared_binary_string1(STR1(data)+left, data->len-left); break; case 2: left = tokenize2(&res,STR2(data), data->len); left_s = make_shared_binary_string2(STR2(data)+left, data->len-left); break; #ifdef PIKE_DEBUG default: Pike_error("Unknown shift size %d.\n", data->size_shift); #endif } UNSET_ONERROR(tmp); pop_n_elems(args); if (!res->size) { free_array(res); push_empty_array(); } else push_array(res); push_string( left_s ); f_aggregate( 2 ); }
/*! @decl int match(string str) *! *! Returns 1 if @[str] matches the regexp bound to the regexp object. *! Zero otherwise. *! *! @decl array(string) match(array(string) strs) *! *! Returns an array containing strings in @[strs] that match the *! regexp bound to the regexp object. *! *! @bugs *! The current implementation doesn't support searching *! in strings containing the NUL character or any *! wide character. *! *! @seealso *! @[split] */ static void regexp_match(INT32 args) { int i; struct regexp *regexp = THIS->regexp; if(args < 1) SIMPLE_TOO_FEW_ARGS_ERROR("Regexp.SimpleRegexp->match", 1); if(Pike_sp[-args].type == T_STRING) { if(Pike_sp[-args].u.string->size_shift) SIMPLE_BAD_ARG_ERROR("Regexp.SimpleRegexp->match", 1, "Expected string (8bit)"); i = pike_regexec(regexp, (char *)STR0(Pike_sp[-args].u.string)); pop_n_elems(args); push_int(i); return; } else if(Pike_sp[-args].type == T_ARRAY) { struct array *arr; int i, n; arr = Pike_sp[-args].u.array; for(i = n = 0; i < arr->size; i++) { struct svalue *sv = ITEM(arr) + i; if(sv->type != T_STRING || sv->u.string->size_shift) SIMPLE_BAD_ARG_ERROR("Regexp.SimpleRegexp->match", 1, "Expected string (8bit)"); if(pike_regexec(regexp, (char *)STR0(sv->u.string))) { ref_push_string(sv->u.string); n++; } } f_aggregate(n); stack_pop_n_elems_keep_top(args); return; } else SIMPLE_BAD_ARG_ERROR("Regexp.SimpleRegexp->match", 1, "string|array(string)"); }
static void image_ttf_face_names(INT32 args) { int n,i; int has[8]={0,0,0,0,0,0,0,0}; /* iso8859=20, unicode=30, any=1 */ char *hasname[8]={"copyright","family","style","full", "expose","version","postscript","trademark"}; struct array *a,*b; image_ttf_face__names(args); if (sp[-1].type!=T_ARRAY) Pike_error("Image.TTF.Face->names(): internal error, weird _names()\n"); a=sp[-1].u.array; n=0; for (i=0; i<a->size; i++) { int ihas=1; int what; b=a->item[i].u.array; what=b->item[3].u.integer; if (what>=8 || what<0) continue; /* weird */ switch (b->item[0].u.integer*100+b->item[1].u.integer) { case 301: /* M$: unicode */ case 300: /* M$: unicode (?) */ ihas=30; break; case 202: /* ISO: iso-8859-1 */ ihas=20; break; } if (ihas<has[what]) continue; /* worse */ push_text(hasname[what]); if (ihas==30) /* unicode, M$ but weird enough correct byteorder */ { ptrdiff_t n = b->item[4].u.string->len/2; struct pike_string *ps=begin_wide_shared_string(n,1); p_wchar1 *d=STR1(ps); p_wchar0 *s=STR0(b->item[4].u.string); while (n--) *(d++)=((p_wchar1)s[0]<<8)|(p_wchar1)s[1],s+=2; push_string(end_shared_string(ps)); } else push_svalue(b->item+4); n++; } f_aggregate_mapping(n*2); stack_swap(); pop_stack(); }
static void image_ilbm___decode(INT32 args) { unsigned char *s; ptrdiff_t len; struct pike_string *str; struct mapping *m; int n; extern void parse_iff(char *, unsigned char *, ptrdiff_t, struct mapping *, char *); get_all_args("__decode", args, "%S", &str); s = (unsigned char *)str->str; len = str->len; pop_n_elems(args-1); for(n=0; n<5; n++) push_int(0); push_mapping(m = allocate_mapping(4)); parse_iff("ILBM", s, len, m, "BODY"); mapping_index_no_free(sp-5, m, &string_[string_BMHD]); mapping_index_no_free(sp-4, m, &string_[string_CMAP]); mapping_index_no_free(sp-3, m, &string_[string_CAMG]); mapping_index_no_free(sp-2, m, &string_[string_BODY]); map_delete(m, &string_[string_BMHD]); map_delete(m, &string_[string_CMAP]); map_delete(m, &string_[string_CAMG]); map_delete(m, &string_[string_BODY]); if(sp[-5].type != T_STRING) Pike_error("Missing BMHD chunk\n"); if(sp[-2].type != T_STRING) Pike_error("Missing BODY chunk\n"); /* Extract image size from BMHD */ s = (unsigned char *)STR0(sp[-5].u.string); len = sp[-5].u.string->len; if(len<20) Pike_error("Short BMHD chunk\n"); free_svalue(sp-7); sp[-7].u.integer = (s[0]<<8)|s[1]; sp[-7].type = T_INT; sp[-7].subtype = NUMBER_NUMBER; sp[-6].u.integer = (s[2]<<8)|s[3]; sp[-6].type = T_INT; sp[-6].subtype = NUMBER_NUMBER; f_aggregate(7); }
/* **! method: string digest() **! Get the result of the hashing operation. **! name: digest, hexdigest - Return the resulting hash **! see_also: Mhash.to_hex **! returns: **! The resulting digest. */ void f_hash_digest(INT32 args) { int len, i; struct pike_string *res; len = get_digest(); res = begin_shared_string(len); for(i = 0; i < len; i++) { STR0(res)[i] = THIS->res[i]; } res = end_shared_string(res); pop_n_elems(args); push_string(res); }
/* php_caudium_set_header() sets a header in the header mapping. Called in a * thread safe manner from php_caudium_sapi_header_handler. */ INLINE static void php_caudium_set_header(char *header_name, char *value, char *p) { struct svalue hsval; struct pike_string *hval, *ind, *hind; struct mapping *headermap; struct svalue *s_headermap, *soldval; int vallen; GET_THIS(); /* hval = make_shared_string(value); */ ind = make_shared_string(" _headers"); hind = make_shared_binary_string(header_name, (int)(p - header_name)); s_headermap = low_mapping_string_lookup(REQUEST_DATA, ind); if(!s_headermap || s_headermap->type != PIKE_T_MAPPING) { struct svalue mappie; mappie.type = PIKE_T_MAPPING; headermap = allocate_mapping(1); mappie.u.mapping = headermap; mapping_string_insert(REQUEST_DATA, ind, &mappie); free_mapping(headermap); hval = make_shared_string(value); } else { headermap = s_headermap->u.mapping; soldval = low_mapping_string_lookup(headermap, hind); vallen = strlen(value); if(soldval != NULL && soldval->type == PIKE_T_STRING && soldval->u.string->size_shift == 0) { /* Existing, valid header. Prepend.*/ hval = begin_shared_string(soldval->u.string->len + 1 + vallen); MEMCPY(hval->str, soldval->u.string->str, soldval->u.string->len); STR0(hval)[soldval->u.string->len] = '\0'; MEMCPY(hval->str+soldval->u.string->len+1, value, vallen); hval = end_shared_string(hval); } else { hval = make_shared_string(value); } } hsval.type = PIKE_T_STRING; hsval.u.string = hval; mapping_string_insert(headermap, hind, &hsval); free_string(hval); free_string(ind); free_string(hind); }
static void f_read( INT32 args ) { char *read_buf; struct svalue *logfun, *file; FD f = -1; int cls, c, my_fd=1, state=0, tzs=0; char *char_pointer; INT32 v=0, yy=0, mm=0, dd=0, h=0, m=0, s=0, tz=0; ptrdiff_t offs0=0, len=0; struct svalue *old_sp; /* #define DYNAMIC_BUF */ #ifdef DYNAMIC_BUF dynamic_buffer buf; #else #define BUFSET(X) do { if(bufpos == bufsize) { bufsize *= 2; buf = realloc(buf, bufsize+1); } buf[bufpos++] = c; } while(0) #define PUSHBUF() do { push_string( make_shared_binary_string( buf,bufpos ) ); bufpos=0; } while(0) char *buf; int bufsize=CLF_BLOCK_SIZE, bufpos=0; #endif if(args>2 && sp[-1].type == T_INT) { offs0 = sp[-1].u.integer; pop_n_elems(1); --args; } old_sp = sp; get_all_args("CommonLog.read", args, "%*%*", &logfun, &file); if(logfun->type != T_FUNCTION) SIMPLE_BAD_ARG_ERROR("CommonLog.read", 1, "function"); if(file->type == T_OBJECT) { f = fd_from_object(file->u.object); if(f == -1) Pike_error("CommonLog.read: File is not open.\n"); my_fd = 0; } else if(file->type == T_STRING && file->u.string->size_shift == 0) { #ifdef PIKE_SECURITY if(!CHECK_SECURITY(SECURITY_BIT_SECURITY)) { if(!CHECK_SECURITY(SECURITY_BIT_CONDITIONAL_IO)) Pike_error("Permission denied.\n"); push_text("read"); push_int(0); ref_push_string(file->u.string); push_text("r"); push_int(00666); safe_apply(OBJ2CREDS(CURRENT_CREDS)->user,"valid_open",5); switch(Pike_sp[-1].type) { case PIKE_T_INT: switch(Pike_sp[-1].u.integer) { case 0: /* return 0 */ errno=EPERM; Pike_error("CommonLog.read(): Failed to open file for reading (errno=%d).\n", errno); case 2: /* ok */ pop_stack(); break; case 3: /* permission denied */ Pike_error("CommonLog.read: permission denied.\n"); default: Pike_error("Error in user->valid_open, wrong return value.\n"); } break; default: Pike_error("Error in user->valid_open, wrong return type.\n"); case PIKE_T_STRING: /* if(Pike_sp[-1].u.string->shift_size) */ /* file=Pike_sp[-1]; */ pop_stack(); } } #endif do { THREADS_ALLOW(); f=fd_open((char *)STR0(file->u.string), fd_RDONLY, 0); THREADS_DISALLOW(); if (f >= 0 || errno != EINTR) break; check_threads_etc(); } while (1); if(f < 0) Pike_error("CommonLog.read(): Failed to open file for reading (errno=%d).\n", errno); } else SIMPLE_BAD_ARG_ERROR("CommonLog.read", 2, "string|Stdio.File"); #ifdef HAVE_LSEEK64 lseek64(f, offs0, SEEK_SET); #else fd_lseek(f, offs0, SEEK_SET); #endif read_buf = malloc(CLF_BLOCK_SIZE+1); #ifndef DYNAMIC_BUF buf = malloc(bufsize); #endif while(1) { do { THREADS_ALLOW(); len = fd_read(f, read_buf, CLF_BLOCK_SIZE); THREADS_DISALLOW(); if (len >= 0 || errno != EINTR) break; check_threads_etc(); } while (1); if(len == 0) break; /* nothing more to read. */ if(len < 0) break; char_pointer = read_buf; while(len--) { offs0++; c = char_pointer[0] & 0xff; char_pointer ++; cls = char_class[c]; #ifdef TRACE_DFA fprintf(stderr, "DFA(%d): '%c' ", state, (c<32? '.':c)); switch(cls) { case CLS_WSPACE: fprintf(stderr, "CLS_WSPACE"); break; case CLS_CRLF: fprintf(stderr, "CLS_CRLF"); break; case CLS_TOKEN: fprintf(stderr, "CLS_TOKEN"); break; case CLS_DIGIT: fprintf(stderr, "CLS_DIGIT"); break; case CLS_QUOTE: fprintf(stderr, "CLS_QUOTE"); break; case CLS_LBRACK: fprintf(stderr, "CLS_LBRACK"); break; case CLS_RBRACK: fprintf(stderr, "CLS_RBRACK"); break; case CLS_SLASH: fprintf(stderr, "CLS_SLASH"); break; case CLS_COLON: fprintf(stderr, "CLS_COLON"); break; case CLS_HYPHEN: fprintf(stderr, "CLS_HYPHEN"); break; case CLS_PLUS: fprintf(stderr, "CLS_PLUS"); break; default: fprintf(stderr, "???"); } fprintf(stderr, " %d items on stack\n", sp-old_sp); #endif switch(state) { case 0: if(sp != old_sp) { if(sp == old_sp+15) { f_aggregate(15); push_int64(offs0); apply_svalue(logfun, 2); pop_stack(); } else pop_n_elems(sp-old_sp); } if(cls > CLS_CRLF) { if(cls == CLS_HYPHEN) { push_int(0); state = 2; break; } #ifdef DYNAMIC_BUF buf.s.str = NULL; initialize_buf( &buf ); low_my_putchar( c, &buf ); #else bufpos = 0; BUFSET(c); #endif state=1; } break; case 1: if(cls > CLS_CRLF) { #ifdef DYNAMIC_BUF low_my_putchar( c, &buf ); #else BUFSET(c); #endif break; } #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); /* remotehost */ #else PUSHBUF(); #endif state = (cls == CLS_WSPACE? 2:0); break; case 2: if(cls > CLS_CRLF) { if(cls == CLS_HYPHEN) { push_int(0); state = 4; break; } #ifdef DYNAMIC_BUF buf.s.str = NULL; initialize_buf( &buf ); low_my_putchar( c, &buf ); #else bufpos = 0; BUFSET(c); #endif state=3; } else if(cls == CLS_CRLF) state=0; break; case 3: if(cls > CLS_CRLF) { #ifdef DYNAMIC_BUF low_my_putchar( c, &buf ); #else BUFSET(c); #endif break; } #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); /* rfc931 */ #else PUSHBUF(); /* rfc931 */ #endif state = (cls == CLS_WSPACE? 4:0); break; case 4: if(cls > CLS_CRLF) { if(cls == CLS_HYPHEN) { push_int(0); state = 6; break; } #ifdef DYNAMIC_BUF buf.s.str = NULL; initialize_buf( &buf ); low_my_putchar( c, &buf ); #else bufpos = 0; BUFSET(c); #endif state=5; } else if(cls == CLS_CRLF) state=0; break; case 5: if(cls > CLS_CRLF) { #ifdef DYNAMIC_BUF low_my_putchar( c, &buf ); #else BUFSET(c); #endif break; } #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); /* authuser */ #else PUSHBUF(); /* authuser */ #endif state = (cls == CLS_WSPACE? 6:0); break; case 6: if(cls == CLS_LBRACK) state = 15; else if(cls == CLS_CRLF) state = 0; else if(cls == CLS_HYPHEN) { push_int(0); push_int(0); push_int(0); state = 7; } break; case 7: if(cls == CLS_QUOTE) { #ifdef DYNAMIC_BUF buf.s.str = NULL; initialize_buf( &buf ); #else bufpos = 0; #endif state = 31; } else if(cls == CLS_CRLF) state = 0; else if(cls == CLS_HYPHEN) { push_int(0); push_int(0); push_int(0); state = 10; } break; case 8: if(cls == CLS_QUOTE) state = 9; else if(cls == CLS_CRLF) { #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); #else PUSHBUF(); #endif state = 0; } else #ifdef DYNAMIC_BUF low_my_putchar( c, &buf ); #else BUFSET(c); #endif break; case 9: if(cls > CLS_CRLF) { #ifdef DYNAMIC_BUF low_my_putchar( '"', &buf); low_my_putchar( c, &buf); #else BUFSET('"'); BUFSET(c); #endif state = 8; break; } #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); /* protocol */ #else PUSHBUF(); /* protoocl */ #endif state = (cls == CLS_CRLF? 0 : 10); break; case 10: if(cls == CLS_DIGIT) { v = c&0xf; state = 11; } else if(cls == CLS_CRLF) state = 0; else if(cls == CLS_HYPHEN) { push_int(0); state = 12; } break; case 11: if(cls == CLS_DIGIT) v = v*10+(c&0xf); else if(cls == CLS_WSPACE) { push_int(v); /* status */ state = 12; } else state = 0; break; case 12: if(cls == CLS_DIGIT) { v = c&0xf; state = 13; } else if(cls == CLS_CRLF) state = 0; else if(cls == CLS_HYPHEN) { push_int(0); state = 14; } break; case 13: if(cls == CLS_DIGIT) v = v*10+(c&0xf); else { push_int(v); /* bytes */ state = (cls == CLS_CRLF? 0:14); } break; case 14: if(cls == CLS_CRLF) state = 0; break; case 15: if(cls == CLS_DIGIT) { dd = c&0xf; state = 16; } else state = (cls == CLS_CRLF? 0:14); break; case 16: /* getting day */ if(cls == CLS_DIGIT) dd = dd*10+(c&0xf); else if(cls == CLS_SLASH) state = 17; else state = (cls == CLS_CRLF? 0:14); break; case 17: if(cls == CLS_DIGIT) { mm = c&0xf; state = 18; } else if(cls == CLS_TOKEN) { mm = c|0x20; state = 21; } else state = (cls == CLS_CRLF? 0:14); break; case 18: /* getting numeric month */ if(cls == CLS_DIGIT) mm = mm*10+(c&0xf); else if(cls == CLS_SLASH) state = 19; else state = (cls == CLS_CRLF? 0:14); break; case 19: if(cls == CLS_DIGIT) { yy = c&0xf; state = 20; } else state = (cls == CLS_CRLF? 0:14); break; case 20: /* getting year */ if(cls == CLS_DIGIT) yy = yy*10+(c&0xf); else if(cls == CLS_COLON) state = 22; else state = (cls == CLS_CRLF? 0:14); break; case 21: /* getting textual month */ if(cls == CLS_TOKEN) mm = (mm<<8)|c|0x20; else if(cls == CLS_SLASH) { state = 19; switch(mm) { case ('j'<<16)|('a'<<8)|'n': mm=1; break; case ('f'<<16)|('e'<<8)|'b': mm=2; break; case ('m'<<16)|('a'<<8)|'r': mm=3; break; case ('a'<<16)|('p'<<8)|'r': mm=4; break; case ('m'<<16)|('a'<<8)|'y': mm=5; break; case ('j'<<16)|('u'<<8)|'n': mm=6; break; case ('j'<<16)|('u'<<8)|'l': mm=7; break; case ('a'<<16)|('u'<<8)|'g': mm=8; break; case ('s'<<16)|('e'<<8)|'p': mm=9; break; case ('o'<<16)|('c'<<8)|'t': mm=10; break; case ('n'<<16)|('o'<<8)|'v': mm=11; break; case ('d'<<16)|('e'<<8)|'c': mm=12; break; default: state = 14; } } break; case 22: if(cls == CLS_DIGIT) { h = c&0xf; state = 23; } else state = (cls == CLS_CRLF? 0:14); break; case 23: /* getting hour */ if(cls == CLS_DIGIT) h = h*10+(c&0xf); else if(cls == CLS_COLON) state = 24; else state = (cls == CLS_CRLF? 0:14); break; case 24: if(cls == CLS_DIGIT) { m = c&0xf; state = 25; } else state = (cls == CLS_CRLF? 0:14); break; case 25: /* getting minute */ if(cls == CLS_DIGIT) m = m*10+(c&0xf); else if(cls == CLS_COLON) state = 26; else state = (cls == CLS_CRLF? 0:14); break; case 26: if(cls == CLS_DIGIT) { s = c&0xf; state = 27; } else state = (cls == CLS_CRLF? 0:14); break; case 27: /* getting second */ if(cls == CLS_DIGIT) s = s*10+(c&0xf); else if(cls == CLS_WSPACE) state = 28; else state = (cls == CLS_CRLF? 0:14); break; case 28: if(cls>=CLS_HYPHEN) { state = 29; tzs = cls!=CLS_PLUS; tz = 0; } else if(cls == CLS_DIGIT) { state = 29; tzs = 0; tz = c&0xf; } else if(cls==CLS_CRLF) state = 0; break; case 29: /* getting timezone */ if(cls == CLS_DIGIT) tz = tz*10+(c&0xf); else { if(tzs) tz = -tz; push_int(yy); push_int(mm); push_int(dd); push_int(h); push_int(m); push_int(s); push_int(tz); if(cls == CLS_RBRACK) state = 7; else state = (cls == CLS_CRLF? 0 : 30); } break; case 30: if(cls == CLS_RBRACK) state = 7; else if(cls == CLS_CRLF) state = 0; break; case 31: if(cls == CLS_QUOTE) { #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); #else PUSHBUF(); #endif push_int(0); push_int(0); state = 10; } else if(cls >= CLS_TOKEN) #ifdef DYNAMIC_BUF low_my_putchar( c, &buf ); #else BUFSET(c); #endif else { #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); /* method */ #else PUSHBUF(); /* method */ #endif state = (cls == CLS_CRLF? 0 : 32); } break; case 32: if(cls == CLS_QUOTE) { push_int(0); push_int(0); state = 10; } else if(cls >= CLS_TOKEN) { #ifdef DYNAMIC_BUF buf.s.str = NULL; initialize_buf( &buf ); low_my_putchar( c, &buf ); #else bufpos = 0; BUFSET(c); #endif state = 33; } else if(cls == CLS_CRLF) state = 0; break; case 33: if(cls == CLS_QUOTE) state = 34; else if(cls == CLS_CRLF) { #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); #else PUSHBUF(); #endif state = 0; } else if(cls == CLS_WSPACE) { #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); /* path */ #else PUSHBUF(); /* path */ #endif state = 35; } else #ifdef DYNAMIC_BUF low_my_putchar( c, &buf ); #else BUFSET(c); #endif break; case 34: if(cls >= CLS_TOKEN) { #ifdef DYNAMIC_BUF low_my_putchar( '"', &buf ); low_my_putchar( c, &buf ); #else BUFSET('"'); BUFSET(c); #endif state = 33; } else if(cls == CLS_CRLF) { #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); #else PUSHBUF(); #endif state = 0; } else { #ifdef DYNAMIC_BUF push_string( low_free_buf( &buf ) ); #else PUSHBUF(); #endif push_text("HTTP/0.9"); state = 10; } break; case 35: if(cls == CLS_QUOTE) { push_text("HTTP/0.9"); state = 10; } else if(cls >= CLS_TOKEN) { #ifdef DYNAMIC_BUF buf.s.str = NULL; initialize_buf( &buf ); low_my_putchar( c, &buf ); #else bufpos = 0; BUFSET(c); #endif state = 8; } else if(cls == CLS_CRLF) state = 0; break; } } }
static void f_ultraparse( INT32 args ) { FD f = -1; int lines=0, cls, c=0, my_fd=1, tzs=0, state=0, next; unsigned char *char_pointer=0; /* array with offsets for fields in the string buffer */ int buf_points[16]; INT32 v=0, offs0=0, len=0, bytes=0, gotdate=0; INT32 last_hour=0, last_date=0, last_year=0, last_month=0, this_date=0, broken_lines=0, tmpinteger=0, field_position=0; time_t start; unsigned char *read_buf; struct svalue *statfun, *daily, *pagexts=0, *file, *refsval, *log_format; unsigned char *buf, *field_buf; #ifdef BROKEN_LINE_DEBUG INT32 broken_line_pos=0; unsigned char *broken_line; #endif INT32 *state_list, *save_field_num, *field_endings, num_states; char *notref = 0; INT32 state_pos=0, bufpos=0, i, fieldnum=0; struct pike_string *url_str = 0, *ref_str = 0, *rfc_str = 0, *hst_str = 0, *tmpagent = 0; struct svalue *url_sval; ONERROR unwind_protect; unsigned INT32 hits_per_hour[24]; unsigned INT32 hosts_per_hour[24]; unsigned INT32 pages_per_hour[24]; unsigned INT32 sessions_per_hour[24]; double kb_per_hour[24]; unsigned INT32 session_length[24]; /* struct mapping *unique_per_hour = allocate_mapping(1);*/ struct mapping *hits_per_error = allocate_mapping(10); struct mapping *error_urls = allocate_mapping(10); struct mapping *error_refs = allocate_mapping(10); struct mapping *user_agents = allocate_mapping(10); struct mapping *directories = allocate_mapping(20); struct mapping *referrers = allocate_mapping(1); struct mapping *refsites = allocate_mapping(1); struct mapping *referredto = allocate_mapping(1); struct mapping *pages = allocate_mapping(1); struct mapping *hosts = allocate_mapping(1); struct mapping *hits = allocate_mapping(1); struct mapping *session_start = allocate_mapping(1); struct mapping *session_end = allocate_mapping(1); struct mapping *hits20x = allocate_mapping(300); struct mapping *hits302 = allocate_mapping(2); struct mapping *sites = allocate_mapping(1); struct mapping *domains = allocate_mapping(1); struct mapping *topdomains = allocate_mapping(1); struct mapping *tmpdest = NULL; /* struct mapping *hits30x = allocate_mapping(2);*/ if(args>6 && sp[-1].type == T_INT) { offs0 = sp[-1].u.integer; pop_n_elems(1); --args; } if(args>5 && sp[-1].type == T_STRING) { notref = sp[-1].u.string->str; pop_n_elems(1); --args; } lmu = 0; get_all_args("UltraLog.ultraparse", args, "%*%*%*%*%*", &log_format, &statfun, &daily, &file, &pagexts); if(log_format->type != T_STRING) error("Bad argument 1 to Ultraparse.ultraparse, expected string.\n"); if(statfun->type != T_FUNCTION) error("Bad argument 2 to Ultraparse.ultraparse, expected function.\n"); if(daily->type != T_FUNCTION) error("Bad argument 3 to Ultraparse.ultraparse, expected function.\n"); if(pagexts->type != T_MULTISET) error("Bad argument 5 to Ultraparse.ultraparse, expected multiset.\n"); if(file->type == T_OBJECT) { f = fd_from_object(file->u.object); if(f == -1) error("UltraLog.ultraparse: File is not open.\n"); my_fd = 0; } else if(file->type == T_STRING && file->u.string->size_shift == 0) { do { f=fd_open(STR0(file->u.string), fd_RDONLY, 0); } while(f < 0 && errno == EINTR); if(errno < 0) error("UltraLog.ultraparse(): Failed to open file for reading (errno=%d).\n", errno); } else error("Bad argument 4 to UltraLog.ultraparse, expected string or object .\n"); state_list = malloc((log_format->u.string->len +3) * sizeof(INT32)); save_field_num = malloc((log_format->u.string->len +3) * sizeof(INT32)); field_endings = malloc((log_format->u.string->len +3) * sizeof(INT32)); num_states = parse_log_format(log_format->u.string, state_list, field_endings, save_field_num); if(num_states < 1) { free(state_list); free(save_field_num); free(field_endings); error("UltraLog.ultraparse(): Failed to parse log format.\n"); } fd_lseek(f, offs0, SEEK_SET); read_buf = malloc(READ_BLOCK_SIZE+1); buf = malloc(MAX_LINE_LEN+2); #ifdef BROKEN_LINE_DEBUG broken_line = malloc(MAX_LINE_LEN*10); #endif MEMSET(hits_per_hour, 0, sizeof(hits_per_hour)); MEMSET(hosts_per_hour, 0, sizeof(hosts_per_hour)); MEMSET(session_length, 0, sizeof(session_length)); MEMSET(pages_per_hour, 0, sizeof(pages_per_hour)); MEMSET(sessions_per_hour, 0, sizeof(sessions_per_hour)); MEMSET(kb_per_hour, 0, sizeof(kb_per_hour)); /*url_sval.u.type = TYPE_STRING;*/ BUFSET(0); field_position = bufpos; buf_points[0] = buf_points[1] = buf_points[2] = buf_points[3] = buf_points[4] = buf_points[5] = buf_points[6] = buf_points[7] = buf_points[8] = buf_points[9] = buf_points[10] = buf_points[11] = buf_points[12] = buf_points[13] = buf_points[14] = buf_points[15] = 0; while(1) { /* THREADS_ALLOW();*/ do { len = fd_read(f, read_buf, READ_BLOCK_SIZE); } while(len < 0 && errno == EINTR); /* THREADS_DISALLOW();*/ if(len <= 0) break; /* nothing more to read or error. */ offs0 += len; char_pointer = read_buf+len - 1; while(len--) { c = char_pointer[-len]; cls = char_class[c]; #if 0 fprintf(stdout, "DFA(%d:%d): '%c' (%d) ", state, state_pos, c, (int)c); switch(cls) { case CLS_WSPACE: fprintf(stdout, "CLS_WSPACE\n"); break; case CLS_CRLF: fprintf(stdout, "CLS_CRLF\n"); break; case CLS_TOKEN: fprintf(stdout, "CLS_TOKEN\n"); break; case CLS_DIGIT: fprintf(stdout, "CLS_DIGIT\n"); break; case CLS_QUOTE: fprintf(stdout, "CLS_QUOTE\n"); break; case CLS_LBRACK: fprintf(stdout, "CLS_LBRACK\n"); break; case CLS_RBRACK: fprintf(stdout, "CLS_RBRACK\n"); break; case CLS_SLASH: fprintf(stdout, "CLS_SLASH\n"); break; case CLS_COLON: fprintf(stdout, "CLS_COLON\n"); break; case CLS_HYPHEN: fprintf(stdout, "CLS_HYPHEN/CLS_MINUS\n"); break; case CLS_PLUS: fprintf(stdout, "CLS_PLUS\n"); break; default: fprintf(stdout, "??? %d ???\n", cls); } #endif #ifdef BROKEN_LINE_DEBUG broken_line[broken_line_pos++] = c; #endif if(cls == field_endings[state_pos]) { /* Field is done. Nullify. */ process_field: /* printf("Processing field %d of %d\n", state_pos, num_states);*/ switch(save_field_num[state_pos]) { case DATE: case HOUR: case MINUTE: case UP_SEC: case CODE: /* BUFSET(0);*/ tmpinteger = 0; for(v = field_position; v < bufpos; v++) { if(char_class[buf[v]] == CLS_DIGIT) tmpinteger = tmpinteger*10 + (buf[v]&0xf); else { goto skip; } } BUFPOINT = tmpinteger; break; case YEAR: tmpinteger = 0; for(v = field_position; v < bufpos; v++) { if(char_class[buf[v]] == CLS_DIGIT) tmpinteger = tmpinteger*10 + (buf[v]&0xf); else { goto skip; } } if(tmpinteger < 100) { if(tmpinteger < 60) tmpinteger += 2000; else tmpinteger += 1900; } BUFPOINT = tmpinteger; break; case BYTES: v = field_position; switch(char_class[buf[v++]]) { case CLS_QUESTION: case CLS_HYPHEN: if(v == bufpos) tmpinteger = 0; else { goto skip; } break; case CLS_DIGIT: tmpinteger = (buf[field_position]&0xf); for(; v < bufpos; v++) { if(char_class[buf[v]] == CLS_DIGIT) tmpinteger = tmpinteger*10 + (buf[v]&0xf); else { goto skip; } } /* printf("Digit: %d\n", tmpinteger);*/ break; default: goto skip; } BUFPOINT = tmpinteger; /* bufpos++;*/ break; case MONTH: /* Month */ /* BUFSET(0);*/ /* field_buf = buf + field_positions[state_pos];*/ switch(bufpos - field_position) { case 2: tmpinteger = 0; for(v = field_position; v < bufpos; v++) { if(char_class[buf[v]] == CLS_DIGIT) tmpinteger = tmpinteger*10 + (buf[v]&0xf); else { goto skip; } } break; case 3: switch(((buf[field_position]|0x20)<<16)|((buf[field_position+1]|0x20)<<8)| (buf[field_position+2]|0x20)) { case ('j'<<16)|('a'<<8)|'n': tmpinteger = 1; break; case ('f'<<16)|('e'<<8)|'b': tmpinteger = 2; break; case ('m'<<16)|('a'<<8)|'r': tmpinteger = 3; break; case ('a'<<16)|('p'<<8)|'r': tmpinteger = 4; break; case ('m'<<16)|('a'<<8)|'y': tmpinteger = 5; break; case ('j'<<16)|('u'<<8)|'n': tmpinteger = 6; break; case ('j'<<16)|('u'<<8)|'l': tmpinteger = 7; break; case ('a'<<16)|('u'<<8)|'g': tmpinteger = 8; break; case ('s'<<16)|('e'<<8)|'p': tmpinteger = 9; break; case ('o'<<16)|('c'<<8)|'t': tmpinteger = 10; break; case ('n'<<16)|('o'<<8)|'v': tmpinteger = 11; break; case ('d'<<16)|('e'<<8)|'c': tmpinteger = 12; break; } break; default: goto skip; } /*printf("Month: %0d\n", mm);*/ if(tmpinteger < 1 || tmpinteger > 12) goto skip; /* Broken Month */ BUFPOINT = tmpinteger; /* bufpos++;*/ break; case ADDR: case REFER: case AGENT: case TZ: case METHOD: case URL: case RFC: case PROTO: BUFSET(0); SETPOINT(); /* printf("Field %d, pos %d, %s\n", save_field_num[state_pos],BUFPOINT,*/ /* buf + BUFPOINT); */ break; } state_pos++; field_position = bufpos; if(cls != CLS_CRLF) continue; } else if(cls != CLS_CRLF) { BUFSET(c); continue; } else { /* printf("Processing last field (%d).\n", state_pos);*/ goto process_field; /* End of line - process what we got */ } /* printf("%d %d\n", state_pos, num_states);*/ /* buf_points[8] = buf_points[9] = buf_points[10] = buf_points[11] = buf;*/ /* buf_points[12] = buf_points[13] = buf_points[14] = buf_points[15] = buf;*/ #if 0 if(!((lines+broken_lines)%100000)) { push_int(lines+broken_lines); push_int((int)((float)offs0/1024.0/1024.0)); apply_svalue(statfun, 2); pop_stack(); /*printf("%5dk lines, %5d MB\n", lines/1000, (int)((float)offs0/1024.0/1024.0));*/ } #endif if(state_pos < num_states) { #ifdef BROKEN_LINE_DEBUG broken_line[broken_line_pos] = 0; printf("too few states (pos=%d): %s\n", state_pos, broken_line); #endif broken_lines++; goto ok; } #define yy buf_points[YEAR] #define mm buf_points[MONTH] #define dd buf_points[DATE] #define h buf_points[HOUR] #define m buf_points[MINUTE] #define s buf_points[UP_SEC] #define v buf_points[CODE] #define bytes buf_points[BYTES] this_date = (yy*10000) + (mm*100) + dd; if(!this_date) { broken_lines++; goto ok; } #if 1 if(!last_date) { /* First loop w/o a value.*/ last_date = this_date; last_hour = h; } else { if(last_hour != h || last_date != this_date) { pages_per_hour[last_hour] += hourly_page_hits(hits20x, pages, hits, pagexts->u.multiset, 200); /* pages_per_hour[last_hour] +=*/ /* hourly_page_hits(hits304, pages, hits, pagexts->u.multiset, 300);*/ /* printf("%5d %5d for %d %02d:00\n",*/ /* pages_per_hour[last_hour], hits_per_hour[last_hour],*/ /*last_date, last_hour);*/ if(m_sizeof(session_start)) { summarize_sessions(last_hour, sessions_per_hour, session_length, session_start, session_end); free_mapping(session_start); free_mapping(session_end); session_start = allocate_mapping(1); session_end = allocate_mapping(1); } hosts_per_hour[last_hour] += m_sizeof(sites); do_map_addition(hosts, sites); free_mapping(sites); sites = allocate_mapping(100); last_hour = h; free_mapping(hits20x); /* Reset this one */ /* free_mapping(hits304); Reset this one */ /* hits304 = allocate_mapping(2);*/ hits20x = allocate_mapping(2); } #if 1 if(last_date != this_date) { /* printf("%d %d\n", last_date, this_date);*/ tmpdest = allocate_mapping(1); summarize_refsites(refsites, referrers, tmpdest); free_mapping(referrers); referrers = tmpdest; tmpdest = allocate_mapping(1); clean_refto(referredto, tmpdest, pagexts->u.multiset); free_mapping(referredto); referredto = tmpdest; summarize_directories(directories, pages); summarize_directories(directories, hits); tmpdest = allocate_mapping(1); http_decode_mapping(user_agents, tmpdest); free_mapping(user_agents); user_agents = tmpdest; tmpdest = allocate_mapping(1); summarize_hosts(hosts, domains, topdomains, tmpdest); free_mapping(hosts); hosts = tmpdest; #if 1 push_int(last_date / 10000); push_int((last_date % 10000)/100); push_int((last_date % 10000)%100); push_mapping(pages); push_mapping(hits); push_mapping(hits302); push_mapping(hits_per_error); push_mapping(error_urls); push_mapping(error_refs); push_mapping(referredto); push_mapping(refsites); push_mapping(referrers); push_mapping(directories); push_mapping(user_agents); push_mapping(hosts); push_mapping(domains); push_mapping(topdomains); for(i = 0; i < 24; i++) { push_int(sessions_per_hour[i]); } f_aggregate(24); for(i = 0; i < 24; i++) { push_int(hits_per_hour[i]); hits_per_hour[i] = 0; } f_aggregate(24); for(i = 0; i < 24; i++) { push_int(pages_per_hour[i]); pages_per_hour[i] = 0; } f_aggregate(24); for(i = 0; i < 24; i++) { /* KB per hour.*/ push_float(kb_per_hour[i]); kb_per_hour[i] = 0.0; } f_aggregate(24); for(i = 0; i < 24; i++) { push_float(sessions_per_hour[i] ? ((float)session_length[i] / (float)sessions_per_hour[i]) / 60.0 : 0.0); sessions_per_hour[i] = 0; session_length[i] = 0; } f_aggregate(24); for(i = 0; i < 24; i++) { push_int(hosts_per_hour[i]); hosts_per_hour[i] = 0; } f_aggregate(24); apply_svalue(daily, 23); pop_stack(); #else free_mapping(error_refs); free_mapping(referredto); free_mapping(refsites); free_mapping(directories); free_mapping(error_urls); free_mapping(hits); free_mapping(hits_per_error); free_mapping(pages); free_mapping(hosts); free_mapping(domains); free_mapping(topdomains); free_mapping(referrers); free_mapping(hits302); #endif user_agents = allocate_mapping(10); hits302 = allocate_mapping(1); hits_per_error = allocate_mapping(10); error_urls = allocate_mapping(10); error_refs = allocate_mapping(10); directories = allocate_mapping(20); referrers = allocate_mapping(1); referredto = allocate_mapping(1); refsites = allocate_mapping(1); pages = allocate_mapping(1); hits = allocate_mapping(1); sites = allocate_mapping(1); hosts = allocate_mapping(1); domains = allocate_mapping(1); topdomains = allocate_mapping(1); last_date = this_date; } #endif } #endif #if 1 process_session(buf+buf_points[ADDR], h*3600+m*60+s, h, sessions_per_hour, session_length, session_start, session_end, sites); url_str = make_shared_binary_string(buf + buf_points[URL], strlen(buf + buf_points[URL])); #if 1 switch(v) { /* Do error-code specific logging. Error urls that are specially treated do not include auth required, service unavailable etc. They are only included in the return code summary. */ case 200: case 201: case 202: case 203: case 204: case 205: case 206: case 207: case 304: mapaddstr(hits20x, url_str); DO_REFERRER(); break; case 300: case 301: case 302: case 303: case 305: mapaddstr(hits302, url_str); DO_REFERRER(); break; case 400: case 404: case 405: case 406: case 408: case 409: case 410: case 411: case 412: case 413: case 414: case 415: case 416: case 500: case 501: DO_ERREF(); map2addint(error_urls, v, url_str); break; } /*rfc_str = http_decode_string(buf + buf_points[RFC]);*/ /*hst_str = make_shared_binary_string(buf, strlen(buf));*/ #endif free_string(url_str); mapaddint(hits_per_error, v); kb_per_hour[h] += (float)bytes / 1024.0; hits_per_hour[h]++; /*#endif*/ if(strlen(buf + buf_points[AGENT])>1) { /* Got User Agent */ tmpagent = make_shared_string(buf + buf_points[AGENT]); mapaddstr(user_agents, tmpagent); free_string(tmpagent); } #endif lines++; #if 0 printf("%s %s %s\n%s %s %s\n%04d-%02d-%02d %02d:%02d:%02d \n%d %d\n", buf + buf_points[ADDR], buf + buf_points[REFER], buf + buf_points[ RFC ], buf + buf_points[METHOD], buf + buf_points[ URL ], buf + buf_points[PROTO], yy, mm, dd, h, m, s, v, bytes); /* if(lines > 10) exit(0);*/ #endif ok: gotdate = /* v = bytes =h = m = s = tz = tzs = dd = mm = yy = */ buf_points[0] = buf_points[1] = buf_points[2] = buf_points[3] = buf_points[4] = buf_points[5] = buf_points[6] = buf_points[7] = /*buf_points[8] = buf_points[9] = buf_points[10] =*/ buf_points[11] = buf_points[12] = buf_points[13] = buf_points[14] = buf_points[15] = bufpos = state_pos = 0; field_position = 1; #ifdef BROKEN_LINE_DEBUG broken_line_pos = 0; #endif BUFSET(0); } } cleanup: free(save_field_num); free(state_list); free(field_endings); free(buf); push_int(lines); push_int((int)((float)offs0 / 1024.0/1024.0)); push_int(1); apply_svalue(statfun, 3); pop_stack(); free(read_buf); #ifdef BROKEN_LINE_DEBUG free(broken_line); #endif if(my_fd) /* If my_fd == 0, the second argument was an object and thus we don't * want to free it. */ fd_close(f); /* push_int(offs0); */ /* printf("Done: %d %d %d ", yy, mm, dd);*/ if(yy && mm && dd) { /* printf("\nLast Summary for %d-%02d-%02d %02d:%02d\n", yy, mm, dd, h, m);*/ pages_per_hour[last_hour] += hourly_page_hits(hits20x, pages, hits, pagexts->u.multiset, 200); if(m_sizeof(session_start)) { summarize_sessions(last_hour, sessions_per_hour, session_length, session_start, session_end); } hosts_per_hour[last_hour] += m_sizeof(sites); do_map_addition(hosts, sites); free_mapping(sites); tmpdest = allocate_mapping(1); summarize_refsites(refsites, referrers, tmpdest); free_mapping(referrers); referrers = tmpdest; summarize_directories(directories, pages); summarize_directories(directories, hits); tmpdest = allocate_mapping(1); clean_refto(referredto, tmpdest, pagexts->u.multiset); free_mapping(referredto); referredto = tmpdest; tmpdest = allocate_mapping(1); http_decode_mapping(user_agents, tmpdest); free_mapping(user_agents); user_agents = tmpdest; tmpdest = allocate_mapping(1); summarize_hosts(hosts, domains, topdomains, tmpdest); free_mapping(hosts); hosts = tmpdest; push_int(yy); push_int(mm); push_int(dd); push_mapping(pages); push_mapping(hits); push_mapping(hits302); push_mapping(hits_per_error); push_mapping(error_urls); push_mapping(error_refs); push_mapping(referredto); push_mapping(refsites); push_mapping(referrers); push_mapping(directories); push_mapping(user_agents); push_mapping(hosts); push_mapping(domains); push_mapping(topdomains); for(i = 0; i < 24; i++) { push_int(sessions_per_hour[i]); } f_aggregate(24); for(i = 0; i < 24; i++) { push_int(hits_per_hour[i]); } f_aggregate(24); for(i = 0; i < 24; i++) { push_int(pages_per_hour[i]); } f_aggregate(24); for(i = 0; i < 24; i++) { push_float(kb_per_hour[i]); } f_aggregate(24); for(i = 0; i < 24; i++) { push_float(sessions_per_hour[i] ? ((float)session_length[i] / (float)sessions_per_hour[i]) / 60.0 : 0.0); } f_aggregate(24); for(i = 0; i < 24; i++) { push_int(hosts_per_hour[i]); hosts_per_hour[i] = 0; } f_aggregate(24); apply_svalue(daily, 23); pop_stack(); } else { free_mapping(error_refs); free_mapping(referredto); free_mapping(refsites); free_mapping(directories); free_mapping(error_urls); free_mapping(hits); free_mapping(hits_per_error); free_mapping(pages); free_mapping(referrers); free_mapping(hits302); free_mapping(user_agents); free_mapping(hosts); free_mapping(domains); free_mapping(topdomains); } free_mapping(hits20x); free_mapping(session_start); free_mapping(session_end); /* free_mapping(hits30x); */ printf("\nTotal lines: %d, broken lines: %d, mapping lookups: %d\n\n", lines, broken_lines, lmu); fflush(stdout); pop_n_elems(args); push_int(offs0); return; skip: broken_lines++; while(1) { while(len--) { #ifdef BROKEN_LINE_DEBUG broken_line[broken_line_pos] = char_pointer[-len]; #endif if(char_class[char_pointer[-len]] == CLS_CRLF) { #ifdef BROKEN_LINE_DEBUG broken_line[broken_line_pos] = 0; printf("Broken Line (pos=%d): %s\n", state_pos, broken_line); #endif goto ok; } } do { len = fd_read(f, read_buf, READ_BLOCK_SIZE); } while(len < 0 && errno == EINTR); if(len <= 0) break; /* nothing more to read. */ offs0 += len; char_pointer = read_buf+len - 1; } goto cleanup; }
} if (!img->img) Pike_error("Image.PVR.encode: no image\n"); if (alpha && !alpha->img) Pike_error("Image.PVR.encode: no alpha image\n"); if (alpha && (alpha->xsize != img->xsize || alpha->ysize != img->ysize)) Pike_error("Image.PVR.encode: alpha and image size differ\n"); if(compress) sz=8+256*4*2+(img->xsize>>1)*(img->ysize>>1); else sz=8+2*img->xsize*img->ysize; res = begin_shared_string(8+sz+(has_gbix? 12:0)); dst = STR0(res); switch(pvr_check_alpha(alpha)) { case 0: alpha = NULL; attr = MODE_RGB565; break; case 1: attr = MODE_ARGB1555; break; case 2: attr = MODE_ARGB4444; break; } if(img->xsize == img->ysize && img->xsize>=8 && img->ysize<=1024 &&