|
|
| version 1.31, 2001/04/03 17:01:03 | version 1.103, 2002/08/01 11:26:51 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: String class part: untaint mechanizm. | Parser: String class part: untaint mechanizm. |
| Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com) | Copyright(c) 2001, 2002 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | |
| Author: Alexander Petrosyan <paf@design.ru>(http://design.ru/paf) | |
| $Id$ | |
| */ | */ |
| #include "pa_config_includes.h" | static const char* IDENT_UNTAINT_C="$Id$"; |
| #include "pa_pool.h" | #include "pa_pool.h" |
| #include "pa_string.h" | #include "pa_string.h" |
| #include "pa_hash.h" | #include "pa_hash.h" |
| #include "pa_exception.h" | #include "pa_exception.h" |
| #include "pa_table.h" | #include "pa_table.h" |
| #include "pa_globals.h" | |
| #include "pa_sql_connection.h" | |
| #include "pa_dictionary.h" | |
| #include "pa_common.h" | |
| #include "pa_charset.h" | |
| //#define DEBUG_STRING_APPENDS_VS_EXPANDS | |
| #ifdef DEBUG_STRING_APPENDS_VS_EXPANDS | |
| ulong string_string_shortcut_economy=0; | |
| #endif | |
| #define escape(action) \ | #define escape(action) \ |
| { \ | { \ |
| Line 24 | Line 32 |
| } | } |
| #define _default default: *dest++=*src; break | #define _default default: *dest++=*src; break |
| #define encode(need_encode_func, prefix) \ | #define encode(need_encode_func, prefix) \ |
| default: \ | |
| if(need_encode_func(*src)) { \ | if(need_encode_func(*src)) { \ |
| static const char *hex="0123456789ABCDEF"; \ | static const char *hex="0123456789ABCDEF"; \ |
| char chunk[3]={prefix}; \ | char chunk[3]={prefix}; \ |
| chunk[1]=hex[((unsigned char)*src)/0x10]; \ | chunk[1]=hex[((unsigned char)*src)/0x10]; \ |
| chunk[2]=hex[((unsigned char)*src)%0x10]; \ | chunk[2]=hex[((unsigned char)*src)%0x10]; \ |
| strncpy(dest, chunk, 3); dest+=3; \ | memcpy(dest, chunk, 3); dest+=3; \ |
| } else \ | } else \ |
| *dest++=*src; \ | *dest++=*src; \ |
| break | break |
| #define to_char(c) *dest++=c | #define to_char(c) *dest++=c |
| #define to_string(b, bsize) \ | #define to_string(b, bsize) \ |
| strncpy(dest, b, bsize); \ | memcpy(dest, b, bsize); \ |
| dest+=bsize; \ | dest+=bsize; \ |
| inline bool need_file_encode(unsigned char c){ | inline bool need_file_encode(unsigned char c){ |
| // theoretical problem with, for instance, "_2B" and "." fragments, | |
| // they would yield the same | |
| // because need_file_encode('_')=false | |
| // but we need to delete such files somehow, getting names from ^index | |
| if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) | if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) |
| return false; | return false; |
| return !strchr( | return !strchr( |
| " _./()-" | |
| #ifdef WIN32 | #ifdef WIN32 |
| ":\\" | ":\\~" |
| #endif | #endif |
| "./", c); | , c); |
| } | } |
| inline bool need_uri_encode(unsigned char c){ | inline bool need_uri_encode(unsigned char c){ |
| if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) | if((c>='0') &&(c<='9') ||(c>='A') &&(c<='Z') ||(c>='a') &&(c<='z')) |
| Line 55 inline bool need_uri_encode(unsigned cha | Line 68 inline bool need_uri_encode(unsigned cha |
| return !strchr("_-./", c); | return !strchr("_-./", c); |
| } | } |
| inline bool need_header_encode(unsigned char c){ | inline bool need_http_header_encode(unsigned char c){ |
| if(strchr(" , :", c)) | if(strchr(" , :", c)) |
| return false; | return false; |
| return need_uri_encode(c); | return need_uri_encode(c); |
| } | } |
| // | |
| static const char * String_Untaint_lang_name[]={ | |
| "U", ///< zero value handy for hash lookup @see untaint_lang_name2enum | |
| "C", ///< clean | |
| "T", ///< tainted, untaint language as assigned later | |
| // untaint languages. assigned by ^untaint[lang]{...} | |
| "P", | |
| /**< | |
| leave language built into string being appended. | |
| just a flag, that value not stored | |
| */ | |
| "A", ///< leave all characters intact | |
| "F", ///< file specification | |
| "H", ///< ext in HTTP response header | |
| "M", ///< text in mail header | |
| "URI", ///< text in uri | |
| "T", ///< ^table:set body | |
| "SQL", ///< ^table:sql body | |
| "JS", ///< JavaScript code | |
| "XML", ///< ^dom:set xml | |
| "HTML" ///< HTML code (for editing) | |
| }; | |
| // String | // String |
| static bool typo_present(Array::Item *value, const void *info) { | /* |
| Array *row=static_cast<Array *>(value); | |
| const char *src=static_cast<const char *>(info); | HTTP-header = field-name ":" [ field-value ] CRLF |
| int partial; | field-name = token |
| row->get_string(0)->cmp(partial, src); | field-value = *( field-content | LWS ) |
| return | |
| partial==0 || // full match | field-content = <the OCTETs making up the field-value |
| partial==1; // typo left column starts 'src' | and consisting of either *TEXT or combinations |
| of token, tspecials, and quoted-string> | |
| word = token | quoted-string | |
| token = 1*<any CHAR except CTLs or tspecials> | |
| tspecials = "(" | ")" | "<" | ">" | "@" | |
| | "," | ";" | ":" | "\" | <"> | |
| | "/" | "[" | "]" | "?" | "=" | |
| | "{" | "}" | SP | HT | |
| SP = <US-ASCII SP, space (32)> | |
| HT = <US-ASCII HT, horizontal-tab (9)> | |
| LWS = [CRLF] 1*( SP | HT ) | |
| TEXT = <any OCTET except CTLs, | |
| but including LWS> | |
| quoted-pair = "\" CHAR | |
| if(strchr("()<>@,;:\\\"/[]?={} \t", *ptr)) | |
| */ | |
| inline bool need_quote_http_header(const char *ptr, size_t size) { | |
| for(; size--; ptr++) | |
| if(strchr(";\\\"= \t" /* excluded ()<>@, :/ ? []{} */, *ptr)) | |
| return true; | |
| return false; | |
| } | } |
| /** | //#include "pa_sapi.h" |
| @test optimize whitespaces for all but 'html' | /** |
| @todo fix theoretical \n mem overrun in TYPO replacements | appends other String, |
| marking all tainted pieces of it with @a lang. | |
| or marking ALL pieces of it with a @a lang when @a forced to, | |
| and propagating OPTIMIZE language bit. | |
| using architecture advantage: after string-to-string-append string never modified. | |
| algorithm: | |
| if no language-change specified and src not yet appended to some other string[last_chunk!=0] | |
| shrinking dest last_chunk[preparing it for linking], | |
| ///shrinking src last_chunk[preparing it to be linked, consequent dest.appends would go there], | |
| linking[dest.last_chunk = src.head.chunk] | |
| if some language-change specified or src already appended to some other string[last_chunk==0] | |
| cloning pieces. | |
| */ | */ |
| char *String::store_to(char *dest, Untaint_lang lang) const { | String& String::append(const String& src, uchar lang, bool forced) { |
| // $MAIN:html-typo table | if(!last_chunk) // growth stopped [we're appended as string to somebody] |
| Table *user_typo_table=static_cast<Table *>(pool().tag()); | throw Exception(0, |
| Table *typo_table=user_typo_table?user_typo_table:default_typo_table; | this, |
| "string growth stopped (append string)"); | |
| const Chunk *chunk=&head; | |
| do { | if(src.is_empty()) |
| const Chunk::Row *row=chunk->rows; | return *this; |
| for(size_t i=0; i<chunk->count; i++, row++) { | |
| if(row==append_here) | // without language-chage, not-appended-before, big[not fitting our tail] string? |
| goto break2; | if(lang==UL_PASS_APPENDED |
| && src.last_chunk | |
| // WARNING: | && (uint(&last_chunk->rows[last_chunk->count]-append_here) < src.used_rows())) { |
| // string can grow only UNTAINT_TIMES_BIGGER | #ifdef DEBUG_STRING_APPENDS_VS_EXPANDS |
| switch(lang==UL_UNKNOWN?row->item.lang:lang) { | string_string_shortcut_economy+=src.used_rows()*sizeof(String::Chunk::Row); |
| case UL_CLEAN: | #endif |
| // clean piece | |
| case UL_TAINTED: | // using fact: |
| // tainted piece, but undefined untaint language | // src.head.chunk.count initally equeals this.head.chunk.count and shrinks-only, |
| // for VString.as_double of tainted values | // so can't be more than this.head.chunk.count, |
| // for ^process{body} evaluation | // which means that we know that |
| case UL_AS_IS: | // src.head.chunk would fit into this.head.chunk |
| // tainted, untaint language: as-is | if(is_empty()) { // our head.chunk is empty |
| memcpy(dest, row->item.ptr, row->item.size); | // they have more than head.chunk? we need all head.chunk : we need only filled-part of head.chunk |
| dest+=row->item.size; | Chunk *src_head_link=src.head.chunk.rows[src.head.chunk.count].link; |
| break; | size_t head_count=src_head_link?src.head.chunk.count:(src.append_here-src.head.chunk.rows); |
| case UL_FILE_NAME: | // "your head.chunk is my head.chunk" |
| // tainted, untaint language: file [name] | memcpy(head.chunk.rows, src.head.chunk.rows, sizeof(Chunk::Row)*(head_count)); |
| escape(switch(*src) { | if(src_head_link) { |
| case ' ': to_char('_'); break; | // "your body is my body" |
| encode(need_file_encode, '-'); | head.chunk.rows[head.chunk.count=head_count].link=src_head_link; |
| }); | // "your last_chunk is mine now" |
| break; | last_chunk=src.last_chunk; |
| case UL_URI: | // "your append_here is mine now" |
| // tainted, untaint language: uri | append_here=src.append_here; |
| escape(switch(*src) { | } else { |
| case ' ': to_char('+'); break; | // "your last_chunk is mine now" |
| encode(need_uri_encode, '%'); | last_chunk=&head.chunk; |
| }); | // "your append_here is recalc-mine now" |
| break; | append_here=head.chunk.rows+head_count; |
| case UL_HEADER: | } |
| // tainted, untaint language: header | } else { // our head.chunk contains something |
| escape(switch(*src) { | // "chopping off my tail-reserve" |
| encode(need_header_encode, '%'); | last_chunk->count=append_here-last_chunk->rows; |
| }); | // "you is my tail" |
| break; | append_here->link=&src.head.chunk; |
| case UL_TABLE: | // "your last_chunk is mine now" |
| // tainted, untaint language: table | last_chunk=src.last_chunk; |
| escape(switch(*src) { | // "your append_here is mine now" |
| case '\t': to_char(' '); break; | append_here=src.append_here; |
| case '\n': to_char(' '); break; | } |
| _default; | |
| }); | // stop-growing mark |
| break; | src.last_chunk=0; |
| case UL_SQL: | return *this; |
| // tainted, untaint language: sql | } |
| // TODO: зависимость от sql сервера | |
| memset(dest, '?', row->item.size); | // manually unrolled code to avoid do{if(const)} constructs |
| if(forced) | |
| STRING_SRC_FOREACH_ROW( | |
| APPEND(row->item.ptr, row->item.size, | |
| lang, //forcing passed lang | |
| row->item.origin.file, row->item.origin.line); | |
| ) | |
| else if(lang==UL_PASS_APPENDED) | |
| STRING_SRC_FOREACH_ROW( | |
| APPEND(row->item.ptr, row->item.size, | |
| row->item.lang, // passing item's lang | |
| row->item.origin.file, row->item.origin.line); | |
| ) | |
| else if(lang&UL_OPTIMIZE_BIT) // main idea here | |
| // tainted piece would get OPTIMIZED bit from 'lang' | |
| // clean piece would be marked OPTIMIZED manually | |
| // pieces with determined languages [not tainted|clean] would retain theirs langs | |
| STRING_SRC_FOREACH_ROW( | |
| APPEND(row->item.ptr, row->item.size, | |
| row->item.lang==UL_TAINTED?lang:( | |
| row->item.lang==UL_CLEAN?UL_CLEAN|UL_OPTIMIZE_BIT: // ORing with OPTIMIZED flag | |
| row->item.lang | |
| ), | |
| row->item.origin.file, row->item.origin.line); | |
| ) | |
| else | |
| STRING_SRC_FOREACH_ROW( | |
| APPEND(row->item.ptr, row->item.size, | |
| row->item.lang==UL_TAINTED?lang:row->item.lang, | |
| row->item.origin.file, row->item.origin.line); | |
| ); | |
| /* | |
| for(Chunk::Row *row=last_chunk->rows; row<append_here; row++) | |
| if(row->link==(void*)0xcdcdcdcd) | |
| _asm int 3;*/ | |
| return *this; | |
| } | |
| size_t String::cstr_bufsize(Untaint_lang lang, | |
| SQL_Connection *connection, | |
| Charset *buf_charset) const { | |
| size_t dest=1; // for terminating 0 | |
| STRING_FOREACH_ROW( | |
| uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang; | |
| switch(to_lang & ~UL_OPTIMIZE_BIT) { | |
| case UL_CLEAN: | |
| case UL_TAINTED: | |
| case UL_AS_IS: | |
| // clean piece | |
| // tainted piece, but undefined untaint language | |
| // for VString.as_double of tainted values | |
| // for ^process{body} evaluation | |
| // tainted, untaint language: as-is | |
| dest+=row->item.size; | |
| break; | |
| case UL_FILE_SPEC: | |
| // tainted, untaint language: file [name] | |
| dest+=row->item.size*3/* worst: Z->%XX */; | |
| break; | |
| case UL_URI: | |
| // tainted, untaint language: uri | |
| dest+=row->item.size*6*3/* worst utf8 x worst Z->%XX */; | |
| break; | |
| case UL_HTTP_HEADER: | |
| // tainted, untaint language: http-field-content-text | |
| dest+=row->item.size*3/* worst: Z->%XX */; | |
| break; | |
| case UL_MAIL_HEADER: | |
| // tainted, untaint language: mail-header | |
| if(buf_charset) { | |
| // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= | |
| dest+= | |
| row->item.size*3+ | |
| buf_charset->name().size()+MAX_STRING/* worst: =?charset?Q?=%XX?= */; | |
| } else | |
| dest+=row->item.size; | dest+=row->item.size; |
| break; | break; |
| case UL_JS: | case UL_TABLE: |
| escape(switch(*src) { | // tainted, untaint language: table |
| case '"': to_string("\\\"", 2); break; | dest+=row->item.size; |
| case '\'': to_string("\\'", 2); break; | break; |
| case '\n': to_string("\\n", 2); break; | case UL_SQL: |
| case '\\': to_string("\\\\", 2); break; | // tainted, untaint language: sql |
| case '\xFF': to_string("\\\xFF", 2); break; | if(connection) |
| _default; | dest+=connection->quote(0, row->item.ptr, row->item.size); |
| }); | break; |
| break; | case UL_JS: |
| case UL_HTML: | escape(switch(*src) { |
| escape(switch(*src) { | case '"': case '\'': case '\n': case '\\': case '\xFF': |
| case '&': to_string("&", 5); break; | dest+=2; break; |
| case '>': to_string(">", 4); break; | default: |
| case '<': to_string("<", 4); break; | dest++; break; |
| case '"': to_string(""", 6); break; | }); |
| //TODO: XSLT case '\'': to_string("'", 6); break; | break; |
| _default; | case UL_XML: |
| }); | escape(switch(*src) { |
| break; | case '&': case '>': case '<': case '"': case '\'': |
| case UL_HTML_TYPO: { | dest+= 6; break; |
| // tainted, untaint language: html-typo | default: |
| char *html_for_typo=(char *)malloc(size()*2/* '\n' -> '\' 'n' */+1); | dest++; break; |
| // note: | }); |
| // there still is a possibility that user | break; |
| // would not replace \n as she supposed to | case UL_HTML: |
| // and rather replace \ and n into huge strings | escape(switch(*src) { |
| // thus causing memory overrun | case '&': |
| // this can be dealed by allocating *2 memory, but that's too expensive | case '>': |
| size_t html_for_typo_size; | case '<': |
| { // local dest | case '"': |
| char *dest=html_for_typo; | dest+=6; break; |
| escape(switch(*src) { | default: |
| // convinient name for typo match "\n" | dest++; break; |
| case '\r': | }); |
| if(typo_table) { | break; |
| *dest++='\\'; *dest++='n'; // \r -> \n | } |
| if(src[1]=='\n') { // \r\n -> remove \n | ); |
| size--; src++; | return dest; |
| } | } |
| } | |
| break; | char *String::store_to(char *dest, Untaint_lang lang, |
| case '\n': | SQL_Connection *connection, |
| if(typo_table) | Charset *store_to_charset, |
| to_string("\\n", 2); | const char *store_to_charset_name) const { |
| break; | // WARNING: |
| //TODO: XSLT case '\'': to_string("'", 6); break; | // before any changes check cstr_bufsize first!!! |
| _default; | bool whitespace=true; |
| }); | STRING_FOREACH_ROW( |
| *dest=0; | uchar to_lang=lang==UL_UNSPECIFIED?row->item.lang:lang; |
| html_for_typo_size=dest-html_for_typo; | |
| } | char *start=dest; |
| // typo table replacements | |
| const char *src=html_for_typo; | switch(to_lang & ~UL_OPTIMIZE_BIT) { |
| do { | case UL_CLEAN: |
| // there is a row where first column starts 'src' | case UL_TAINTED: |
| if(Table::Item *item=typo_table->first_that(typo_present, src)) { | case UL_AS_IS: |
| // get a=>b values | // clean piece |
| const String& a=*static_cast<Array *>(item)->get_string(0); | |
| const String& b=*static_cast<Array *>(item)->get_string(1); | // tainted piece, but undefined untaint language |
| // empty 'a' | 'b' checks | // for VString.as_double of tainted values |
| if(a.size()==0 || b.size()==0) { | // for ^process{body} evaluation |
| pool().set_tag(default_typo_table); // avoid recursion | |
| THROW(0, 0, | // tainted, untaint language: as-is |
| typo_table->origin_string(), | memcpy(dest, row->item.ptr, row->item.size); |
| "typo table column elements must not be empty"); | dest+=row->item.size; |
| } | break; |
| // overflow check: | case UL_FILE_SPEC: |
| // b allowed to be max UNTAINT_TIMES_BIGGER then a | // tainted, untaint language: file [name] |
| if(b.size()>UNTAINT_TIMES_BIGGER*a.size()) { | escape( |
| pool().set_tag(default_typo_table); // avoid recursion | encode(need_file_encode, '_'); |
| THROW(0, 0, | ); |
| &b, | break; |
| "is %g times longer then '%s', " | case UL_URI: |
| "while maximum, handled by Parser, is %d", | // tainted, untaint language: uri |
| ((double)b.size())/a.size(), | const void *client_ptr; |
| a.cstr(), | size_t client_size; |
| UNTAINT_TIMES_BIGGER); | Charset::transcode(pool(), |
| pool().get_source_charset(), row->item.ptr, row->item.size, | |
| pool().get_client_charset(), client_ptr, client_size); | |
| { | |
| const char *src=(const char *)client_ptr; | |
| for(int size=client_size; size--; src++) | |
| switch(*src) { | |
| case ' ': to_char('+'); break; | |
| default: encode(need_uri_encode, '%'); | |
| }; | |
| } | |
| break; | |
| case UL_HTTP_HEADER: | |
| // tainted, untaint language: http-field-content-text | |
| escape(switch(*src) { | |
| case ' ': to_char('+'); break; | |
| default: encode(need_uri_encode, '%'); | |
| }); | |
| break; | |
| case UL_MAIL_HEADER: | |
| // tainted, untaint language: mail-header | |
| if(store_to_charset && store_to_charset_name) { | |
| const void *mail_ptr; | |
| size_t mail_size; | |
| Charset::transcode(pool(), | |
| pool().get_source_charset(), row->item.ptr, row->item.size, | |
| *store_to_charset, mail_ptr, mail_size); | |
| // Subject: Re: parser3: =?koi8-r?Q?=D3=C5=CD=C9=CE=C1=D2?= | |
| const char *src=(const char *)mail_ptr; | |
| bool to_quoted_printable=false; | |
| for(int size=mail_size; size--; src++) { | |
| if((*src & 0x80) // starting quote-printable-encoding on first 8bit char | |
| || (to_quoted_printable && (*src=='?' || *src=='=')) // additionally encoding '?' and '|' | |
| ) { | |
| if(!to_quoted_printable) { | |
| dest+=sprintf(dest, "=?%s?Q?", store_to_charset_name); | |
| to_quoted_printable=true; | |
| } | } |
| dest+=sprintf(dest, "=%02X", *src & 0xFF); | |
| // skip 'a' in 'src' | } else { |
| src+=a.size(); | *dest++=*src; |
| // write 'b' to 'dest' | } |
| b.store_to(dest); | |
| dest+=b.size(); | |
| } else | |
| *dest++=*src++; | |
| } while(*src); | |
| break; | |
| } | } |
| default: | if(to_quoted_printable) // close |
| THROW(0, 0, | dest+=sprintf(dest, "?="); |
| this, | |
| "unknown untaint language #%d of %d piece", | } else { |
| static_cast<int>(row->item.lang), | memcpy(dest, row->item.ptr, row->item.size); |
| i); | dest+=row->item.size; |
| } | } |
| break; | |
| case UL_TABLE: | |
| // tainted, untaint language: table | |
| escape(switch(*src) { | |
| case '\t': to_char(' '); break; | |
| case '\n': to_char(' '); break; | |
| _default; | |
| }); | |
| break; | |
| case UL_SQL: | |
| // tainted, untaint language: sql | |
| if(connection) | |
| dest+=connection->quote(dest, row->item.ptr, row->item.size); | |
| else | |
| throw Exception(0, | |
| this, | |
| "untaint in SQL language failed - no connection specified"); | |
| break; | |
| case UL_JS: | |
| escape(switch(*src) { | |
| case '"': to_string("\\\"", 2); break; | |
| case '\'': to_string("\\'", 2); break; | |
| case '\n': to_string("\\n", 2); break; | |
| case '\\': to_string("\\\\", 2); break; | |
| case '\xFF': to_string("\\\xFF", 2); break; | |
| _default; | |
| }); | |
| break; | |
| case UL_XML: | |
| escape(switch(*src) { | |
| case '&': to_string("&", 5); break; | |
| case '>': to_string(">", 4); break; | |
| case '<': to_string("<", 4); break; | |
| case '"': to_string(""", 6); break; | |
| case '\'': to_string("'", 6); break; | |
| _default; | |
| }); | |
| break; | |
| case UL_HTML: | |
| escape(switch(*src) { | |
| case '&': to_string("&", 5); break; | |
| case '>': to_string(">", 4); break; | |
| case '<': to_string("<", 4); break; | |
| case '"': to_string(""", 6); break; | |
| _default; | |
| }); | |
| break; | |
| default: | |
| throw Exception(0, | |
| this, | |
| "unknown untaint language #%d", | |
| static_cast<int>(row->item.lang)); // sould never | |
| break; // never | |
| } | } |
| chunk=row->link; | |
| } while(chunk); | if(to_lang & UL_OPTIMIZE_BIT) { |
| break2: | // optimizing whitespace |
| char *stop=dest; dest=start; | |
| for(char *src=start; src<stop; src++) | |
| switch(*src) { | |
| // of all consequent white space chars leaving only first one | |
| case ' ': case '\r': case '\n': case '\t': | |
| if(!whitespace) { | |
| *dest++=*src; | |
| whitespace=true; | |
| } | |
| break; | |
| default: | |
| whitespace=false; | |
| *dest++=*src; | |
| break; | |
| }; | |
| } else // piece without optimization | |
| whitespace=false; | |
| ); | |
| return dest; | return dest; |
| } | } |
| char *String::cstr_debug_origins() const { | |
| //_asm int 3; | |
| char *result=(char *)malloc(size()+used_rows()*MAX_STRING*2); | |
| char *dest=result; | |
| STRING_FOREACH_ROW( | |
| IFNDEF_NO_STRING_ORIGIN( | |
| if(row->item.origin.file) | |
| dest+=sprintf(dest, ORIGIN_FILE_LINE_FORMAT, | |
| row->item.origin.file, | |
| 1+row->item.origin.line); | |
| else | |
| dest+=sprintf(dest, "<unknown>"); | |
| ); | |
| uchar show_lang=row->item.lang & ~UL_OPTIMIZE_BIT; | |
| if(show_lang>=sizeof(String_Untaint_lang_name)/sizeof(String_Untaint_lang_name[0])) | |
| throw Exception(0, | |
| this, | |
| "unknown untaint language #%d", | |
| static_cast<int>(show_lang)); // sould never | |
| dest+=sprintf(dest, "#%s%s: ", | |
| String_Untaint_lang_name[show_lang], | |
| row->item.lang & UL_OPTIMIZE_BIT?".O":""); | |
| char *dest_after_origins=dest; | |
| memcpy(dest, row->item.ptr, row->item.size); | |
| dest+=row->item.size; | |
| remove_crlf(dest_after_origins, dest); | |
| to_char('\n'); | |
| ); | |
| *dest=0; | |
| return result; | |
| } |