|
|
| version 1.10, 2001/03/20 06:45:19 | version 1.29, 2001/04/03 05:23:41 |
|---|---|
| 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 ArtLebedev Group(http://www.artlebedev.com) |
| Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf) | Author: Alexander Petrosyan <paf@design.ru>(http://design.ru/paf) |
| $Id$ | $Id$ |
| */ | */ |
| #include <string.h> | #include "pa_config_includes.h" |
| #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" | |
| #define escape(cases) \ | #define escape(action) \ |
| { \ | { \ |
| const char *ptr=row->item.ptr; \ | const char *src=row->item.ptr; \ |
| for (int size=row->item.size; size--; ptr++) \ | for(int size=row->item.size; size--; src++) \ |
| switch(*ptr) { \ | action \ |
| cases \ | |
| } \ | |
| } | } |
| #define escape_value(a, c) case a: *copy_here++=c; break | #define _default default: *dest++=*src; break |
| #define escape_default default: *copy_here++=*ptr; break | #define encode(need_encode_func, prefix) \ |
| #define escape_subst(a, b, bsize) \ | |
| case a: \ | |
| strncpy(copy_here, b, bsize); \ | |
| copy_here+=bsize; \ | |
| break | |
| #define escape_encode(need_encode_func, prefix) \ | |
| default: \ | default: \ |
| if(need_encode_func(*ptr)) { \ | 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)*ptr)/0x10]; \ | chunk[1]=hex[((unsigned char)*src)/0x10]; \ |
| chunk[2]=hex[((unsigned char)*ptr)%0x10]; \ | chunk[2]=hex[((unsigned char)*src)%0x10]; \ |
| strncpy(copy_here, chunk, 3); copy_here+=3; \ | strncpy(dest, chunk, 3); dest+=3; \ |
| } else \ | } else \ |
| *copy_here++=*ptr; \ | *dest++=*src; \ |
| break | break |
| #define to_char(c) *dest++=c | |
| #define to_string(b, bsize) \ | |
| strncpy(dest, b, bsize); \ | |
| dest+=bsize; \ | |
| inline bool need_file_encode(unsigned char c){ | inline bool need_file_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')) |
| return false; | return false; |
| return !strchr("./\\", c); | return !strchr("./", 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')) |
| return false; | return false; |
| return !strchr("_-./", c); | return !strchr("_-./", c); |
| } | } |
| inline bool need_header_encode(unsigned char c){ | inline bool need_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); |
| Line 63 inline bool need_header_encode(unsigned | Line 60 inline bool need_header_encode(unsigned |
| // String | // String |
| /// @todo optimize whitespaces for all but 'html' | static bool typo_present(Array::Item *value, const void *info) { |
| char *String::cstr() const { | Array *row=static_cast<Array *>(value); |
| char *result=(char *)malloc(size()*UNTAINT_TIMES_BIGGER+1); | const char *src=static_cast<const char *>(info); |
| int partial; | |
| row->get_string(0)->cmp(partial, src); | |
| return | |
| partial==0 || // full match | |
| partial==1; // typo left column starts 'src' | |
| } | |
| /// @test optimize whitespaces for all but 'html' | |
| char *String::store_to(char *dest) const { | |
| // $MAIN:html-typo table | |
| Table *user_typo_table=static_cast<Table *>(pool().tag()); | |
| Table *typo_table=user_typo_table?user_typo_table:default_typo_table; | |
| char *copy_here=result; | |
| const Chunk *chunk=&head; | const Chunk *chunk=&head; |
| do { | do { |
| const Chunk::Row *row=chunk->rows; | const Chunk::Row *row=chunk->rows; |
| for(int i=0; i<chunk->count; i++) { | for(size_t i=0; i<chunk->count; i++, row++) { |
| if(row==append_here) | if(row==append_here) |
| goto break2; | goto break2; |
| // WARNING: | // WARNING: |
| // string can grow only UNTAINT_TIMES_BIGGER | // string can grow only UNTAINT_TIMES_BIGGER |
| switch(row->item.lang) { | switch(row->item.lang) { |
| case NO: | case UL_CLEAN: |
| // clean piece | // clean piece |
| case YES: | case UL_TAINTED: |
| // tainted piece, but undefined untaint language | // tainted piece, but undefined untaint language |
| // for VString.get_double of tainted values | // for VString.as_double of tainted values |
| // for ^process{body} evaluation | // for ^process{body} evaluation |
| case AS_IS: | case UL_AS_IS: |
| // tainted, untaint language: as-is | // tainted, untaint language: as-is |
| memcpy(copy_here, row->item.ptr, row->item.size); | memcpy(dest, row->item.ptr, row->item.size); |
| copy_here+=row->item.size; | dest+=row->item.size; |
| break; | break; |
| case FILE_NAME: | case UL_FILE_NAME: |
| // tainted, untaint language: file [name] | // tainted, untaint language: file [name] |
| escape( | escape(switch(*src) { |
| escape_value(' ', '_'); | case ' ': to_char('_'); break; |
| escape_encode(need_file_encode, '-'); | encode(need_file_encode, '-'); |
| ); | }); |
| break; | break; |
| case URI: | case UL_URI: |
| // tainted, untaint language: uri | // tainted, untaint language: uri |
| escape( | escape(switch(*src) { |
| escape_value(' ', '+'); | case ' ': to_char('+'); break; |
| escape_encode(need_uri_encode, '%'); | encode(need_uri_encode, '%'); |
| ); | }); |
| break; | break; |
| case HEADER: | case UL_HEADER: |
| // tainted, untaint language: header | // tainted, untaint language: header |
| escape( | escape(switch(*src) { |
| escape_encode(need_header_encode, '%'); | encode(need_header_encode, '%'); |
| ); | }); |
| break; | break; |
| case TABLE: | case UL_TABLE: |
| escape( | // tainted, untaint language: table |
| escape_value('\t', ' '); | escape(switch(*src) { |
| escape_value('\n', ' '); | case '\t': to_char(' '); break; |
| escape_default; | case '\n': to_char(' '); break; |
| ); | _default; |
| }); | |
| break; | break; |
| case SQL: | case UL_SQL: |
| // tainted, untaint language: sql | // tainted, untaint language: sql |
| // TODO: зависимость от sql сервера | // TODO: зависимость от sql сервера |
| memset(copy_here, '?', row->item.size); | memset(dest, '?', row->item.size); |
| copy_here+=row->item.size; | dest+=row->item.size; |
| break; | break; |
| case JS: | case UL_JS: |
| escape( | escape(switch(*src) { |
| escape_subst('"', "\\\"", 2); | case '"': to_string("\\\"", 2); break; |
| escape_subst('\'', "\\'", 2); | case '\'': to_string("\\'", 2); break; |
| escape_subst('\n', "\\n", 2); | case '\n': to_string("\\n", 2); break; |
| escape_subst('\r', "\\r", 2); | case '\\': to_string("\\\\", 2); break; |
| escape_subst('\\', "\\\\", 2); | case '\xFF': to_string("\\\xFF", 2); break; |
| escape_subst('я', "\\я", 2); | _default; |
| escape_default; | }); |
| ); | break; |
| break; | case UL_HTML: |
| case HTML: | escape(switch(*src) { |
| escape( | case '&': to_string("&", 5); break; |
| escape_subst('&', "&", 5); // BEFORE consequent relpaces yelding '&' | case '>': to_string(">", 4); break; |
| escape_subst('>', ">", 4); | case '<': to_string("<", 4); break; |
| escape_subst('<', "<",4); | case '"': to_string(""", 6); break; |
| escape_subst('"', """,6); | //TODO: XSLT case '\'': to_string("'", 6); break; |
| escape_value('\t', ' '); | _default; |
| //TODO: XSLT escape_subst('\'', "'", 6) | }); |
| escape_default; | |
| ); | |
| break; | break; |
| case HTML_TYPO: | case UL_HTML_TYPO: { |
| // tainted, untaint language: html-typo | // tainted, untaint language: html-typo |
| escape( | char *html_for_typo=(char *)malloc(size()*2/* '\n' -> '\' 'n' */+1); |
| escape_subst('&', "&", 5); // BEFORE consequent relpaces yelding '&' | // note: |
| escape_subst('>', ">", 4); | // there still is a possibility that user |
| escape_subst('<', "<",4); | // would not replace \n as she supposed to |
| escape_subst('"', """,6); | // and rather replace \ and n into huge strings |
| escape_value('\t', ' '); | // thus causing memory overrun |
| //TODO: $MAIN:html-type table replace, max length(b)==UNTAINT_TIMES_BIGGER*length(a) | // this can be dealed by allocating *2 memory, but that's too expensive |
| escape_default; | // todo parser4: fix that |
| ); | size_t html_for_typo_size; |
| { // local dest | |
| char *dest=html_for_typo; | |
| escape(switch(*src) { | |
| // convinient name for typo match "\n" | |
| case '\r': | |
| if(typo_table) { | |
| *dest++='\\'; *dest++='n'; // \r -> \n | |
| if(src[1]=='\n') { // \r\n -> remove \n | |
| size--; src++; | |
| } | |
| } | |
| break; | |
| case '\n': | |
| if(typo_table) | |
| to_string("\\n", 2); | |
| break; | |
| //TODO: XSLT case '\'': to_string("'", 6); break; | |
| _default; | |
| }); | |
| *dest=0; | |
| html_for_typo_size=dest-html_for_typo; | |
| } | |
| // typo table replacements | |
| const char *src=html_for_typo; | |
| do { | |
| // there is a row where first column starts 'src' | |
| if(Table::Item *item=typo_table->first_that(typo_present, src)) { | |
| // get a=>b values | |
| const String& a=*static_cast<Array *>(item)->get_string(0); | |
| const String& b=*static_cast<Array *>(item)->get_string(1); | |
| // empty 'a' | 'b' checks | |
| if(a.size()==0 || b.size()==0) { | |
| pool().set_tag(default_typo_table); // avoid recursion | |
| THROW(0, 0, | |
| typo_table->origin_string(), | |
| "typo table column elements must not be empty"); | |
| } | |
| // overflow check: | |
| // b allowed to be max UNTAINT_TIMES_BIGGER then a | |
| if(b.size()>UNTAINT_TIMES_BIGGER*a.size()) { | |
| pool().set_tag(default_typo_table); // avoid recursion | |
| THROW(0, 0, | |
| &b, | |
| "is %g times longer then '%s', " | |
| "while maximum, handled by Parser, is %d", | |
| ((double)b.size())/a.size(), | |
| a.cstr(), | |
| UNTAINT_TIMES_BIGGER); | |
| } | |
| // skip 'a' in 'src' | |
| src+=a.size(); | |
| // write 'b' to 'dest' | |
| b.store_to(dest); | |
| dest+=b.size(); | |
| } else | |
| *dest++=*src++; | |
| } while(*src); | |
| break; | break; |
| } | |
| default: | default: |
| THROW(0,0, | THROW(0, 0, |
| this, | this, |
| "unknown untaint language #%d of %d piece", | "unknown untaint language #%d of %d piece", |
| static_cast<int>(row->item.lang), | static_cast<int>(row->item.lang), |
| i); | i); |
| } | } |
| row++; | |
| } | } |
| chunk=row->link; | chunk=row->link; |
| } while(chunk); | } while(chunk); |
| break2: | break2: |
| *copy_here=0; | return dest; |
| return result; | |
| } | } |