Parse-RPN (V 2.xx) - Introduction Parse::RPN - Is a minimalist RPN parser/processor (a little like FORTH) $Revision: 50 $ SYNOPSIS use Parse::RPN; $result=rpn(string ...); @results=rpn(string ...); $error=rpn_error(); string... is a list of RPN operator and value separated by a coma in scalar mode RPN return the result of the calculation (If the stack contain more then one element, you receive a warning and the top value on the stack) in array mode, you receive the content of the stack after evaluation DESCRIPTION rpn() receive in entry a scalar of one or more elements coma separated and evaluate as an RPN (Reverse Polish Notation) command. The function split all elements and put in the stack. The operator are case sensitive. The operator are detect as is, if they are alone in the element of the stack. Extra space before or after are allowed (e.g "3,4,ADD" here ADD is an opeartor but it is not the case in "3,4,ADD 1") If element is not part of the predefined operator (dictionary), the element is push as a litteral. If you would like to put a string which is part of the dictionary, put it between quote or double-quote (e.g "3,4,'ADD'" here ADD is a literal and the evaluation reurn ADD and a warning because the stack is not empty) If the string contain a coma, you need also to quote or double-quote the string. (be care to close your quoted or double-quoted string) The evaluation follow the rule of RPN or FORTH or POSTCRIPT or pockect calcutor HP. Look on web for documentation about the use of RPN notation. I use this module in a application where the final user need to create an maintain a configuration file with the possibility to do calculation on variable returned from application. The idea of this module is comming from Math::RPN of Owen DeLong, owen@delong.com that I used for more then a year before some of my customer would like more ... I correct a bug (interversion of > and >=), add the STRING function, pattern search and some STACK functions. rpn_error() return the last error from the evaluation (illegal division by 0, error from the PERL function execution...) each time that rpn() is call the rpn_error() is reinitianised. MATHEMATIC operators . a b + return the result of 'a' + 'b' a b - return the result of 'a' - 'b' a b * return the result of 'a' * 'b' a b / return the result of 'a' / 'b' if b =0 return '' (to prevent exception raise) a b ** return the result of 'a' ** 'b' (exponant) a 1+ return the result of 'a' +1 a 1- return the result of 'a' -1 a 2- return the result of 'a' -2 a 2+ return the result of 'a' +2 a b MOD return the result of 'a' % 'b' a ABS return the result of abs 'a' a INT return the result of INT 'a' a +- return the result negate value of 'a' (- 'a' ) a REMAIN return the result of 'a' - int 'a' (fractional part of 'a' ) a SIN return the result of sin 'a' ('a' in RADIAN) a COS return the result of cos 'a' ('a' in RADIAN) a TAN return the result of tan 'a' ('a' in RADIAN) a CTAN return the result of cotan 'a' ('a' in RADIAN) a LN return the result of ln 'a' if = 0 return '' (to prevent exception raise) a EXP return the result of 'e' ** 'a' PI return the value of PI (3.14159265358979) relational operators . a b < return the result of 'a' < 'b' ( BOOLEAN value ) a b <= return the result of 'a' <= 'b' ( BOOLEAN value ) a b > return the result of 'a' > 'b' ( BOOLEAN value ) a b >= return the result of 'a' >= 'b' ( BOOLEAN value ) a b == return the result of 'a' == 'b' ( BOOLEAN value ) 1 if a == b else 0 a b <=> return the result of 'a' <=> 'b' ( BOOLEAN value ) -1 if a < b ,0 if a == b, 1 if a > b a b != return the result of 'a' != 'b' ( BOOLEAN value ) 0 if a == b else 1 LOGICAL operators a b OR return the 1 one of the 2 argument are not equal to 0 a b AND return the 0 one of the 2 argument are equal to 0 a b XOR return the 0 if the 2 argument are equal a NOT return the 0 if the argument is not eqauk to 0 return the 1 if the argument is eqauk to 0 a TRUE return the 1 if the top of stack is !=0 and if stack not empty a FALSE return the 0 if the top of stack is !=0 MISC operators a b >> bitwise shift to the right shift the bits in a to the left of b level a b << bitwise shift to the left shift the bits in a to the left of b level a b MIN return the result smallest of the 2 arguments a b MAX return the result greatest of the 2 arguments a VAL,RET, "operator" LOOKUP test with the "operator" the [a] value on each elements of VAL and if test succeed return the value from array RET with the same index the "operator" must be quoted to prevent evaluation a VAL,RET, "operator" LOOKUPP test with the perl "operator" the [a] value on each elements of VAL and if test succeed return the value from array RET with the same index the "operator" must be quoted to prevent evaluation a VAL,RET,OPE LOOKUP loop on each item of array VAL and test the value [ a ] with the operator from ope ARRAY against the corresponding value in array VAL and return the value from array RET with the same index a VAL,RET,OPE LOOKUPP loop on each item of array VAL and test the value [ a ] with the perl operator from ope ARRAY against the corresponding value in array VAL and return the value from array RET with the same index TICK return the current time in ticks a LTIME return the localtime coresponding to the ticks value 'a' the format is 'sec' 'min' 'hour' 'day_in_the_month' 'month' 'year' 'day_in_week' 'day_year' 'dayloight_saving' 'year' is the elapsed year since 1900 'month' start to 0 The format is the same as localtime() in perl a GTIME return the gmtime coresponding to the ticks value 'a' the format is 'sec' 'min' 'hour' 'day_in_the_month' 'month' 'year' 'day_in_week' 'day_year' 'dayloight_saving' 'year' is the elapsed year since 1900 'month' start to 0 The format is the same as gmtime() in perl a HLTIME return the localtime coresponding to the ticks value 'a' in a human readable format a HGTIME return the gmtime coresponding to the ticks value 'a' in a human readable format a HTTPTIME return the ticks coresponding to the time value in a format accepted by HTTP::Date RAND return a random value in the range [0,1[ a LRAND return a random value in the range [0,'a'[ a SPACE return the number 'a' formated with space each 3 digits a DOT return the number 'a' formated with . (dot) each 3 digits a NORM return the number 'a' normalize by slice of 1000 with extra power value "K", "M", "G", "T", "P" (or nothing if lower than 1000) a NORM2 return the number 'a' normalize by slice of 1024 with extra power value "K", "M", "G", "T", "P" (or nothing if lower than 1024) a UNORM return the number from a 'a' with a sufix "K", "M", "G", "T", "P" (or nothing if lower than 1000) and calculate the real value base 1000 ( e.g 7k = 7000) a UNORM2 return the number from a 'a' with a sufix "K", "M", "G", "T", "P" (or nothing if lower than 1024) and calculate the real value base 1024 ( e.g 7k = 7168) a OCT return the decimal value for the HEX, BINARY or OCTAL value 'a' OCTAL is like '0nn' where n is in the range of 0-7 BINARY is like '0bnnn...' where n is in the range of 0-1 HEX is like '0xnnn' where n is in the range of 0-9A-F if no specific format convert as an hexadecimal by default a OCTSTR2HEX return a HEX string from a OCTETSTRING. useful when receiving an SNMP ASN.1 OCTETSTRING like mac address a HEX2OCTSTR return a OCTETSTRING string from a HEX useful when you need to check if an SNMP ASN.1 OCTETSTRING if matching the hex value provided a DDEC2STR return a string from a dotted DEC string useful when you need to manipulate an SNMP extension with 'exec' a STR2DDEC return a dotted DEC string to a string useful when you need to manipulate an SNMP extension with 'exec' STRING operators . a b EQ return the result of 'a' EQ 'b' ( BOOLEAN value ) a b NE return the result of 'a' NE 'b' ( BOOLEAN value ) a b LT return the result of 'a' LT 'b' ( BOOLEAN value ) a b GT return the result of 'a' GT 'b' ( BOOLEAN value ) a b LE return the result of 'a' LE 'b' ( BOOLEAN value ) a b GE return the result of 'a' GE 'b' ( BOOLEAN value ) a b CMP return the result of 'a' CMP 'b' ( BOOLEAN value ) a LEN return the length of 'a' EQ 'b' a b CAT return the concatenation 'a' and 'b' a b CATALL return the concatenation all element on the stack a b REP return the result of 'a' x 'b' duplicate 'a' by the number of 'x' a REV return the reverse of 'a' EQ 'b' a b c SUBSTR return the substring of 'c' starting at 'b' with the length of 'a' a UC return 'a' in uppercase a LC return 'a' in lowercase a UCFIRST return 'a' with the first letter in uppercase a LCFIRST return 'a' with the first letter in lowercase a b SPLIT return all splitted item of 'a' by the separator 'b' 'b' is a REGEX !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry !!! if the split match on the beginning of string, SPLIT return the matched value WITHOUT the empty string of the beginning a b SPLITI return all splitted item of 'a' by the separator 'b' 'b' is a REGEX case insensitive !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry !!! if the split match on the beginning of string, SPLIT return the matched value WITHOUT the empty string of the beginning a b PAT return one or more occurance of 'b' in 'a' 'b' is a REGEX !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry a b PATI return one or more occurance of 'b' in 'a' 'b' is a REGEX case insensitive !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry a b TPAT test if the pattern 'b' is in 'a' 'b' is a REGEX !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry a b TPATI test if the pattern 'b' is in 'a' 'b' is a REGEX !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry a b c SPAT substitute the pattern 'b' by the pattern 'a' in 'c' 'b' and 'c' are a REGEX !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry a b c SPATG substitute the pattern 'b' by the pattern 'a' in 'c' as many time as possible (g flag in REGEX) 'b' and 'c' are a REGEX !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry a b c SPATI substitute the pattern 'b' by the pattern 'a' in 'c'case insensitive (i flag in REGEX) 'b' and 'c' are a REGEX !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry a b c SPATGI substitute the pattern 'b' by the pattern 'a' in 'c' as many time as possible (g flag in REGEX) and case insensitive (1 flag in REGEX) 'b' and 'c' are a REGEX !!! becare, if you need to use : as a regex, you need to backslash to prevent overlap with new dictionary entry a ... z PRINTF use the format 'z' to print the value(s) on the stack 7,3,/,10,3,/,%d %f,PRINTF -> 2 3.333333 see printf in perl a b PACK pack the value 'a' with the format 'b' 2004,06,08,a4 a2 a2,PACK -> 20040608 see pack in perl a b UNPACK unpack the value 'a' with the format 'b' 20040608,a4 a2 a2,PACK -> 2004,06,08 see unpack in perl a b ISNUM test if top of the stack is a number return 1 if if it is a NUMBER otherwise return 0 a b ISNUMD test if top of the stack is a number delete the top element on the statck and return 1 if it is a NUMBER otherwise return 0 a b ISINT test if top of the stack is a integer (natural number) return 1 if if it is a INTEGER otherwise return 0 a b ISINTD test if top of the stack is a integer (natural number) delete the top element on the statck and return 1 if it is a INTEGER otherwise return 0 a b ISHEX test if top of the stack is a hexadecimal value (starting with 0x or 0X or # ) return 1 if if it is a HEXADECIMAL otherwise return 0 a b ISHEXD test if top of the stack is a hexadecimal value (starting with 0x or 0X or # ) delete the top element on the statck and return 1 if it is a HEXADECIMAL otherwise return 0 STACK operators . a b SWAP return 'b' 'a' a b OVER return 'a' 'b' 'a' a DUP return 'a' 'a' a b DDUP return 'a' 'b' 'a' 'b' a b c ROT return 'b' 'c' 'a' a b c RROT return 'c' 'a' 'b' DEPTH return the number of elements on the stack a b POP remove the last element on the stack a ... z POPN remove the 'z' last element(s) from the stack a b c d e n ROLL rotate the stack on 'n' element a,b,c,d,e,4,ROLL -> a c d e b if n = 3 <=> ROT a PICK copy element from depth 'a' to the stack a GET get (remove) element from depth 'a' and put on top of stack a b PUT put element 'a' at the level 'b' of the stack if 'b' gretaer than the stack put at first place if 'b' < 0 start to the reverse order of the stack a b DEL delete 'b' element on the stack from level 'a' 'a' and 'b' is get in absolute value a FIND get the level of stack containing the exact value 'a' a SEARCH get the first level of stack containing the REGEX 'a' a SEARCHI get the first level of stack containing the REGEX 'a' (cas insensitive) a SEARCHK keep all level of stack containing the REGEX 'a' (cas sensitive) a SEARCHIK keep all level of stack containing the REGEX 'a' (cas insensitive) a KEEP delete all element on the stack except the level 'a' if 'a' is deeper then stack, keep the stack untouched a b KEEPN keep 'b' element on the stack from level 'a' and delete all other element 'a' and 'b' is get in absolute value a b PRESERVE keep element on the stack from level 'a' to level 'b' and delete all other element 'a' and 'b' is get in absolute value if 'a' > 'b' keep the reverse of selection (boustrophedon) a b COPY copy element on the stack from level 'a' to level 'b' 'a' and 'b' is get in absolute value if 'a' > 'b' keep the reverse of selection (boustrophedon) a SUM sum the a element on top of the stack remove these a element and return the result value on the stack a STATS STATS the a element on top of the stack remove these a element the new variable _SUM_, _MULT_, _ARITH_MEAN_, _GEOM_MEAN_, _QUAD_MEAN_ (= _RMS_), _HARM_MEAN_, _STD_DEV_, _SAMPLE_STD_DEV_, _VARIANCE_, DICTIONARY and VARS operators . WORDS return as one stack element the list of WORD in DICT separated by a | VARS return as one stack element the list of VARS separated by a | INC incremente (+ 1) the value of the variable on the statck DEC decremente (- 1) the value of the variable on the statck VARIABLE xxx deckare the variable 'xxx' (reserve memory) xx var ! set the value xx to the variable 'var' x1 x2 x3 ... n var !! put and delete 'n' element(s) from the stack in the variable 'var' 'n' is in absolute value x1 x2 x3 ... n var !!C copy 'n' element(s) from the stack in the variable 'var' 'n' is in absolute value x1 x2 x3 ... b a var !!! put and delete ' element(s) from the stack in the variable 'var' starting at element 'a' to element 'b' 'a' and 'b' in absolute value if 'a' > 'b' keep the reverse of selection (boustrophedon) x1 x2 x3 ... b a var !!!C copy element(s) on the stack in the variable 'var' starting at element 'a' to element 'b' 'a' and 'b' in absolute value if 'a' > 'b' keep the reverse of selection (boustrophedon) var @ return the value of the variable 'var' :xxx name1 ; create a new entry in the dictionary whith name name1 and store the progam xxx : xxx yyy name1 PERL execute the PERL code with parameter(s) xxx yyy !!! be care if the perl code need to use a coma (,) you need to enclose the line inside double quote if you need double quote in code use qq{ ... } : xxx name1 PERLFUNC execute the PERL function name1 with the parameter xxx the default name space is "main::" It is possible tu use a specific name space the paramter are "stringified" e.g. ':,5,filename,save,PERLFUNC' call the function save("filename", 5); a >R put 'a' on the return stack R> remove first element from the return stack and copy on the normal stack RL return the depth of the return stack R@ copy return stack on normal stack LOOP and DECISION operators . a IF xxx THEN test the element on top of stack if == 1 execute 'xxx' block The loop is executed always one time a IF zzz ELSE xxx THEN test the element on top of stack if == 1 execute 'xxx' block if != 1 execute 'zzz' block The loop is executed always one time BEGIN xxx WHILE zzz REPEAT execute 'xxx' block test the element on top of stack if == 0 execute 'zzz' block and branch again at 'BEGIN' if != 0 end the loop The loop is executed always one time end start DO,block,LOOP process 'block' with iterator from value 'start' until 'end' value,with increment of 1; The iterator variable is the second value on the stack (start argument) end start increment DO,block,+LOOP process 'block' with iterator from value 'start' untill 'end' value,with increment of 'increment' This allow rational or negative value The iterator variable is the second value on the stack (start argument) Useful functions for the module (not related to the RPN language) . =head2 rpn_error() function which return the debug info from the calculation (like a division by 0) rpn_separator( 'sep' ) function to set a specific separator for the returned stack (default = space) This is useful when the result of rpn() is use inside another rpn() call OPERATORS The operators get value from the stack and push the result on top In the following explanation, the stack is represented as a pair of brackets () and each elements by a pair of square barcket [] The left part is the state before evalutation and the right part is the state of the stack after evaluation Arithmetic operators --------------------- + ([a][b]) ([a+b]) - ([a][b]) ([a-b]) * ([a][b]) ([a*b]) / ([a][b]) ([a/b]) Becare if division by null return a blank value ** ([a][b]) ([a**b]) 1+ ([a]) ([a+1]) 1- ([a]) ([a-1]) 2+ ([a]) ([a+2]) 2- ([a]) ([a-2]) MOD ([a][b]) ([a%b]) ABS ([a]) ([ABS a]) INT ([a]) ([INT a]) +- ([a]) ([-a]) REMAIN ([a]) ([a- INT a]) Rationnal operators ------------------- SIN ([a]) ([SIN a]) Unit in radian COS ([a]) ([COS a]) Unit in radian TAN ([a]) ([TAN a]) Unit in radian CTAN ([a]) ([CTAN a]) Unit in radian LN ([a]) ([LOG a]) EXP ([a]) ([EXP a]) PI ([3.14159265358979]) Relational operator ---------------- < ([a][b]) ([1]) if [a]<[b] else ([0]) <= ([a][b]) ([1]) if [a]<=[b] else ([0]) > ([a][b]) ([1]) if [a]>[b] else ([0]) >= ([a][b]) ([1]) if [a]>=[b] else ([0]) == ([a][b]) ([1]) if [a]==[b] else ([0]) <=> ([a][b]) ([-1]) if [a]>[b],([1]) if [a]<[b], ([0])if [a]==[b] != ([a][b]) ([0]) if [a]==[b] else ([1]) TRUE ([a]) Return 1 if [a]>0 and exist FALSE ([a]) Return 0 if [a]>0 Logical operator ---------------- OR ([a][b]) ([1]) if [a] or [b] >0 AND ([a][b]) ([1]) if [a] and [b] >0 XOR ([a][b]) ([1]) if [a] and [b] are >0 or ==0 NOT ([a]) Return 0 if [a]>0, Return 1 if[a]==0, Other operator ---------------- >> ([a][b]) shift to the right the bits from [a] of [b] rank << ([a][b]) shift to the left the bits from [a] of [b] rank MIN ([a][b]) ([a]) if [a]<[b] else ([b]) MAX ([a][b]) ([a]) if [a]>[b] else ([b]) LOOKUP ([a] V R [ope] ) test [ a ] on all value of array V with the operator [ope] if succeed, return the value from array R at the succesfull indice LOOKUPP ([a] V R [ope] ) test [ a ] on all value of array V with the perl operator [ope] if succeed, return the value from array R at the succesfull indice LOOKUPOP ([a] V R O] ) test [ a ] on all value of array V with the operator from the array OPE with the same indice LOOKUPOPP ([a] V R O] ) test [ a ] on all value of array V with the perl operator from the array OPE with the same indice if succeed, return the value from array R at the succesfull indice TICK () ([time]) time in ticks LTIME ([a]) ([min][hour][day_in_the_month][month][year][day_in_week][day_year][daylight_saving] localtime of [a] like PERL GTIME ([a]) ([min][hour][day_in_the_month][month][year][day_in_week][day_year][daylight_saving] ([a]) gmtime of [a] like PERL HLTIME ([a]) ([a]) localtime human readeable HGTIME ([a]) gmtime human readeable RAND () ([rand]) a random numder between 0 and 1 LRAND ([a]) ([rand]) a random numder between 0 and [a] SPACE ([a]) Return [a] with space between each 3 digits DOT ([a]) Return [a] with dot (.) between each 3 digits NORM ([a]) Return [a] normalized by 1000 (K,M,G = 1000 * unit) NORM2 ([a]) Return [a] normalized by 1000 (K,M,G = 1024 * unit) OCT (|a|) Return the DECIMAL value from HEX,OCTAL or BINARY value |a| (see oct from perl) OCTSTR2HEX (|a|) Return a HEX string from a OCTETSTRING OCTSTR2HEX (|a|) Return a OCTETSTRING string from a HEX DDEC2STR (|a|) Return a string from a dotted DEC string STR2DDEC (|a|) Return a dotted DEC string to a string String operators ---------------- EQ ([a][b]) ([1]) if [a] eq [b] else ([0]) NE ([a][b]) ([1]) if [a] ne [b] else ([0]) LT ([a][b]) ([1]) if [a] lt [b] else ([0]) GT ([a][b]) ([1]) if [a] gt [b] else ([0]) LE ([a][b]) ([1]) if [a] le [b] else ([0]) GE ([a][b]) ([1]) if [a] ge [b] else ([0]) CMP ([a][b]) ([-1]) if [a] gt [b],([1]) if [a] lt [b], ([0])if [a] eq [b] LEN ([a]) ([LENGTH a]) CAT ([a][b]) ([ab]) String concatenation CATALL ([a][b]...[z]) ([ab...z]) String concatenation of all elements on the stack REP ([a][b]) ([a x b]) repeat [b] time the motif [a] REV ([a]) ([REVERSE a]) SUBSTR ([a][b][c]) ([SUBSTR [a], [b], [c]) get substring of [a] starting from [b] untill [c] UC ([a]) ([UC a]) LC ([a]) ([LC a]) UCFIRST ([a]) ([UCFIRST a]) LCFIRST ([a]) ([LCFIRST a]) PAT ([a][b]) ([r1]...) use the pattern [b] on the string [a] and return result if more then one result like $1, $2 ... return all the results PATI ([a][b]) ([r1]...) use the pattern CASE INSENSITIVE [b] on the string [a] and return result if more then one result like $1, $2 ... return all the results TPAT ([a][b]) ([r]) use the pattern [b] on the string [a] and return 1 if pattern macth otherwise return 0 TPATI ([a][b]) ([r]) use the pattern CASE INSENSITIVE [b] on the string [a] and return 1 if pattern macth otherwise return 0 SPLIT ([a][b]) split ([a]) using the pattern ([b]) and return all elements on stack SPLITI split ([a]) using the pattern CASE INSENSITIVE ([b])) and return all elements on stack SPAT ([a][b][c]) Do a pattern subsititution following this rule I<[c] =~s/[a]/[b]/> SPATG ([a][b][c]) Do a pattern subsititution following this rule I<[c] =~s/[a]/[b]/g> SPATI ([a][b][c]) Do a pattern subsititution following this rule I<[c] =~s/[a]/[b]/i> (case insensitive) SPATGI ([a][b][c]) Do a pattern subsititution following this rule I<[c] =~s/[a]/[b]/gi> (case insensitive) PRINTF ([a][b]...[x]) use the format present in [a] to print the value [b] to [x] the format is the same as (s)printf PACK ([a][b]...[x]) Do an unpack on variable [b] to [x] using format [b] UNPACK ([a][b]) Do an unpack on variable [b] using format [a] ISNUM ([a]) Test if a is a NUMBER return 1 if success ( [a] [1|0] ) Keep the value on the stack ISNUMD ([a]) Test if a is a NUMBER return 1 if success ( [1|0] ) Remove the value from the stack ISINT ([a]) Test if a is a INTEGER (natural number ) Return 1 if success ( [a] [1|0] ) Keep the value on the stack ISINTD ([a]) Test if a is a INTEGER (natural number ) Return 1 if success ( [1|0] ) Remove the value from the stack ISHEX ([a]) Test if a is a HEXADECIMAL (hex starting with 0x or 0X or # ) Return 1 if success ( [a] [1|0] ) Keep the value on the stack ISHEXD ([a]) Test if a is a HEXADECIMAL (hex starting with 0x or 0X or # ) Return 1 if success ( [1|0] ) Remove the value from the stack Stack operators --------------- SWAP ([a][b]) ([b][a]) OVER ([a][b]) ([a][b][a]) DUP ([a]) ([a][a]) DDUP ([a][b]) ([a][b][a][b]) ROT ([a][b][c]) ([b][c][a]) RROT ([a][b][c]) ([c][a][b]) DEPTH ([r1]...) ([re1]...[nbr]) Return the number of elements in the statck POP ([a][b]) ([a]) POPN ([a][b][c]...[x]) ([l]...[x]) remove [b] element from the stack (starting at [c]) SWAP2 ([a][b][c]) ([a][c][b]) ROLL ([a][b][c][d][e][n]) ([a][c][d][e][b]) rotate the [n] element of the stack (here [n]=4) if [n] =3 it is equivalent to ROT PICK ([a][b][c][d][e][n]) ([a][b][c][d][e][b]) copy element from depth [n] on top GET ([a][b][c][d][e][n]) ([a][b][c][d][e][b]) get element from depth [n] and put on top PUT ([a][b][c][d][v][n]) ([a][v][b][c][d]) put element [v] at level [n] (here [n]=3) DEL ([a][b]) delete [b] element on the stack from level [a] [a] and [b] is get in absolute value KEEPN ([a][b]) keep [b] element(s) on the stack from level [a] (and delete all other elements) [a] and [b] is get in absolute value PRESERVE ([a][b]) keep element(s) on the stack from level [a] to level [b] (and delete all other elements) [a] and [b] is get in absolute value COPY ([a][b]) copy element(s) on the stack from level [a] to level [b] [a] and [b] is get in absolute value FIND ([a]) get the level of stack containing [a] SEARCH ([a]) get the level of stack containing the REGEX [a] SEARCHI ([a]) get the level of stack containing the REGEX [a] ( case insensitive ) SEARCHK ([a]) keep only level of stack matching the REGEX [a] SEARCHIK ([a]) keep only level of stack matching the REGEX [a] ( case insensitive ) KEEP ([a][b][c][d][e][n]) remove all elements of the stack except the element at deepth |n| Dictionary operators -------------------- WORDS () ([a])return as one stack element the list of WORD in DICT separated by a | VARS () ([a])return as one stack element the list of VARIABLE in VAR separated by a | INC ([a]) () increment (+1) the value of variable [a] DEC ([a]) () decrement (-1) the value of variable [a] VARIABLE ([a]) () create a entry in VAR for the variable [a] ! ([a][b]) store the value [a] in the variable [b] !! ([a][b][c]...[n] [var]) put and delete 'n' element(s) from the stack in the variable 'var' 'n' is in absolute value !!C ([a][b][c]...[n] [var]) copy 'n' element(s) from the stack in the variable 'var' 'n' is in absolute value !!! ([a][b][c]...[n1] [n2] [var]) put and delete element(s) from the stack in the variable 'var' starting at element 'a' to element 'b' 'a' and 'b' in absolute value if 'a' > 'b' keep the reverse of selection (boustrophedon) !!!C ([a][b][c]...[n] [var]) copy 'element(s) from the stack in the variable 'var' starting at element 'a' to element 'b' 'a' and 'b' in absolute value if 'a' > 'b' keep the reverse of selection (boustrophedon) @ ([a]) ([a]) return the value of the variable [a] : xxx yyy ; create a new word (sub) into the dictionary with the xxx "code" and name yyy : xxx yyy PERLFUNC execute the PERL function yyy with parameter(s) yyy the default name space is "main::" It is possible tu use a specific name space : xxx yyy PERL execute the PERL code xxx ; yyy Return Stack operators ---------------------- >R ([a]) put ^a$ on the return stack R> () remove first element from the return stack and copy on the normal RL () return the depth of the return stack R@ () copy return stack ion normal stack LOOP and DECISION operators --------------------------- [a] IF [..xxx] THEN Test the element on top of stack if ==0, execute 'xxx' block The loop is executed always one time [a] IF [...zzz...] ELSE [..xxx...] THEN Test the element on top of stack if ==0, execute 'xxx' block if != 0 execute 'zzz' block The loop is executed always one time BEGIN xxx WHILE zzz REPEAT Execute 'xxx' block Test the element on top of stack if ==0 execute 'zzz' block and branch again to BEGIN if != 0 end the loop The loop is executed always one time [a] [b] DO [...xxx...] LOOP ([a][b]) process block [...xxx...] with iterator from value [b] untill [a] value, with increment of 1; The iterator variable is '_I_' (read only and scoop only the DO ... LOOP block) [a] [b] DO [...xxx...] [c] +LOOP ([a][b]) process block [...xxx...] with iterator from value [b] untill [a] value, with increment of [c]; The iterator variable is '_I_' (read only and scoop only the DO ... LOOP block) EXAMPLES use Parse::RPN; $test ="3,5,+"; $ret = rpn($test); # $ret = 8 $test = "Hello World,len,3,+"; $ret = rpn($test); # $ret = 14 $test = "'Hello,World',len,3,+"; $ret = rpn($test); # $ret = 14 $test = "'Hello,World,len,3,+"; ---------^-----------^- $ret = rpn($test); # $ret = 8 with a warning because the stack is not empty ([Hello] [8]) # be care to close your quoted string $test = "'Hello world','or',PAT,'or',EQ,IF,'string contain or',ELSE,'No or in string',THEN" $ret = rpn($test); # $ret = "Contain a coma" $test = "'Hello world','or',TPAT,IF,'string contain or',ELSE,'No or in string',THEN"; $ret = rpn($test); # $ret = "string contain or" $test = "3,10,/,5,+,82,*,%b,PRINTF"; $ret = rpn($test); # $ret = "110110010" $test = "3,10,/,5,+,82,*,%016b,PRINTF"; $ret = rpn($test); # $ret = "0000000110110010" $test = "55,N,pack,B32,unpack,^0+(?=\d), ,spat,'+',ds"; $ret = rpn($test); # $ret = 110111 $test = "7,3,/,10,3,/,%d %f,PRINTF"; @ret = rpn($test); # @ret = 2 3.333333 $test = "VARIABLE,a,0,a,!,##,b,BEGIN,bbbb,a,INC,a,@,4,<,WHILE,####,a,@,****,REPEAT"; @ret =rpn($test); # @ret = ## b bbbb #### 1 **** bbbb #### 2 **** bbbb #### 3 **** bbbb $test = "VARIABLE,a,0,a,!,z,0,5,-1,DO,a,INC,6,1,2,DO,A,_I_,+LOOP,#,+LOOP,##,a,@"; @ret =rpn($test); # @ret = z A 3 A 5 A 7 # A 3 A 5 A 7 # A 3 A 5 A 7 # A 3 A 5 A 7 # A 3 A 5 A 7 # A 3 A 5 A 7 # ## 6 $test = "1,2,3,4,5,6,7,8,9,3,KEEP"; ret =rpn($test); # @ret = 7 $test = "1,2,3,4,5,6,7,8,9,30,KEEP"; ret =rpn($test); # @ret = 1,2,3,4,5,6,7,8,9 $test = "h,g,f,e,d,c,b,a,4,3,DEL"; ret =rpn($test); # @ret = h,c,b,a $test = "h,g,f,e,d,c,b,a,4,3,KEEPN""; ret =rpn($test); # @ret = g,f,e,d sub Test { my $a = shift; my $b = shift; my $c = $a/$b; print "a=$a\tb=$b\ttotal=$c\n"; return $c; } $test = ":,5,6,Test,PERLFUNC"; @ret =rpn($test); # call the function "Test" from the main package (the caller) with parameter 5,6 and return result (in @ret) $test = ":,05,11,01,0,0,0,Time::Local::timelocal,PERLFUNC"; @ret =rpn($test); # @ret = 1133391600 $test = "1,2,3,+,:, my $b=7, "open LOG , qq{ >/tmp/log }",print LOG time,PERL"; @ret =rpn($test); # @ret = 1,5 and the file /tmp/log contain a line with the tick time. $test = "11,55,*,5,2,401,+,:,my $b=,SWAP,CAT, "open LOG , qq{ >/tmp/log }",print LOG $b.qq{ \n },PERL" @ret =rpn($test); # @ret =1 2 3 1 (the latest 1 is the succes result return) and the file /tmp/log contain a line with 403 + a cariage return $test = 'mb,tb,gb,mb,kb,4,V,!!,12,9,6,3,4,R,!!,V,R,"TPATI",LOOKUP' @ret =rpn($test); # @ret = 6 $test = '5,1,2,3,4,5,5,V,!!," "," ",ok," ",nok,5,R,!!,V,R,"<=",LOOKUPP' @ret =rpn($test); # @ret = nok $test = '3,1,2,3,4,5,5,V,!!,a,b,ok,d,nok,5,R,!!,"<","<","<","<","<",5,O,!!,V,R,O,LOOKUPOPP' @ret =rpn($test); # @ret = d The small tool 'RPN.pl' provide an easy interface to test quickly an RPN. This include two test functions named 'save' and 'restore' Try RPN.pl to get a minimal help. Take a look to the minimalistic code, and put RPN.pl in your path. Sample of use: RPN.pl -r '1,2,3,:,123,100,+,7,*,test,save,PERLFUNC' save in file '/tmp/test' the value '1561' (whithout CR/LF) and return 1 2 3 1 AUTHOR Fabrice Dulaunoy It is a full rewrite from the version 1.xx to allow DICTIONNARY use and STRUCTURE control Thanks to the module Math::RPN from Owen DeLong, for the idea of using RPN in a config file SEE ALSO Math-RPN from Owen DeLong, TODO Error processing, stack underflow... CREDITS Thank's to Stefan Moser for the idea to call a perl function from the rpn() and also for pin-pointing an error in stack return. LICENSE Under the GNU GPL2 This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA Parse::RPN Copyright (C) 2004 2005 2006 2007 DULAUNOY Fabrice Parse::RPN comes with ABSOLUTELY NO WARRANTY; for details See: L This is free software, and you are welcome to redistribute it under certain conditions;