github.com/vmware/govmomi@v0.51.0/ovf/parser.go

// © Broadcom. All Rights Reserved.
// The term “Broadcom” refers to Broadcom Inc. and/or its subsidiaries.
// SPDX-License-Identifier: Apache-2.0

package ovf

import (
	"math"
	"regexp"
	"strconv"
	"strings"
)

// These are used to validate the overall structure of the string being parsed and to differentiate tokens as we are
// processing them
var (
	blankRegexp         = regexp.MustCompile(`[[:blank:]]`)
	validIntegerRegexp  = regexp.MustCompile(`^([1-9]\d*)$`)
	validExponentRegexp = regexp.MustCompile(`^([1-9]\d*\^[1-9]\d*)$`)
	validByteUnitRegexp = regexp.MustCompile(`((^|kilo|kibi|mega|mebi|giga|gibi)byte(s?)$)`)
	validCapacityRegexp = regexp.MustCompile(`^[[:blank:]]*((([1-9]\d*\^[1-9]\d*)|([1-9]\d*))($|[[:blank:]]*\*[[:blank:]]*))*(([a-zA-Z]*(b|B)(y|Y)(t|T)(e|E)(s|S)?)($|([[:blank:]]*\*[[:blank:]]*(([1-9]\d*\^[1-9]\d*)|([1-9]\d*)))*))?$`)
)

// We only handle kilo, kibi, mega, mebi, giga, gibi prefixes due to size constraints of int64/uint64, but more
// importantly because prefixes larger than giga & gibi don't make sense for our use-case
var prefixMultipliers = map[string]int64{
	"byte":     1,                      // byte
	"kilobyte": 1 * 1000,               // byte * 1000
	"kibibyte": 1 * 1024,               // byte * 1024
	"megabyte": 1 * 1000 * 1000,        // byte * 1000 * 1000 = kilobyte * 1000
	"mebibyte": 1 * 1024 * 1024,        // byte * 1024 * 1024 = kibibyte * 1024
	"gigabyte": 1 * 1000 * 1000 * 1000, // byte * 1000 * 1000 * 1000 = kilobyte * 1000 * 1000 = megabyte * 1000
	"gibibyte": 1 * 1024 * 1024 * 1024, // byte * 1024 * 1024 * 1024 = kibibyte * 1024 * 1024 = mebibyte * 1024
}

// ParseCapacityAllocationUnits validates the string s is a valid programmatic unit with respect to the base unit 'byte'
// and parses the string to return the number of bytes s represents
func ParseCapacityAllocationUnits(s string) int64 {
	// Any strings which don't match against the regular expression are deemed invalid and zero is returned as the result
	if !validCapacityString(s) {
		return 0
	}
	var capacityBytes int64 = 1
	// Remove any whitespace in s and lowercase any alphabetic characters. Removal of whitespace is done after
	// validating against the regular expression because whitespace is valid for the most part, but is not valid
	// for exponential terms, e.g 2 ^ 10
	s = strings.ToLower(blankRegexp.ReplaceAllString(s, ""))
	// Split s on multiplication operator (*) so that we can just calculate integer multipliers. Each token will
	// then be either an integer, an exponential term to be converted to an integer, or a unit term to be converted
	// to an integer
	tokens := strings.Split(s, "*")

	// Loop through all tokens and convert any to integers if necessary and use to compute a running product
	for _, token := range tokens {
		switch {
		// "" should be treated identically to "byte". capacityBytes is already set to 1 so there is nothing to do
		case len(token) == 0:
			continue
		case validByteUnitString(token):
			capacityBytes = capacityBytes * prefixMultipliers[strings.TrimSuffix(token, "s")]
		case validExponentString(token):
			p := strings.Split(token, "^")
			b, _ := strconv.ParseInt(p[0], 10, 64)
			e, _ := strconv.ParseInt(p[1], 10, 64)
			capacityBytes = capacityBytes * int64(math.Pow(float64(b), float64(e)))
		case validIntegerString(token):
			n, _ := strconv.ParseInt(token, 10, 64)
			capacityBytes = capacityBytes * n
		default:
			// This should be unreachable. validCapacityString should have filtered out anything that cannot be
			// matched by the non-default cases
			capacityBytes = 0
		}
	}
	return capacityBytes
}

// validIntegerString matches the string s against the regular expression `^([1-9]\d*)$`; i.e. s should be of the form:
// any non-zero digit ([1-9]), followed by zero or more digits (\d*)
func validIntegerString(s string) bool {
	return validIntegerRegexp.MatchString(s)
}

// validExponentString matches the string s against the regular expression `^([1-9]\d*\^[1-9]\d*)$`; i.e. s should be of
// the form: any non-zero digit ([1-9]), followed by a caret (^) followed by any non-zero digit ([1-9]), followed by zero
// or more digits (\d*)
func validExponentString(s string) bool {
	return validExponentRegexp.MatchString(s)
}

// validByteUnitString matches the string s against a regular expression which only allows a unit of byte
// (optionally plural) with a valid decimal or binary prefix. See prefixMultipliers
func validByteUnitString(s string) bool {
	return validByteUnitRegexp.MatchString(s)
}

// validCapacityString matches the string s against the regular expression validCapacityRegexp and verifies that s is a
// valid programmatic unit with respect to the base unit 'byte'.
//
// Per the OVF schema defined in DSP8023: "If not specified default value is bytes. Value shall match a recognized value
// for the UNITS qualifier in DSP0004"
//
// DSP004 defines a programmatic unit as:
//
// programmatic-unit = [ sign ] *S unit-element *( *S unit-operator *S unit-element )
// sign = HYPHEN
// unit-element = number / [ prefix ] base-unit [ CARET exponent ]
// unit-operator = "*" / "/"
// number = floatingpoint-number / exponent-number
//
// ; An exponent shall be interpreted as a floating point number
// ; with the specified decimal base and exponent and a mantissa of 1
// exponent-number = base CARET exponent
// base = integer-number
// exponent = [ sign ] integer-number
//
// ; An integer shall be interpreted as a decimal integer number
// integer-number = NON-ZERO-DIGIT *( DIGIT )
//
// ; A float shall be interpreted as a decimal floating point number
// floatingpoint-number = 1*( DIGIT ) [ "." ] *( DIGIT )
//
// ; A prefix for a base unit (e.g. "kilo"). The numeric equivalents of
// ; these prefixes shall be interpreted as multiplication factors for the
// ; directly succeeding base unit. In other words, if a prefixed base
// ; unit is in the denominator of the overall programmatic unit, the
// ; numeric equivalent of that prefix is also in the denominator
// prefix = decimal-prefix / binary-prefix
//
// ; SI decimal prefixes as defined in ISO 1000
// decimal-prefix =
//
//	  "deca" ; 10^1
//	/ "hecto" ; 10^2
//	/ "kilo" ; 10^3
//	/ "mega" ; 10^6
//	/ "giga" ; 10^9
//	/ "tera" ; 10^12
//	/ "peta" ; 10^15
//	/ "exa" ; 10^18
//	/ "zetta" ; 10^21
//	/ "yotta" ; 10^24
//	/ "deci" ; 10^-1
//	/ "centi" ; 10^-2
//	/ "milli" ; 10^-3
//	/ "micro" ; 10^-6
//	/ "nano" ; 10^-9
//	/ "pico" ; 10^-12
//	/ "femto" ; 10^-15
//	/ "atto" ; 10^-18
//	/ "zepto" ; 10^-21
//	/ "yocto" ; 10^-24
//
// ; IEC binary prefixes as defined in IEC 80000-13
// binary-prefix =
//
//	  "kibi" ; 2^10
//	/ "mebi" ´ ; 2^20
//	/ "gibi" ; 2^30
//	/ "tebi" ; 2^40
//	/ "pebi" ; 2^50
//	/ "exbi" ; 2^60
//	/ "zebi" ; 2^70
//	/ "yobi" ; 2^80
//
// ; The name of a base unit
// base-unit = standard-unit / extension-unit
//
// ; The name of a standard base unit
// standard-unit = UNIT-IDENTIFIER
//
// ; The name of an extension base unit. If UNIT-IDENTIFIER begins with a
// ; prefix (see prefix ABNF rule), the meaning of that prefix shall not be
// ; changed by the extension base unit (examples of this for standard base
// ; units are "decibel" or "kilogram")
// ; extension-unit = org-id COLON UNIT-IDENTIFIER
//
// ; org-id shall include a copyrighted, trademarked, or otherwise unique
// ; name that is owned by the business entity that is defining the
// ; extension unit, or that is a registered ID assigned to the business
// ; entity by a recognized global authority. org-id shall not begin with
// ; a prefix (see prefix ABNF rule)
// org-id = UNIT-IDENTIFIER
// UNIT-IDENTIFIER = FIRST-UNIT-CHAR [ *( MID-UNIT-CHAR )
// LAST-UNIT-CHAR ]
// FIRST-UNIT-CHAR = UPPERALPHA / LOWERALPHA / UNDERSCORE
// LAST-UNIT-CHAR = FIRST-UNIT-CHAR / DIGIT / PARENS
// MID-UNIT-CHAR = LAST-UNIT-CHAR / HYPHEN / S
//
// DIGIT = ZERO / NON-ZERO-DIGIT
// ZERO = "0"
// NON-ZERO-DIGIT = "1"-"9"
// HYPHEN = U+002D ; "-"
// CARET = U+005E ; "^"
// COLON = U+003A ; ":"
// UPPERALPHA = U+0041-005A ; "A" ... "Z"
// LOWERALPHA = U+0061-007A ; "a" ... "z"
// UNDERSCORE = U+005F ; "_"
// PARENS = U+0028 / U+0029 ; "(", ")"
// S = U+0020 ; " "
//
// This definition is further restricted as such a broad definition by the above grammar does not make sense in the
// context of virtual disk capacity.
//
// We do not allow for negative values, division operations, floating-point numbers, negative exponents, nor the use of
// multiple units. Furthermore, we limit the allowed decimal and binary prefixes. This gives us:
//
// programmatic-unit =
//
//	   number
//		/ [prefix] base-unit
//		/ number *( *S unit-operator *S number) *S unit-operator *S [prefix] base-unit
//		/ [prefix] base-unit *( *S unit-operator *S number)
//		/ number *( *S unit-operator *S number) *S unit-operator *S [prefix] base-unit *( *S unit-operator *S number)
//
// unit-operator = "*"
// number = integer-number / exponent-number
// exponent-number = base CARET exponent
// base = integer-number
// exponent = integer-number
// integer-number = NON-ZERO-DIGIT *( DIGIT )
// prefix = decimal-prefix / binary-prefix
//
// decimal-prefix =
//
//	  "kilo" ; 10^3
//	/ "mega" ; 10^6
//	/ "giga" ; 10^9
//
// binary-prefix =
//
//	  "kibi" ; 2^10
//	/ "mebi" ; 2^20
//	/ "gibi" ; 2^30
//
// This function and the regular expression validCapacityRegexp are used to verify that the string we are parsing follows
// our above restricted grammar
func validCapacityString(s string) bool {
	// Integer followed by a trailing '*' is not handled by the regular expression and so is explicitly checked
	return validCapacityRegexp.MatchString(s) && !strings.HasSuffix(s, "*")
}