return searchRes, true, nil
		}
	}

	// Not under any configured base DN so return false.
	return searchRes, false, nil
}

// Bind - binds to ldap, searches LDAP and returns the distinguished name of the
// user and the list of groups.
func (l *Config) Bind(username, password string) (*xldap.DNSearchResult, []string, error) {
	conn, err := l.LDAP.Connect()
	if err != nil {
		return nil, nil, err
	}

	inclusive := mant[0]&2 == 0 // test bit 1 since original mantissa was shifted by 1

	// Now we can figure out the minimum number of digits required.
	// Walk along until d has distinguished itself from upper and lower.
	for i, m := range d.mant {
		l := lower.at(i)
		u := upper.at(i)

		// Okay to round down (truncate) if lower has a different digit

// An optional base-10 “e” or base-2 “p” (or their upper-case variants)
// exponent may be provided as well, except for hexadecimal floats which
// only accept an (optional) “p” exponent (because an “e” or “E” cannot
// be distinguished from a mantissa digit). If the exponent's absolute value
// is too large, the operation may fail.
// The entire string, not just a prefix, must be valid for success. If the

	// and we know that rounding up falls within the bound.
	var upperdelta uint8

	// Now we can figure out the minimum number of digits required.
	// Walk along until d has distinguished itself from upper and lower.
	for ui := 0; ; ui++ {
		// lower, d, and upper may have the decimal points at different
		// places. In this case upper is the longest, so we iterate from

	// encodes as ^uint64(0x42_43_44).
	// LMS-substrings can never start or end with 0xFF.
	// Adding 1 ensures the encoded byte sequence never
	// starts or ends with 0x00, so that present bytes can be
	// distinguished from zero-padding in the top bits,
	// so the length need not be separately encoded.
	// Inverting the bytes increases the chance that a
	// 4-byte encoding will still be ≥ len(text).

// substrings have matching “length,” the text need not be read at all.
// The definition of “very short” is that the text bytes must pack into a uint32,
// and the unsigned encoding e must be ≥ len(text), so that it can be
// distinguished from a valid length.
func length_8_32(text []byte, sa []int32, numLMS int) {
	end := 0 // index of current LMS-substring end (0 indicates final LMS-substring)

	// The encoding of N text bytes into a “length” word

	}
	klog.V(3).Infof("Triggered API priority and fairness config reloading because priority level %s is undesired and idle", plName)
	cfgCtlr.configQueue.Add(0)
}

// computeFlowDistinguisher extracts the flow distinguisher according to the given method
func computeFlowDistinguisher(rd RequestDigest, method *flowcontrol.FlowDistinguisherMethod) string {
	if method == nil {
		return ""
	}
	switch method.Type {

// The caller should report this error instead of continuing to probe
// other possible module paths.
//
// TODO(golang.org/issue/31730): See if we can invert this. (Return a
// distinguished error for “repo not found” and treat everything else
// as terminal.)
type VCSError struct {
	Err error
}

func (e *VCSError) Error() string { return e.Err.Error() }

====

[[targeting-different-platforms]]
== Targeting different platforms

It is common for a library to target different platforms.
In the Java ecosystem, we often see different artifacts for the same library, distinguished by a different _classifier_.
A typical example is Guava, which is published as this:

- `guava-jre` for JDK 8 and above
- `guava-android` for JDK 7

			return nil, err
		}

		// The kernel relocation symbol (the mapping start address) can be either
		// _text or _stext. When profiles are generated by `perf`, which one was used is
		// distinguished by the mapping name for the kernel image:
		// '[kernel.kallsyms]_text' or '[kernel.kallsyms]_stext', respectively. If we haven't
		// been able to parse it from the mapping, we default to _stext.
		if relocationSymbol == "" {