//
	// It should be a comma-separated list of CIDRs, e.g. `0.0.0.0/0` to
	// allow full access (the default) or `18.0.0.0/8,56.0.0.0/8` to allow
	// access only from the CIDRs currently allocated to MIT & the USPS.
	//
	// Not all cloud providers support this annotation, though AWS & GCE do.

	//
	// It should be a comma-separated list of CIDRs, e.g. `0.0.0.0/0` to
	// allow full access (the default) or `18.0.0.0/8,56.0.0.0/8` to allow
	// access only from the CIDRs currently allocated to MIT & the USPS.
	//
	// Not all cloud providers support this annotation, though AWS & GCE do.

#tf_type.shape<>, #tf_type.shape<>, #tf_type.shape<>, #tf_type.shape<>, #tf_type.shape<>, #tf_type.shape<>], device = "/job:localhost/replica:0/task:0/device:CPU:0", operandSegmentSizes = array<i32: 1, 1, 0, 8, 8>, resultSegmentSizes = array<i32: 0, 0, 0, 8>} : (tensor<?x!tf_type.string>, tensor<0x!tf_type.string>, tensor<!tf_type.string>, tensor<!tf_type.string>, tensor<!tf_type.string>, tensor<!tf_type.string>, tensor<!tf_type.string>, tensor<!tf_type.string>, tensor<!tf_type.string>, tensor<!tf_type.string>,...

			rawNodeIPs:     nil,
			bindAddress:    "0.0.0.0",
			expectedFamily: v1.IPv4Protocol,
			expectedIPv4:   "127.0.0.1",
			expectedIPv6:   "::1",
		},
		{
			name:           "Bind address 0.0.0.0 and node with IPv4 InternalIP set",
			rawNodeIPs:     []net.IP{netutils.ParseIPSloppy("192.168.1.1")},
			bindAddress:    "0.0.0.0",
			expectedFamily: v1.IPv4Protocol,
			expectedIPv4:   "192.168.1.1",

	ALDEORAD:  3<<30 | 0x1c5<<21 | 0x08<<10,
	ALDEORAW:  2<<30 | 0x1c5<<21 | 0x08<<10,
	ALDEORAH:  1<<30 | 0x1c5<<21 | 0x08<<10,
	ALDEORAB:  0<<30 | 0x1c5<<21 | 0x08<<10,
	ALDEORALD: 3<<30 | 0x1c7<<21 | 0x08<<10,
	ALDEORALW: 2<<30 | 0x1c7<<21 | 0x08<<10,
	ALDEORALH: 1<<30 | 0x1c7<<21 | 0x08<<10,
	ALDEORALB: 0<<30 | 0x1c7<<21 | 0x08<<10,
	ALDEORD:   3<<30 | 0x1c1<<21 | 0x08<<10,
	ALDEORW:   2<<30 | 0x1c1<<21 | 0x08<<10,

# ADR-0002 - Avoid using Java serialization

## Date

2012-12-01

## Context

In Gradle we often need to serialize in-memory objects for caching, or to transmit them across process barriers, etc.
Java serialization is one way to implement this, however, despite its simplicity of implementation, it has several drawbacks:

- **Performance:**
Java's built-in serialization mechanism is often slower compared to other serialization solutions.

# ADR-0004 - Introduce a UI architecture module to the core platform

## Date

2024-02-07

## Context

The Gradle core platform provides many services to the Gradle platforms and builds logic. One such group of services allows logic to interact with the build user, to provide diagnostics, progress information, prompt for questions, and so on. Currently, these services are part of the core platform runtime architecture module.

# ADR-0001 - Use Architectural Decision Records

## Date

2023-12-01

## Context

In a distributed team with many subteams, the best solution to communicate decisions is to use a format accessible by everyone in charge of development.

We use *Specification* and *Discovery* documents stored in Google Drive, but they present some downsides:

00000330  00 0a 0d 00 00 06 03 01  02 40 00 00 16 03 01 00  |.........@......|
00000340  04 0e 00 00 00                                    |.....|
>>> Flow 3 (client to server)
00000000  16 03 01 01 fd 0b 00 01  f9 00 01 f6 00 01 f3 30  |...............0|
00000010  82 01 ef 30 82 01 58 a0  03 02 01 02 02 10 5c 19  |...0..X.......\.|
00000020  c1 89 65 83 55 6f dc 0b  c9 b9 93 9f e9 bc 30 0d  |..e.Uo........0.|

    %0 = "quantfork.stats"(%arg0) {layerStats = dense<[0.000000e+00, 0.1]> : tensor<2xf32>} : (tensor<2x2xf32>) -> tensor<2x2xf32>
    %1 = "tf.PartitionedCall"(%0, %cst) {_tfl_quant_trait = "fully_quantizable", config = "", config_proto = "", executor_type = "", f = @composite_matmul_fn_2} : (tensor<2x2xf32>, tensor<2x2xf32>) -> tensor<2x2xf32>
    %2 = "quantfork.stats"(%1) {layerStats = dense<[0.000000e+00, 0.2]> : tensor<2xf32>} : (tensor<2x2xf32>) -> tensor<2x2xf32>