Add many response types

runtime/vm/service/protocol.md

 # Dart VM Service Protocol
 
 Description
 How to start
 JSON
 Websocket
 
 ## Types
 
 Every response returned by the VM Service has the <code>type</code> property.  This allows the client distinguish between different kinds of responses.  For example, global information about the VM is encoded in an response of type [VM](#VM) and information about an isolate is encoded in an response of type [Isolate](#Isolate).
@@ skipping 12 matching lines @@
 
     {
       type: "@Isolate",
       id: "isolates/123",
       name: "worker"
       entry: ...
       heaps: ...
       topFrame: ...
       ...
     }
- 
+
 ## IDs
 
 Most responses returned by the VM Service have an <code>id</code> property.  An id is used to request an object from the VM.
 
 An id is either _global_ or _relative_.  Global ids can be requested from the VM directly by requesting the uri <code>/{global id}</code>.
 
 The following is a list of known,  fixed global ids:
 
 | id | uri | type
 | --- | --- | ---
@@ skipping 53 matching lines @@
 | pauseOnExit | bool |
 | pauseEvent? | [DebuggerEvent](#DebuggerEvent) |
 | rootLib | [@Library](#atLibrary) |
 | timers | ??? |
 | tagCounters | ??? |
 | error? | [Error](#Error) |
 | canonicalTypeArguments | | kill? |
 | libs | List of [@Library](#atLibrary) |
 | features | List of String |
 
-
 ### <a name="atLibrary"></a>@Library
 
 | keys | values | comments
 | --- | --- | ---
 | type | "@Library" |
 | id | String |
 | user_name | String |
 | name | String
 | url | String
 
 ### <a name="Library"></a>Library
 
 | keys | values | comments
 | --- | --- | ---
 | type | "Library" |
 | id | String |
 | user_name | String |
 | name | String
 | classes | List of [@Class](#atClass) |
 | imports | List of [@Library](#atLibrary) |
 | variables | List of ... |
 | functions | List of [@Function](#atFunction) |
 | scripts | List of [@Script](#atScript) |
 
+### <a name="atClass"></a>@Class
+| keys | values | comments
+| --- | --- | ---
+| type | "@Class" |
+| id | String |
+| user_name | String |
+| name | String |
+
+### <a name="Class"></a>Class
+| keys | values | comments
+| --- | --- | ---
+| type | "@Class" |
+| id | String |
+| user_name | String |
+| name | String |
+| error | [Error](#Error) | (optional) Error encountered during class finalization

[turnidge, 2014/08/21 00:33:01]

I have been using ? for optional properties.  What do you think?

[Cutch, 2014/08/26 20:29:28]

Done.

+| implemented | bool |
+| abstract | bool |
+| patch | bool |
+| finalized | bool |
+| const | bool |
+| super | [@Class](#atClass) | (optional) Super class
+| library | [@Library](#atLibrary) | Owning library
+| script | [@Script](#atScript) | (optional) Script containing class source
+| tokenPos | int | starting token position of class source in script
+| endTokenPos | int | end token position of class source in script
+| interfaces | List of [@Class](#atClass) | interfaces this class has implemented
+| fields | List of [@Field](#atField) |
+| functions | List of [@Function](#atFunction) |
+| subclasses | List of [@Class](#atClass) | classes which extend this class.
+| canonicalTypes | [@TypeList] | kill?
+| allocationStats | ClassHeapStats |
+
+### <a name="atFunction"></a>@Function
+| keys | values | comments
+| --- | --- | ---
+| type | "@Function" |
+| id | String |
+| user_name | String |
+| name | String |
+| owningLibrary | [@Library](#atLibrary) | Set for non-top level functions

[turnidge, 2014/08/21 00:33:01]

optional.

[Cutch, 2014/08/26 20:29:28]

Done.

+| owningClass | [@Class](#atClass) | Set for non-top level functions

[turnidge, 2014/08/21 00:33:01]

optional.

[Cutch, 2014/08/26 20:29:28]

Done.

+| parent | [@Function](#atFunction) | (optional) Parent function
+| kind | String |
+
+### <a name="Function"></a>Function
+| keys | values | comments
+| --- | --- | ---
+| type | "@Function" |
+| id | String |
+| user_name | String |
+| name | String |
+| owningLibrary | [@Library](#atLibrary) | Set for non-top level functions
+| owningClass | [@Class](#atClass) | Set for non-top level functions

[turnidge, 2014/08/21 00:33:01]

ditto these two.

[Cutch, 2014/08/26 20:29:28]

Done.

+| parent | [@Function](#atFunction) | (optional) Parent function
+| kind | String |
+| is_static | bool |
+| is_const | bool |
+| is_optimizable | bool |
+| is_inlinable | bool |
+| usage_counter | int |
+| optimized_call_site_count | int |
+| deoptimizations | int |
+| script | [@Script](#atScript) | (optional) Script containing function source
+| tokenPos | int | starting token position of function source in script
+| endTokenPos | int | end token position of function source in script
+| unoptimized_code | [@Code](#atCode) |
+| code | [@Code](#atCode) | Current code
+
+### <a name="atField"></a>@Field
+| keys | values | comments
+| --- | --- | ---
+| type | "@Field" |
+| id | String |
+| user_name | String |
+| name | String |
+| value | Instance | (optional) value associated with static field <-- do we want to include this in a field reference?

[turnidge, 2014/08/21 00:33:01]

Good catch, probably not.

[Cutch, 2014/08/26 20:29:28]

Acknowledged.

+| owner | [@Library](#atLibrary) | Owning library <-- handling of owner is inconsistent with Function
+| owner | [@Class](#atClass) | Owning class <-- handling of owner is inconsistent with Function

[turnidge, 2014/08/21 00:33:01]

I have been thinking of using a comma when multiple types are allowed:

| owner | [@Library](#atLibrary), [@Class](#atClass)

[Cutch, 2014/08/26 20:29:28]

Done.

+| declared_type | [@AbstractType](#atAbstractType) |
+| static | bool |
+| final | bool |
+| const | bool |
+
+### <a name="Field"></a>Field
+| keys | values | comments
+| --- | --- | ---
+| type | "Field" |
+| id | String |
+| user_name | String |
+| name | String |
+| value | Instance | (optional) value associated with static field <-- do we want to include this in a field reference?
+| owner | [@Library](#atLibrary) | Owning library <-- handling of owner is inconsistent with Function
+| owner | [@Class](#atClass) | Owning class <-- handling of owner is inconsistent with Function
+| declared_type | [@AbstractType](#atAbstractType) |
+| static | bool |
+| final | bool |
+| const | bool |
+| guard_nullable | bool | can this field hold a null?
+| guard_class | String OR [@Class](#atClass) | "unknown", "dynamic", or a class
+| guard_length | String OR int | "unknown", "variable", or length of array
+| script | [@Script](#atScript) | (optional) Script containing field source
+| tokenPos | int | starting token position of field source in script
+
+### <a name="atCode"></a>@Code
+| keys | values | comments
+| --- | --- | ---
+| type | "@Code" |
+| id | String |
+| user_name | String |
+| name | String |
+| start | String | starting address of code
+| end | String | ending address of code
+| isOptimized | bool |
+| isAlive | bool |
+| kind | String
+| function | [@Function](#atFunction) |
+
+### <a name="Code"></a>Code
+| keys | values | comments
+| --- | --- | ---
+| type | "@Code" |
+| id | String |
+| user_name | String |
+| name | String |
+| start | String | starting address of code
+| end | String | ending address of code
+| isOptimized | bool |
+| isAlive | bool |
+| kind | String
+| function | [@Function](#atFunction) |
+| object_pool | List of [@Object](Object) |
+| disassembly | List of String | See note below on disassembly list format
+
+*Disassembly list format*
+| index | value | description
+| --- | --- | --- |
+| 0 | String | Address of instruction
+| 1 | String | Hex encoding of instruction
+| 2 | String | Human encoding of instruction
+| 0 + (3 * K) | String | Address of Kth instruction
+| 1 + (3 * K) | String | Hex encoding of instruction of Kth instruction
+| 2 + (3 * K) | String | Human encoding of instruction of Kth instruction
+
+### <a name="Error"></a>Error
+| keys | values | comments
+| --- | --- | ---
+| type | "Error" |
+| id | String | always empty
+| kind | String |
+| message | String |
+
+### <a name="ClassHeapStats"></a>ClassHeapStats
+| keys | values | comments
+| --- | --- | ---
+| type | "ClassHeapStats" |
+| id | String |
+| class | [@Class](#atClass) |
+| new | List of int | Allocation statistics for new space. See note below on allocation statistics list format.
+| old | List of int | Allocation statistics for old space. See note below on allocation statistics list format.
+| promotedInstances | int | number of instances promoted at last new-space GC.
+| promotedBytes | int | number of bytes promoted at last new-space GC.
+
+*Allocation statistics list format*
+| index | value | description
+| --- | --- | --- |
+| 0 | int | Instances allocated before last GC |
+| 1 | int | Bytes allocated before last GC |
+| 2 | int | Instances alive after last GC |
+| 3 | int | Bytes alive after last GC |
+| 4 | int | Instances allocated since last GC |
+| 5 | int | Bytes allocated since last GC |
+| 6 | int | Instances allocated since last accumulator reset |
+| 7 | int | Bytes allocated since last accumulator reset |

[turnidge, 2014/08/21 00:33:01]

You should do the same thing with TokenLine down below -- say the type is a List
of int and then break out the format separately.

[Cutch, 2014/08/26 20:29:28]

Done.

+
+### <a name="atAbstractType"></a>@AbstractType
+
+### <a name="AbstractType"></a>AbstractType
+
+### <a name="PcDescriptor"></a>PcDescriptor
+
+
 ### <a name="atScript"></a>@Script
 | keys | values | comments | example |
 | --- | --- | ---
 | type | "@Script" |
 | id | String
 | user_name | String
 | name | String
 | kind | String
 
 ### <a name="Script"></a>Script
@@ skipping 10 matching lines @@
 
 ### <a name="TokenLine"></a>TokenLine
 | index | value | comments
 | --- | --- | ---
 | 0   | integer | line number
 | 1   | integer | first token position
 | 2   | integer | first column number
 | ... | ... | ...
 | 1 + (2 * k) | integer | kth token position
 | 2 + (2 * k) | integer | kth column number