MozillaWiki

Remote Debugging Protocol: Difference between revisions

Page Discussion

Remote Debugging Protocol (view source)

Revision as of 05:22, 12 January 2019

47,675 bytes removed ,  12 January 2019
bold
(Move "Source Locations" and "Lexical Environments" into their own top-level sections at the top.)
(bold)
 
(72 intermediate revisions by 9 users not shown)
Line 1: Line 1:
<i>(Note: this page is a draft design of work not yet completed. It is written in the present tense to be easily promoted to documentation when implemented, and also to simplify the grammar.)</i>
This page moved to the mozilla-central repository with [https://bugzilla.mozilla.org/show_bug.cgi?id=1354587 bug 1354587].


The Mozilla debugging protocol allows a debugger to connect to a browser, discover what sorts of things are present to debug or inspect, select JavaScript threads to watch, and observe and modify their execution. The protocol provides a unified view of JavaScript, DOM nodes, CSS rules, and the other technologies used in client-side web applications. The protocol ought to be sufficiently general to be extended for use with other sorts of clients (profilers, say) and servers (mail readers; random XULrunner applications).
Read it now at '''https://docs.firefox-dev.tools/backend/protocol.html'''.


All communication between debugger (client) and browser (server) is in the form of JSON objects. This makes the protocol directly readable by humans, capable of graceful evolution, and easy to implement using stock libraries. In particular, it should be easy to create mock implementations for testing and experimentation.
Source code, at http://searchfox.org/mozilla-central/source/devtools/docs/backend/protocol.md
 
The protocol operates at the JavaScript level, not at the C++ or machine level, and assumes that the JavaScript implementation itself is healthy and responsive. The JavaScript program being executed may well have gone wrong, but the JavaScript implementation's internal state must not be corrupt. Bugs in the implementation may cause the debugger to fail; bugs in the interpreted program must not.
 
= Actors =
 
An <b>actor</b> is something on the server that can exchange JSON packets with the client. Every packet from the client specifies the actor to which it is directed, and every packet from the server indicates which actor sent it.
 
Each server has a root actor, with which the client first interacts. The root actor can explain what sort of thing the server represents (browser; mail reader; etc.), and enumerate things available to debug: tabs, chrome, and so on. Each of these, in turn, is represented by an actor to which requests can be addressed. Both artifacts of the program being debugged, like JavaScript objects and stack frames, and artifacts of the debugging machinery, like breakpoints and watchpoints, are actors with whom packets can be exchanged.
 
For example, a debugger might connect to a browser, ask the root actor to list the browser's tabs, and present this list to the developer. If the developer chooses some tabs to debug, then the debugger can send <tt>attach</tt> requests to the actors representing those tabs, to begin debugging.
 
To allow the server to reuse actor names and the resources they require, actors have limited lifetimes. All actors in a server form a tree, whose root is the root actor. Closing communications with an actor automatically closes communications with its descendants. For example, the actors representing a thread's stack frames are children of the actor representing the thread itself, so that when a debugger detaches from a thread, which closes the thread's actor, the frames' actors are automatically closed. This arrangement allows the protocol to mention actors liberally, without making the client responsible for explicitly closing every actor that has ever been mentioned.
 
When we say that some actor <i>A</i> is a <b>child</b> of some actor <i>B</i>, we mean that <i>A</i> is a direct child of <i>B</i>, not a grandchild, great-grandchild, or the like. Similarly, <b>parent</b> means "direct parent". We use the terms <b>ancestor</b> and <b>descendent</b> to refer to those looser relationships.
 
The root actor has no parent, and lives as long as the underlying connection to the client does; when that connection is closed, all actors are closed.
 
Note that the actor hierarchy does not, in general, correspond to any particular hierarchy appearing in the debuggee. For example, although web workers are arranged in a hierarchy, the actors representing web worker threads are all children of the root actor: one might want to detach from a parent worker while continuing to debug one of its children, so it doesn't make sense to close communications with a child worker simply because one has closed communications with its parent.
 
<i>(We are stealing the "actor" terminology from Mozilla's [[IPDL]], to mean, roughly, "things participating in the protocol". However, IPDL does much more with the idea than we do: it treats both client and server as collections of actors, and uses that detail to statically verify properties of the protocol. In contrast, the debugging protocol simply wants a consistent way to indicate the entities to which packets are directed.)</i>
 
= Packets =
 
The protocol is carried by a reliable, bi-directional byte stream; data sent in both directions consists of JSON objects, called packets. A packet is a top-level JSON object, not contained inside any other value.
 
Every packet sent from the client has the form:
 
  { "to":<i>actor</i>, "type": <i>type</i>, ... }
 
where <i>actor</i> is the actor to whom the packet is directed&mdash;actor names are always natural numbers&mdash;and <i>type</i> is a string specifying what sort of packet it is. Additional properties may be present, depending on <i>type</i>.
 
Every packet sent from the server has the form:
 
  { "from":<i>actor</i>, ... }
 
where <i>actor</i> is the name of the actor that sent it. The packet may have additional properties, depending on the situation.
 
If a packet is directed to an actor that no longer exists, the server sends a packet to the client of the following form:
 
  { "from":null, "type":"no-such-actor" }
 
Clients should silently ignore properties they do not recognize. We expect that, as the protocol evolves, we will specify new properties that can appear in existing packets, and experimental implementations will do the same.
 
= Requests and Replies =
 
In this protocol description, a <b>request</b> is a packet sent from the client which always elicits a single packet from the recipient, the <b>reply</b>. These terms indicate a simple pattern of communication: at any given time, either the client or actor is permitted to send a packet, but never both.
 
The client's communication with each actor is treated separately: the client may send a request to one actor, and then send a request to a different actor before receiving a reply from the first.
 
Any actor can send reply to a request it is unable to carry out with an <b>error reply</b> of the form:
 
  { "from":<i>actor</i>, "error":<i>name</i>, "message":<i>message</i> }
 
where <i>name</i> is a JSON string naming what went wrong, and <i>message</i> is an English error message. Error <i>name</i>s are specified by the protocol; the client can use the name to identify which error condition arose. The <i>message</i> may vary from implementation to implementation, and should only be displayed to the user as a last resort, as the server lacks enough information about the user interface context to provide appropriate messages.
 
Packets not described as requests or replies are part of some more complicated interaction, which should be spelled out in more detail.
 
= Grips =
 
A grip is a JSON value that refers to a specific JavaScript value in the debuggee. Grips appear anywhere an arbitrary value from the debuggee needs to be conveyed to the client: stack frames, object property lists, lexical environments, <tt>paused</tt> packets, and so on.
 
For mutable values like objects and arrays, grips do not merely convey the value's current state to the client. They also act as references to the original value, by including an actor to which the client can send messages to modify the value in the debuggee.
 
A grip has one of the following forms:
 
  <i>value</i>
 
where <i>value</i> is a string, a number, or a boolean value. For these types of values, the grip is simply the JSON form of the value.
 
  { "type":"null" }
 
This represents the JavaScript <tt>null</tt> value. (The protocol does not represent JavaScript <tt>null</tt> simply by the JSON <tt>null</tt>, for the convenience of clients implemented in JavaScript: this representation allows such clients to use <tt>typeof(<i>grip</i>) == "object"</tt> to decide whether the grip is simple or not.)
 
  { "type":"undefined" }
 
This represents the JavaScript <tt>undefined</tt> value. (<tt>undefined</tt> has no direct representation in JSON.)
 
  { "type":"object", "class":<i>class-name</i>, "actor":<i>actor</i> }
 
This represents a JavaScript object whose class is <i>class-name</i>. (Arrays and functions are treated as objects for the sake of forming grips.) <i>Actor</i> can be consulted for the object's contents, as explained below.
 
  { "type":"long-string", "initial":<i>initial</i>, "length":<i>length</i>, "actor":<i>actor</i> }
 
This represents a very long string, where "very long" is defined at the server's discretion. <i>Initial</i> is some initial portion of the string, <i>length</i> is the string's full length, and <i>actor</i> can be consulted for the rest of the string, as explained below.
 
For example, the following table shows some JavaScript expressions and the grips that would represent them in the protocol:
 
{| frame="box" rules="all" cellpadding="8"
! JavaScript Expression
! Grip
|-
| 42
| 42
|-
| true
| true
|-
| "nasu"
| "nasu"
|-
| (void 0)
| { "type":"undefined" }
|-
| ({x:1})
| { "type":"object", "class":"Object", "actor":24 }
|}
 
Garbage collection will never free objects visible to the client via the protocol. Thus, actors representing JavaScript objects are effectively garbage collection roots.
 
== Objects ==
 
While a thread is paused, the client can send requests to the actors appearing in object grips to examine the objects they represent in more detail.
 
=== Property Descriptors ===
 
Protocol requests that describe objects' properties to the client often use <b>descriptors</b>, JSON values modeled after ECMAScript 5's property descriptors, to describe individual properties.
 
A descriptor has the form:
 
  { "enumerable":<i>enumerable</i>, "configurable":<i>configurable</i>, ... }
 
where <i>enumerable</i> and <i>configurable</i> are boolean values indicating whether the property is enumerable and configurable, and additional properties are present depending on what sort of property it is.
 
A descriptor for a data property has the form:
 
  { "enumerable":<i>enumerable</i>, "configurable":<i>configurable</i>,
    "value":<i>value</i>, "writeable":<i>writeable</i> }
 
where <i>value</i> is a grip on the property's value, and <i>writeable</i> is a boolean value indicating whether the property is writeable.
 
A descriptor for an accessor property has the form:
 
  { "enumerable":<i>enumerable</i>, "configurable":<i>configurable</i>,
    "get":<i>getter</i>, "set":<i>setter</i> }
 
where <i>getter</i> and <i>setter</i> are grips on the property's getter and setter functions. These may be <tt>{ "type":"undefined" }</tt> if the property lacks the given accessor function.
 
For example, if the JavaScript program being debugged evaluates the expression:
 
  ({x:10, y:"kaiju", get a() { return 42; }})
 
then a grip on this value would have the form:
 
  { "type":"object", "class":"Object", "actor":<i>actor</i> }
 
and sending a [[#Finding_An_Object's_Prototype_And_Properties|"prototype-and-properties"]] request to <i>actor</i> would produce the following reply:
 
  { "from":<i>actor</i>, "prototype": { "type":"object", "class":"Object", "actor":<i>objproto-actor</i> },
    "own-properties": { "x":{ "enumerable":true, "configurable":true, "writeable":true, "value":10 },
                        "y":{ "enumerable":true, "configurable":true, "writeable":true, "value":"kaiju" },
                        "a":{ "enumerable":true, "configurable":true,
                              "get":{ "type":"object", "class":"Function", "actor":<i>getter-actor</i> }
                            }
                      }
  }
 
=== Finding An Object's Prototype And Properties ===
 
To examine an object's prototype and properties, a client can send the object's grip's actor a request of the form:
 
  { "to":<i>grip-actor</i>, "type":"prototype-and-properties" }
 
to which the grip actor replies:
 
  { "from":<i>grip-actor</i>, "prototype":<i>prototype</i>,
    "own-properties":<i>own-properties</i> }
 
where <i>prototype</i> is a grip on the object's prototype (possibly <tt>{ "type":"null" }</tt>), and <i>own-properties</i> has the form:
 
  { <i>name</i>:<i>descriptor</i>, ... }
 
with a <i>name</i>:<i>descriptor</i> pair for each of the object's own properties.
 
<i>TODO: What about objects with many properties?</i>
 
=== Finding an Object's Prototype ===
 
To find an object's prototype, a client can send the object's grip's actor a request of the form:
 
  { "to":<i>grip-actor</i>, "type":"prototype" }
 
to which the grip actor replies:
 
  { "from":<i>grip-actor</i>, "prototype":<i>prototype</i> }
 
where <i>prototype</i> is a grip on the object's prototype (possibly <tt>{ "type":"null" }</tt>).
 
=== Listing an Object's Own Properties' Names ===
 
To list an object's own properties' names, a client can send the object's grip's actor a request of the form:
 
  { "to":<i>grip-actor</i>, "type":"own-property-names" }
 
to which the grip actor replies:
 
  { "from":<i>grip-actor</i>, "own-property-names": [ <i>name<i>, ... ] }
 
where each <i>name</i> is a string naming an own property of the object.
 
=== Finding Descriptors For Single Properties ===
 
To obtain a descriptor for a particular property of an object, a client can send the object's grip's actor a request of the form:
 
  { "to":<i>grip-actor</i>, "type":"property" }
 
to which the grip actor replies:
 
  { "from":<i>grip-actor</i>, "descriptor":<i>descriptor</i> }
 
where <i>descriptor</i> is a descriptor for the given property, or <tt>null</tt> if there is no such property on the object.
 
<i>TODO: assign to value property</i>
 
<i>TODO: special stuff for arrays</i>
 
<i>TODO: special stuff for functions</i>
 
<i>TODO: find function's source position</i>
 
<i>TODO: descriptors for Harmony proxies</i>
 
== Long Strings ==
 
The client can find the full contents of a long string by sending a request to the long string grip actor of the form:
 
  { "to":<i>grip-actor</i>, "type":"substring", "start":<i>start</i>, "length":<i>length</i> }
 
where <i>start</i> and <i>length</i> are integers. This requests the substring <i>length</i> characters long, starting at the <i>start</i>'th character. The actor replies as follows:
 
  { "from":<i>grip-actor</i>, "substring":<i>string</i> }
 
where <i>string</i> is the requested portion of the string the actor represents.
 
Like object grip actors, long string grip actors must only receive messages from the client while the thread is in the <b>Paused</b> state.
 
== Grip Lifetimes ==
 
Most grips are <b>pause-lifetime</b> grips: they last only while the JavaScript thread is paused, and become invalid as soon as the debugger allows the thread to resume execution. (The actors in pause grips are children of an actor that is closed when the thread resumes, or is detached from.) This arrangement allows the protocol to use grips freely in responses without requiring the client to remember and close them all.
 
However, in some cases the client may wish to retain a reference to an object while the debuggee runs. For example, a panel displaying objects selected by the user must update its view of the objects each time the debuggee pauses. To carry this out, the client can promote a pause-lifetime grip to a <b>thread-lifetime</b> grip, which lasts until the thread is detached from or exits. Actors in thread-lifetime grips are children of the thread actor. When the client no longer needs a thread-lifetime grip, it can explicitly release it. Both kinds of grips are garbage collection roots.
 
To promote a pause-lifetime grip to a thread-lifetime grip, the client sends a packet of the form:
 
  { "to":<i>grip-actor</i>, "type":"thread-grip" }
 
where <i>grip-actor</i> is the actor from the existing pause-lifetime grip. The grip actor will reply:
 
  { "from":<i>grip-actor</i>, "thread-grip":<i>thread-grip</i> }
 
where <i>thread-grip</i> is a new grip on the same object, but whose actor is parented by the thread actor, not the pause actor.
 
The client can release a thread-lifetime grip by sending the grip actor a request of the form:
 
  { "to":<i>grip-actor</i>, "type":"release" }
 
The grip actor will reply, simply:
 
  { "from":<i>grip-actor</i> }
 
This closes the grip actor.
 
Regardless of the lifetime of a grip, the client may only send messages to grip actors while the thread to which they belong is paused; the client's interaction with values cannot take place concurrently with the thread.
 
= Source Locations =
 
Descriptions of source locations (written as <i>location</i> in packet descriptions) can take one of the following forms:
 
  { "url":<i>url</i>, "line":<i>line</i>, "column":<i>column</i> }
 
This refers to line <i>line</i>, column <i>column</i> of the source code loaded from <i>url</i>. Line and column numbers start with 1. If <i>column</i> or <i>line</i> are omitted, they default to 1.
 
  { "eval":<i>location</i>, "id":<i>id</i>, "line":<i>line</i>, "column":<i>column</i> }
 
This refers to line <i>line</i>, column <i>column</i> of the source code passed to the call to eval at <i>location</i>. To distinguish the different texts passed to eval, each is assigned a unique integer, <i>id</i>.
 
  { "function":<i>location</i>, "id":<i>id</i>, "line":<i>line</i>, "column":<i>column</i> }
 
This refers to line <i>line</i>, column <i>column</i> of the source code passed to the call to the <tt>Function</tt> constructor at <i>location</i>. To distinguish the different texts passed to eval, each is assigned a unique integer, <i>id</i>.
 
As indicated, locations can be nested: a location like this one:
 
  { "eval":{ "eval":{ "url":"file:///home/example/sample.js", "line":20 }
            "line":30 }
    "line":40 }
 
refers to line 40 of the code passed to the call to eval occurring on line 30 of the code passed to the call to eval on line 20 of <tt>file:///home/example/sample.js</tt>.
 
= Lexical Environments =
 
A lexical environment (written as <i>environment</i> in packet descriptions) records the identifier bindings visible at a particular point in the program. An environment has one of the following forms:
 
  { "type":"object", "actor":<i>actor</i>, "object":<i>object</i>, "parent":<i>parent-environment</i> }
 
This represents a scope chain element whose identifier bindings reflect the properties of <i>object</i> (a grip). This could be the global object (<tt>window</tt> in a browser), or a DOM element (for event handler content attributes, which have the input element, form, and document on their scope chain along with the <tt>window</tt>).
 
<i>Actor</i> is the name of an actor representing this lexical environment. The requests it can answer are described below.
 
<i>Parent-environment</i> describes the next enclosing lexical environment; the <tt>parent</tt> property is omitted on the outermost environment.
 
  { "type":"function", "actor":<i>actor</i>, "function":<i>function</i>, "function-name":<i>function-name</i>,
    "bindings":<i>bindings</i>, "parent":<i>parent-environment</i> }
 
This represents the variable environment created by a call to <i>function</i> (a grip), whose name is <i>function-name</i> (a string). <i>Bindings</i> describes the bindings in scope, including the function's arguments, the <tt>arguments</tt> object, and local <tt>var</tt> and function bindings; its form is described in detail below. The <tt>function-name</tt> property is omitted if the function is anonymous. The other properties are as described above.
 
  { "type":"with", "actor":<i>actor</i>, "object":<i>object</i>, "parent":<i>parent-environment</i> }
 
This represents bindings introduced by a <tt>with</tt> statement whose operand is <i>object</i> (a grip). The other properties are as described above.
 
  { "type":"block", "actor":<i>actor</i>, "bindings":<i>bindings</i>, "parent":<i>parent-environment</i> }
 
This represents bindings introduced by a <tt>let</tt> block, <tt>for-in</tt> statement, <tt>catch</tt> block, or the like. The properties are as described above.
 
A <i>bindings</i> value has the form:
 
  { "mutable":{ <i>name</i>:<i>value</i>, ... },
    "immutable":{ <i>name</i>:<i>value</i>, ... } }
 
where each <i>name</i> is the name of a bound identifier, and each <i>value</i> is a grip on that identifier's value. Mutable bindings appear in the <tt>mutable</tt> object, and immutable bindings appear in the <tt>immutable</tt> object. If either category has no bindings, the property may be omitted entirely.
 
Note that language implementations may omit some environment records from a function's scope if it can determine that the function would not use them. This means that it may be impossible for a debugger to find all the variables that ought to be in scope.
 
To fully enumerate the bindings introduced by any lexical environment, the client can send a request of the following form to the environment's actor:
 
  { "to":<i>env-actor</i>, "type":"bindings" }
 
The actor will reply as follows:
 
  { "from":<i>env-actor</i>, "bindings":<i>bindings</i> }
 
To change the value of a variable bound in a particular lexical environment, the client can send a request to the environment's actor:
 
  { "to":<i>env-actor</i>, "type":"assign", "name":<i>name</i>, "value":<i>value</i> }
 
This changes the value of the identifier whose name is <i>name</i> (a string) to that represented by <i>value</i> (a grip). The actor will reply as follows, simply:
 
  { "from":<i>env-actor</i> }
 
If the named identifier is immutable, the actor will send an error reply of the form:
 
  { "from":<i>env-actor</i>, "error":"immutable-binding", "message":<i>message</i> }
 
= The Root Actor =
 
When the connection to the server is opened, the root actor opens the conversation with the following packet:
 
  { "from":0, "application-type":<i>app-type</i>, "traits":<i>traits</i>, ...}
 
The root actor's name is always zero. <i>app-type</i> is a string indicating what sort of program the server represents. There may be more properties present, depending on <i>app-type</i>.
 
<i>traits</i> is an object describing protocol variants this server supports that are not convenient for the client to detect otherwise. The property names present indicate what traits the server has; the properties' values depend on their names. This version of the protocol defines no traits, so <i>traits</i> must be an object with no properties, <tt>{}</tt>.
 
For web browsers, the introductory packet should have the following form:
 
  { "from":0, "application-type":"browser", "traits":<i>traits</i> }
 
== Listing Top-Level Browsing Contexts ==
 
To get a list of the top-level browsing contexts (tabs) present in a browser, a client should send a request like the following to the root actor:
 
  { "to":0, "type":"list-contexts" }
 
The reply should have the form:
 
  { "from":0, "contexts":[<i>context</i>...], selected:<i>index</i> }
 
The <tt>contexts</tt> property's value is an array with one element for each top-level browsing context present in the browser, and <i>index</i> is the index within that list of the browsing context the user is currently interacting with. Each <i>context</i> has the following form:
 
  { "actor":<i>actor</i>, "title":<i>title</i>, "url":<i>url</i> }
 
<i>actor</i> is the actor representing that top-level browsing context; <i>title</i> is the context's document's title, and <i>url</i> is the context's document's URL.
 
Clients should send "list-contexts" requests only to root actors that have identified themselves as browsers.
 
Actor names given in a list-contexts reply are children of the root actor. They remain valid at least until the next list-contexts request is received. If the client attaches to a context actor, its name is valid at least until the client detaches from the context and receives a "detached" packet from the context, or until the client sends a "release" packet to the context. (These packets are described in detail in [[#Interacting_with_Thread-Like_Actors|Interacting with Thread-Like Actors]].)
 
For example, upon connection to a web browser visiting two pages at example.com, the root actor's introductory packet might look like this:
 
  { "from":0, "application-type":"browser",
    "contexts": [ { "actor":1, "title":"Fruits",
                    "url":"http://www.example.com/fruits/" },
                  { "actor":2, "title":"Bats",
                    "url":"http://www.example.com/bats/" }]}
 
<i>(This may not be the right information to provide in these packets; suggestions very welcome. The point here is to give the debugger enough information to select which context it would like to debug without having to do too many round trips. Round trips are bad for UI responsiveness, but large packets are probably not a problem, so whatever would help to add, we should add.)</i>
 
= Interacting with Thread-Like Actors =
 
Actors representing independent threads of JavaScript execution, like browsing contexts and web workers, are collectively known as "threads". Interactions with actors representing threads follow a more complicated communication pattern.
 
A thread is always in one of the following states:
 
* <b>Detached</b>: the thread is running freely, and not presently interacting with the debugger. Detached threads run, encounter errors, and exit without exchanging any sort of messages with the debugger. A debugger can attach to a thread, putting it in the <b>Running</b> state. Or, a detached thread may exit on its own, entering the <b>Exited</b> state.
 
* <b>Running</b>: the thread is running under the debugger's observation, executing JavaScript code or possibly blocked waiting for input. It will report exceptions, breakpoint hits, watchpoint hits, and other interesting events to the client, and enter the <b>Paused</b> state. The debugger can also interrupt a running thread; this elicits a response and puts the thread in the <b>Paused</b> state. A running thread may also exit, entering the <b>Exited</b> state.
 
* <b>Paused</b>: the thread has reported a pause to the client and is awaiting further instructions. In this state, a thread can accept requests and send replies. If the client asks the thread to continue or step, it returns to the <b>Running</b> state.
 
* <b>Exited</b>: the thread has ceased execution, and will disappear. The resources of the underlying thread may have been freed; this state really indicates that the actor's name is not yet available for reuse. When the actor receives a "release" packet, the name may be reused.
 
[[File:thread-states.png]]
 
These interactions are meant to have certain properties:
 
* At no point may either client or server send an unbounded number of packets without receiving a packet from its counterpart. This avoids deadlock without requiring either side to buffer an arbitrary number of packets per actor.
 
* In states where a transition can be initiated by either the debugger or the thread, it is always clear to the debugger which state the thread actually entered, and for what reason.<p>For example, if the debugger interrupts a running thread, it cannot be sure whether the thread stopped because of the interruption, paused of its own accord (to report a watchpoint hit, say), or exited. However, the next packet the debugger receives will either be "interrupted", "paused", or "exited", resolving the ambiguity.</p><p>Similarly, when the debugger attaches to a thread, it cannot be sure whether it has succeeded in attaching to the thread, or whether the thread exited before the "attach" packet arrived. However, in either case the debugger can expect a disambiguating response: if the attach suceeded, it receives an "attached" packet; and in the second case, it receives an "exit" packet.</p><p>To support this property, the thread ignores certain debugger packets in some states (the "interrupt" packet in the <b>Paused</b> and <b>Exited</b> states, for exmple). These cases all handle situations where the ignored packet was preempted by some thread action.</p>
 
Note that the rules here apply to the client's interactions with each thread actor separately. A client may send an "interrupt" to one thread actor while awaiting a reply to a request sent to a different thread actor.
 
<i>TODO: What about user selecting nodes in displayed content? Should those be eventy things the client can receive in the "paused" state? What does that mean for the "request"/"reply" pattern?</i>
 
== Attaching To a Thread ==
 
To attach to a thread, the client sends a packet of the form:
 
  { "to":<i>thread</i>, "type":"attach", "pause-for":<i>pause-types</i> }
 
Here, <i>thread</i> is the actor representing the thread, perhaps a browsing context from a "list-contexts" reply. This tells the thread to continue to run, but asks it to pause if any of the events described by <i>pause-types</i> occurs. The form of <i>pause-types</i> is described in [[#Pause_Types|Pause Types]].
 
The thread responds in one of two ways:
 
  { "from":<i>thread</i>, "type":"attached" }
 
This indicates that the thread received the <tt>attach</tt> packet, and will continue to run, reporting events of interest to the debugger. The thread is now in the <b>Running</b> state. The actor name <i>thread</i> remains valid until the client detaches from the thread or acknowledges a thread exit.
 
  { "from":<i>thread</i>, "type":"exited" }
 
This indicates that the thread exited before receiving the <tt>attach</tt> packet. The thread is now in the <b>Exited</b> state. The client must respond to this with a <tt>release</tt> packet; see [[#Exiting_Threads|Exiting Threads]].
 
== Detaching From a Thread ==
 
To detach from a thread, the client sends a packet of the form:
 
  { "to":<i>thread</i>, "type":"detach" }
 
The thread responds in one of three ways:
 
  { "from":<i>thread</i>, "type":"detached" }
 
This indicates that the client has detached from the thread. The thread is now in the <b>Detached</b> state: it can run freely, and no longer reports events to the client. The actor name <i>thread</i> is released and available for reuse.
 
  { "from":<i>thread</i>, "type":"paused", ... }
  { "from":<i>thread</i>, "type":"detached" }
 
This series of packets indicates that the thread paused of its own accord (for the reason given by the additional properties of the "paused" packet), and only then received the "detach" packet. As above, this indicates that the thread is in the <b>Detached</b> state, and the actor name is available for reuse.
 
  { "from":<i>thread</i>, "type":"exited" }
 
This indicates that the thread exited on its own before receiving the "detach" packet. The client should follow by sending a "release" packet; see [[#Exiting_Threads|Exiting Threads]], below.
 
== Running Threads ==
 
Once the client has attached to a thread, it is in the <b>Running</b> state. In this state, four things can happen:
 
* The thread can hit a breakpoint or watchpoint, or encounter some other condition of interest to the client.
* The thread can exit.
* The client can detach from the thread.
* The client can interrupt the running thread.
 
Note that a client action can occur simultaneously with a thread action. The protocol is designed to avoid ambiguities when both client and thread act simultaneously.
 
== Thread Pauses ==
 
If the thread pauses to report an interesting event to the client, it sends a packet of the form:
 
  { "from":<i>thread</i>, "type":"paused", "actor":<i>actor</i>, "frame":<i>frame</i>, "why":<i>reason</i> }
 
This indicates that the thread has entered the <b>Paused</b> state, and explains where and why.
 
<i>Actor</i> is a "pause actor", representing this specific pause of the thread; it lives until the thread next leaves the <b>Paused</b> state. The pause actor parents actors actors referring to stack frames, values, and other entities uncovered during this pause; when the thread resumes, those actors are automatically closed. This relieves the client from the responsibility to explicitly close every actor mentioned during the pause.
 
Since actors in value grips are parented by the pause actor as well, this means that those grips become invalid when the thread resumes; it is not possible to take a grip from one pause and use it in the next. To create a grip that remains valid between pauses, see [[#Holding_Values_Between_Pauses|Holding Values Between Pauses]].
 
<i>Frame</i> describes the top frame on the JavaScript stack; see [[#Listing_Stack_Frames|Listing Stack Frames]], below.
 
The <i>reason</i> value describes why the thread paused. It has one of the following forms:
 
  { "type":"breakpoint", "actors":[<i>actor</i>...] }
 
The thread stopped at the breakpoints represented by the given <i>actor</i>s.
 
  { "type":"watchpoint", "actors":[<i>watchpoint</i>...] }
 
The thread stopped at the watchpoints represented by the given <i>actor</i>s.
 
<i>TODO: This should provide more details about the watchpoint in the packet, instead of incurring another round-trip before we can display anything helpful.</i>
 
  { "type":"stepped" }
 
The client had asked the thread to step to the next statement, and the thread completed that step.
 
  { "type":"pre-call" }
 
The client had asked the thread to pause before making each function call, and the thread is about to call a function. Single-stepping the thread will a place it at the head of the function's code, with all arguments, local variables, and local functions bound.
 
  { "type":"pre-return" }
 
The client had asked the thread to pause before returning from functions, and the thread is about to return from a function. Single-stepping the thread will return the thread to the calling frame.
 
  { "type":"pre-throw", "exception":<i>grip</i> }
 
The client had asked the thread to pause before throwing an exception; <i>grip</i> is a grip on the exception value being thrown. Control is still at the point of the throw; it has not yet passed to a catch clause. Single-stepping this thread will report either a "caught" or "uncaught" pause.
 
  { "type":"caught", "exception":<i>grip</i> }
 
The client stepped the thread from a "pre-throw" pause, and a catch clause has been found for the exception referred to by <i>grip</i>; control is stopped at the head of the catch clause, with catch variable bindings made. If the catch is conditional, control is at the beginning of the condition.
 
  { "type":uncaught", "exception":<i>grip</i> }
 
The client stepped the thread from a "pre-throw" pause, and no catch clause was found for the exception. <i>Grip</i> is as above. <i>(I'm not sure which code the thread is executing at this point; we might as well reveal SpiderMonkey's natural behavior.)</i>
 
  { "type":"pre-throw-by-guard", "exception":<i>grip</i> }
 
The thread had been stopped in a conditional guard, and the client asked the thread to continue but pause before throwing an exception. The guard condition evaluated to false, and the thread is about to re-throw the exception value, <i>grip</i>.
 
  { "type":"client-evaluated", "value":<i>grip</i> }
 
The client's prior <tt>client-evaluate</tt> command has completed normally; <i>grip</i> is a grip on the expression's value. See [[#Evaluating_Source-Language_Expressions|Evaluating Source-Language Expressions]] for details.
 
  { "type":"client-evaluated", "exception":<i>grip</i> }
 
The client's prior <tt>client-evaluate</tt> command has completed abruptly; <i>grip</i> is the uncaught exception value. See [[#Evaluating_Source-Language_Expressions|Evaluating Source-Language Expressions]] for details.
 
=== Pause Types ===
 
The <tt>pause-for</tt> property of an <tt>attach</tt> or <tt>resume</tt> packet (written as <i>pause-types</i>) is an object whose properties' names indicate which events the client is interested in. The vocabulary of names is the same as that of pause <tt>reason</tt> types, listed [[#Thread_Pauses|above]]. If a property's value in a <i>pause-type</i> is <tt>true</tt>, then the event is of interest.
 
For example, the following <i>resume</i> packet would instruct <i>thread</i> to continue until the thread is about to return from a function or is about to throw an exception:
 
  { "to":<i>thread</i>, "type":"resume",
    "pause-for": { "pre-return":true, "pre-throw":true } }
 
Certain pause types cause other pause types to be included automatically, if those pause types are not mentioned explicitly in <i>pause-types</i>:
 
{| frame="box" rules="all" cellpadding="8"
! Pause type
! Also implies, if not mentioned explicitly
|-
| <tt>pre-throw</tt>
| <tt>pre-throw-by-guard</tt>
|}
 
For some pause types, the property's value can provide additional detail.
 
To request that the thread pause before returning from a specific frame, the <i>pause-types</i> object can include a property of the form:
 
  { ... "pre-return":<i>frame-actor</i> ... }
 
where <i>frame-actor</i> is the name of the actor representing a specific stack frame, as given in the reply to a <tt>frames</tt> request.
 
=== Extending The Protocol With New Pause Types ===
 
The format of <i>reason</i> values in <tt>paused</tt> packets should be treated as a point of extension, with application-specific pause types added as the need arises. DOM events, DOM mutation, and web worker message transmission might all be interesting pause types. Debuggers unprepared for a particular event type will simply not request that type in the <tt>pause-for</tt> properties of its <tt>attach</tt> and <tt>resume</tt> packets, so extending the events a server supports should not disturb existing clients.
 
== Resuming a Thread ==
 
If a thread is in the <b>Paused</b> state, the client can resume it by sending a packet of the following form:
 
  { "to":<i>thread</i>, "type":"resume", "pause-for":<i>pause-types</i> }
 
This puts the thread in the <b>Running</b> state, but asks it to pause if any of the event described by <i>pause-types</i> occurs. The form of <i>pause-types</i> is described in [[#Pause_Types|Pause Types]].
 
A "resume" packet closes the pause actor the client provided in the "paused" or "interrupted" packet that began the pause.
 
== Interrupting a Thread ==
 
If a thread is in the <b>Running</b> state, the client can cause it to pause where it is by sending a packet of the following form:
 
  { "to":<i>thread</i>, "type":"interrupt" }
 
The thread responds in one of three ways:
 
  { "from":<i>thread</i>, "type":"interrupted", "actor":<i>actor</i>, "frame":<i>frame</i> }
 
This indicates that the thread stopped due to the client's <tt>interrupt</tt> packet, and is now in the <b>Paused</b> state.
 
  { "from":<i>thread</i>, "type":"paused", "frame":<i>frame</i>, "why":<i>reason</i> }
 
This indicates that the thread stopped of its own accord before receiving the client's <tt>interrupt</tt> packet, and is now in the <b>Paused</b> state. The meanings of the "paused" packet's properties are as for an ordinary pause. The thread will ignore the client's interrupt packet when it receives it.
 
  { "from":<i>thread</i>, "type":"exited" }
 
This indicates that the thread exited before receiving the client's <tt>interrupt</tt> packet, and is now in the <b>Exited</b> state. See [[#Exiting_Threads|Exiting Threads]], below.
 
== Exiting Threads ==
 
When a thread in the <b>Running</b> state exits, it sends a packet of the following form:
 
  { "from":<i>thread</i>, "type":"exited" }
 
At this point, the thread can no longer be manipulated by the client, and most of the thread's resources may be freed; however, the thread actor name must remain alive, to handle stray <tt>interrupt</tt> and <tt>detach</tt> packets. To allow the last trace of the thread to be freed, the client should send a packet of the following form:
 
  { "to":<i>thread</i>, "type":"release" }
 
This acknowledges the exit and allows the thread actor name, <i>thread</i>, to be reused for other actors.
 
= Inspecting Paused Threads =
 
When a thread is in the <b>Paused</b> state, the debugger can make requests to inspect its stack, lexical environment, and values.
 
== Listing Stack Frames ==
 
To inspect the thread's JavaScript stack, the client can send the following request:
 
  { "to":<i>thread</i>, "type":"frames", "start":<i>start</i>, "count":<i>count</i> }
 
The <tt>start</tt> and <tt>count</tt> properties are optional. If present, <i>start</i> gives the number of the youngest stack frame the reply should describe, where the youngest frame on the stack is frame number zero; if absent, <i>start</i> is taken to be zero. If present, <i>count</i> specifies the maximum number of frames the reply should describe; if absent, it is taken to be infinity. (Clients should probably avoid sending <tt>frames</tt> requests with no <i>count</i>, to avoid being flooded by frames from unbounded recursion.)
 
The thread replies as follows:
 
  { "from":<i>thread</i>, "frames":[<i>frame</i> ...] }
 
where each <i>frame</i> has the form:
 
  { "actor":<i>actor</i>, "depth":<i>depth</i>, "id":<i>id</i>,
    "type":<i>type</i>, ... }
 
where:
 
* <i>actor</i> is the name of an actor representing this frame;
* <i>depth</i> is the number of this frame, starting with zero for the youngest frame on the stack;
* <i>id</i> is a unique number assigned to this stack frame, to help the client identify frames across pauses; and
* <i>type</i> is a string indicating what sort of frame this is.
 
The frame may have other properties, depending on <i>type</i>.
 
All actors mentioned in the frame or grips appearing in the frame (<i>actor</i>, <i>callee</i>, <i>environment</i>, and so on) are parented by the current pause actor, as given in the "paused" or "interrupted" packet.
 
=== Global Code Frames ===
 
A frame for global code has the form:
 
  { "actor":<i>actor</i>, "depth":<i>depth</i>, "id":<i>id</i>,
    "type":"global", "where":<i>location</i>, "environment",<i>environment</i> }
 
where:
   
* <i>location</i> is the source location of the current point of execution in the global code (see [[#Source_Locations|Source Locations]]);
* <i>environment</i> is a value representing the lexical environment of the current point of execution (see [[#Lexical_Environments|Lexical Environments]]);
 
and other properties are as above.
 
=== Function Call Frames ===
 
A frame for an ordinary JavaScript function call has the form:
 
  { "actor":<i>actor</i>, "depth":<i>depth</i>, "id":<i>id</i>,
    "type":"call",
    "where":<i>location</i>, "environment",<i>environment</i>,
    "callee":<i>callee</i>, "callee-name":<i>callee-name</i>,
    "this",<i>this</i>, "arguments":<i>arguments</i> }
 
where:
 
* <i>where</i> is the current point of execution within the callee;
* <i>callee</i> is a grip on the function value being called;
* <i>callee-name</i> is the name of the callee, a string (this property is omitted for anonymous functions);
* <i>this</i> is a grip on the value of <tt>this</tt> for this call;
* <i>arguments</i> is an array of grips on the actual values passed to the function;
 
and other properties are as above.
 
The argument list may be incomplete or inaccurate, for various reasons. If the program has assigned to its formal parameters, the original values passed may have been lost, and compiler optimizations may drop some argument values.
 
=== Host Function Call Frames ===
 
A frame for a call to a host function (a function provided by the JavaScript implementation, usually written in the same language as the implementation itself) has the form:
 
  { "actor":<i>actor</i>, "depth":<i>depth</i>, "id":<i>id</i>,
    "type":"host-call",
    "callee":<i>callee</i>, "callee-name":<i>callee-name</i>,
    "this",<i>this</i>, "arguments":<i>arguments</i> }
 
where the properties are as defined above. (Compared to an ordinary call, the <tt>where</tt> and <tt>environment</tt> properties are missing.)
 
=== Eval Frames ===
 
A frame for a call to <tt>eval</tt> has the form:
 
  { "actor":<i>actor</i>, "depth":<i>depth</i>, "id":<i>id</i>,
    "type":"eval", "where":<i>location</i>, "environment",<i>environment</i> }
 
where the properties are as defined above.
 
=== Client Evaluation Frames ===
 
When the client evaluates an expression with an <tt>client-evaluate</tt> packet, the evaluation appears on the stack as a special kind of frame, of the form:
 
  { "actor":<i>actor</i>, "depth":<i>depth</i>, "id":<i>id</i>,
    "type":"client-evaluate", "where":<i>location</i>, "environment",<i>environment</i> }
 
where the properties are as defined above. In this case, <i>where</i> will be a location inside the expression provided by the debugger.
 
== Popping Stack Frames ==
 
The client can remove frames from the stack by sending a request of the form:
 
  { "to":<i>frame-actor</i>, "type":"pop", "value":<i>value</i> }
 
where <i>frame-actor</i> is the actor representing the stack frame to pop, and <i>value</i> is a grip on the value that should be returned as the value of the frame. All younger stack frames are also popped. The frame actor will reply:
 
  { "from":<i>frame-actor</i>, "watches":[<i>watch-actor</i> ...] }
 
where each <i>watch-actor</i> is the name of a frame pop watch actor that has been triggered in the process of popping the given frame. If no frame pop watches are triggered, the <tt>watches</tt> property may be omitted.
 
<i>TODO: specify the error to return if the frame cannot be popped --- can host (C++) function frames be popped?</i>
 
== Evaluating Source-Language Expressions ==
 
To evaluate a source-language expression in a thread, the client sends a specialized <tt>resume</tt> packet of the form:
 
  { "to":<i>thread</i>, "client-evaluate":<i>expr</i>, "frame":<i>frame</i>, "pause-for":<i>pause-types</i> }
 
This resumes the thread just as an ordinary <tt>resume</tt> packet does, but rather than continuing execution where the pause took place, has the thread begin evaluation of the source-language expression given by <i>expr</i>, a string. The evaluation takes place in a new [[#Client_Evaluation_Frames|Client Evaluation Frame]], pushed on the stack. When evaluation of <i>expr</i> completes, the client will report an <tt>client-evaluate</tt> pause containing the expression's value.
 
If evaluating <i>expr</i> completes abruptly, this outcome is still reported via an <tt>client-evaluated</tt> pause, so it is not necessary for the client to take explicit steps to catch exceptions thrown by the expression.
 
<i>TODO: evaluate with given grips bound to given identifiers</i>
 
= Breakpoints =
 
While a thread is paused, a client can set breakpoints in the thread's code by sending requests of the form:
 
  { "to":<i>thread</i>, "type":"set-breakpoint", "location":<i>location</i> }
 
where <i>location</i> is a [[#Source_Locations|source location]]. If the thread is able to establish a breakpoint at the given location, it replies:
 
  { "from":<i>thread</i>, "actor":<i>actor</i>, "actual-location":<i>actual-location</i> }
 
where <i>actor</i> is an actor representing the breakpoint (a child of the thread actor), and <i>actual-location</i> is the location at which the breakpoint was really set. If <i>location</i> and <i>actual-location</i> are the same, then the <tt>actual-location</tt> property can be omitted.
 
If the thread cannot find the script referred to in <i>location</i>, it sends an error reply of the form:
 
  { "from":<i>thread</i>, "error":"no-script" }
 
If <i>location</i> refers to a line and column at which the given script has no program code, and no reasonable alternative location can be chosen (say, by skipping forward), then the thread sends an error reply of the form:
 
  { "from":<i>thread</i>, "error":"no-code-at-line-column" }
 
To delete a breakpoint, the client can send the breakpoint's actor a message of the form:
 
  { "to":<i>breakpoint-actor</i>, "type":"delete" }
 
to which the breakpoint actor will reply, simply:
 
  { "from":<i>breakpoint-actor</i> }
 
This closes communications with <i>breakpoint-actor</i>.
 
= Watchpoints =
= Frame Pop Watches =
 
<i>TODO: DOM node inspection, highlighting</i>
Krinkle
6

edits

Retrieved from "https://wiki.allizom.org/Special:MobileDiff/241616...1206107"