sc.api — vvvvalvalval/scope-capture 0.3.3

brk clj/s macro

(brk)

(brk expr)

(brk opts expr)

Like spy, but will block the executing Thread until you release it from the REPL.

There are 3 ways of resuming execution at an Execution Point where it was suspended by brk:

• (sc.api/loose ep-id): resumes execution by evaluating the wrapped expression.
• (sc.api/loose-with ep-id value): resumes execution by yielding value instead of evaluating the wrapped expression.
• (sc.api/loose-with-err ep-id err): resumes execution by throwing err instead of evaluating the wrapped expression.

Similarly to spy, you can customize the behaviour of brk by passing a literal map of options; the accepted keys are:

• :sc/brk-cs-logger-id
• :sc/brk-ep-pre-eval-logger
• :sc/brk-ep-post-eval-logger
• :sc/dynamic-vars
• :sc/called-from

You can completely disable the logging, recording, and blocking behaviour of spy at a given Code Site by using (cs.api/disable! cs-id).

You can discard the recorded information about a given Execution Point (typically for freeing memory) by using (sc.api/dispose! ep-id). Execution Points which are being blocked will see their execution resumed by throwing an Exception.

Since you will likely need your REPL to un-block, you should never execute a (brk ...) block directly from your REPL thread!

brk-emit clj/s

(brk-emit opts expr amp-env amp-form)

Helper function for implementing (your own version of) the brk macro. See the source of brk for a usage example.

brkqt clj/s macro

(brkqt)

(brkqt expr)

(brkqt opts expr)

'BReaK QuieTly' - like sc.api/brk but less verbose. Will not print the :sc.ep/value, :sc.ep/error nor :sc.cs/expr; useful when dealing with large values.

calling-from clj/s macro

(calling-from from-v & body)

Binds sc.api.from/caller to from-v; to be used in combination with the :sc/called-from option.

When a Code Site has the :sc/called-from option set to a non-nil value v, (e.g via a SPY call of the form (sc.api/spy{:sc/called-from :foo} MY-EXPR)) logging and recording of Execution Points at this Code Site will only occur when the Dynamic Varsc.api.from/calleris bound tov`.

Any non-nil value is legal for from-v; you'll typically use a keyword, number or boolean.

Use this when you're only interested in the executions of a Code Site that are invoked from a given other place in your code (which you wrap with (sc.api/calling-from ...))

cs-info clj/s

(cs-info cs-id)

Given a Code Site Id, retrieves a map of information about that Code Site, featuring the following keys:

• :sc.cs/id (negative integer): the ID of the Code Site
• :sc.cs/expr (code form): the wrapped code expression
• :sc.cs/local-names (vector of symbols): the names of the locals that are in the lexical environment of the Code Site.
• :sc.cs/dynamic-vars (vector of symbols); the names of the dynamic Vars which bindings will be recorded at that Code Site.
• :sc.cs/file (String): the path of the source file of the wrapped expr, when available
• :sc.cs/line (integer): the line number of the spy or brk call in the source file, when available
• :sc.cs/column (integer): the column number of the spy or brk call in the source file, when available
• :sc.cs/disabled (boolean): whether this Code Site is in a disabled state.

defsc clj/s macro

(defsc ep-id)

Given an Execution Point Id, expands to several (def ...) expressions, defining Vars which names match the names of the locals captured by the spy or brk macros at the Execution Point.

In practice, this enables you to make some local, temporary bindings available globally and durably in a namespace (as Vars), which is convenient for evaluating forms using these bindings at the REPL.

Note: defsc will not recreate the dynamic Var bindings at this Execution Point.

You can un-define (via ns-unmap) the Vars defined by defsc for a code site by using sc.api/undefsc.

ep-id may be provided either as a positive integer literal (the Execution Point Id, e.g 8), or as a vector literal of one positive integer (the Execution Point Id) and one negative integer (the Code Site Id), e.g [8 -3]. Note that in Clojure environments where compilation and execution don't happen in the same address space (which is the case for JVM-compiled ClojureScript), only the second form is accepted, as there is no way of inferring the Code Site Id from the Execution Point Id at macro-expansion.

disable! clj/s

(disable! cs-id)

Disables all logging, recording, and blocking behaviour at a Code Site. No more Execution Points will be created at that Code Site, but the previously collected ones remain fully available. Idempotent. The Code Site can be re-enabled by calling sc.api/enable!

dispose! clj/s

(dispose! ep-id)

Frees up resources held by the given Execution Point. Useful for freeing memory.

If the Execution Point is in a suspened state (as by brk), will resume execution by throwing an Exception.

dispose-all! clj/s

(dispose-all!)

Disposes of all Execution Points (as per sc.api/dispose!)

enable! clj/s

(enable! cs-id)

Undoes the action of sc.api/disable!. Idempotent.

ep-info clj/s

(ep-info)

(ep-info ep-id)

Given an Execution Point Id, returns a map of information about that Execution Point, and the associated Code Site.

Returns a map featuring the following keys:

• :sc.ep/id (positive integer) The id of the Execution Point
• :sc.ep/code-site A map of data about the associated Code Site, see sc.api/cs-info
• :sc.ep/value The value the wrapped expression evaluated to (if any)
• :sc.ep/error The exception thrown by the evaluation of the wrapped exception (if any)
• :sc.ep/local-bindings A map which keys are the names of the locals in the scope of the wrapped expression, and which values are the values the locals were bound to at this Execution Point.
• :sc.ep/dynamic-var-bindings A map which keys are the fully-qualified names of the dynamic Vars recorded for the Code Site, and which values are the values they were bound to at the Execution Point.

As a convenience, when ep-id is not supplied, uses (sc.api/last-ep-id).

ep-value clj/s

(ep-value)

(ep-value ep-id)

Returns the value that was saved at the Execution Point identified by ep-id (or nil if none was saved). Shorthand for (:sc.ep/value (sc.api/ep-info ep-id)).

As a convenience, when ep-id is not supplied, uses (sc.api/last-ep-id).

gen-ep-id clj/s

(gen-ep-id)

Generates a unique Execution Point Id, a positive integer.

last-ep-id clj/s

(last-ep-id)

Returns the id of the last saved Execution Point, in [ep-id cs-id] form. Throws an Exception if no Execution Point has been saved.

letsc clj/s macro

(letsc ep-id & body)

Given an Execution Point Id ep-id, will expand to let and binding forms wrapping body which will reproduce the local and dynamic Var bindings that were recorded by the spy and brk macro at the Execution Point.

In practice, this enables you to run body in the same local environment where the Execution Point was recorded.

ep-id may be provided either as a positive integer literal (the Execution Point Id, e.g 8), or as a vector literal of one positive integer (the Execution Point Id) and one negative integer (the Code Site Id), e.g [8 -3]. Note that in Clojure environments where compilation and execution don't happen in the same address space (which is the case for JVM-compiled ClojureScript), only the second form is accepted, as there is no way of inferring the Code Site Id from the Execution Point Id at macro-expansion.

loose clj/s

(loose)

(loose ep-id)

Given an Execution Point Id ep-id for an Execution Point created by brk which is currently in a suspended state, resumes execution by evaluating the wrapped form.

As a convenience, when ep-id is not supplied, uses (sc.api/last-ep-id).

loose-with clj/s

(loose-with v)

(loose-with ep-id v)

Same as sc.api/loose, except that execution is resumed by yielding the provided value v instead of evaluating the wrapped expression.

As a convenience, when ep-id is not supplied, uses (sc.api/last-ep-id).

loose-with-err clj/s

(loose-with-err err)

(loose-with-err ep-id err)

Same as sc.api/loose, except that execution is resumed by throwing err instead of evaluating the wrapped expression.

As a convenience, when ep-id is not supplied, uses (sc.api/last-ep-id).

save-ep clj/s

(save-ep ep-id ep-data)

Creates or updates an Execution Point. ep-data should follow the same schema as what is returned by ep-info; only the :sc.ep/local-bindings and :sc.ep/code-site keys are required, and only :sc.cs/id will be taken in account for the Code Site. Returns ep-id.

Useful for artificially creating an Execution Point that is a slight variation of another one, e.g:

(sc.api/save-ep (sc.api/gen-ep-id) (-> (sc.api/ep-info 17) (select-keys [:sc.ep/local-bindings :sc.ep/code-site]) (assoc-in [:sc.ep/local-bindings 'x] 32)))

spy clj/s macro

(spy)

(spy expr)

(spy opts expr)

Records the scope (i.e bindings of local names) and value of the expression it wraps at runtime; yields the value of the wrapped expression.

At macro-expansion time, a Code Site Id (a negative integer) will be created and logged. Using this id, static information about the Code Site may be retrieved using (sc.api/cs-info cs-id).

Each time the expression is evaluated, an Execution Point Id (a positive integer) will be created and logged. Dynamic information about this Execution Point (and the associated Code Site) may be retrieved using (sc.api/ep-info ep-id).

Having noted the Execution Point Id, you can then 'recreate' the local runtime environment of the expression (its scope) using the letsc and defsc macros:

(letsc ep-id ...)

(defsc ep-id)

The behaviour of this macro may be customized by passing a map literal as a first argument, containing some of the following keys:

• :sc/spy-cs-logger-id (keyword, defaults to :sc.api.logging/log-spy-cs) The id of the logger function that will be called at macro-expansion time with Code Site information. This function must have been registered using sc.api/register-cs-logger.

• :sc/spy-ep-pre-eval-logger-fn (expression, defaults to `sc.api.logging/log-spy-ep-pre-eval) An expression which evaluates to a function (typically a Var name), which will be invoked at runtime before the wrapped expression is evaluated with Execution Point information - typically to log the Execution Point Id.

• :sc/spy-ep-post-eval-logger (expression, defaults to `sc.api.logging/log-spy-ep-post-eval) Like :sc/spy-ep-pre-eval-logger-fn, but the function will be invoked after the wrapped expression is evaluated - typically to log the value of the expression.

• :sc/dynamic-vars (Seq of symbols, defaults to []) The names of the dynamic Vars which bindings should be captured at evaluation time.

• :sc/called-from (expression, defaults to nil) When set, recording and logging will only occur when the Dynamic Var sc.api.from/from holds the same value as the :sc/called-from expression (typically a keyword). Useful in generic functions, for only spying downstream of some caller. See also: sc.api/calling-from.

You may also provide your own defaults for these options by defining your own version of the spy macro: to do that, use sc.api/spy-emit.

You can completely disable the logging and recording behaviour of spy at a given Code Site by using (cs.api/disable! cs-id).

You can discard the recorded information about a given Execution Point (typically for freeing memory) by using (sc.api/dispose! ep-id).

spy-emit clj/s

(spy-emit opts expr amp-env amp-form)

Helper function for implementing (your own version of) the spy macro. See the source of spy for a usage example.

spyqt clj/s macro

(spyqt)

(spyqt expr)

(spyqt opts expr)

'Spy QuieTly' - like sc.api/spy but less verbose. Will not print the :sc.ep/value, :sc.ep/error nor :sc.cs/expr; useful when dealing with large values.

undefsc clj/s macro

(undefsc ep-or-cs-id)

Given an identifier ep-or-cs-id for a Code Site, undoes the effect of defsc by undefing the Vars defined by defsc (via ns-unmap).

ep-or-cs-id may be provided in one of the following forms:

• a positive integer literal, identifying an Execution Point, e.g 8
• a negative integer literal, identifying a Code Site, e.g -3
• an [ep-id cs-id] vector literal, e.g [8 -3]