final class ParsingRun[+T] extends AnyRef
Models an in-progress parsing run; contains all the mutable state that may be necessary during the parse, in order to avoid the individual parsers needing to perform their own allocations and instantiations, and greatly improving performance
There are a few patterns that let us program with these mutable variables in a sort-of-pure-functional way:
- If a parser that wishes to ignore changes to a field within their child parsers, a common pattern is to save the value of the field before the wrapped parser runs, and then re-set the field. e.g. this can be used to backtrack index after a lookahead parser finishes
- If a parser wants to read the value of the field "returned" by multiple child parsers, make sure to read the field into a local variable after each child parser is complete to make sure the value you want from an earlier child isn't stomped over by a later child
In general, for a parser to "return" a value in a mutable field, it is sufficient to simply set the value of that field before returning. It is the parent-parser's responsibility to make sure it reads out the value of the field to a local variable before running another child parser that will over-write the mutable field
- Alphabetic
- By Inheritance
- ParsingRun
- AnyRef
- Any
- Hide All
- Show All
- Public
- All
Instance Constructors
-
new
ParsingRun(input: ParserInput, startIndex: Int, originalParser: (ParsingRun[_]) ⇒ ParsingRun[_], traceIndex: Int, instrument: Instrument, terminalMsgs: Msgs, aggregateMsgs: Msgs, shortMsg: Msgs, lastFailureMsg: Msgs, failureStack: List[(String, Int)], isSuccess: Boolean, logDepth: Int, index: Int, cut: Boolean, successValue: Any, verboseFailures: Boolean, noDropBuffer: Boolean, misc: Map[Any, Any])
- input
The input to the parsing run, as a ParserInput.
- startIndex
Where the parse initially started, so as to allow
.result.tracedto re-create it with tracing enabled.- originalParser
The original parsing function we used to start this run, so as to allow
.result.tracedto re-create it with tracing enabled.- traceIndex
The index we wish to trace if tracing is enabled, else -1. Used to find failure messages to aggregate into
terminalMsgs- instrument
Callbacks that can be injected before/after every
P(...)parser.- terminalMsgs
When tracing is enabled, this collects up all the upper-most failures that happen at traceIndex (in Lazy wrappers) so they can be shown to the user at end-of-parse as suggestions for what could make the parse succeed. For terminal parsers like LiteralStr, it just aggregate's the string representation. For composite parsers like
a ~ bor!awhich may fail at traceIndex even without any of their wrapped terminal parsers failing there, it makes use of the shortMsg as the string representation of the composite parser.- shortMsg
When tracing is enabled, this contains string representation of the last parser to run. Since parsers aren't really objects, we piece together the string in the parser body and return store it here, and an enclosing parser may fetch it back out to help build its own string representation. If the last parser started before the
traceIndex, we only aggregate the portion of the parser msg that takes place aftertraceIndex- failureStack
The stack of named
P(...)parsers in effect when the failure occured; only constructed when tracing is enabled viatraceIndex != -1- isSuccess
Whether or not the parse is currently successful
- logDepth
How many nested
.logcalls are currently surrounding us. Used to nicely indent the log output so you can see which parsers are nested within which other parsers; -1 means logging is disabled- index
The current index of the parse
- cut
Has the current parse been prevented from backtracking? This field starts as
trueat top-level, since there is nowhere to backtrack to. Upon entering a parser that can backtrack, such as|or.?or.rep, it is set tofalse, and re-set totrueupon encountering a./or~/cut operator that prevents backtracking.- successValue
The currently returned success value
- verboseFailures
Whether or not we are currently collecting lastFailureMsgs; defaults to false, unless traceIndex is set OR we are inside a LogByNameOps.log call which necessitates tracing in order to print out failure traces during logging
- noDropBuffer
Flag that prevents the parser from dropping earlier input. Used for the
.!capture operator that needs the input around to return as a string, theNoCutoperator that forces backtracking (regardless of internal cuts), or whitespace consumption which implicitly backtracks if the parser on the RHS of the whitespace fails or consumes 0 characters. The value for this field is lexically scoped, but it is up to the individual parser method implementations to set the values and remember to un-set it to the previous value after they finish. Forgetting to re-set it to the previous value can cause strange behavior or crashes.- misc
Additional key-value metadata that a user can attach to a parsing run, and manage however they like. Not as high performance as the built-in fields of ParsingRun, but perhaps sufficient to satisfy ad-hoc needs e.g. keeping track of indentation or other contextual information
Value Members
-
final
def
!=(arg0: Any): Boolean
- Definition Classes
- AnyRef → Any
-
final
def
##(): Int
- Definition Classes
- AnyRef → Any
-
final
def
==(arg0: Any): Boolean
- Definition Classes
- AnyRef → Any
- var aggregateMsgs: Msgs
-
final
def
asInstanceOf[T0]: T0
- Definition Classes
- Any
- def augmentFailure(index: Int, cut: Boolean): ParsingRun[Nothing]
- def augmentFailure(index: Int): ParsingRun[Nothing]
- def checkForDrop(): Boolean
-
def
clone(): AnyRef
- Attributes
- protected[lang]
- Definition Classes
- AnyRef
- Annotations
- @throws( ... ) @native()
- var cut: Boolean
-
final
def
eq(arg0: AnyRef): Boolean
- Definition Classes
- AnyRef
-
def
equals(arg0: Any): Boolean
- Definition Classes
- AnyRef → Any
- var failureStack: List[(String, Int)]
-
def
finalize(): Unit
- Attributes
- protected[lang]
- Definition Classes
- AnyRef
- Annotations
- @throws( classOf[java.lang.Throwable] )
- def freshFailure(startPos: Int): ParsingRun[Nothing]
- def freshFailure(): ParsingRun[Nothing]
- def freshSuccess[V](value: V, index: Int, cut: Boolean): ParsingRun[V]
- def freshSuccess[V](value: V, cut: Boolean): ParsingRun[V]
- def freshSuccess[V](value: V, index: Int): ParsingRun[V]
- def freshSuccess[V](value: V): ParsingRun[V]
- def freshSuccessUnit(index: Int): ParsingRun[Unit]
- def freshSuccessUnit(): ParsingRun[Unit]
-
final
def
getClass(): Class[_]
- Definition Classes
- AnyRef → Any
- Annotations
- @native()
-
def
hashCode(): Int
- Definition Classes
- AnyRef → Any
- Annotations
- @native()
- var index: Int
- val input: ParserInput
- val instrument: Instrument
-
final
def
isInstanceOf[T0]: Boolean
- Definition Classes
- Any
- var isSuccess: Boolean
- var lastFailureMsg: Msgs
- var logDepth: Int
- val misc: Map[Any, Any]
-
final
def
ne(arg0: AnyRef): Boolean
- Definition Classes
- AnyRef
- var noDropBuffer: Boolean
-
final
def
notify(): Unit
- Definition Classes
- AnyRef
- Annotations
- @native()
-
final
def
notifyAll(): Unit
- Definition Classes
- AnyRef
- Annotations
- @native()
- val originalParser: (ParsingRun[_]) ⇒ ParsingRun[_]
- def reportAggregateMsg(newshortMsg: Msgs, newAggregateMsgs: Msgs, forceAggregate: Boolean): Unit
- def reportAggregateMsg(newshortMsg: Msgs, forceAggregate: Boolean): Unit
- def reportAggregateMsg(newshortMsg: Msgs, newAggregateMsgs: Msgs): Unit
-
def
reportAggregateMsg(newshortMsg: Msgs): Unit
Called by non-terminal parsers after completion, success or failure
Called by non-terminal parsers after completion, success or failure
This needs to be called for both successful and failed parsers, as we need to record the msg of a successful parse in case it forms part of a larger failed parse later.
For example:
- Using "a" ~ ("b" ~ "c" | "d") to parse "abe" - We report that the the parser ("b" ~ "c" | "d") failed at index 1 - That msg contains the msg of the parse "b" even though it was successful
Overloaded to minimize the amount of callsite bytecode, since we do a ton of inlining in Fastparse, and large amounts of bytecode inlined in a method can cause JVM performance problems (e.g. JIT compilation may get disabled)
- def reportParseMsg0(newshortMsg: Msgs, newAggregateMsgs: Msgs, forceAggregate: Boolean, setShortMsg: Boolean): Unit
-
def
reportTerminalMsg(startIndex: Int, newshortMsg: Msgs): Unit
Called by any terminal parser; these are parsers for which displaying sub-failures does not make sense these include:
Called by any terminal parser; these are parsers for which displaying sub-failures does not make sense these include:
- Individual strings or characters - Parsers like negation
!por.filterwhere the entire parser failing is not caused by sub-failure - Parsers like.opaque, where sub-failures are intentionally hidden and not shown to the userThese "terminal" failures will be stored in the
terminalMsgsin case a user wants to know what could have been placed at the failure point to let the parse progress - var shortMsg: Msgs
- val startIndex: Int
- var successValue: Any
-
final
def
synchronized[T0](arg0: ⇒ T0): T0
- Definition Classes
- AnyRef
- var terminalMsgs: Msgs
-
def
toString(): String
- Definition Classes
- AnyRef → Any
- val traceIndex: Int
- var verboseFailures: Boolean
-
final
def
wait(): Unit
- Definition Classes
- AnyRef
- Annotations
- @throws( ... )
-
final
def
wait(arg0: Long, arg1: Int): Unit
- Definition Classes
- AnyRef
- Annotations
- @throws( ... )
-
final
def
wait(arg0: Long): Unit
- Definition Classes
- AnyRef
- Annotations
- @throws( ... ) @native()