public final class Formatter extends Object
This formatter uses the javac parser to generate an AST. Because the AST loses information about the non-tokens in the input (including newlines, comments, etc.), and even some tokens (e.g., optional commas or semicolons), this formatter lexes the input again and follows along in the resulting list of tokens. Its lexer splits all multi-character operators (like ">>") into multiple single-character operators. Each non-token is assigned to a token---non-tokens following a token on the same line go with that token; those following go with the next token--- and there is a final EOF token to hold final comments.
The formatter walks the AST to generate a Greg Nelson/Derek Oppen-style list of formatting
Op
s [1--2] that then generates a structured Doc
. Each AST node type has a visitor
to emit a sequence of Op
s for the node.
Some data-structure operations are easier in the list of Op
s, while others become
easier in the Doc
. The Op
s are walked to attach the comments. As the Op
s
are generated, missing input tokens are inserted and incorrect output tokens are dropped,
ensuring that the output matches the input even in the face of formatter errors. Finally, the
formatter walks the Doc
to format it in the given width.
This formatter also produces data structures of which tokens and comments appear where on the input, and on the output, to help output a partial reformatting of a slightly edited input.
Instances of the formatter are immutable and thread-safe.
[1] Nelson, Greg, and John DeTreville. Personal communication.
[2] Oppen, Derek C. "Prettyprinting". ACM Transactions on Programming Languages and Systems, Volume 2 Issue 4, Oct. 1980, pp. 465–483.
Constructor and Description |
---|
Formatter()
A new Formatter instance with default options.
|
Formatter(JavaFormatterOptions options) |
Modifier and Type | Method and Description |
---|---|
void |
formatSource(CharSource input,
CharSink output)
Format the given input (a Java compilation unit) into the output stream.
|
String |
formatSource(String input)
Format an input string (a Java compilation unit) into an output string.
|
String |
formatSource(String input,
Collection<Range<Integer>> characterRanges)
Format an input string (a Java compilation unit), for only the specified character ranges.
|
String |
formatSourceAndFixImports(String input)
Formats an input string (a Java compilation unit) and fixes imports.
|
ImmutableList<Replacement> |
getFormatReplacements(String input,
Collection<Range<Integer>> characterRanges)
Emit a list of
Replacement s to convert from input to output. |
static RangeSet<Integer> |
lineRangesToCharRanges(String input,
RangeSet<Integer> lineRanges)
Converts zero-indexed, [closed, open) line ranges in the given source file to character ranges.
|
public Formatter()
public Formatter(JavaFormatterOptions options)
public void formatSource(CharSource input, CharSink output) throws FormatterException, IOException
FormatterException
- if the input cannot be parsedIOException
public String formatSource(String input) throws FormatterException
Leaves import statements untouched.
input
- the input stringFormatterException
- if the input string cannot be parsedpublic String formatSourceAndFixImports(String input) throws FormatterException
Fixing imports includes ordering, spacing, and removal of unused import statements.
input
- the input stringFormatterException
- if the input string cannot be parsedpublic String formatSource(String input, Collection<Range<Integer>> characterRanges) throws FormatterException
input
- the input stringcharacterRanges
- the character ranges to be reformattedFormatterException
- if the input string cannot be parsedpublic ImmutableList<Replacement> getFormatReplacements(String input, Collection<Range<Integer>> characterRanges) throws FormatterException
Replacement
s to convert from input to output.input
- the input compilation unitcharacterRanges
- the character ranges to reformatReplacement
s, sorted from low index to high index, without overlapsFormatterException
- if the input string cannot be parsedCopyright © 2015–2017 Google Inc.. All rights reserved.