RFC: Block String

src/language/__tests__/blockStringValue-test.js

@@ -0,0 +1,110 @@
+/**
+ * Copyright (c) 2015-present, Facebook, Inc.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+import { expect } from 'chai';
+import { describe, it } from 'mocha';
+import blockStringValue from '../blockStringValue';
+
+describe('blockStringValue', () => {
+
+  it('removes uniform indentation from a string', () => {
+    const rawValue = [
+      '',
+      '    Hello,',
+      '      World!',
+      '',
+      '    Yours,',
+      '      GraphQL.',
+    ].join('\n');
+    expect(blockStringValue(rawValue)).to.equal([
+      'Hello,',
+      '  World!',
+      '',
+      'Yours,',
+      '  GraphQL.',
+    ].join('\n'));
+  });
+
+  it('removes empty leading and trailing lines', () => {
+    const rawValue = [
+      '',
+      '',
+      '    Hello,',
+      '      World!',
+      '',
+      '    Yours,',
+      '      GraphQL.',
+      '',
+      '',
+    ].join('\n');
+    expect(blockStringValue(rawValue)).to.equal([
+      'Hello,',
+      '  World!',
+      '',
+      'Yours,',
+      '  GraphQL.',
+    ].join('\n'));
+  });
+
+  it('removes blank leading and trailing lines', () => {
+    const rawValue = [
+      '  ',
+      '        ',
+      '    Hello,',
+      '      World!',
+      '',
+      '    Yours,',
+      '      GraphQL.',
+      '        ',
+      '  ',
+    ].join('\n');
+    expect(blockStringValue(rawValue)).to.equal([
+      'Hello,',
+      '  World!',
+      '',
+      'Yours,',
+      '  GraphQL.',
+    ].join('\n'));
+  });
+
+  it('retains indentation from first line', () => {
+    const rawValue = [
+      '    Hello,',
+      '      World!',
+      '',
+      '    Yours,',
+      '      GraphQL.',
+    ].join('\n');
+    expect(blockStringValue(rawValue)).to.equal([
+      '    Hello,',
+      '  World!',
+      '',
+      'Yours,',
+      '  GraphQL.',
+    ].join('\n'));
+  });
+
+  it('does not alter trailing spaces', () => {
+    const rawValue = [
+      '               ',
+      '    Hello,     ',
+      '      World!   ',
+      '               ',
+      '    Yours,     ',
+      '      GraphQL. ',
+      '               ',
+    ].join('\n');
+    expect(blockStringValue(rawValue)).to.equal([
+      'Hello,     ',
+      '  World!   ',
+      '           ',
+      'Yours,     ',
+      '  GraphQL. ',
+    ].join('\n'));
+  });
+
+});

src/language/__tests__/kitchen-sink.graphql

@@ -46,7 +46,11 @@ subscription StoryLikeSubscription($input: StoryLikeSubscribeInput) {
 }
 
 fragment frag on Friend {
-  foo(size: $size, bar: $b, obj: {key: "value"})
+  foo(size: $size, bar: $b, obj: {key: "value", block: """
+
+      block string uses \"""
+
+  """})
 }
 
 {

src/language/__tests__/lexer-test.js

@@ -289,6 +289,121 @@ describe('Lexer', () => {
     );
   });
 
+  it('lexes block strings', () => {
+
+    expect(
+      lexOne('"""simple"""')
+    ).to.containSubset({
+      kind: TokenKind.BLOCK_STRING,
+      start: 0,
+      end: 12,
+      value: 'simple'
+    });
+
+    expect(
+      lexOne('" white space "')
+    ).to.containSubset({
+      kind: TokenKind.STRING,
+      start: 0,
+      end: 15,
+      value: ' white space '
+    });
+
+    expect(
+      lexOne('"""contains " quote"""')
+    ).to.containSubset({
+      kind: TokenKind.BLOCK_STRING,
+      start: 0,
+      end: 22,
+      value: 'contains " quote'
+    });
+
+    expect(
+      lexOne('"""contains \\""" triplequote"""')
+    ).to.containSubset({
+      kind: TokenKind.BLOCK_STRING,
+      start: 0,
+      end: 31,
+      value: 'contains """ triplequote'
+    });
+
+    expect(
+      lexOne('"""multi\nline"""')
+    ).to.containSubset({
+      kind: TokenKind.BLOCK_STRING,
+      start: 0,
+      end: 16,
+      value: 'multi\nline'
+    });
+
+    expect(
+      lexOne('"""multi\rline\r\nnormalized"""')
+    ).to.containSubset({
+      kind: TokenKind.BLOCK_STRING,
+      start: 0,
+      end: 28,
+      value: 'multi\nline\nnormalized'
+    });
+
+    expect(
+      lexOne('"""unescaped \\n\\r\\b\\t\\f\\u1234"""')
+    ).to.containSubset({
+      kind: TokenKind.BLOCK_STRING,
+      start: 0,
+      end: 32,
+      value: 'unescaped \\n\\r\\b\\t\\f\\u1234'
+    });
+
+    expect(
+      lexOne('"""slashes \\\\ \\/"""')
+    ).to.containSubset({
+      kind: TokenKind.BLOCK_STRING,
+      start: 0,
+      end: 19,
+      value: 'slashes \\\\ \\/'
+    });
+
+    expect(
+      lexOne(`"""
+
+        spans
+          multiple
+            lines
+
+        """`)
+    ).to.containSubset({
+      kind: TokenKind.BLOCK_STRING,
+      start: 0,
+      end: 68,
+      value: 'spans\n  multiple\n    lines'
+    });
+
+  });
+
+  it('lex reports useful block string errors', () => {
+
+    expect(
+      () => lexOne('"""')
+    ).to.throw('Syntax Error GraphQL request (1:4) Unterminated string.');
+
+    expect(
+      () => lexOne('"""no end quote')
+    ).to.throw('Syntax Error GraphQL request (1:16) Unterminated string.');
+
+    expect(
+      () => lexOne('"""contains unescaped \u0007 control char"""')
+    ).to.throw(
+      'Syntax Error GraphQL request (1:23) Invalid character within String: "\\u0007".'
+    );
+
+    expect(
+      () => lexOne('"""null-byte is not \u0000 end of file"""')
+    ).to.throw(
+      'Syntax Error GraphQL request (1:21) Invalid character within String: "\\u0000".'
+    );
+
+  });
+
   it('lexes numbers', () => {
 
     expect(

src/language/__tests__/parser-test.js

@@ -326,6 +326,22 @@ describe('Parser', () => {
       });
     });
 
+    it('parses block strings', () => {
+      expect(parseValue('["""long""" "short"]')).to.containSubset({
+        kind: Kind.LIST,
+        loc: { start: 0, end: 20 },
+        values: [
+          { kind: Kind.STRING,
+            loc: { start: 1, end: 11},
+            value: 'long',
+            block: true },
+          { kind: Kind.STRING,
+            loc: { start: 12, end: 19},
+            value: 'short',
+            block: false } ]
+      });
+    });
+
   });
 
   describe('parseType', () => {

src/language/__tests__/printer-test.js

@@ -127,7 +127,9 @@ describe('Printer', () => {
       }
 
       fragment frag on Friend {
-        foo(size: $size, bar: $b, obj: {key: "value"})
+        foo(size: $size, bar: $b, obj: {key: "value", block: """
+        block string uses \"""
+        """})
       }
 
       {

src/language/__tests__/visitor-test.js

@@ -590,6 +590,12 @@ describe('Visitor', () => {
       [ 'enter', 'StringValue', 'value', 'ObjectField' ],
       [ 'leave', 'StringValue', 'value', 'ObjectField' ],
       [ 'leave', 'ObjectField', 0, undefined ],
+      [ 'enter', 'ObjectField', 1, undefined ],
+      [ 'enter', 'Name', 'name', 'ObjectField' ],
+      [ 'leave', 'Name', 'name', 'ObjectField' ],
+      [ 'enter', 'StringValue', 'value', 'ObjectField' ],
+      [ 'leave', 'StringValue', 'value', 'ObjectField' ],
+      [ 'leave', 'ObjectField', 1, undefined ],
       [ 'leave', 'ObjectValue', 'value', 'Argument' ],
       [ 'leave', 'Argument', 2, undefined ],
       [ 'leave', 'Field', 0, undefined ],

src/language/ast.js

@@ -66,6 +66,7 @@ type TokenKind = '<SOF>'
                | 'Int'
                | 'Float'
                | 'String'
+               | 'BlockString'
                | 'Comment';
 
 /**
@@ -288,6 +289,7 @@ export type StringValueNode = {
   kind: 'StringValue';
   loc?: Location;
   value: string;
+  block?: boolean;
 };
 
 export type BooleanValueNode = {

src/language/blockStringValue.js

@@ -0,0 +1,64 @@
+/**
+ * Copyright (c) 2015-present, Facebook, Inc.
+ *
+ * This source code is licensed under the MIT license found in the
+ * LICENSE file in the root directory of this source tree.
+ *
+ * @flow
+ */
+
+/**
+ * Produces the value of a block string from its parsed raw value, similar to
+ * Coffeescript's block string, Python's docstring trim or Ruby's strip_heredoc.
+ *
+ * This implements the GraphQL spec's BlockStringValue() static algorithm.
+ */
+export default function blockStringValue(rawString: string): string {
+  // Expand a block string's raw value into independent lines.
+  const lines = rawString.split(/\r\n|[\n\r]/g);
+
+  // Remove common indentation from all lines but first.
+  let commonIndent = null;
+  for (let i = 1; i < lines.length; i++) {
+    const line = lines[i];
+    const indent = leadingWhitespace(line);
+    if (
+      indent < line.length &&
+      (commonIndent === null || indent < commonIndent)
+    ) {
+      commonIndent = indent;
+      if (commonIndent === 0) {
+        break;
+      }
+    }
+  }
+
+  if (commonIndent) {
+    for (let i = 1; i < lines.length; i++) {
+      lines[i] = lines[i].slice(commonIndent);
+    }
+  }
+
+  // Remove leading and trailing blank lines.
+  while (lines.length > 0 && isBlank(lines[0])) {
+    lines.shift();
+  }
+  while (lines.length > 0 && isBlank(lines[lines.length - 1])) {
+    lines.pop();
+  }
+
+  // Return a string of the lines joined with U+000A.
+  return lines.join('\n');
+}
+
+function leadingWhitespace(str) {
+  let i = 0;
+  while (i < str.length && (str[i] === ' ' || str[i] === '\t')) {
+    i++;
+  }
+  return i;
+}
+
+function isBlank(str) {
+  return leadingWhitespace(str) === str.length;
+}