[Parse] Disable support for multiline/extended escaping string literal in attribute message

[rintaro]

Strings of diagnostics message processed in EncodedDiagnosticMessage aren't necessarily from parsed Swift source code. That means, they might not have quotes around it. Furthermore, memory around them might not be managed. The logic in Lexer::getEncodedStringSegment() used to cause
access violation.

For now, disable multiline string literal and extended escaping in string literal for attribute message position. Considering the message might be from Clang, we cannot simply enable this.

rdar://problem/44228891

[rintaro]

@swift-ci Please ASAN test

[rintaro]

@swift-ci Please smoke test

[johnno1962]

Hi @rintaro, fair enough, reading off the end of the buffer was not clever here. I’m not too fussed about raw strings in depreciation messages but multiline strings could be useful to provide more detailed info.

How about the following in Lexer::getEncodedStringSegment():

  // Special case when being called from EncodedDiagnosticMessage(...)
  // This allows multiline strings to work as attribute messages.
  if (IndentToStrip == ~0u) {
    const char *Trailing = Bytes.end() - 1;
    while (Trailing > BytesPtr && (*Trailing == ' ' || *Trailing == '\t'))
      Trailing--;

    IndentToStrip = Trailing >= BytesPtr && *Trailing == '\n' ?
      Bytes.end() - Trailing - 1 : 0;
  }

[rintaro]

@johnno1962 Sounds good. Please post a new PR with test cases after I merge this change.

[benlangmuir]

@swift-ci please smoke test macOS

[johnno1962]

Followed up with #19219 which gets multiline messages working while respecting the bounds of the StringRef passed in.