Please review a big, but conceptually simple, change to convert JavaDoc
tests to use text blocks where reasonable.
This is an "automated" change, albeit using a custom new utility program
written for this specific purpose. There are no manual edits to any
test files. There should be no functional change in any test file.
The conversion program scans source files looking for candidates to
convert into text blocks. Candidates are long strings containing '\n' or
'"' characters.
Some tests contain very long strings, that previously were manually
wrapped at arbitrary points in the string. The converter program wraps
these very long strings (> 240 characters) every 80 characters. The
fixed length of each "chunk" and trailing '\' helps provide a visual
distinctive block of text for the long line. Note, this algorithm
assumes there is no risk of breaking low-level Unicode sequences within
the strings in the test files.
-- Jon
JBS: https://bugs.openjdk.java.net/browse/JDK-8242532
Webrev: http://cr.openjdk.java.net/~jjg/8242532/webrev.01/
