Content token BUILD_LOG_REGEX matches the given regex against each line of the build log.

I propose to add a new content token BUILD_LOG_MULTILINE_REGEX, just like BUILD_LOG_REGEX, but with two modifications:

1. Matches are against the full content of the build log, rather than line-at-a-time. As a result, the regex can recognize multiple log lines in a single match, either by direct specification, or using the DOTALL modifier, e.g. (?s:.*).
2. No context arguments (linesBefore and linesAfter) - if the user wants this behavior, their regex can do the job.