docs(15.9): list every log file in the English guide and note that stopwords match per token

[marevol]

Summary

Two small corrections to the 15.9 tree that a release test turned up after #553.

• English log file guide: en/15.9/admin/log-guide.rst listed only six log files. The other six languages also describe fess-urls.log, searchlog.log and gc-crawler.log, which Fess writes to the logs directory and offers on the Log Files page. docs(15.9): correct eight places where the docs disagree with the product #553 corrected search.log to searchlog.log in those languages, but the English page had no entry for it. This PR adds the three entries.
• Stopwords note: docs(15.9): correct eight places where the docs disagree with the product #553 added a note saying that a word should also be added to the stopwords dictionary for the documents' language. A stopword is compared with each token the analyzer produces, though. A word that the analyzer splits into several tokens (for example, a word that mixes letters and digits, which the Japanese analyzer splits into ud / 5 / noise) is not removed when added as it is, so following the note does not stop it from matching. The note now says so and points to the _analyze API.

All seven languages for the stopwords note; English only for the log guide. 15.9 tree only.

Verification

• The three log files exist in the logs directory of a 15.9 installation after a crawl and a search. fess-urls.log comes from the crawler's log4j2.xml, and gc-crawler.log comes from the -Xlog:gc option in fess.in.sh.
• Adding a letters-and-digits word to ja/stopwords.txt, reloading the dictionary and recrawling still leaves it matching through content_ja. _analyze with the Japanese analyzer splits it into three tokens.
• docutils parses the changed pages with no new warnings.