This can happen, e.g., when printing "UNRESOLVED: regexp.test: TEST-NAME",
where TEST-NAME contains characters that cannot be converted to the
encoding of the current output or error port.
* test-suite/guile-test (main): Set the conversion strategy for the
output and error ports to `escape'.
* test-suite/tests/regexp.test (current-output-port): Likewise.
As a side effect, it fixes tests on platforms with no 8-bit locale and
where executing regexps on characters >= 128 can lead to errors such as
`cannot convert to output locale "US-ASCII": ""\x80""'.
This commit partially reverts 7583976b ("More setlocale robustness in
regexp tests").
* test-suite/tests/regexp.test (mysetlocale, set-latin-1): Remove.
("regexp-quote"): Use `with-latin1-locale' instead of the above
procedures.
* test-suite/tests/regexp.test (set-latin-1): also add .ISO8859-1 as a
possible extension for latin-1 locales. Use set-latin-1 exclusively
to set the locale for tests.
* test-suite/tests/regexp.test (mysetlocale, set-latin-1): new functions
(with-latin1-locale): removed
(regexp-quote tests): try to print test names in locale but run tests
in ISO-8859-1.
Since the regex library expects 8-bit clean characters and
an 8-bit locale, tests of 8-bit characters need to occur within
the context of an 8-bit locale.
* test-suite/tests/regexp.test (regexp-quote tests): wrap them in an
ISO-8859-1 locale
* ice-9/regex.scm (fold-matches): If FLAGS is non-null, use
`(car flags)', not `flags'.
* test-suite/tests/regexp.test ("fold-matches"): New test prefix.
* NEWS: Update.
