Fix regression tests of unaccent to work without UTF8 support The tests of unaccent rely on UTF8 characters, and unlike any other test suite in the tree (fuzzystrmatch, citext, hstore, etc.), they would fail if run on a database that does not support UTF8 encoding.
This commit fixes the tests of unaccent so as these are skipped when run on a database without UTF8 support, using the same method as the other test suits based on \if, getdatabaseencoding() and an alternate output file. This has been broken for a long time, but nobody has complained about that either, so no backpatch is done. This can be reproduced with something like REGRESS_OPTS="--no-locale --encoding=sql_ascii", for instance. To defend against that, this module's Makefile and meson.build enforced a UTF8 encoding without locales, but it did not offer protection for options given by REGRESS_OPTS. This switch makes this regression test suite more consistent with all the others, as well. Reviewed-by: Peter Eisentraut Discussion: https://postgr.es/m/ziq1huniv2ksw...@paquier.xyz Branch ------ master Details ------- https://git.postgresql.org/pg/commitdiff/44e73a498c5f3ad7e39dc70821c2d8316d3207f2 Modified Files -------------- contrib/unaccent/Makefile | 4 ---- contrib/unaccent/expected/unaccent.out | 16 +++++++++------- contrib/unaccent/expected/unaccent_1.out | 8 ++++++++ contrib/unaccent/meson.build | 1 - contrib/unaccent/sql/unaccent.sql | 14 +++++++++++--- 5 files changed, 28 insertions(+), 15 deletions(-)