diff options
Diffstat (limited to 'tests/testsuite/utf8basic.sh')
| -rwxr-xr-x | tests/testsuite/utf8basic.sh | 10 |
1 files changed, 10 insertions, 0 deletions
diff --git a/tests/testsuite/utf8basic.sh b/tests/testsuite/utf8basic.sh new file mode 100755 index 0000000..625f1ff --- /dev/null +++ b/tests/testsuite/utf8basic.sh @@ -0,0 +1,10 @@ +#/bin/sh + +# utf8basic.good originally generated with: +# uconv --from-code UTF-8 --to-code UTF-8 --from-callback substitute UTF-8-test.txt > utf8basic.good +# but modified to ignore UTF-16 surrogates which are apparently illegal. We return multiple replacement +# characters there, but the spec apparently says we are only supposed to return 1 per UTF-16 surrogate +# there are comments in the spec about "security vulnerability" but we always check if we're at the +# end of our buffer before continuing processing each byte (shouldn't all decoders do this?), so there +# shouldn't be a problem. Ignoring the UTF-16 non-conformance for now. +../utf8norm < UTF-8-test.txt |