sed-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: tests


From: Jim Meyering
Subject: Re: tests
Date: Sun, 5 Jun 2016 22:25:45 -0700

On Sun, Jun 5, 2016 at 7:32 PM, Assaf Gordon <address@hidden> wrote:
> Hello,
>
>> On Jun 4, 2016, at 01:15, Jim Meyering <address@hidden> wrote:
>> Thanks for still more tests.
>> I don't have time for the full review today, but may have spotted a
>> nit. Is this normalize-text.sh comment obsolete, because the bug it
>> refers to is the one you fixed recently?
>
> Thank you for spotting that.
>
> Attached improved 3 tests (where '\c\\' is tested).

Nice catch.
Looking at the TODO in your first patch in "tests: test text
normalization (\X sequences)",

  +# \c at end of (valid) text - normalize_text() stops, returns
control to caller.
  +# TODO: is this a bug?
  +#       compare with 'y/a/\d/' and 'y/a/\x/'
  +cat <<\EOF >exp-err-c || framework_failure_
  +sed: -e expression #1, char 7: strings for `y' command are different lengths
  +EOF
  +returns_ 1 sed 'y/a/\c/' </dev/null 2>err-c || fail=1

I agree that the current behavior seems incorrect.
What do you think about making that elicit a diagnostic complaining
about the unterminated \c directive instead?



reply via email to

[Prev in Thread] Current Thread [Next in Thread]