[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: sed --posix
From: |
Paolo Bonzini |
Subject: |
Re: sed --posix |
Date: |
Wed, 27 May 2009 14:52:54 +0200 |
> My claim that sed has a bug is that you chose the first behavior without
> - --posix, and the second behavior with --posix; from the script-writer's
> perspective, we want the first behavior without --posix, but the third
> behavior with --posix. You are correct that there is no conformance bug
> regardless of which of the three options you use. Rather, it is a QoI bug
> in that you are not friendly to the user given the intent of what --posix
> is documented as doing.
Okay. I see the point of implementing what you asked for. In the
meanwhile, if you want to patch the documentation to point out the
deficiency of --posix, that's worthwhile.
>>> My preference would be that
>>> 'sed --posix' failed with a verbose error message when undefined input
>>> like \| is detected, rather than silently changing behavior.
>>
>> This is hard, since regex does not have a flag for "fail on this
>> escape", just for "enable/disable this escape".
>
> Hard or not, that is exactly what I'm asking for.
Yes, I was just pointing out that unfortunately it's a gnulib/glibc
missing feature rather than a sed one.
Paolo