Re: unset IFS and ${v=$*} CTLNUL leakage

2019-03-13 Thread Chet Ramey 
On 3/8/19 6:09 PM, Grisha Levit wrote:
> A few more that produce \177 in the output, though I'm not sure if
> these have defined output:
> 
> $ IFS=$' \t\n'   # or any other IFS
> $ set -- ''
> $ recho ${v= "$*" }
> <$'\177'>
> 
> $ IFS=''
> $ set -- '' ''
> $ recho ${v= "$*" }
> <$' \177 '>

These were changed last week in the 20190306 update.

> And also variations like the below all have \177 as well, but these
> I'm pretty sure are undefined:

These are all undefined, but there's nothing wrong with doing something
more reasonable than leaving the quoted nulls in there.

Chet

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, UTech, CWRUc...@case.eduhttp://tiswww.cwru.edu/~chet/

Re: Minor mistake in Bash manual

2019-03-13 Thread Chet Ramey 
On 3/12/19 4:26 PM, Todd Lehman wrote:
> On Tue, 12 Mar 2019, Chet Ramey wrote:
>> On 3/11/19 4:55 PM, Todd Lehman wrote:
>>> Hi!
>>>
>>> In the version dated 2019-Jan-07, I noticed the following mistake in the
>>> GNU Bash manual:
>>>
>>> | Associative arrays are created using
>>> | declare -A name.
>>>
>>> The period (".") after "name" should not be there.
>>
>> Is this the man page or the texinfo manual? And where are you getting it,
>> since the versions in the bash-5.0 distribution are dated 7 December 2018?
> 
> Ah!  Sorry, sorry, sorry.  Forgot to include the URL in my original mail.
> It appears here:
> 
>    https://www.gnu.org/software/bash/manual/bash.html#Arrays
> 
> in the fourth verbatim text block.  I'm assuming that's the latest &
> greatest documentation, yes?  It's dated 7 December 2018.

Thanks. This was fixed in the devel branch around a month ago.

Chet
-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, UTech, CWRUc...@case.eduhttp://tiswww.cwru.edu/~chet/