Re: trap output piped to another program gives incomplete results

[Chet Ramey]

On 4/10/10 12:18 AM, Siddhesh Poyarekar wrote:
> On Sat, Apr 10, 2010 at 8:51 AM, Chet Ramey <chet.ramey@case.edu> wrote:
>> On 3/31/10 1:45 AM, Siddhesh Poyarekar wrote:
>>
>>> Bash Version: 4.1
>>> Patch Level: 2
>>> Release Status: release
>>>
>>> Description:
>>>               When output of trap is piped to another command, say, less or
>>>               cat -, it shows only ignored signals.
>>
>> As the standard is currently written, this is the correct behavior.
>>
>> The standard considers traps in subshells in two places: where it talks
>> about "subshell environments" and where it describes the behavior of trap.
>>
>> The former says only that all commands in a pipeline are run in a subshell
>> environment, and that in a subshell environment traps "are set to the
>> default values."  The latter says that "when a subshell is entered, traps
>> that are not being ignored are set to the default actions."  This is an
>> obvious contradiction, and most shells implement the latter.
>>
>> Bash and some other shells special-case command substitution, because the
>> standard includes an example illustrating the use of command substitution
>> to save and restore traps.  This is not univeral -- dash, for instance,
>> does not do it.
>>
> 
> Thanks, will you consider fixing this or accept a patch to fix this? A
> fix along the lines of the standards change Eric pointed out ought to
> work I think. I'm not sure at what point the change becomes
> effective/enforceable.

I'll implement whatever the standard eventually ends up saying in bash-4.2.
I've already asked for a clarification of the points above.

Chet

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
                 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRU    chet@case.edu    http://cnswww.cns.cwru.edu/~chet/