On 1/14/12, Måns Rullgård <[email protected]> wrote:
> Sean McGovern <[email protected]> writes:
>
>> On Thursday, December 29, 2011, Måns Rullgård <[email protected]> wrote:
>>> Sean McGovern <[email protected]> writes:
>>>
>>>> On Thursday, December 29, 2011, Måns Rullgård <[email protected]> wrote:
>>>>>
>>>>> What happens if you remove the () around the check_exec call?
>>>>>
>>>>
>>>> No difference other than the Terminated line ending up in config.log
>>>> instead of on the terminal.
>>>>
>>>>> The more interesting question is why this is sending a fatal signal to
>>>>> the shell in the first place.
>>>>>
>>>>
>>>> Most likely in the suncc case it is actually crashing.
>>>
>>> Doesn't matter, the *shell* should never receive the signal.
>>
>> I finally got an answer[1] on this -- believe it or not this is actually
>> _expected behaviour_ for Korn shell *and any of its derivates* (of which
>> /usr/xpg4/bin/sh is one as it is based on ksh88).
>>
>> So I guess we should trap SIGTERM and process it accordingly to indicate
>> the test failed? Or do signal traps propagate down to subshells making
>> this
>> impossible?
>>
>> [1] See the last paragraph of this recent submission on autoconf-patches
>> ML:
>> http://lists.gnu.org/archive/html/autoconf-patches/2011-09/msg00005.html
>
> That's just absurd.
>

I agree, but I think we should put this to rest.

Especially since Solaris 11 uses bash as /bin/sh -- I just needed a
coping strategy until I make the jump. Thanks for the quick fix. :)

-- Sean McG.
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to