Re: Clarification needed on signal spec EXIT
Thanks for answering, however...
On Sat, Oct 20, 2012 at 9:01 PM, Chet Ramey wrote:
> On 10/16/12 4:00 PM, Bob Proulx wrote:
>> Francis Moreau wrote:
>>> --
>>> main_cleanup () { echo main cleanup; }
>>> submain_cleanup () { echo sub cleanup; }
>>>
>>> trap main_cleanup EXIT
>>>
>>> task_in_background () {
>>> echo "subshell $BASHPID"
>>>
>>> while :; do
>>> # echo "FOO"
>>> sleep 1
>>> done
>>> echo "subshell exiting..."
>>> }
>>>
>>> {
>>> trap submain_cleanup EXIT
>>> trap
>>> task_in_background
>>> } &
>>>
>>> echo exiting...
>>> --
>>>
>>> Sending TERM signal to the subshell doesn't make "submain_cleanup()"
>>> to be called.
>>
>> And it does in ksh93. Hmm... And it does if I comment out the line
>> "trap main_cleanup EXIT". It seems to only set the trap if no trap
>> handler was previously set.
>
> Yes, this is a bug in bash-4.2. The subshell doesn't properly reinitialize
> the traps if a trap has already been set in the parent. This was fixed in
> late July and will be in the next version of bash.
That doesn't explain the following (taken from a previous post):
Even weirder: if the subshell exits by calling "exit 1", then the trap
handler is not called at all regardless of whether or not the "trap
main_cleanup EXIT" is commented you.
To make that case work, you need to use the "( ... ) &" construct
instead of "{ ... } &" one.
Thanks.
--
Francis
Re: Clarification needed on signal spec EXIT
On 10/21/12 7:35 AM, Francis Moreau wrote:
>> Yes, this is a bug in bash-4.2. The subshell doesn't properly reinitialize
>> the traps if a trap has already been set in the parent. This was fixed in
>> late July and will be in the next version of bash.
>
> That doesn't explain the following (taken from a previous post):
>
> Even weirder: if the subshell exits by calling "exit 1", then the trap
> handler is not called at all regardless of whether or not the "trap
> main_cleanup EXIT" is commented you.
>
> To make that case work, you need to use the "( ... ) &" construct
> instead of "{ ... } &" one.
I think I understand this, but let me make sure. Where are you putting
the `exit 1'? If you put it as the last statement in task_in_background,
it works fine. If you put it after the call to task_in_background in the
lambda subshell, it works fine. (Both cases with the parent's EXIT trap
commented out and using a brace group.)
Do you mean to replace the call to task_in_background in the lambda
subshell with `exit 1'? If so, you're right that it doesn't work in
bash-4.2. It's the same fix from the same bug report:
http://lists.gnu.org/archive/html/bug-bash/2012-07/msg00084.html
The fixes for that report are all in the bash-20120727 snapshot, if you
want to look at that in the devel branch of the git tree on savannah.
Chet
--
``The lyf so short, the craft so long to lerne.'' - Chaucer
``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRUc...@case.eduhttp://cnswww.cns.cwru.edu/~chet/
