Heya!

Sorry, but I don't think that forking after initializing uv loop is
supported. What `cleanup` function are you talking about?

On Tue, Feb 11, 2014 at 8:49 PM, booch <[email protected]> wrote:
>
>
> On Monday, February 10, 2014 1:51:20 PM UTC-7, booch wrote:
>>
>> Hi all,
>>
>> I'm wondering if anyone knows an easy way around an issue I'm seeing.
>>
>> I have some code using libuv for its threadpool.c feature. Unfortunately
>> this code forks. When the child exits we run the destructor code
>> theadpool.c:cleanup...since the child inherited most of the parent it has an
>> allocated threads array, and initialized == 1, but since it's a fork it has
>> no other started threads. With no other threads the uv_thread_join fails and
>> the library aborts when unloading.
>>
>> Has anyone ran into this before and found a workaround? Is there a
>> notify_of_fork method that I'm missing?
>>
>> Thanks,
>> MY
>
>
> Update: I worked around this in the code I was refactoring, but am
> interested in people's thoughts on this still. Is the shutdown behavior of
> threadpool expected; as in, works as designed and libuv thread pooling is
> incompatible with forked processes. Or is this a bug and a reset_after_fork
> or some other utility necessary?
>
> Thanks,
> MY
>
> --
> You received this message because you are subscribed to the Google Groups
> "libuv" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at http://groups.google.com/group/libuv.
> For more options, visit https://groups.google.com/groups/opt_out.

-- 
You received this message because you are subscribed to the Google Groups 
"libuv" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/libuv.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to