[Yum-devel] [PATCH/RFC] Don't exit immediately on lock held

[Charlie Brady]

On Sat, 28 Jul 2007, seth vidal wrote:

> On Fri, 2007-07-27 at 23:16 -0400, Charlie Brady wrote:
>> On Fri, 27 Jul 2007, Jeremy Katz wrote:
>>
>>> On Fri, 27 Jul 2007 20:29:00 -0400 (EDT)
>>> Charlie Brady <charlieb-yum-devel at budge.apana.org.au> wrote:
>>>
>>>> I don't see any added value in printing the message multiple times.
>>>
>>> The value is mostly avoiding the impression of things being hung.
>>
>> If I see a message which says "Waiting for exclusive lock ..." and nothing
>> more, I assume that the software is still waiting for an exclusive lock.
>
> Except when that suffer is interacting with rpm. Then you might think it
> is wedged in a FUTEX_WAIT for the rest of eternity.

Isn't that "still waiting for an exclusive lock"?

[Is suffer the same as software? Or is it specific to software built on 
top of rpm libraries? :-)]

If the software is wedged for eternity, is a new message every two seconds 
any better than one message and then nothing more? I would rather have the 
latter...