On 09.08.2011 18:04, Alvaro Herrera wrote:
> Excerpts from Heikki Linnakangas's message of mar ago 09 08:32:43 -0400 2011:
>> On 09.08.2011 13:25, Heikki Linnakangas wrote:
>>> On 08.08.2011 22:11, Alvaro Herrera wrote:
>>>> Perhaps the easiest way to fix it is as you suggest, by declaring the
>>>> struct to take a pointer rather than the value directly. Not sure how
>>>> to make both cases work sanely; the add_string_reloption path will need
>>>> updates.
>>>
>>> Agreed, I propose the attached patch to do that.
>>
>> Committed this.
>
> Thanks.
>
> I think I vaguely remember that the reason for doing it this way is that
> the copy into the relcache worked, i.e. if the originals went away,
> there was no dangling pointer. Did you test this?
These strings are never freed, so I don't think it's possible to end up
with a dangling pointer.
-- Heikki Linnakangas EnterpriseDB http://www.enterprisedb.com