On Thu, May 5, 2011 at 12:14 PM, Alex Hunsaker <badalex@gmail.com> wrote:
> On Thu, May 5, 2011 at 06:51, Alvaro Herrera <alvherre@commandprompt.com>=
wrote:
>> Excerpts from Alex Hunsaker's message of mi=E9 may 04 23:53:34 -0300 201=
1:
>>
>>> After playing with it a bit more I see 2 clear options:
>>> 1) make $_TD global like %_SHARED. This should not cause any problems
>>> as we make $_TD private via local() before each trigger call. Also pre
>>> 9.1 non trigger functions could still access and check the definedness
>>> of $_TD so if someone was relying on that (for whatever unknown
>>> reason) that will work again.
>>
>> This is strange. =A0Are you saying that there's no decent way to make a
>> variable global in C code?
>
> Im sure we could... I don't see any reason to do it in C. (performance
> or otherwise)
>
> In other news I found another bug with this-- it was trying to
> local($_TD) by using SAVESPTR() when it seems it really should be
> using save_item(). Currently its not really localizing $_TD, which at
> the very least means recursive triggers might modify the callers $_TD.
> Ugh.
>
> Fixed in the attached plus added regression tests for both issues (use
> strict; && Global symbol "$_TD" requires explicit package name, test
> recursive trigger calls). Although Ill admit, given the point we are
> in the release I could see a revert also being justified.
>
> Greg, big thanks for testing! keep it up! :)
Do we need to apply this patch?
--=20
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company