> I looked at this and can see some of the argument on both sides, but
> if it's setting off static-analyzer warnings for some people, that
> seems like a sufficient reason to change it. We certainly make more
> significant changes than this in order to silence warnings.
>
> I rewrote the comment
Andres Freund writes:
> On 2016-04-05 11:38:27 -0300, Alvaro Herrera wrote:
>> The current arrangement looks bizantine to me, for no reason. If we
>> think that one of the two branches might do something additional to the
>> list deletion, surely that will be in a separate stanza with its own
>>
On 2016-04-05 11:38:27 -0300, Alvaro Herrera wrote:
> IMO the code is wrong.
I'm a bit confused how an intentionally duplicated block makes code
wrong...
But whatever, I found it to be clerarer that way, but apparently I'm alone.
> The current arrangement looks bizantine to me, for no reason.
> > IMO the code is wrong. There should be a single block comment
> > saying something like "Remove the node from its containing list.
> > In the FOO case, the list corresponds to BAR and therefore we
> > delete it because BAZ. In the QUUX case the list is PLUGH and we
> > delete in because THUD.
On Tue, Apr 5, 2016 at 10:38 AM, Alvaro Herrera
wrote:
> IMO the code is wrong. There should be a single block comment saying
> something like "Remove the node from its containing list. In the FOO
> case, the list corresponds to BAR and therefore we delete it because
> BAZ. In the QUUX case the
Simon Riggs wrote:
> On 5 April 2016 at 10:12, Andres Freund wrote:
>
> > On 2016-04-05 12:07:40 +0300, Aleksander Alekseev wrote:
> > > > I recall discussing this code with Andres, and I think that he has
> > > > mentioned me this is intentional, because should things be changed for
> > > > a re
On 5 April 2016 at 10:12, Andres Freund wrote:
> On 2016-04-05 12:07:40 +0300, Aleksander Alekseev wrote:
> > > I recall discussing this code with Andres, and I think that he has
> > > mentioned me this is intentional, because should things be changed for
> > > a reason or another in the future,
On 2016-04-05 12:07:40 +0300, Aleksander Alekseev wrote:
> > I recall discussing this code with Andres, and I think that he has
> > mentioned me this is intentional, because should things be changed for
> > a reason or another in the future, we want to keep in mind that a list
> > of TXIDs and a li
> I recall discussing this code with Andres, and I think that he has
> mentioned me this is intentional, because should things be changed for
> a reason or another in the future, we want to keep in mind that a list
> of TXIDs and a list of sub-TXIDs should be handled differently.
I see. If this it
On Tue, Apr 5, 2016 at 1:03 AM, Aleksander Alekseev
wrote:
> There is weird peace of code in reorderbuffer.c:
>
> ```
> /* delete from list of known subxacts */
> if (txn->is_known_as_subxact)
> {
> /* NB: nsubxacts count of parent will be too high now */
> dlist_delete
Hello
There is weired peace of code in reorderbuffer.c:
```
/* delete from list of known subxacts */
if (txn->is_known_as_subxact)
{
/* NB: nsubxacts count of parent will be too high now */
dlist_delete(&txn->node);
}
/* delete from LSN ordered list of
11 matches
Mail list logo
