| From: | Antonin Houska <antonin(dot)houska(at)gmail(dot)com> |
|---|---|
| To: | Heikki Linnakangas <hlinnakangas(at)vmware(dot)com> |
| Cc: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: GIN improvements part 1: additional information |
| Date: | 2013-06-29 11:09:48 |
| Message-ID: | 51CEC07C.4000408@gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 06/29/2013 11:00 AM, Heikki Linnakangas wrote:
> On 27.06.2013 17:20, Antonin Houska wrote:
>> I was curious about the new layout of the data page, so I spent a while
>> looking into the code.
>> It's interesting, but I suspect 2 things are not o.k.:
>>
>> * gindatapage.c:dataIsEnoughSpace() - 'i++' in the for loop should
>> probably be 'j++', otherwise it loops forever
>
> Hmm. It won't loop forever, i is counting the number of entries that
> fit on the page, while j is used to loop through the items being
> added. However, i isn't actually used for anything (and isn't
> initialized either), so it's just dead code.
>
> - Heikki
You're right. While thinking about possible meaning of the 'i' I didn't
notice that j++ is in the 'for' construct. Stupid mistake on my side.
Tony
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Erik Rijkers | 2013-06-29 11:55:28 | Re: psql and pset without any arguments |
| Previous Message | Pavel Stehule | 2013-06-29 11:07:16 | proposal: simple date constructor from numeric values |