Re: Use of 8192 as BLCKSZ in xlog.c

Lists: pgsql-hackers
From: Mark Kirkwood <markir(at)paradise(dot)net(dot)nz>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Use of 8192 as BLCKSZ in xlog.c
Date: 2005-11-22 02:38:34
Message-ID: 438284AA.2040106@paradise.net.nz
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

In two of the sections covered by #ifdef WAL_DEBUG there are
declarations like:

char buf[8192];

It seems to me that these should be:

char buf[BLCKSZ];

- or have I misunderstood what is going on here?

I realize that it's probably not terribly significant, as most people
will do development with BLCKSZ=8192 anyway - I'm just trying to
understand the code ... :-).

regards

Mark

From: "Qingqing Zhou" <zhouqq(at)cs(dot)toronto(dot)edu>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Use of 8192 as BLCKSZ in xlog.c
Date: 2005-11-22 02:44:54
Message-ID: dlu0m8fr7dlu0m8$2fr7$1@news.hub.org@news.hub.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers


"Mark Kirkwood" <markir(at)paradise(dot)net(dot)nz> wrote
> In two of the sections covered by #ifdef WAL_DEBUG there are declarations
> like:
>
> char buf[8192];
>
> It seems to me that these should be:
>
> char buf[BLCKSZ];
>

Those two 8192 have nothing to do with BLCKSZ, it is just an arbitrary
buffer size as long as it is big enough to hold debug information.

Regards,
Qingqing

From: Michael Glaesemann <grzm(at)myrealbox(dot)com>
To: "Qingqing Zhou" <zhouqq(at)cs(dot)toronto(dot)edu>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Use of 8192 as BLCKSZ in xlog.c
Date: 2005-11-22 02:57:52
Message-ID: CF1E0051-1729-4A64-B054-4D442D1823D3@myrealbox.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers


On Nov 22, 2005, at 11:44 , Qingqing Zhou wrote:

>
> "Mark Kirkwood" <markir(at)paradise(dot)net(dot)nz> wrote
>> In two of the sections covered by #ifdef WAL_DEBUG there are
>> declarations
>> like:
>>
>> char buf[8192];
>>
> Those two 8192 have nothing to do with BLCKSZ, it is just an arbitrary
> buffer size as long as it is big enough to hold debug information.

Would it make sense to abstract that out so it's clear that it's
*not* related to BLCKSZ? Or maybe just a comment would be enough.

Michael Glaesemann
grzm myrealbox com

From: Qingqing Zhou <zhouqq(at)cs(dot)toronto(dot)edu>
To: Michael Glaesemann <grzm(at)myrealbox(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Use of 8192 as BLCKSZ in xlog.c
Date: 2005-11-22 03:01:47
Message-ID: Pine.LNX.4.58.0511212200260.17240@eon.cs
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

On Mon, 21 Nov 2005, Michael Glaesemann wrote:

>
> Would it make sense to abstract that out so it's clear that it's
> *not* related to BLCKSZ? Or maybe just a comment would be enough.
>

"Insprite of incremental improvement", I think rename "buf" to "str" would
work,

Regards,
Qingqing

From: Mark Kirkwood <markir(at)paradise(dot)net(dot)nz>
To: Qingqing Zhou <zhouqq(at)cs(dot)toronto(dot)edu>
Cc: pgsql-hackers(at)postgresql(dot)org, grzm(at)myrealbox(dot)com
Subject: Re: Use of 8192 as BLCKSZ in xlog.c
Date: 2005-11-22 03:59:46
Message-ID: 438297B2.9040305@paradise.net.nz
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

Qingqing Zhou wrote:
> "Mark Kirkwood" <markir(at)paradise(dot)net(dot)nz> wrote
>
>>In two of the sections covered by #ifdef WAL_DEBUG there are declarations
>>like:
>>
>>char buf[8192];
>
>
>
> Those two 8192 have nothing to do with BLCKSZ, it is just an arbitrary
> buffer size as long as it is big enough to hold debug information.
>

Thanks - of course, different sort of buffer!

It is a bit more obvious now that I'm running with WAL_DEBUG enabled,
and can see that nature of the output. As has been suggested, maybe a
comment about the size and nature of 'buf' might be a nice addition.

cheers

Mark