Re: memory leak in libxml2 - fix

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
Cc: Itagaki Takahiro <itagaki(dot)takahiro(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: memory leak in libxml2 - fix
Date: 2010-11-26 19:14:08
Message-ID: 14418.1290798848@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com> writes:
> 2010/11/26 Itagaki Takahiro <itagaki(dot)takahiro(at)gmail(dot)com>:
>> Why did you change doctree and ctxt to global variables?
>> I'm not sure why /* xmlFreeDoc(doctree); */ is commented out
>> at the end of pgxml_xpath(), but is it enough to enable the code?

> I am thinking, so you must not to call xmlFreeDoc(doctree) early.
> Probably xmlXPathCastToXXX reading a doctree.

Those static variables are really ugly, and what's more this patch only
stops some of the leakage. Per experimentation, the result object from
pgxml_xpath has to be freed too, once it's been safely converted to
whatever the end result type is. You can see this by watching

select sum(xpath_number('<data>' || generate_series || '</data>','/data')) from
generate_series(1,500000);

which still shows leakage with the submitted patch. I cleaned it up
as per attached, which doesn't show any leakage.

regards, tom lane

Attachment Content-Type Size
xml2-leakage-2.patch text/x-patch 10.0 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Bruce Momjian 2010-11-26 19:19:51 Re: Re: [BUGS] BUG #5650: Postgres service showing as stopped when in fact it is running
Previous Message Heikki Linnakangas 2010-11-26 19:06:18 Re: Assertion failure on hot standby