Re: malloc in xlog.c - Mailing list pgsql-hackers

From Tom Lane
Subject Re: malloc in xlog.c
Date
Msg-id 9794.1040150351@sss.pgh.pa.us
Whole thread Raw
In response to malloc in xlog.c  (Bruce Momjian <pgman@candle.pha.pa.us>)
List pgsql-hackers
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> I see a few malloc's in backend/access/transam/xlog.c that don't check
> to see if malloc returns NULL/failure.  I think there should be at least
> an Assert() in there.

It'll dump core just fine without the help of an Assert ;-).  I don't
see that an Assert adds much of anything.

In practice those mallocs are done during startup, so it seems very
unlikely that they could fail, anyway.

> Also, seems we use malloc() a few other places where palloc should be
> used, like variable.c.  Is that correct?

Those are correct because of guc.c's conventions.  Note the comments.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Matthew Kirkwood
Date:
Subject: Re: Suggestion; "WITH VACUUM" option
Next
From: "Roderick A. Anderson"
Date:
Subject: Re: Update on replication