Re: Fwd: Core dump with nested CREATE TEMP TABLE

<div dir="ltr"><br /><div class="gmail_extra"><br /><div class="gmail_quote">On Fri, Sep 4, 2015 at 12:04 PM, Tom Lane
<spandir="ltr"><<a href="mailto:tgl@sss.pgh.pa.us" target="_blank">tgl@sss.pgh.pa.us</a>></span> wrote:<br
/><blockquoteclass="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">I
wrote:<br/> > Hmm.  I am not feeling especially comfortable about this: it's not clear<br /> > that there's
anythingpreventing a suspended portal from containing a<br /> > dangerous reference.  However, a quick trial did not
showthat it was<br /> > possible to break it --- in the cases I tried, we detected that a cached<br /> > plan was
nolonger valid, tried to rebuild it, and noticed the missing<br /> > object at that point.  So maybe it's OK.<br
/><br/></span>After further thought I concluded that this probably is safe.  The<br /> portal's original query was
createdand planned when it was opened,<br /> so it cannot contain any references to objects of the failed<br />
subtransaction. We have a hazard from queries within functions,<br /> but if the portal is suspended then no such
functionsare in progress.<br /><br /> Attached is an updated patch with comments to this effect and some<br /> minor
othercode cleanup (mainly, not assuming that CurrentResourceOwner<br /> is the right thing to use in
AtSubAbort_Portals).<br/></blockquote></div><br /></div><div class="gmail_extra">Thanks! This looks good to me.<br />--
<br/><div class="gmail_signature">Michael<br /></div></div></div>