Horton

Re: [postgresql 10 beta3] unrecognized node type: 90

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Tom Lane <tgl@sss.pgh.pa.us>, Rushabh Lathia <rushabh.lathia@gmail.com>
Cc: Robert Haas <robertmhaas@gmail.com>, pgsql-bugs@postgresql.org
Date: 2017-08-11T20:26:29Z
Lists: pgsql-bugs, pgsql-hackers 
On 2017-08-11 16:20:51 -0400, Tom Lane wrote:
> I wrote:
> > So, apparently somebody wrote ExecReScanGatherMerge, but never bothered
> > to plug it into ExecReScan.  From which we may draw depressing conclusions
> > about how much it's been tested.
> 
> While I'm bitching ... the code coverage report at
> 
> https://coverage.postgresql.org/src/backend/executor/nodeGatherMerge.c.gcov.html
> 
> also leaves one with less than a warm feeling about the extent of test
> coverage on this file.  heap_compare_slots isn't invoked even once?

I complained about this before at
http://archives.postgresql.org/message-id/20170401022605.4wag26gtyzhny7ue%40alap3.anarazel.de
but I just noticed that Rushabh appears to have sent a patch adding
coverage. Missed that somehow, will apply.

- Andres

Commits

  1. Separate reinitialization of shared parallel-scan state from ExecReScan.

  2. Restore test case from a2b70c89ca1a5fcf6181d3c777d82e7b83d2de1b.

  3. Force rescanning of parallel-aware scan nodes below a Gather[Merge].

  4. Fix ExecReScanGatherMerge.

  5. Add missing call to ExecReScanGatherMerge.