pgsql: Improve index AMs' opclass validation procedures.

Started by Tom Laneover 10 years ago3 messagescomitters
Jump to latest
#1Tom Lane
tgl@sss.pgh.pa.us

Improve index AMs' opclass validation procedures.

The amvalidate functions added in commit 65c5fcd353a859da were on the
crude side. Improve them in a few ways:

* Perform signature checking for operators and support functions.

* Apply more thorough checks for missing operators and functions,
where possible.

* Instead of reporting problems as ERRORs, report most problems as INFO
messages and make the amvalidate function return FALSE. This allows
more than one problem to be discovered per run.

* Report object names rather than OIDs, and work a bit harder on making
the messages understandable.

Also, remove a few more opr_sanity regression test queries that are
now superseded by the amvalidate checks.

Branch
------
master

Details
-------
http://git.postgresql.org/pg/commitdiff/be44ed27b86ebd165bbedf06a4ac5a8eb943d43c

Modified Files
--------------
src/backend/access/brin/brin_validate.c | 260 ++++++++++++++++++++-------
src/backend/access/gin/ginvalidate.c | 233 ++++++++++++++++++------
src/backend/access/gist/gistvalidate.c | 241 +++++++++++++++++++------
src/backend/access/hash/hashvalidate.c | 289 +++++++++++++++++++++++-------
src/backend/access/index/Makefile | 2 +-
src/backend/access/index/amvalidate.c | 246 +++++++++++++++++++++++++
src/backend/access/nbtree/nbtvalidate.c | 274 ++++++++++++++++------------
src/backend/access/spgist/spgvalidate.c | 211 +++++++++++++++++-----
src/backend/utils/cache/lsyscache.c | 23 +++
src/include/access/amvalidate.h | 36 ++++
src/include/utils/lsyscache.h | 1 +
src/test/regress/expected/opr_sanity.out | 204 +--------------------
src/test/regress/sql/opr_sanity.sql | 171 +-----------------
13 files changed, 1417 insertions(+), 774 deletions(-)

--
Sent via pgsql-committers mailing list (pgsql-committers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-committers

#2David Rowley
dgrowleyml@gmail.com
In reply to: Tom Lane (#1)
Re: pgsql: Improve index AMs' opclass validation procedures.

On 22 January 2016 at 13:47, Tom Lane <tgl@sss.pgh.pa.us> wrote:

Improve index AMs' opclass validation procedures.

The amvalidate functions added in commit 65c5fcd353a859da were on the
crude side. Improve them in a few ways:

* Perform signature checking for operators and support functions.

* Apply more thorough checks for missing operators and functions,
where possible.

* Instead of reporting problems as ERRORs, report most problems as INFO
messages and make the amvalidate function return FALSE. This allows
more than one problem to be discovered per run.

* Report object names rather than OIDs, and work a bit harder on making
the messages understandable.

Also, remove a few more opr_sanity regression test queries that are
now superseded by the amvalidate checks.

I'm getting 3 new warnings with this change which mention about 32-bit
bit shifting then bitwise anding to a 64-bit variable.

The attached fixes the warnings

--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

Attachments:

amvalidate_64bit_bitshift.patchapplication/octet-stream; name=amvalidate_64bit_bitshift.patchDownload+5-3
#3Tom Lane
tgl@sss.pgh.pa.us
In reply to: David Rowley (#2)
Re: pgsql: Improve index AMs' opclass validation procedures.

David Rowley <david.rowley@2ndquadrant.com> writes:

I'm getting 3 new warnings with this change which mention about 32-bit
bit shifting then bitwise anding to a 64-bit variable.
The attached fixes the warnings

Pushed, thanks.

regards, tom lane

--
Sent via pgsql-committers mailing list (pgsql-committers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-committers