Horton

Re: BitmapHeapScan streaming read user and prelim refactoring

Melanie Plageman <melanieplageman@gmail.com>

View thread View on postgr.es
From: Melanie Plageman <melanieplageman@gmail.com>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Tomas Vondra <tomas.vondra@enterprisedb.com>, Thomas Munro <thomas.munro@gmail.com>, Heikki Linnakangas <hlinnaka@iki.fi>, Andres Freund <andres@anarazel.de>, Pg Hackers <pgsql-hackers@postgresql.org>, Nazir Bilal Yavuz <byavuz81@gmail.com>
Date: 2024-06-19T18:20:59Z
Lists: pgsql-hackers 
Thanks for taking a look at my patches, Álvaro!

On Wed, Jun 19, 2024 at 12:51 PM Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
>
> On 2024-Jun-14, Melanie Plageman wrote:
>
> > Subject: [PATCH v21 12/20] Update variable names in bitmap scan descriptors
> >
> > The previous commit which added BitmapTableScanDesc and
> > BitmapHeapScanDesc used the existing member names from TableScanDescData
> > and HeapScanDescData for diff clarity. This commit renames the members
> > -- in many cases by removing the rs_ prefix which is not relevant or
> > needed here.
>
> *Cough*  Why?  It makes grepping for struct members useless.  I'd rather
> keep these prefixes, as they allow easier code exploration.  (Sometimes
> when I need to search for uses of some field with a name that's too
> common, I add a prefix to the name and let the compiler guide me to
> them.  But that's a waste of time ...)

If we want to make it possible to use no tools and only manually grep
for struct members, that means we can never reuse struct member names.
Across a project of our size, that seems like a very serious
restriction. Adding prefixes in struct members makes it harder to read
code -- both because it makes the names longer and because people are
more prone to abbreviate the meaningful parts of the struct member
name to make the whole name shorter.

While I understand we as a project want to make it possible to hack on
Postgres without an IDE or a set of vim plugins a mile long, I also
think we have to make some compromises for readability. Most commonly
used text editors have LSP (language server protocol) support and
should allow for meaningful identification of the usages of a struct
member even if it has the same name as a member of another struct.

That being said, I'm not unreasonable. If we have decided we can not
reuse struct member names, I will change my patch.

- Melanie

Commits

  1. Fix bitmapheapscan incorrect recheck of NULL tuples

  2. Increase default maintenance_io_concurrency to 16

  3. Separate TBM[Shared|Private]Iterator and TBMIterateResult

  4. Improve read_stream.c advice for dense streams.

  5. Increase default effective_io_concurrency to 16

  6. Delay extraction of TIDBitmap per page offsets

  7. Add lossy indicator to TBMIterateResult

  8. Move BitmapTableScan per-scan setup into a helper

  9. Add and use BitmapHeapScanDescData struct

  10. Fix bitmap table scan crash on iterator release

  11. Bitmap Table Scans use unified TBMIterator

  12. Add common interface for TBMIterators

  13. Make table_scan_bitmap_next_block() async-friendly

  14. Move EXPLAIN counter increment to heapam_scan_bitmap_next_block

  15. Refactor tidstore.c iterator buffering.

  16. BitmapHeapScan: Remove incorrect assert and reset field

  17. Change BitmapAdjustPrefetchIterator to accept BlockNumber

  18. BitmapHeapScan: Use correct recheck flag for skip_fetch

  19. BitmapHeapScan: Push skip_fetch optimization into table AM

  20. BitmapHeapScan: postpone setting can_skip_fetch

  21. BitmapHeapScan: begin scan after bitmap creation

  22. Fix EXPLAIN Bitmap heap scan to count pages with no visible tuples

  23. Remove redundant snapshot copying from parallel leader to workers

  24. Remove some obsolete smgrcloseall() calls.

  25. Remove the "snapshot too old" feature.

  26. Compute XID horizon for page level index vacuum on primary.