On Fri, Jan 27, 2023 at 12:29 PM Jeff Davis wrote:
> I am using these GUCs for testing the various collation paths in my
> collation refactoring branch.
Speaking of testing, has anyone ever tried porting Tom's random test
program[1] to ICU?
[1]
On Fri, Jan 27, 2023 at 12:34 PM Jeff Davis wrote:
> It is non-deterministic, but I tried with two generated files, and got
> similar results.
Jeff and I coordinated off-list. It turned out that the
nondeterministic nature of the program to generate test data was
behind my initial inability to
On Fri, 2023-01-27 at 11:41 -0800, Peter Geoghegan wrote:
> I cannot recreate the issue you describe.
Interesting. For my test:
glibc 2.35 ICU 70.1
gcc11.3.0LLVM 14.0.0
> It's not impossible that the perl program you wrote produces
> non-deterministic output
It is
On Thu, Jan 26, 2023 at 3:29 PM Jeff Davis wrote:
> On Thu, 2023-01-26 at 22:39 +0100, Peter Eisentraut wrote:
> > Maybe an easier way to enable or disable it in the source code with a
> > #define would serve this. Making it a GUC right away seems a bit
> > heavy-handed. Further exploration and
On Wed, Jan 25, 2023 at 4:16 PM Jeff Davis wrote:
> $ perl text_generator.pl 1000 10 > /tmp/strings.txt
>
> CREATE TABLE s (t TEXT);
> COPY s FROM '/tmp/strings.txt';
> VACUUM FREEZE s;
> CHECKPOINT;
> SET work_mem='10GB';
> SET max_parallel_workers = 0;
> SET max_parallel_workers_per_gather
On Thu, 2023-01-26 at 22:39 +0100, Peter Eisentraut wrote:
> Maybe an easier way to enable or disable it in the source code with a
> #define would serve this. Making it a GUC right away seems a bit
> heavy-handed. Further exploration and tweaking might well require
> further source code
On 25.01.23 22:16, Jeff Davis wrote:
I am highlighting this case because the existence of a single non-
contrived case or regression suggests that we may want to explore
further and tweak heuristics. That's quite natural when the heuristics
are based on a complex dependency like a collation
- AWS
From 39ed011cc51ba3a4af5e3b559a7b8de25fb895a5 Mon Sep 17 00:00:00 2001
From: Jeff Davis
Date: Sat, 21 Jan 2023 12:44:07 -0800
Subject: [PATCH v2] Introduce GUCs to control abbreviated keys sort
optimization.
The setting sort_abbreviated_keys turns the optimization on or off
ove
On Tue, 2023-01-24 at 21:42 -0500, Robert Haas wrote:
> I find it a bit premature to include this comment in the very first
> email what if other people don't like the idea?
The trust_strxfrm GUC was pulled from the larger collation refactoring
patch, which has been out for a while. The
On Sat, Jan 21, 2023 at 8:16 PM Jeff Davis wrote:
> This is fairly simple, so I plan to commit soon.
I find it a bit premature to include this comment in the very first
email what if other people don't like the idea?
I would like to hear about the cases where abbreviated keys resulted
in a
On Sat, Jan 21, 2023 at 05:16:01PM -0800, Jeff Davis wrote:
> + xreflabel="sort_abbreviated_keys">
> + sort_abbreviated_keys (boolean)
> +
> + sort_abbreviated_keys configuration
> parameter
> +
> +
> +
> +
> +Enables or disables the use of
Davis
Date: Sat, 21 Jan 2023 12:44:07 -0800
Subject: [PATCH v1] Introduce GUCs to control abbreviated keys sort
optimization.
The setting sort_abbreviated_keys turns the optimization on or off
overall. The optimization relies on collation providers, which are
complex dependencies, and the performance
12 matches
Mail list logo