Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Mime-Version: 1.0 (1.0)
Subject: Re: Collation version tracking for macOS
From: Jeremy Schneider <schneider@ardentperf.com>
In-Reply-To: <0867fe37-abbd-77ba-aafc-572074978bb0@amazon.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Thomas Munro <thomas.munro@gmail.com>,
 Peter Eisentraut <peter.eisentraut@enterprisedb.com>,
 pgsql-hackers <pgsql-hackers@postgresql.org>
Date: Mon, 6 Jun 2022 17:42:41 -0700
Message-Id: <127AA2FF-F097-464D-B8D6-761141EF504D@ardentperf.com>
References: <0867fe37-abbd-77ba-aafc-572074978bb0@amazon.com>
To: Jim Nasby <nasbyj@amazon.com>
Archived-At: 
 <https://www.postgresql.org/message-id/127AA2FF-F097-464D-B8D6-761141EF504D%40ardentperf.com>
Precedence: bulk


> On Jun 6, 2022, at 17:10, Jim Nasby <nasbyj@amazon.com> wrote:
> =EF=BB=BFIgnoring broken backups, segfaults and data corruption as a "rant=
" implies that we simply throw in the towel and tell users to suck it up or s=
witch engines.


Well now, let=E2=80=99s be clear, I was the one who called my email a =E2=80=
=9Crant=E2=80=9D.  =F0=9F=99=82

And I do apologize for that - it was grumpy and impulsive and Tom isn=E2=80=99=
t wrong that rants don=E2=80=99t usually help move things forward.

Thomas - thanks for the link back to one of the threads. I spent some time r=
eading through that and it=E2=80=99s a lot of material; I haven=E2=80=99t re=
ad the whole thread yet. If you have some others that would also be particul=
arly good background, let me know. I=E2=80=99m doing a chunk of this in my s=
pare time at the moment, but I do want to keep getting more up to speed. I w=
as pulled into a bunch of various things related to PostgreSQL and ICU and c=
ollation and OS=E2=80=99s over the past couple years, so I learned a lot fro=
m on-the-ground experience and I am interested in trying to get a little mor=
e involved in the conversation here.

Personally, I really do think there should at least be an *option* to tell t=
he DB to fully error rather than just warn on version mismatch. Correctness m=
atters to many users, and being able to *trust* string comparisons are corre=
ct is pretty damn fundamental all throughout a database. It really doesn=E2=80=
=99t get any more basic and the potential for bad things to happen is pretty=
 astronomical, if you can=E2=80=99t trust those. I understand the consternat=
ion about dealing with upgrades of large & busy databases, but I=E2=80=99m s=
till surprised that the community consensus arrived at the present behavior,=
 and I have a lot of reading to do, to really understand how that happened a=
nd where the dialogue is today.

Multiple versions of ICU sounds nice for users who need real linguistic coll=
ation (like what Oracle and DB2 offer), but I still feel like there needs to=
 be a super simple basic =E2=80=9Cpseudo-linguistic=E2=80=9D collation baked=
 in, that=E2=80=99s =E2=80=9Cgood enough=E2=80=9D for 99% of users and that i=
s guaranteed to be the same everywhere on every platform and just won=E2=80=99=
t ever change. I think glibc needs to be phased out somehow. At a minimum, n=
ot the default for new users=E2=80=A6 to stop the bleeding. If MySQL wasn=E2=
=80=99t GPL then I=E2=80=99d say to just copy their collations. I=E2=80=99d b=
e reluctant to spend too much time on a POC now though, it feels like my ide=
a is the outlier and the general PG hacker consensus would be to reject this=
 idea. (But maybe I=E2=80=99m wrong?)

Anyway, again, apologies for my pants-on-fire email last week. I hope I can e=
njoy a few beers someday - or coffee for the non-drinkers - with a few other=
 PG collation nerds (which I never set out to be, but it may have befallen m=
e <g>).

-Jeremy


Sent from my TI-83