Re: Type assertions without GCC builtins - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: Type assertions without GCC builtins
Date
Msg-id 2a965ac6-fa42-4054-bee0-b1618e7729d6@eisentraut.org
Whole thread
In response to Re: Type assertions without GCC builtins  ("Jelte Fennema-Nio" <postgres@jeltef.nl>)
Responses Re: Type assertions without GCC builtins
List pgsql-hackers
[Replace __builtin_types_compatible_p with _Generic]

Here is an updated and expanded patch set.

The first patch replaces __builtin_types_compatible_p with _Generic.  It 
is similar to the previous patches in this thread.  I added some more 
code comments about what is supposed to work and how.

I did not include any C++ support here.  I feel this ought to be done as 
a separate patch.  This patch already carries portability risk and will 
require some buildfarm animal changes, so it seems better not to overlay 
any C++ issues.

As we had discussed before, this does not work with VS 2019.  We no 
longer have CI coverage for VS 2019, and I raised this issue at the CI 
session at the pgconf.dev unconference, and the consensus was that we 
should drop VS 2019 support.  So I'm including a patch for that here. 
This is just a documentation patch.  But we can also drop some support 
for early VS 2022 versions (_MSC_VER < 1933), since it's unlikely that 
those would be relevant.

And then, since we are dropping support for gcc 4.8, which means RHEL 7, 
we also drop support for some old ICU and Tcl versions.  I have included 
patches for that there, too.

Required buildfarm updates:

VS 2019:
- 
https://buildfarm.postgresql.org/cgi-bin/show_history.pl?nm=drongo&br=master

RHEL 7:
- 
https://buildfarm.postgresql.org/cgi-bin/show_history.pl?nm=siskin&br=master
- 
https://buildfarm.postgresql.org/cgi-bin/show_history.pl?nm=rhinoceros&br=master
- https://buildfarm.postgresql.org/cgi-bin/show_history.pl?nm=buri&br=master

Attachment

pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: Reject negative max_retention_duration values
Next
From: Chao Li
Date:
Subject: Re: Reject negative max_retention_duration values