Re: Synchronize with imath upstream

Noah Misch <noah@leadboat.com>

From: Noah Misch <noah@leadboat.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Andres Freund <andres@anarazel.de>, pgsql-hackers@lists.postgresql.org, Alvaro Herrera <alvherre@2ndquadrant.com>, David Fetter <david@fetter.org>, Andrew Gierth <andrew@tao11.riddles.org.uk>, pgsql-hackers@postgresql.org, daniel@yesql.se, michael@paquier.xyz
Date: 2019-02-16T19:10:21Z
Lists: pgsql-hackers
On Thu, Feb 07, 2019 at 11:56:18PM -0500, Tom Lane wrote:
> Noah Misch <noah@leadboat.com> writes:
> > On Thu, Feb 07, 2019 at 10:12:05AM -0500, Tom Lane wrote:
> >> I have no particular opinion on whether pgindent should be part of the
> >> mix for imath, but I do strongly recommend setting up and documenting a
> >> reproducible import process, as I did for src/timezone.
> 
> > Good idea.  I've modified the imath.c header comment to take the form of an
> > import process; see attached imath1.29-pgedits-v2.patch.
> 
> Maybe easier to keep the instructions in a separate README file?

Most imath.c patches have cause to update those lines, and patches to other
files almost never do.  Hence, I think hackers are more likely to find and
update those lines in imath.c.  I would choose a README if concerns weren't
concentrated in one file.

> (I don't see a need to put a PG copyright in this file, when the
> mods from upstream are this small.)

The tree has been inconsistent about that, which is okay.  I was mimicking
src/port/strlcpy.c.  Others, e.g. src/port/crypt.c, haven't added a copyright.


Commits

  1. Import changes from IMath versions (1.3, 1.29].

  2. Cherry-pick security-relevant fixes from upstream imath library.