Early-adopter report for macOS 10.13

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: pgsql-hackers(at)postgreSQL(dot)org
Subject: Early-adopter report for macOS 10.13
Date: 2017-09-28 02:16:19
Message-ID: 29926.1506564979@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

I mentioned last week that updating to Xcode 9.0 while still on macOS
Sierra wasn't such a hot idea. I've now updated longfin's host to
macOS High Sierra, and that fixed the little problem with a missing
library dependency ... but all is not entirely well, because the new
version of Xcode command line tools doesn't bother to install a
symlink in /usr/lib/tclConfig.sh, causing our configure script to
fail if building --with-tcl.

I have a bug report filed with Apple, and we'll see what they say
about that. As a short-term workaround to get longfin off the ground,
I've adjusted its configure arguments to include
--with-tclconfig=/System/Library/Frameworks/Tcl.framework/Versions/8.5
which is where the missing symlink ought to point.

If Apple say "that's intentional and we aren't changing it", we're going
to have to think about adapting PGAC_PATH_TCLCONFIGSH for their brave new
world. I will try to get them to elucidate exactly what they think a
reasonable search strategy for finding tclConfig.sh ought to be. But
I don't have a lot of faith that they'll provide a usable answer;
cross-platform solutions are not their strong suit.

I did a bit of googling on this topic and came across
http://wiki.tcl.tk/12968
which suggests that searching auto_path as we do now is fundamentally
wrong. However, I don't have much faith in the solutions proposed
there either.

Perhaps Apple will say "oops" and this problem will go away quickly.
Just putting out a heads-up that there's something to worry about here.

regards, tom lane

Browse pgsql-hackers by date

  From Date Subject
Next Message Masahiko Sawada 2017-09-28 03:15:20 Re: Transactions involving multiple postgres foreign servers
Previous Message Andrew Dunstan 2017-09-28 01:11:54 Re: Arrays of domains