Thread: BUG #6465: postgresql-pltcl refuses to compile with 9.1, and FreeBSD does support threaded python

The following bug has been logged on the website:

Bug reference:      6465
Logged by:          Chris Rees
Email address:      crees@FreeBSD.org
PostgreSQL version: 9.1.2
Operating system:   FreeBSD
Description:=20=20=20=20=20=20=20=20

Hey all,

Two patches:

https://redports.org/browser/crees/databases/postgresql91-server/files/patc=
h-configure.in

>From a user (Pierre David) who has reported that the patch fixed compilati=
on
of pltcl (http://bugs.freebsd.org/164796)

https://redports.org/browser/crees/databases/postgresql91-server/files/patc=
h-config-python-m4

Girgen has disabled the threaded python check in the port for a while now,
so this could go in.

Thanks!

crees@FreeBSD.org writes:
> Two patches:

> https://redports.org/browser/crees/databases/postgresql91-server/files/patch-configure.in

This seems like a rather bad idea: it is not configure's responsibility
to create that symlink.  The bug report you reference looks like the
user has attempted to build pltcl without building the rest of the
source tree, something we do not consider supported.  In the case at
hand, the potential failure mode is that the derived errcodes.h file
that the symlink is pointing at might not exist or might not be
up-to-date.  There are other possible problems with a subtree build too,
and it would take a lot more work than just creating this one symlink
earlier to avoid them all.

> https://redports.org/browser/crees/databases/postgresql91-server/files/patch-config-python-m4

Hmm ... we have multiple buildfarm machines routinely building
--with-python on FreeBSD, and none of them have shown any sign of not
liking threaded python:
http://buildfarm.postgresql.org/cgi-bin/show_status.pl

I think we'd need more information before arbitrarily disabling this.

            regards, tom lane

Hi Tom,

Thanks for the prompt reply!

On 18 Feb 2012 18:49, "Tom Lane" <tgl@sss.pgh.pa.us> wrote:
>
> crees@FreeBSD.org writes:
> > Two patches:
>
> >
https://redports.org/browser/crees/databases/postgresql91-server/files/patch-configure.in
>
> This seems like a rather bad idea: it is not configure's responsibility
> to create that symlink.  The bug report you reference looks like the
> user has attempted to build pltcl without building the rest of the
> source tree, something we do not consider supported.  In the case at
> hand, the potential failure mode is that the derived errcodes.h file
> that the symlink is pointing at might not exist or might not be
> up-to-date.  There are other possible problems with a subtree build too,
> and it would take a lot more work than just creating this one symlink
> earlier to avoid them all.

I think I'll remove the port for the docs, and have it as an option for the
client install then.

> >
https://redports.org/browser/crees/databases/postgresql91-server/files/patch-config-python-m4
>
> Hmm ... we have multiple buildfarm machines routinely building
> --with-python on FreeBSD, and none of them have shown any sign of not
> liking threaded python:
> http://buildfarm.postgresql.org/cgi-bin/show_status.pl
>
> I think we'd need more information before arbitrarily disabling this.

The patch is to disable the error!  It says that threaded python is
absolutely fine for FreeBSD; the check has been removed in the FreeBSD port
for a while, anyway.

Chris

Chris Rees <crees@freebsd.org> writes:
> On 18 Feb 2012 18:49, "Tom Lane" <tgl@sss.pgh.pa.us> wrote:
>>> https://redports.org/browser/crees/databases/postgresql91-server/files/patch-config-python-m4

>> I think we'd need more information before arbitrarily disabling this.

> The patch is to disable the error!

Doh ... I read it backwards.  Sorry for the confusion.

            regards, tom lane

I wrote:
> Chris Rees <crees@freebsd.org> writes:
>> On 18 Feb 2012 18:49, "Tom Lane" <tgl@sss.pgh.pa.us> wrote:
>>> https://redports.org/browser/crees/databases/postgresql91-server/files/patch-config-python-m4

>>> I think we'd need more information before arbitrarily disabling this.

>> The patch is to disable the error!

> Doh ... I read it backwards.  Sorry for the confusion.

Applied, will be in next week's releases.

            regards, tom lane

On 20 February 2012 21:23, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> I wrote:
>> Chris Rees <crees@freebsd.org> writes:
>>> On 18 Feb 2012 18:49, "Tom Lane" <tgl@sss.pgh.pa.us> wrote:
>>>> https://redports.org/browser/crees/databases/postgresql91-server/files=
/patch-config-python-m4
>
>>>> I think we'd need more information before arbitrarily disabling this.
>
>>> The patch is to disable the error!
>
>> Doh ... I read it backwards. =A0Sorry for the confusion.
>
> Applied, will be in next week's releases.
>

Thank you.  The pltcl advice (build more of the tree) worked great, so
I've fixed that on our (ports) side.  We can consider this bug closed
:)

Chris