Re: Flex output missing from 9.1a4 tarballs?

I'm seeing this failure on a build machine with an old (and therefore
unusable) version of flex:

gcc -O2 -Wall -Wmissing-prototypes -Wpointer-arith
-Wdeclaration-after-statement -Wendif-labels -Wformat-security
-fno-strict-aliasing -fwrapv -I../../../src/include -D_GNU_SOURCE
-I/usr/local/include/libxml2 -I/usr/local/include -c -o basebackup.o
basebackup.c
/usr/bin/bison -d -o repl_gram.c repl_gram.y
***
ERROR: `flex' is missing on your system. It is needed to create the
file `repl_scanner.c'. You can either get flex from a GNU mirror site
or download an official distribution of PostgreSQL, which contains
pre-packaged flex output.
***
make[3]: *** [repl_scanner.c] Error 1
make[3]: Leaving directory
`/mnt/buildfarm/pginstaller/server/source/postgres.linux/src/backend/replication'
make[2]: Leaving directory
`/mnt/buildfarm/pginstaller/server/source/postgres.linux/src/backend'
make[2]: *** [replication-recursive] Error 2
make[1]: Leaving directory
`/mnt/buildfarm/pginstaller/server/source/postgres.linux/src'
make[1]: *** [all-backend-recurse] Error 2
make: *** [all-src-recurse] Error 2

Looks like we're missing the pre-build output from the tarball.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On 11.03.2011 17:41, Dave Page wrote:
> Looks like we're missing the pre-build output from the tarball.

Yes. Tom spotted and fixed this yesterday:

commit 174f65ab00bb8de0f119a6a60d562b516ba71bba
Author: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Date: Thu Mar 10 00:03:26 2011 -0500

Fix some oversights in distprep and maintainer-clean targets.

At least two recent commits have apparently imagined that a comment in
a Makefile stating that something would be included in the distribution
tarball was sufficient to make it so. They hadn't bothered to hook
into the upper maintainer-clean targets either. Per bug #5923 from
Charles Johnson, in which it emerged that the 9.1alpha4 tarballs are
short a few files that should be there.

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com

On Fri, Mar 11, 2011 at 9:15 PM, Heikki Linnakangas
<heikki(dot)linnakangas(at)enterprisedb(dot)com> wrote:
> On 11.03.2011 17:41, Dave Page wrote:
>>
>> Looks like we're missing the pre-build output from the tarball.
>
> Yes. Tom spotted and fixed this yesterday:

I really should pay more attention to the committers list. Thanks!

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On Fri, 2011-03-11 at 17:45 +0200, Heikki Linnakangas wrote:
>
> On 11.03.2011 17:41, Dave Page wrote:
> > Looks like we're missing the pre-build output from the tarball.
>
> Yes. Tom spotted and fixed this yesterday:

I believe we need an alpha5 for post-alpha-4 fixes, including syncrep
ones.

Regards,
--
Devrim GÜNDÜZ
Principal Systems Engineer @ EnterpriseDB: http://www.enterprisedb.com
PostgreSQL Danışmanı/Consultant, Red Hat Certified Engineer
Community: devrim~PostgreSQL.org, devrim.gunduz~linux.org.tr
http://www.gunduz.org Twitter: http://twitter.com/devrimgunduz

2011/3/11 Devrim GÜNDÜZ <devrim(at)gunduz(dot)org>:
> On Fri, 2011-03-11 at 17:45 +0200, Heikki Linnakangas wrote:
>>
>> On 11.03.2011 17:41, Dave Page wrote:
>> > Looks like we're missing the pre-build output from the tarball.
>>
>> Yes. Tom spotted and fixed this yesterday:
>
> I believe we need an alpha5 for post-alpha-4 fixes, including syncrep
> ones.

Might be useful. My build scripts fell over because of the parallel
build issues too.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

2011/3/11 Dave Page <dpage(at)pgadmin(dot)org>:
> 2011/3/11 Devrim GÜNDÜZ <devrim(at)gunduz(dot)org>:
>> On Fri, 2011-03-11 at 17:45 +0200, Heikki Linnakangas wrote:
>>>
>>> On 11.03.2011 17:41, Dave Page wrote:
>>> > Looks like we're missing the pre-build output from the tarball.
>>>
>>> Yes. Tom spotted and fixed this yesterday:
>>
>> I believe we need an alpha5 for post-alpha-4 fixes, including syncrep
>> ones.
>
> Might be useful. My build scripts fell over because of the parallel
> build issues too.

Have those been fixed?

My vote would be forget about building installers for alpha4 and
instead wrap an alpha5 next week. By that time Tom will have
hopefully also finished hacking on the collation stuff.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

2011/3/11 Robert Haas <robertmhaas(at)gmail(dot)com>:
> 2011/3/11 Dave Page <dpage(at)pgadmin(dot)org>:
>> 2011/3/11 Devrim GÜNDÜZ <devrim(at)gunduz(dot)org>:
>>> On Fri, 2011-03-11 at 17:45 +0200, Heikki Linnakangas wrote:
>>>>
>>>> On 11.03.2011 17:41, Dave Page wrote:
>>>> > Looks like we're missing the pre-build output from the tarball.
>>>>
>>>> Yes. Tom spotted and fixed this yesterday:
>>>
>>> I believe we need an alpha5 for post-alpha-4 fixes, including syncrep
>>> ones.
>>
>> Might be useful. My build scripts fell over because of the parallel
>> build issues too.
>
> Have those been fixed?

Dunno - I removed the -j option from my build scripts as a workaround.
I have seen discussion of that issue though, so I assumed it was in
hand and didn't bother reporting it.

--
Dave Page
Blog: http://pgsnake.blogspot.com
Twitter: @pgsnake

EnterpriseDB UK: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On Fri, 2011-03-11 at 11:40 -0500, Robert Haas wrote:

> My vote would be forget about building installers for alpha4 and
> instead wrap an alpha5 next week.

If you need more votes for this: +1 from me. I'd like to wrap the
packages before I leave for PGEast.

Regards,
--
Devrim GÜNDÜZ
Principal Systems Engineer @ EnterpriseDB: http://www.enterprisedb.com
PostgreSQL Danışmanı/Consultant, Red Hat Certified Engineer
Community: devrim~PostgreSQL.org, devrim.gunduz~linux.org.tr
http://www.gunduz.org Twitter: http://twitter.com/devrimgunduz

2011/3/15 Devrim GÜNDÜZ <devrim(at)gunduz(dot)org>:
> On Fri, 2011-03-11 at 11:40 -0500, Robert Haas wrote:
>
>> My vote would be forget about building installers for alpha4 and
>> instead wrap an alpha5 next week.
>
> If you need more votes for this: +1 from me. I'd like to wrap the
> packages before I leave for PGEast.

My only hesitation about this is that it seems like sync rep and
collation support are both still pretty broken. Should we just not
worry about that for alpha?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> 2011/3/15 Devrim GNDZ <devrim(at)gunduz(dot)org>:
>> If you need more votes for this: +1 from me. I'd like to wrap the
>> packages before I leave for PGEast.

> My only hesitation about this is that it seems like sync rep and
> collation support are both still pretty broken. Should we just not
> worry about that for alpha?

FWIW, collations are probably still several days away from being
noticeably less broken than they were in alpha4. I have mixed feelings
about whether an alpha5 right now is useful.

regards, tom lane

On Tue, Mar 15, 2011 at 10:00 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
>> 2011/3/15 Devrim GÜNDÜZ <devrim(at)gunduz(dot)org>:
>>> If you need more votes for this: +1 from me. I'd like to wrap the
>>> packages before I leave for PGEast.
>
>> My only hesitation about this is that it seems like sync rep and
>> collation support are both still pretty broken.  Should we just not
>> worry about that for alpha?
>
> FWIW, collations are probably still several days away from being
> noticeably less broken than they were in alpha4.  I have mixed feelings
> about whether an alpha5 right now is useful.

I think we should wait another week.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On Tue, 2011-03-15 at 22:00 -0400, Tom Lane wrote:
>
> > My only hesitation about this is that it seems like sync rep and
> > collation support are both still pretty broken. Should we just not
> > worry about that for alpha?
>
> FWIW, collations are probably still several days away from being
> noticeably less broken than they were in alpha4. I have mixed
> feelings
> about whether an alpha5 right now is useful.

Fair enough. Looks like we will skip next week, too.

Regards,
--
Devrim GÜNDÜZ
Principal Systems Engineer @ EnterpriseDB: http://www.enterprisedb.com
PostgreSQL Danışmanı/Consultant, Red Hat Certified Engineer
Community: devrim~PostgreSQL.org, devrim.gunduz~linux.org.tr
http://www.gunduz.org Twitter: http://twitter.com/devrimgunduz

2011/3/16 Devrim GÜNDÜZ <devrim(at)gunduz(dot)org>:
> On Tue, 2011-03-15 at 22:00 -0400, Tom Lane wrote:
>>
>> > My only hesitation about this is that it seems like sync rep and
>> > collation support are both still pretty broken.  Should we just not
>> > worry about that for alpha?
>>
>> FWIW, collations are probably still several days away from being
>> noticeably less broken than they were in alpha4.  I have mixed
>> feelings
>> about whether an alpha5 right now is useful.
>
> Fair enough. Looks like we will skip next week, too.

Yeah, probably best. I think we're making good progress, though, so I
propose we wrap alpha5 on Monday, March 28. I don't think we'll be
all the way there yet by then, but I think we will have made enough
progress that it makes sense to get another snapshot out the door and
into the hands of testers.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

On Tue, Mar 15, 2011 at 10:00 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
>> 2011/3/15 Devrim GÜNDÜZ <devrim(at)gunduz(dot)org>:
>>> If you need more votes for this: +1 from me. I'd like to wrap the
>>> packages before I leave for PGEast.
>
>> My only hesitation about this is that it seems like sync rep and
>> collation support are both still pretty broken.  Should we just not
>> worry about that for alpha?
>
> FWIW, collations are probably still several days away from being
> noticeably less broken than they were in alpha4.

How much less broken are they now?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Tue, Mar 15, 2011 at 10:00 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> FWIW, collations are probably still several days away from being
>> noticeably less broken than they were in alpha4.

> How much less broken are they now?

I think the core code is only locally broken now, with issues like LIKE
index optimization probably not doing the right thing. The main problem
at this point is probably plpgsql. To get far with that, we're going to
need to extend the parser's Param interfaces to allow specification of a
collation for a Param. And SPI too. (Yay, still another generation of
SPI interfaces that we'll have to support forever...)

So, another couple days minimum before it's reasonable to ship. At
least if you want stuff like this to work:
http://archives.postgresql.org/message-id/20110314194507.GC13507@huehner.biz

regards, tom lane