Beta going well

I don't think so ... let's talk beta2 in mid November, with a much broader
announcement then just -hackers, with maybe an rc1 around the end of
November, schedualing a release for January 1st or there abouts ...

Few ppl ever jump onto Beta1 of anything ... there have been some changes
since Beta, so a Beta2 is warranted ... rc1 is when we've always stated
that we are confident with it for release, so that more ppl start to jump
on ...

On Sat, 3 Nov 2001, Bruce Momjian wrote:

On Saturday 03 November 2001 07:51 pm, Marc G. Fournier wrote:

I don't think so ... let's talk beta2 in mid November, with a much broader
announcement then just -hackers, with maybe an rc1 around the end of
November, schedualing a release for January 1st or there abouts ...

FYI, I went to the developers page on the web site and clicked on the link
"Beta versions of PostgreSQL". That took me to
http://developer.postgresql.org/beta.php which said, "No beta software
currently available" Since beta 1 has been released, is the site just out of
date, or is it intentionally not listed there?

I don't think so ... let's talk beta2 in mid November, with a much broader
announcement then just -hackers, with maybe an rc1 around the end of
November, schedualing a release for January 1st or there abouts ...

Few ppl ever jump onto Beta1 of anything ... there have been some changes
since Beta, so a Beta2 is warranted ... rc1 is when we've always stated
that we are confident with it for release, so that more ppl start to jump
on ...

I am afraid you may be correct. It bothers me that we will spend
another two months doing little but waiting, and considering post-final,
there could be three months of downtime here. Yuck.

Folks, can we shorten this up? If no one is reporting on beta1, let's
roll a beta2, and if we don't get anything major in a week, can't we go
to rc1?

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

On Saturday 03 November 2001 07:51 pm, Marc G. Fournier wrote:

I don't think so ... let's talk beta2 in mid November, with a much broader
announcement then just -hackers, with maybe an rc1 around the end of
November, schedualing a release for January 1st or there abouts ...

Few ppl ever jump onto Beta1 of anything ... there have been some changes
since Beta, so a Beta2 is warranted ... rc1 is when we've always stated
that we are confident with it for release, so that more ppl start to jump
on ...

In addition to my previous post, I have been trying unsuccessfully for the
last 10 - 15 mintues to find somewhere that I can download the beta software
from. ftp.postgresql.org is at it's user limit and the message references
the mirror list found at

http://www.postgresql.org/sites.html

However when I go to this link I am redirected to

http://www.postgresql.org/

"Matthew T. O'Connor" wrote:

<snip>

http://www.postgresql.org/sites.html

Just tried this link (using Netscape) and got a :

Not Found

The requested URL /css/depts.css was not found on this server.

Seems to be a missing CSS. :(

+ Justin

However when I go to this link I am redirected to

http://www.postgresql.org/

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql.org so that your
message can get through to the mailing list cleanly

--
"My grandfather once told me that there are two kinds of people: those
who work and those who take the credit. He told me to try to be in the
first group; there was less competition there."
- Indira Gandhi

Folks, can we shorten this up? If no one is reporting on beta1, let's
roll a beta2, and if we don't get anything major in a week, can't we go
to rc1?

I've asked it before and I'll ask it again, is there somewhere TO report
success or only failures?

- Brandon

----------------------------------------------------------------------------
c: 646-456-5455 h: 201-798-4983
b. palmer, bpalmer@crimelabs.net pgp:crimelabs.net/bpalmer.pgp5

On Sat, Nov 03, 2001 at 09:08:54PM -0600, Matthew T. O'Connor wrote:

[beta1 talk]

In addition to my previous post, I have been trying unsuccessfully for the
last 10 - 15 mintues to find somewhere that I can download the beta software
from. ftp.postgresql.org is at it's user limit and the message references
the mirror list found at

http://www.postgresql.org/sites.html

However when I go to this link I am redirected to

http://www.postgresql.org/

Worse, that's been the case for at least a few weeks. I've only satisfied
my repeated downloads of 7.1.3 by having known about ftp.us.postgresql.org.
The web site has been looping, in various states of broken, as you say.

-jeremy
_____________________________________________________________________
jeremy wohl ..: http://igmus.org

Folks, can we shorten this up? If no one is reporting on beta1, let's
roll a beta2, and if we don't get anything major in a week, can't we go
to rc1?

I've asked it before and I'll ask it again, is there somewhere TO report
success or only failures?

That is a good question. Right now we only get problem reports.
However, considering we sort of stopped adding stuff around
mid-September, we really have been in beta for 1.5 months now so it is
no surprise things are looking very stable.

I just hate to wait around if there is a low probability that something
major will be reported.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

bpalmer wrote:

Folks, can we shorten this up? If no one is reporting on beta1, let's
roll a beta2, and if we don't get anything major in a week, can't we go
to rc1?

I've asked it before and I'll ask it again, is there somewhere TO report
success or only failures?

We have a place where we can report platforms which pass the regression
tests, called the "Regression Test Database" :

http://developer.postgresql.org/regress/

Is this what you're after?

:-)

Regards and best wishes,

Justin Clift

- Brandon

----------------------------------------------------------------------------
c: 646-456-5455 h: 201-798-4983
b. palmer, bpalmer@crimelabs.net pgp:crimelabs.net/bpalmer.pgp5

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql.org so that your
message can get through to the mailing list cleanly

--
"My grandfather once told me that there are two kinds of people: those
who work and those who take the credit. He told me to try to be in the
first group; there was less competition there."
- Indira Gandhi

I don't think so ... let's talk beta2 in mid November, with a much broader
announcement then just -hackers, with maybe an rc1 around the end of
November, schedualing a release for January 1st or there abouts ...

I am afraid you may be correct. It bothers me that we will spend
another two months doing little but waiting, and considering post-final,
there could be three months of downtime here. Yuck.

Folks, can we shorten this up?

In the past we've usually targeted a one-month beta cycle, haven't we?
Often it took longer, but with so few trouble reports I can't see a
justification for suddenly changing the target to ten weeks.

I would almost say that we could plan to do beta2 this week, rc1 the
week of Thanksgiving, and final about Dec 1 (barring major trouble
reports of course).

But ... there are a couple of flies in the ointment. One is that with
the kinks still not completely worked out of the new server setup,
I don't have much confidence that there are really a lot of people doing
beta testing. (If Marc ever put out an actual announcement of beta1,
I didn't get it. And we know some people have been unable to download
the beta.) The other is that with the holiday season coming up, many
people will have less spare time than usual to spend on Postgres. So
maybe Marc's unaggressive schedule proposal is appropriate.

On the whole, though, I agree with Bruce. We've been in "almost beta"
mode for two months now, we shouldn't need another two months to get to
release.

regards, tom lane

"Matthew T. O'Connor" <matthew@zeut.net> writes:

In addition to my previous post, I have been trying unsuccessfully for the
last 10 - 15 mintues to find somewhere that I can download the beta software
from.

Try ftp://ftp.us.postgresql.org/beta/

ftp.postgresql.org is at it's user limit and the message references
the mirror list found at
http://www.postgresql.org/sites.html
However when I go to this link I am redirected to
http://www.postgresql.org/

Hmm, you're right; the index-of-mirrors page has disappeared from view.
Vince, can you straighten this out?

regards, tom lane

On Sat, 3 Nov 2001, Matthew T. O'Connor wrote:

On Saturday 03 November 2001 07:51 pm, Marc G. Fournier wrote:

I don't think so ... let's talk beta2 in mid November, with a much broader
announcement then just -hackers, with maybe an rc1 around the end of
November, schedualing a release for January 1st or there abouts ...

FYI, I went to the developers page on the web site and clicked on the link
"Beta versions of PostgreSQL". That took me to
http://developer.postgresql.org/beta.php which said, "No beta software
currently available" Since beta 1 has been released, is the site just out of
date, or is it intentionally not listed there?

None of the above, I forgot to uncomment the listing in the php code!
It's there now.

Vince.
--
==========================================================================
Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net
56K Nationwide Dialup from $16.00/mo at Pop4 Networking
Online Campground Directory http://www.camping-usa.com
Online Giftshop Superstore http://www.cloudninegifts.com
==========================================================================

On Sat, 3 Nov 2001, Matthew T. O'Connor wrote:

On Saturday 03 November 2001 07:51 pm, Marc G. Fournier wrote:

I don't think so ... let's talk beta2 in mid November, with a much broader
announcement then just -hackers, with maybe an rc1 around the end of
November, schedualing a release for January 1st or there abouts ...

Few ppl ever jump onto Beta1 of anything ... there have been some changes
since Beta, so a Beta2 is warranted ... rc1 is when we've always stated
that we are confident with it for release, so that more ppl start to jump
on ...

In addition to my previous post, I have been trying unsuccessfully for the
last 10 - 15 mintues to find somewhere that I can download the beta software
from. ftp.postgresql.org is at it's user limit and the message references
the mirror list found at

http://www.postgresql.org/sites.html

However when I go to this link I am redirected to

http://www.postgresql.org/

The machine the database is on doesn't appear to be available. I guess
things will be disrupted until the game of musical machines is over :(

Vince.
--
==========================================================================
Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net
56K Nationwide Dialup from $16.00/mo at Pop4 Networking
Online Campground Directory http://www.camping-usa.com
Online Giftshop Superstore http://www.cloudninegifts.com
==========================================================================

We have a place where we can report platforms which pass the regression
tests, called the "Regression Test Database" :

http://developer.postgresql.org/regress/

Sure, can this list be reset for 7.2b1?

- Brandon

----------------------------------------------------------------------------
c: 646-456-5455 h: 201-798-4983
b. palmer, bpalmer@crimelabs.net pgp:crimelabs.net/bpalmer.pgp5

On Sun, 4 Nov 2001, Justin Clift wrote:

"Matthew T. O'Connor" wrote:

<snip>

http://www.postgresql.org/sites.html

Just tried this link (using Netscape) and got a :

Not Found

The requested URL /css/depts.css was not found on this server.

Seems to be a missing CSS. :(

fixed. lynx seems to be acting differently than it was before and rather
than doing a redirect it's downloading the redirected source.

Vince.
--
==========================================================================
Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net
56K Nationwide Dialup from $16.00/mo at Pop4 Networking
Online Campground Directory http://www.camping-usa.com
Online Giftshop Superstore http://www.cloudninegifts.com
==========================================================================

In the past we've usually targeted a one-month beta cycle, haven't we?
Often it took longer, but with so few trouble reports I can't see a
justification for suddenly changing the target to ten weeks.

That was my reading of events.

I would almost say that we could plan to do beta2 this week, rc1 the
week of Thanksgiving, and final about Dec 1 (barring major trouble
reports of course).

It would be great if we could do this. If there were server problems,
that would be a good reason to get those all ironed out, if they aren't
already, and announce a big beta2, and tell people they have ~1 week and
if we don't hear anything, we are going RC1. That will get their
attention. :-)

But ... there are a couple of flies in the ointment. One is that with
the kinks still not completely worked out of the new server setup,
I don't have much confidence that there are really a lot of people doing
beta testing. (If Marc ever put out an actual announcement of beta1,
I didn't get it. And we know some people have been unable to download
the beta.) The other is that with the holiday season coming up, many
people will have less spare time than usual to spend on Postgres. So
maybe Marc's unaggressive schedule proposal is appropriate.

Actually, the Christmas holiday season is often busy with development,
especially the week between Christmas and New Years because most offices
are slow during the period. Also, lots of people like to upgrade during
that period for the same reason, which makes releaseing a final prior to
Christmas a double win -- we get development time and they can upgrade
during a slow period.

On the whole, though, I agree with Bruce. We've been in "almost beta"
mode for two months now, we shouldn't need another two months to get to
release.

In previois betas, we have had the occasional "Wow, I am glad that
didn't get into final" bugs, but I haven't seen _any_ yet and I am
doubting if I will.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

On Sunday 04 November 2001 01:08 pm, Bruce Momjian wrote:

In the past we've usually targeted a one-month beta cycle, haven't we?
Often it took longer, but with so few trouble reports I can't see a
justification for suddenly changing the target to ten weeks.

That was my reading of events.

I am concerned in that I have yet to receive a request for RPMs of the beta.

In previous release cycles, those requests have come hot and heavy shortly
after the beta announcements. Unless the interest in RPM has drastically
reduced in the last couple of months, my gut feel is that our beta1 has had
very little exposure outside this list -- and the netizens of this list
aren't typically RPM addicts.

I have not had time in the past couple of weeks to do much with them, though.
Looking for a break in the action this week, hopefully.

I would, however, guard against rushing a release -- we _do_ have a 'solid
release' reputation to protect.
--
Lamar Owen
WGCR Internet Radio
1 Peter 4:11

Lamar Owen <lamar.owen@wgcr.org> writes:

... my gut feel is that our beta1 has had
very little exposure outside this list

Probably not, considering it has not been announced anywhere outside
this list. Ahem.

Since we've made a number of fixes in the past two weeks, I think
our next step should be to roll a beta2, and then actually announce it
[as in pgsql-announce]. We can argue more about schedule after that's
been out for a week or so.

Anyone have stuff that they need to get in there before beta2?

regards, tom lane

But ... there are a couple of flies in the ointment. One is that with
the kinks still not completely worked out of the new server setup,
I don't have much confidence that there are really a lot of people doing
beta testing. (If Marc ever put out an actual announcement of beta1,
I didn't get it. And we know some people have been unable to download
the beta.) The other is that with the holiday season coming up, many
people will have less spare time than usual to spend on Postgres. So
maybe Marc's unaggressive schedule proposal is appropriate.

I never got a beta notification, it's not mentioned on the PostgreSQL page,
and I personally have no idea from where to download it. I looked though
the Australian mirror FTP site, but I could find no trace of a beta
download.

Chris

On 04 Nov 2001 at 19:59 (-0500), Tom Lane wrote:
| Lamar Owen <lamar.owen@wgcr.org> writes:
| > ... my gut feel is that our beta1 has had
| > very little exposure outside this list
|
| Probably not, considering it has not been announced anywhere outside
| this list. Ahem.
|
| Since we've made a number of fixes in the past two weeks, I think
| our next step should be to roll a beta2, and then actually announce it
| [as in pgsql-announce]. We can argue more about schedule after that's
| been out for a week or so.
|
| Anyone have stuff that they need to get in there before beta2?

I've an in progress (er, stalled ATM) attempt to squash a bug
in ALTER TABLE RENAME, where the args to referenced in function
calls are updtead to reflect the new name[1]. There is also the
outstanding resolution to David Ford's libpq EINTR issues[2], on which
I agree with your assessment of handling EINTR the same as EINPROGRESS
for connect(), though [AFAICT] David has not confirmed that your
proposed solution solves his problem.

I'll do my best to get a patch to the list tonight for the ALTER
TABLE RENAME fixes. If I don't wrap this up tonight, it will be
next weekend before I'll have time to dive back in. I don't consider
this pending work any reason to hold up any beta2/rc1 scheduling,
since the change is very localized, but this fix should be in before
7.2 is released.

cheers,
Brent

1) http://fts.postgresql.org/db/mw/msg.html?mid=1038272
2) http://fts.postgresql.org/db/mw/msg.html?mid=1041165

--
"Develop your talent, man, and leave the world something. Records are
really gifts from people. To think that an artist would love you enough
to share his music with anyone is a beautiful thing." -- Duane Allman

Probably not, considering it has not been announced anywhere outside
this list. Ahem.

Since we've made a number of fixes in the past two weeks, I think
our next step should be to roll a beta2, and then actually announce it
[as in pgsql-announce]. We can argue more about schedule after that's
been out for a week or so.

Anyone have stuff that they need to get in there before beta2?

Yes. doesn't compile on AIX 5L. I would like to fix it before beta2
(see attached pacthes below).

However I'm not sure if it's a correct solution. Problem is, AIX 5L
has sys/inttypes.h where int8, int16, int32 and int64 are
defined. Should we detect them in configure? Also, I'm afraid it would
break other AIX versions. Comments?
--
Tatsuo Ishii

*** include/c.h.orig	Mon Oct 29 11:58:33 2001
--- include/c.h	Mon Oct 29 12:08:13 2001
***************
*** 205,213 ****
--- 205,215 ----
   *		frontend/backend protocol.
   */
  #ifndef __BEOS__				/* this shouldn't be required, but is is! */
+ #if !defined(_AIX)
  typedef signed char int8;		/* == 8 bits */
  typedef signed short int16;		/* == 16 bits */
  typedef signed int int32;		/* == 32 bits */
+ #endif /* _AIX */
  #endif	 /* __BEOS__ */

  /*
***************
*** 275,281 ****
--- 277,285 ----
  #else
  #ifdef HAVE_LONG_LONG_INT_64
  /* We have working support for "long long int", use that */
+ #if !defined(_AIX)
  typedef long long int int64;
+ #endif /* _AIX */
  typedef unsigned long long int uint64;

#else

Tatsuo Ishii <t-ishii@sra.co.jp> writes:

However I'm not sure if it's a correct solution. Problem is, AIX 5L
has sys/inttypes.h where int8, int16, int32 and int64 are
defined. Should we detect them in configure? Also, I'm afraid it would
break other AIX versions. Comments?

Perhaps have configure test for the presence of <sys/inttypes.h>
and then let c.h do

#ifdef HAVE_SYS_INTTYPES_H
#include <sys/inttypes.h>
#else
typedef signed char int8;
... etc
#endif

Could this substitute for the ugly #ifndef __BEOS__ as well?

regards, tom lane

I never got a beta notification, it's not mentioned on the PostgreSQL page,
and I personally have no idea from where to download it. I looked though
the Australian mirror FTP site, but I could find no trace of a beta
download.

I think Tom's suggestion was best. Let's package beta2, publicise it,
and see what things look like one week after that.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

Anyone have stuff that they need to get in there before beta2?

Yes. doesn't compile on AIX 5L. I would like to fix it before beta2
(see attached pacthes below).

IIRC not all (old) versions of AIX define those, thus your patch would
break those that havent :-( I am not sure we care about those old
versions
anymore though.
Also I actually see those defines as a bug in AIX, since a comment
states,
that BSD requires them, I certainly havent heard BSD'ers complain about
redefines ?

Also I don't understand why your compiler stops ? Mine only give a
warning.

These defines are only included if _ALL_SOURCE is defined. I think this
define
stems from another system header file, that gets included somewhere
else.
Maybe a better fix would be to #undef _ALL_SOURCE before including
inttypes.h ?

Andreas

Perhaps have configure test for the presence of <sys/inttypes.h>
and then let c.h do

It is directly in /usr/include/inttypes.h in AIX 4.3.2 :-(

Andreas

IIRC not all (old) versions of AIX define those, thus your patch would
break those that havent :-( I am not sure we care about those old
versions
anymore though.
Also I actually see those defines as a bug in AIX, since a comment
states,
that BSD requires them, I certainly havent heard BSD'ers complain about
redefines ?

Also I don't understand why your compiler stops ? Mine only give a
warning.

These defines are only included if _ALL_SOURCE is defined. I think this
define
stems from another system header file, that gets included somewhere
else.
Maybe a better fix would be to #undef _ALL_SOURCE before including
inttypes.h ?

Ok, let me see if it fixes the problems.
--
Tatsuo Ishii

Thus spake Tatsuo Ishii

Anyone have stuff that they need to get in there before beta2?

Yes. doesn't compile on AIX 5L. I would like to fix it before beta2
(see attached pacthes below).

However I'm not sure if it's a correct solution. Problem is, AIX 5L
has sys/inttypes.h where int8, int16, int32 and int64 are
defined. Should we detect them in configure? Also, I'm afraid it would
break other AIX versions. Comments?

I see the same problem on rs6000-ibm-aix4.3.3.0. Your patch fixes it
there as well.

-- 
D'Arcy J.M. Cain <darcy@{druid|vex}.net>   |  Democracy is three wolves
http://www.druid.net/darcy/                |  and a sheep voting on
+1 416 425 1212     (DoD#0082)    (eNTP)   |  what's for dinner.

"Tom" == Tom Lane <tgl@sss.pgh.pa.us> writes:

Tom> Tatsuo Ishii <t-ishii@sra.co.jp> writes:

However I'm not sure if it's a correct solution. Problem is,
AIX 5L has sys/inttypes.h where int8, int16, int32 and int64
are defined. Should we detect them in configure? Also, I'm
afraid it would break other AIX versions. Comments?

Tom> Perhaps have configure test for the presence of
Tom> <sys/inttypes.h> and then let c.h do

Tom> #ifdef HAVE_SYS_INTTYPES_H #include <sys/inttypes.h> #else
Tom> typedef signed char int8; ... etc #endif

Tom> Could this substitute for the ugly #ifndef __BEOS__ as well?

Hmm, AIX 4.2 and 4.1 define these in sys/ltypes.h. How that affects
the result of this discussion I have no idea.

Sincerely,

Adrian Phillips

--
Your mouse has moved.
Windows NT must be restarted for the change to take effect.
Reboot now? [OK]

On Sunday 04 November 2001 07:59 pm, Tom Lane wrote:

Lamar Owen <lamar.owen@wgcr.org> writes:

... my gut feel is that our beta1 has had
very little exposure outside this list

Probably not, considering it has not been announced anywhere outside
this list. Ahem.

Me needs another podondectomy..... :-)
--
Lamar Owen
WGCR Internet Radio
1 Peter 4:11

On Mon, 5 Nov 2001, Christopher Kings-Lynne wrote:

But ... there are a couple of flies in the ointment. One is that with
the kinks still not completely worked out of the new server setup,
I don't have much confidence that there are really a lot of people doing
beta testing. (If Marc ever put out an actual announcement of beta1,
I didn't get it. And we know some people have been unable to download
the beta.) The other is that with the holiday season coming up, many
people will have less spare time than usual to spend on Postgres. So
maybe Marc's unaggressive schedule proposal is appropriate.

I never got a beta notification, it's not mentioned on the PostgreSQL page,
and I personally have no idea from where to download it. I looked though
the Australian mirror FTP site, but I could find no trace of a beta
download.

Which "PostgreSQL page" and there is no official Australian mirror ftp
(or web for all that matter) site so they wouldn't have a copy of the
beta.

Vince.
--
==========================================================================
Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net
56K Nationwide Dialup from $16.00/mo at Pop4 Networking
Online Campground Directory http://www.camping-usa.com
Online Giftshop Superstore http://www.cloudninegifts.com
==========================================================================

On Sun, 4 Nov 2001, Lamar Owen wrote:

I am concerned in that I have yet to receive a request for RPMs of
the beta.

Sorry Lamar. I thought I saw a message saying you were putting the RPMs
together so I was waiting to hear they were ready.

I'd like the RPMs!!!

I have a RHL7.1 system that I can use to test the RPMs and probably do
the regression tests and friends on also.

Waiting for my new SDSL line so it will be a couple of days before I
report back as I have to downlaod to a fast-connected system, move onto
tape, and take home for loading and testing.

Rod
--
Let Accuracy Triumph Over Victory

Zetetic Institute
"David's Sling"
Marc Stiegler

"Zeugswetter Andreas SB SD" <ZeugswetterA@spardat.at> writes:

Maybe a better fix would be to #undef _ALL_SOURCE before including
inttypes.h ?

Is it possible to avoid including inttypes.h altogether?

regards, tom lane

On Monday 05 November 2001 10:35 am, Roderick A. Anderson wrote:

On Sun, 4 Nov 2001, Lamar Owen wrote:

I am concerned in that I have yet to receive a request for RPMs of
the beta.

Sorry Lamar. I thought I saw a message saying you were putting the RPMs
together so I was waiting to hear they were ready.

I'd like the RPMs!!!

Ah, a customer... :-)

I'll wait for 7.2b2, which fixes some things that currently prevent RPM
building without patching from CVS.

Plus, my time has been extremely tight the last couple of months, but a few
days off actually looks possible now -- and building this set shouldn't be
too difficult, thanks to PeterE's patchwork.

RPM's will be built at this time on RHL7.1and RHL7.2 -- although the 7.1
option will disappear here when I upgrade my devel server to 7.2. As stated
before, I don't currently have other machines (except LER's OpenUnix system)
on which to built and test -- so you having 7.1 will be nice further down the
cycle.
--
Lamar Owen
WGCR Internet Radio
1 Peter 4:11

Maybe a better fix would be to #undef _ALL_SOURCE before including
inttypes.h ?

Is it possible to avoid including inttypes.h altogether?

Looks like we get it from arpa/inet.h. I don't see any content why we
would need inttypes.h directly.

Andreas

On Mon, 5 Nov 2001, Lamar Owen wrote:

Ah, a customer... :-)

Too bad it wasn't the millionth so I could win a door prize or
something.

I'll wait for 7.2b2, which fixes some things that currently prevent RPM
building without patching from CVS.

Sounds good to me.

Plus, my time has been extremely tight the last couple of months, but a few
days off actually looks possible now -- and building this set shouldn't be
too difficult, thanks to PeterE's patchwork.

When I get the DSL line in I'd like to help more with the RPM packaging.
Dialup works OK but only OK.

RPM's will be built at this time on RHL7.1and RHL7.2 -- although the 7.1
option will disappear here when I upgrade my devel server to 7.2. As stated
before, I don't currently have other machines (except LER's OpenUnix system)
on which to built and test -- so you having 7.1 will be nice further down the
cycle.

Yeah I'm a late migrator. I can do 7.1 for quite awhile.

Let us(me) know when and where.

Cheers,
Rod
--
Let Accuracy Triumph Over Victory

Zetetic Institute
"David's Sling"
Marc Stiegler

On Tuesday 06 November 2001 02:35, Roderick A. Anderson wrote:

On Sun, 4 Nov 2001, Lamar Owen wrote:

I am concerned in that I have yet to receive a request for RPMs of
the beta.

Sorry Lamar. I thought I saw a message saying you were putting the RPMs
together so I was waiting to hear they were ready.

I'd like the RPMs!!!

I have a RHL7.1 system that I can use to test the RPMs and probably do
the regression tests and friends on also.

And so would I! I have a RH7.2 system plus a Mandrake 8.1 system eager to
start testing.

Horst

It is directly in /usr/include/inttypes.h in AIX 4.3.2 :-(

Isn't it linked to /usr/include/sys/inttypes.h ?
--
Tatsuo Ishii

Which "PostgreSQL page" and there is no official Australian mirror ftp
(or web for all that matter) site so they wouldn't have a copy of the
beta.

On www.postgresql.org and the Australian(-ish) mirror site is:

http://postgresql.planetmirror.com/

Chris

On Tue, 6 Nov 2001, Christopher Kings-Lynne wrote:

Which "PostgreSQL page" and there is no official Australian mirror ftp
(or web for all that matter) site so they wouldn't have a copy of the
beta.

On www.postgresql.org and the Australian(-ish) mirror site is:

http://postgresql.planetmirror.com/

www.postgresql.org is a list of mirror sites. As to the mirror you
refer to, I timestamp ALL mirrors:

http://postgresql.planetmirror.com/timestamp.txt

shows: Thu Oct 18 00:01:00 EDT 2001 not exactly up to date.

http://www.at.postgresql.org/timestamp.txt is the Austrian mirror,
it shows: Mon Nov 5 00:01:00 EST 2001

The stamp is written at midnite local time. As I said above, there
is currently no official PostgreSQL mirror in Australia. Beta info
is found on the Developer's website, I keep it separated because of
our history of stable releases, many people come to expect it out of
the betas as well and aren't happy if it doesn't have release version
stability. The developer's website is: http://developer.postgresql.org/

Vince.
--
==========================================================================
Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net
56K Nationwide Dialup from $16.00/mo at Pop4 Networking
Online Campground Directory http://www.camping-usa.com
Online Giftshop Superstore http://www.cloudninegifts.com
==========================================================================

Also I don't understand why your compiler stops ? Mine only give a
warning.

Maybe due to the difference of the compiler verion. xlc coming with
AIX 5L stops. Note that gcc doesn't. What is your OS version?

These defines are only included if _ALL_SOURCE is defined. I think this
define
stems from another system header file, that gets included somewhere
else.
Maybe a better fix would be to #undef _ALL_SOURCE before including
inttypes.h ?

It seems inttypes.h is included by types.h which is included by
stdio.h which is included by c.h. I inserted #undef into c.h but it
does not help at all:-<
--
Tatsuo Ishii

The stamp is written at midnite local time. As I said above, there
is currently no official PostgreSQL mirror in Australia. Beta info
is found on the Developer's website, I keep it separated because of
our history of stable releases, many people come to expect it out of
the betas as well and aren't happy if it doesn't have release version
stability. The developer's website is: http://developer.postgresql.org/

May I ask why there isn't an active Australian mirror? I, for one, would
find it quite helpful. I did notice when it was removed from the
www.postgresql.org 'flags' list tho.

Chris

"Zeugswetter Andreas SB SD" <ZeugswetterA@spardat.at> writes:

Maybe a better fix would be to #undef _ALL_SOURCE before including
inttypes.h ?

Is it possible to avoid including inttypes.h altogether?

It seems not possible. inttypes.h is included by types.h which is
included by stdio.h which is included by c.h.
--
Tatsuo Ishii

On Tue, 6 Nov 2001, Christopher Kings-Lynne wrote:

The stamp is written at midnite local time. As I said above, there
is currently no official PostgreSQL mirror in Australia. Beta info
is found on the Developer's website, I keep it separated because of
our history of stable releases, many people come to expect it out of
the betas as well and aren't happy if it doesn't have release version
stability. The developer's website is: http://developer.postgresql.org/

May I ask why there isn't an active Australian mirror? I, for one, would
find it quite helpful. I did notice when it was removed from the
www.postgresql.org 'flags' list tho.

We don't solicit mirrors. If someone wants to mirror, they contact us.
Recently we changed the rsync host and requirements for mirror hosts,
planetmirror hasn't responded.

Vince.
--
==========================================================================
Vince Vielhaber -- KA8CSH email: vev@michvhf.com http://www.pop4.net
56K Nationwide Dialup from $16.00/mo at Pop4 Networking
Online Campground Directory http://www.camping-usa.com
Online Giftshop Superstore http://www.cloudninegifts.com
==========================================================================

It is directly in /usr/include/inttypes.h in AIX 4.3.2 :-(

Isn't it linked to /usr/include/sys/inttypes.h ?

Yes, I didn't see that, sorry.

Andreas

Peter Eisentraut <peter_e@gmx.net> writes:

The correct way would be to check for the existance of int8, int16, etc.

Good in theory ... but ... are you sure you have included the correct
set of system headers before checking this? (It's not at all clear
to me that we know what "correct" is in this context.)

regards, tom lane

Tom Lane writes:

Tatsuo Ishii <t-ishii@sra.co.jp> writes:

However I'm not sure if it's a correct solution. Problem is, AIX 5L
has sys/inttypes.h where int8, int16, int32 and int64 are
defined. Should we detect them in configure? Also, I'm afraid it would
break other AIX versions. Comments?

Perhaps have configure test for the presence of <sys/inttypes.h>
and then let c.h do

#ifdef HAVE_SYS_INTTYPES_H
#include <sys/inttypes.h>
#else
typedef signed char int8;
... etc
#endif

This is not correct, since we don't have any portable guarantees about the
content of <sys/inttypes.h>.

The correct way would be to check for the existance of int8, int16, etc.
The Autoconf <=2.13 macros for existance of types have been deprecated in
2.50 because they're broken. (If the type is missing the replacement type
is #define'd, not typedef'd, which is claimed to be incorrect. I don't
know why offhand, but let's not start finding out now.)

A possible workaround until we update our Autoconf (not now) would be

AC_CHECK_SIZEOF(int8)

#if SIZEOF_INT8 == 0
typedef signed char int8;
#endif

because the sizeof check results in 0 if the type doesn't exist.

I have attached a patch to this effect which I ask the affected people to
try out.

This patch could theoretically be insufficient if 'long int' is 64 bits
but int64 exists and is actually 'long long int'. You would probably get
warnings about mismatched printf arguments, but I don't think there would
be actual problems.

--
Peter Eisentraut peter_e@gmx.net

Tom Lane writes:

The correct way would be to check for the existance of int8, int16, etc.

Good in theory ... but ... are you sure you have included the correct
set of system headers before checking this?

I'm sure I haven't, that's why someone is supposed to check this.

(It's not at all clear to me that we know what "correct" is in this
context.)

If the compiler is complaining that int8 is defined twice we need to check
if its already defined once and avoid a second declaration. The problem
is setting up an appropriate environment to be relatively sure about the
result of "already defined once". That's the usual procedure in autoconf
programming.

--
Peter Eisentraut peter_e@gmx.net

From: Jean-Michel POURE <jm.poure@freesurf.fr>
Subject: Java's Unicode Notation
Date: Thu, 08 Nov 2001 14:12:04 +0100
Message-ID: <4.2.0.58.20011108141018.00a59dc0@pop.freesurf.fr>

Dear Tatsuo,

Could it be possible to use the Java Unicode Notation to define UTF-8
strings in PostgreSQL 7.2.

No. It's too late. We are in the beta freeze stage.

The correct way would be to check for the existance of int8,

int16, etc.

Just tried to check your patch, but failed because I don't have a
running autoconf
:-(

Tried manual edit of pg_config:
It is important that all lines #undef SIZEOF_INTxx get replaced.
Wouldn't it be better to default those lines to #define SIZEOF_INT8 0

The check for int64 and uint64 has to be separated, my AIX
has: int8, int16, int32, int64
but not: uint8, uint16, uint32, uint64

Would you be so kind, as to supply me another patch (maybe including
configure) I can test before beta3 ?

Thanks in advance
Andreas
2.