Major change to CVS effective immediately ...

----- Original Message -----
From: Marc G. Fournier <scrappy@hub.org>
Sent: Sunday, September 16, 2001 1:05 PM

Now, I don't imagine it being *that* simple to move it over, so please let
me know if anyone sees any errors on commits or stuff like that ...

CVSweb seems to be screwed up.
It gives an error:

------------8<------------
Error
Error: $CVSROOT not found!
The server on which the CVS tree lives is probably down. Please try again in a few minutes.
------------8<------------

when I access it with my browser.

Serguei

CVSWeb is going to be broken for a day or two, while Vince and I work out
some issues as regards moving the main www site over to the same server
... but thanks for pointing it out, as I hadn't thought about it ..

On Sun, 16 Sep 2001, Serguei Mokhov wrote:

----- Original Message -----
From: Marc G. Fournier <scrappy@hub.org>
Sent: Sunday, September 16, 2001 1:37 PM

CVSWeb is going to be broken for a day or two, while Vince and I work out
some issues as regards moving the main www site over to the same server
... but thanks for pointing it out, as I hadn't thought about it ..

It's just because the web interface is the only way for me for now
I can access the CVS repository. Otherwise, I wouldn't have noticed the problem
most likely...

Serguei

Marc G. Fournier writes:

Now, I don't imagine it being *that* simple to move it over, so please let
me know if anyone sees any errors on commits or stuff like that ...

cvs commit
...
cvs server: failed to create lock directory for \
`/cvsroot/pgsql/doc/src/sgml' (/cvsroot/pgsql/doc/src/sgml/#cvs.lock): \
Permission denied
cvs server: lock failed - giving up
cvs [server aborted]: lock failed - giving up

For those with already checked out repositories, from everything I've
read, all you have to do is change the value of the CVS/Root file to point
to the new Root ...

CVS/Repository as well.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

On Sun, 16 Sep 2001, Peter Eisentraut wrote:

Marc G. Fournier writes:

Now, I don't imagine it being *that* simple to move it over, so please let
me know if anyone sees any errors on commits or stuff like that ...

cvs commit
...
cvs server: failed to create lock directory for \
`/cvsroot/pgsql/doc/src/sgml' (/cvsroot/pgsql/doc/src/sgml/#cvs.lock): \
Permission denied
cvs server: lock failed - giving up
cvs [server aborted]: lock failed - giving up

Okay, everything looks well to me on the server itself ... first stupid
question, what is the IP of cvs.postgresql.org for you? it should be
85.28, but you might still have the old one cached ...

For those with already checked out repositories, from everything I've
read, all you have to do is change the value of the CVS/Root file to point
to the new Root ...

CVS/Repository as well.

Wouldn't you have to apply this change in *every* /CVS subdirectory of
the tree? A fresh checkout seems easier.

However, I find that the permissions problem stymies a checkout too...
and yes, my DNS cache is up to date:

$ nslookup cvs.postgresql.org
Server: localhost
Address: 127.0.0.1

Non-authoritative answer:
Name: mail.postgresql.org
Address: 216.126.85.28
Aliases: cvs.postgresql.org

regards, tom lane

Tom Lane writes:

Wouldn't you have to apply this change in *every* /CVS subdirectory of
the tree? A fresh checkout seems easier.

I just ran this on my tree:

find -name Root -exec perl -pi -e 's,:pserver:([a-z]+)\@postgresql.org:/home/projects/pgsql/cvsroot,:pserver:\1\@cvs.postgresql.org:/cvsroot,' '{}' ';'
find -name Repository -exec perl -pi -e 's,/home/projects/pgsql/cvsroot/(.*),/cvsroot/\1,' '{}' ';'

and it looks like it didn't botch.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

Thus spake Marc G. Fournier

This will most likely screw some ppl up, and fix others ...

CVSROOT has now moved to the new machine, finally ... and I've cleaned up
pathing ... and CVS_RSH=ssh now works again too ...

So, now CVSROOT is accessible as:

:pserver:<userid>@cvs.postgresql.org:/cvsroot

-or-

:ext:<userid>@cvs.postgresql.org:/cvsroot

- where CVS_RSH is set to ssh

None of this helps me. I still can't get into that system. Can you
please check this and get me back in. If I can't get in I will have
to move PyGreSQL back to my own CVS repository and I think it is good
for both projects to leave it where it is. Will you be in today? I
can call you later to discuss this and hopefully resolve it.

-- 
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.

* "Marc G. Fournier" <scrappy@hub.org> wrote:
|
| anoncvs.postgresql.org is going to be out of sync until, most likely,
| tomorrow, for anyone trying to use that ... anoncvs is *no longer*
| available through the main cvs repository either ...

Is anoncvs.postgresql.org working yet ?

I just tried :

# cvs -d :pserver:anoncvs@anoncvs.postgresql.org:/home/projects/pgsql/cvsroot login

and

# cvs -d :pserver:anoncvs@anoncvs.postgresql.org:/cvsroot login

with "postgresql" as password. In both cases the response was this :

cvs [login aborted]: authorization failed: server anoncvs.postgresql.org rejected access

regards,

Gunnar

--
Gunnar R�nning - gunnar@polygnosis.com
Senior Consultant, Polygnosis AS, http://www.polygnosis.com/

:pserver:anoncvs@anoncvs.postgresql.org:/projects/cvsroot

On 18 Sep 2001, Gunnar [iso-8859-1] R���nning wrote:

can you ssh into cvs.postgresql.org?

On Tue, 18 Sep 2001, D'Arcy J.M. Cain wrote:

On Wed, Sep 19, 2001 at 10:14:44AM -0400, Marc G. Fournier wrote:

:pserver:anoncvs@anoncvs.postgresql.org:/projects/cvsroot

While trying a cvs update, I get

? ChangeLogs/libecpg.so.3.1.1
? ChangeLogs/HTML
? ChangeLogs/GTAGS
? ChangeLogs/GPATH
? ChangeLogs/GRTAGS
? ChangeLogs/GSYMS
? ChangeLogs/libpqpp.h
cannot create_adm_p /tmp/cvs-serv27285/ChangeLogs

Cheers,

Patrick

Thus spake Marc G. Fournier

can you ssh into cvs.postgresql.org?

Yes! I could not do that before. Did you fix something?

I will be sending some PyGreSQL changes over shortly. Thanks.

-- 
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.

While checking out TOT pgsql today onto an HFS+ file system (case-preserving, case-insensitive), I hit the following CVS conflict:

pgsql/src/backend/utils/mb/Unicode/utf8_to_alt.map pgsql/src/backend/utils/mb/Unicode/utf8_to_ALT.map

HFS+ can not store two differerent files in a path that differs only by case.

Mac OS X users will be grateful if you can find a way to rename one of these files.

-pmb

Peter Bierman writes:

While checking out TOT pgsql today onto an HFS+ file system (case-preserving, case-insensitive), I hit the following CVS conflict:

pgsql/src/backend/utils/mb/Unicode/utf8_to_alt.map pgsql/src/backend/utils/mb/Unicode/utf8_to_ALT.map

HFS+ can not store two differerent files in a path that differs only by case.

Remove both of these files and update again. The files were recently
renamed to have a consistent case-ness.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

At 10:47 PM +0200 9/19/01, Peter Eisentraut wrote:

Peter Bierman writes:

While checking out TOT pgsql today onto an HFS+ file system (case-preserving, case-insensitive), I hit the following CVS conflict:

pgsql/src/backend/utils/mb/Unicode/utf8_to_alt.map pgsql/src/backend/utils/mb/Unicode/utf8_to_ALT.map

HFS+ can not store two differerent files in a path that differs only by case.

Remove both of these files and update again. The files were recently
renamed to have a consistent case-ness.

This was from an anoncvs HEAD/TOT checkout I did into an empty directory less than an hour ago.

If it's already been fixed (yay!), the fix isn't at anoncvs yet.

-pmb

----- Original Message -----
From: Peter Bierman <bierman@apple.com>
Sent: Wednesday, September 19, 2001 3:41 PM

While checking out TOT pgsql today onto an HFS+ file system (case-preserving, case-insensitive), I hit the following CVS conflict:

pgsql/src/backend/utils/mb/Unicode/utf8_to_alt.map pgsql/src/backend/utils/mb/Unicode/utf8_to_ALT.map

I thought the latter was supposed to go (it was supposed to be renamed to the former, wasn't it?)

S.

Peter Bierman:

While checking out TOT pgsql today onto an HFS+ file system

(case-preserving, case-insensitive), I hit the following CVS conflict:

pgsql/src/backend/utils/mb/Unicode/utf8_to_alt.map

pgsql/src/backend/utils/mb/Unicode/utf8_to_ALT.map

HFS+ can not store two differerent files in a path that differs only by

case.

Mac OS X users will be grateful if you can find a way to rename one of

these files.

I had that problem today to -- I work under Cygwin on Windows 2000; NTFS is
also case-preserving but case-insensitive:

U pgsql/src/backend/utils/mb/Unicode/utf8_to_ALT.map
cvs checkout: move away pgsql/src/backend/utils/mb/Unicode/utf8_to_alt.map;
it is in the way
C pgsql/src/backend/utils/mb/Unicode/utf8_to_alt.map

Cheers,

Colin

Peter Bierman writes:

At 10:47 PM +0200 9/19/01, Peter Eisentraut wrote:

Peter Bierman writes:

While checking out TOT pgsql today onto an HFS+ file system (case-preserving, case-insensitive), I hit the following CVS conflict:

pgsql/src/backend/utils/mb/Unicode/utf8_to_alt.map pgsql/src/backend/utils/mb/Unicode/utf8_to_ALT.map

HFS+ can not store two differerent files in a path that differs only by case.

Remove both of these files and update again. The files were recently
renamed to have a consistent case-ness.

This was from an anoncvs HEAD/TOT checkout I did into an empty directory less than an hour ago.

Indeed, someone forgot to remove the old file. I just removed it a second
ago, so you should be fine now.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

Peter Bierman <bierman@apple.com> writes:

If it's already been fixed (yay!), the fix isn't at anoncvs yet.

I think there is some lag between the master CVS and anoncvs now.
Marc, is that correct? How much lag?

regards, tom lane

should be four hours, but I haven't had a chance, with the newest
worm/virus going around right now having killed our core router yesterday,
to redirect the sync'ag with the new server ... will do that first thing
tomorrow ...

On Wed, 19 Sep 2001, Tom Lane wrote:

"Marc G. Fournier" wrote:

should be four hours, but I haven't had a chance, with the newest
worm/virus going around right now having killed our core router yesterday,
to redirect the sync'ag with the new server ... will do that first thing
tomorrow ...

On Wed, 19 Sep 2001, Tom Lane wrote:

Peter Bierman <bierman@apple.com> writes:

If it's already been fixed (yay!), the fix isn't at anoncvs yet.

I think there is some lag between the master CVS and anoncvs now.
Marc, is that correct? How much lag?

regards, tom lane

It's definitely more than 16 hours. I still can't see M. Meskes' commits
(16:09 MEST, 10:09 EDT)

While you're at it, could you please fix this error:

~/pgsql-cvs/pgsql > cvs -z3 update -dP
cannot create_adm_p /tmp/cvs-serv2966/ChangeLogs
Permission denied

for i in `find -type d ! -name CVS ` ; do (cd $i ; cvs -z3 update -l -d )
done
cvs server: Updating .
cvs server: Updating .
[.....]

This works somehow but is really ugly and bandwidth-wasting. This even occurs
with a fresh checkout:

~/pgsql-cvs/tmp > cvs -d
:pserver:anoncvs@anoncvs.postgresql.org:/projects/cvsroot co pgsql/ChangeLogs

cvs server: Updating pgsql/ChangeLogs
U pgsql/ChangeLogs/ChangeLog-7.1-7.1.1
U pgsql/ChangeLogs/ChangeLog-7.1RC1-to-7.1RC2
U pgsql/ChangeLogs/ChangeLog-7.1RC2-to-7.1RC3
U pgsql/ChangeLogs/ChangeLog-7.1RC3-to-7.1rc4
U pgsql/ChangeLogs/ChangeLog-7.1beta1-to-7.1beta3
U pgsql/ChangeLogs/ChangeLog-7.1beta3-to-7.1beta4
U pgsql/ChangeLogs/ChangeLog-7.1beta4-to-7.1beta5
U pgsql/ChangeLogs/ChangeLog-7.1beta5-to-7.1beta6
U pgsql/ChangeLogs/ChangeLog-7.1beta6-7.1RC1
U pgsql/ChangeLogs/ChangeLog-7.1rc4-7.1
~/pgsql-cvs/tmp > cd pgsql/
~/pgsql-cvs/tmp/pgsql > cvs update
cannot create_adm_p /tmp/cvs-serv4350/ChangeLogs
Permission denied

Yours
Christof

While you're at it, could you please fix this error:

~/pgsql-cvs/pgsql > cvs -z3 update -dP
cannot create_adm_p /tmp/cvs-serv2966/ChangeLogs
Permission denied

Instead of checking out over your existing checkout, checkout to a new dir
and there's no problem.

Chris

Christopher Kings-Lynne wrote:

While you're at it, could you please fix this error:

~/pgsql-cvs/pgsql > cvs -z3 update -dP
cannot create_adm_p /tmp/cvs-serv2966/ChangeLogs
Permission denied

Instead of checking out over your existing checkout, checkout to a new dir
and there's no problem.

Sorry, I want to update (only the differences cross the wire) or diff, not
check out all again twice a day (which works).

Christof

Okay, its updated effective a few minutes ago ... and the upate should
work as well ...

On Thu, 20 Sep 2001, Christof Petig wrote:

"Marc G. Fournier" wrote:

Okay, its updated effective a few minutes ago ... and the upate should
work as well ...

Should ...

~/pgsql-cvs/pgsql/src/interfaces/ecpg/preproc > cvs status preproc.y
cvs server: failed to create lock directory for
`/projects/cvsroot/pgsql/src/interfaces/ecpg/preproc'
(/projects/cvsroot/pgsql/src/interfaces/ecpg/preproc/#cvs.lock): Permission denied

cvs server: failed to obtain dir lock in repository
`/projects/cvsroot/pgsql/src/interfaces/ecpg/preproc'
cvs [server aborted]: read lock failed - giving up

~/pgsql-cvs/pgsql > cvs update
cannot create_adm_p /tmp/cvs-serv48812/ChangeLogs
Permission denied

~/pgsql-cvs/pgsql > cvs update -l
cvs server: Updating .
cvs server: failed to create lock directory for `/projects/cvsroot/pgsql'
(/projects/cvsroot/pgsql/#cvs.lock): Permission denied
cvs server: failed to obtain dir lock in repository `/projects/cvsroot/pgsql'
cvs [server aborted]: read lock failed - giving up

... but it does not, yet.

Christof

On Wed, 19 Sep 2001, D'Arcy J.M. Cain wrote:

Thus spake Marc G. Fournier

can you ssh into cvs.postgresql.org?

Yes! I could not do that before. Did you fix something?

Nope :(

Thus spake Marc G. Fournier

can you ssh into cvs.postgresql.org?

Yes! I could not do that before. Did you fix something?

Nope :(

Just weird. Oh well. All's well, etc.

-- 
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.