Removal of pg_variable, pg_inheritproc, pg_ipl
The attached patch completes the following TODO item:
* -Remove unused pg_variable, pg_inheritproc, pg_ipl tables
initdb everyone. Sorry the ealier posting didn't have the patch
attached.
--
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 19026Attachments:
/bjm/diff2text/plainDownload+70-264
Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...
regards, tom lane
Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...
I patched the clients also. Is that a problem?
--
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 19026Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...
I realize what you are saying now. Older versions of ODBC still
reference indisclustered, even though it was bogus. I will put the
column into pg_index and mark it to be removed at some future date.
--
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 19026Bruce Momjian <pgman@candle.pha.pa.us> writes:
Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...
I patched the clients also. Is that a problem?
Yes. People frequently try to use clients that aren't the same rev
as the server. There's no good reason to remove this column at all,
and creating a version incompatibility for two major clients seems like
sufficient reason not to.
I recommend reverting that change.
regards, tom lane
Bruce Momjian <pgman@candle.pha.pa.us> writes:
Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...I patched the clients also. Is that a problem?
Yes. People frequently try to use clients that aren't the same rev
as the server. There's no good reason to remove this column at all,
and creating a version incompatibility for two major clients seems like
sufficient reason not to.I recommend reverting that change.
Doing it now, but just to the system catalogs. No reason to keep the
backend code.
--
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 19026Bruce Momjian <pgman@candle.pha.pa.us> writes:
I recommend reverting that change.
Doing it now, but just to the system catalogs. No reason to keep the
backend code.
That's debatable; someday we might want the thing (after we fix CLUSTER).
At least revert the client changes too.
regards, tom lane
Bruce Momjian <pgman@candle.pha.pa.us> writes:
I recommend reverting that change.
Doing it now, but just to the system catalogs. No reason to keep the
backend code.That's debatable; someday we might want the thing (after we fix CLUSTER).
At least revert the client changes too.
But the client is testing a useless value, and the test controls whether
it reports the index as HASH or OTHER. Doesn't even seem relevant.
--
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 19026Bruce Momjian <pgman@candle.pha.pa.us> writes:
I recommend reverting that change.
Doing it now, but just to the system catalogs. No reason to keep the
backend code.That's debatable; someday we might want the thing (after we fix CLUSTER).
At least revert the client changes too.
In fact, the only reason I started poking around here is because I am
evaluating a JDBC patch that was going to reference indisclustered. The
user was guessing that it related to HASH, just like the ODBC coder.
--
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 Mon, 14 May 2001, Bruce Momjian wrote:
Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...I realize what you are saying now. Older versions of ODBC still
reference indisclustered, even though it was bogus. I will put the
column into pg_index and mark it to be removed at some future date.
why is it being removed again? I think I missed that discussion, sorry ;(
On Mon, 14 May 2001, Bruce Momjian wrote:
Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...I realize what you are saying now. Older versions of ODBC still
reference indisclustered, even though it was bogus. I will put the
column into pg_index and mark it to be removed at some future date.why is it being removed again? I think I missed that discussion, sorry ;(
It is not being used by anything and is confusing client interface
developers because they think it is being set somewhere. It is not, but
we will keep it because ODBC references it, meaningless as it is.
--
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
Bruce Momjian wrote:
Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...I patched the clients also. Is that a problem?
Please don't reflect the server side changes in clients
directly. Even clients in beta have to be able to talk
to existent servers. As for removing indisclustered from
current psqlodbc, it seems to have no problem but I'm not
sure.
regards,
Hiroshi Inoue
[ Charset US-ASCII unsupported, converting... ]
Bruce Momjian wrote:
Why did you remove indisclustered?
Useless it may be, but gratuitously breaking at least two extant clients
doesn't seem like a good idea ...I patched the clients also. Is that a problem?
Please don't reflect the server side changes in clients
directly. Even clients in beta have to be able to talk
to existent servers. As for removing indisclustered from
current psqlodbc, it seems to have no problem but I'm not
sure.
It isn't a problem. The new ODBC version doesn't reference the column
anymore. The value was meaningless anyway. When we finally remove it
from pg_index, it would affect old clients, but that removal may be
years away.
--
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