Re: [HACKERS] NUMERIC's transcendental functions

2002-09-28 Thread Bruce Momjian
Tom Lane wrote: > Bruce Momjian <[EMAIL PROTECTED]> writes: > > Is this an open item? > > Yes. (Fooling with it right now, in fact, in a desultory way ...) OK, added. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 +

Re: [HACKERS] NUMERIC's transcendental functions

2002-09-28 Thread Tom Lane
Bruce Momjian <[EMAIL PROTECTED]> writes: > Is this an open item? Yes. (Fooling with it right now, in fact, in a desultory way ...) regards, tom lane ---(end of broadcast)--- TIP 6: Have you searched our list archives? ht

Re: [HACKERS] NUMERIC's transcendental functions

2002-09-28 Thread Bruce Momjian
Is this an open item? --- Tom Lane wrote: > Jan Wieck <[EMAIL PROTECTED]> writes: > > One problem is, that division already has an inherently inexact > > result. Do you intend to rip that out too while at it? (Just > > kidd

Re: BETA2 HOLD: was Re: [HACKERS] NUMERIC's transcendental functions

2002-09-23 Thread Bruce Momjian
Tom Lane wrote: > Bruce Momjian <[EMAIL PROTECTED]> writes: > > Seems we need to resolve this before beta2. > > Not really. It's just a bug; we have others. Oh, it doesn't effect initdb? -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] |

Re: BETA2 HOLD: was Re: [HACKERS] NUMERIC's transcendental functions

2002-09-23 Thread Tom Lane
Bruce Momjian <[EMAIL PROTECTED]> writes: > Seems we need to resolve this before beta2. Not really. It's just a bug; we have others. regards, tom lane ---(end of broadcast)--- TIP 6: Have you searched our list archives? h

Re: BETA2 HOLD: was Re: [HACKERS] NUMERIC's transcendental functions

2002-09-23 Thread Jan Wieck
Bruce Momjian wrote: > > Seems we need to resolve this before beta2. I'd go with making the NUMERIC default precision 16 for v7.3, so we are backwards compatible on this release (except that it is now a predictable 16 digit precision instead of an hardware implementation dependent one). For v7.

BETA2 HOLD: was Re: [HACKERS] NUMERIC's transcendental functions

2002-09-23 Thread Bruce Momjian
Seems we need to resolve this before beta2. --- Tom Lane wrote: > Jan Wieck <[EMAIL PROTECTED]> writes: > > One problem is, that division already has an inherently inexact > > result. Do you intend to rip that out too while

Re: [HACKERS] NUMERIC's transcendental functions

2002-09-23 Thread Tom Lane
Jan Wieck <[EMAIL PROTECTED]> writes: > One problem is, that division already has an inherently inexact > result. Do you intend to rip that out too while at it? (Just > kidding) No, but that too is now delivering less precision than it used to: regression=# select 10.1/7.0; ?column? -

Re: [HACKERS] NUMERIC's transcendental functions

2002-09-23 Thread Jan Wieck
Tom Lane wrote: > > I have noticed a change in behavior following the recent changes for > casting of numeric constants. In prior releases, we got > > regression=# select log(10.1); >log > -- > 1.00432137378264 > (1 row) > > CVS tip gives > > regression=# select log(1

Re: [HACKERS] NUMERIC's transcendental functions

2002-09-22 Thread Bruce Momjian
When you say: > So proposal #1 is to rip out the numeric versions of these functions. you mean remove the ability to do transendentals on numerics to prevent such unusual auto-casting? Are you suggesting that in all other cases, autocast to numeric is OK? I am a little confused.

[HACKERS] NUMERIC's transcendental functions

2002-09-21 Thread Tom Lane
I have noticed a change in behavior following the recent changes for casting of numeric constants. In prior releases, we got regression=# select log(10.1); log -- 1.00432137378264 (1 row) CVS tip gives regression=# select log(10.1); log -- 1.0043213738