Modulo syntax

Started by Oleg Broytmannabout 27 years ago5 messageshackers
Jump to latest
#1Oleg Broytmann
phd@sun.med.ru

Hello!

I ran the query
update producers SET cor_id = producer_id % 9 + 1;

and found that result is eqiuvalent to
update producers SET cor_id = producer_id % 9;

I added parens:
update producers SET cor_id = (producer_id % 9) + 1;

and got what I needed.

Is it a bug, a feature, or I just misinterpreted the syntax?

Oleg.
----
Oleg Broytmann http://members.xoom.com/phd2/ phd2@earthling.net
Programmers don't die, they just GOSUB without RETURN.

#2Bruce Momjian
bruce@momjian.us
In reply to: Oleg Broytmann (#1)
Re: [HACKERS] Modulo syntax

Hello!

I ran the query
update producers SET cor_id = producer_id % 9 + 1;

and found that result is eqiuvalent to
update producers SET cor_id = producer_id % 9;

I added parens:
update producers SET cor_id = (producer_id % 9) + 1;

and got what I needed.

Looks like a bug. We have associativity for +, -, * and /, but not %.

From gram.y:

%left '+' '-'
%left '*' '/'

I will add '%' to that.

-- 
  Bruce Momjian                        |  http://www.op.net/~candle
  maillist@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
#3Bruce Momjian
bruce@momjian.us
In reply to: Oleg Broytmann (#1)
Re: [HACKERS] Modulo syntax

Fixed now.

Hello!

I ran the query
update producers SET cor_id = producer_id % 9 + 1;

and found that result is eqiuvalent to
update producers SET cor_id = producer_id % 9;

I added parens:
update producers SET cor_id = (producer_id % 9) + 1;

and got what I needed.

Is it a bug, a feature, or I just misinterpreted the syntax?

Oleg.
----
Oleg Broytmann http://members.xoom.com/phd2/ phd2@earthling.net
Programmers don't die, they just GOSUB without RETURN.

-- 
  Bruce Momjian                        |  http://www.op.net/~candle
  maillist@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
#4Thomas Lockhart
lockhart@alumni.caltech.edu
In reply to: Bruce Momjian (#2)
Re: [HACKERS] Modulo syntax

I ran the query
update producers SET cor_id = producer_id % 9 + 1;
and found that result is eqiuvalent to
update producers SET cor_id = producer_id % 9;
I added parens:
update producers SET cor_id = (producer_id % 9) + 1;
and got what I needed.

Looks like a bug. We have associativity for +, -, * and /, but not %.
From gram.y:
%left '+' '-'
%left '*' '/'
I will add '%' to that.

This will not fix the associativity problem, unless you *also* go
through and add the explicit syntax *throughout* gram.y, as is
currently done for '+', '-', etc.

I'm pretty sure that we don't want to do this, since there are way too
many other operators which would need the same treatment.

The correct solution will be to identify the operator as a particular
class in scan.l, include that class in the associativity declarations,
and then handle that class in the body of gram.y. Sort of like we do
for generic operators already, but with some discrimination between
them. To be done right, we should look up the precedence in a db
table, to allow new operators to participate in the scheme. In any
case, gram.y will become more complex...

Unless we are going to solve this, I would suggest backing out the
change in gram.y.

- Tom

#5Bruce Momjian
bruce@momjian.us
In reply to: Thomas Lockhart (#4)
Re: [HACKERS] Modulo syntax

I ran the query
update producers SET cor_id = producer_id % 9 + 1;
and found that result is eqiuvalent to
update producers SET cor_id = producer_id % 9;
I added parens:
update producers SET cor_id = (producer_id % 9) + 1;
and got what I needed.

Looks like a bug. We have associativity for +, -, * and /, but not %.
From gram.y:
%left '+' '-'
%left '*' '/'
I will add '%' to that.

This will not fix the associativity problem, unless you *also* go
through and add the explicit syntax *throughout* gram.y, as is
currently done for '+', '-', etc.

I'm pretty sure that we don't want to do this, since there are way too
many other operators which would need the same treatment.

I did this for %. I felt it was common enough and similar to / that
people should expect it to have / associativity. I did not play with
any other operators.

The correct solution will be to identify the operator as a particular
class in scan.l, include that class in the associativity declarations,
and then handle that class in the body of gram.y. Sort of like we do
for generic operators already, but with some discrimination between
them. To be done right, we should look up the precedence in a db
table, to allow new operators to participate in the scheme. In any
case, gram.y will become more complex...

Yikes. Don't think we want to go there.

Unless we are going to solve this, I would suggest backing out the
change in gram.y.

I would like to keep % as a special case like /.

-- 
  Bruce Momjian                        |  http://www.op.net/~candle
  maillist@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