trying to work my way around a few other problems with booleans i've run across.
I'd much prefer to correctly type them consistantly throughout the app, i've tried,
but ...
MySQL for instance doesn't support a boolean data type, at least not in the current
version. It chokes on incoming boolean values. Dang!
In CFMX Q of Q's, boolean values are converted to 1's and 0's, and you run into
trouble using true or false in the Where clauses of QofQ's (in MX), you need to use 1
or 0 (at least sometimes you do) ...
So i decided for the sake of across-the-board compatibility to give up trying to use
booleans in CF in and out of the datastore at least) and just consistantly use the
numeric values of 1 or 0. That way i should be able to use any database without
needing to change a lot of SQL, since CF evaluates 0 as false and all other integers
as true (at least i thought it did!).
Now the newest dilemmna is that using the cfsqltype="cf_sql_bit" in the cfqueryparam
tag evaluates incoming values of both true and 1 to 255 - and hence *some* of my
boolean tests are rendering inconsistent results.
Hey, when it rains it pours ...
So i'm in the middle of changing all my cfsqltype's to numeric for my boolean fields
and that "solves" the problem ... then cfqueryparam evaluates an incoming value of
true into 1 and all my boolean tests work, no matter how i've coded them.
I guess this qualifies as a "hack".
Anyone dealt with this issue in a better way? The goal is to have "booleans" work
consistently across DBMS's, in CFSQL and within CF.
-----Original Message-----
From: Barney Boisvert [mailto:[EMAIL PROTECTED]
Sent: Friday, February 13, 2004 8:42 PM
To: CF-Talk
Subject: RE: Strange inconsistancy evaluating True/False
Ah, the wonders of weakly-typed languages. Not to mention the 50/50
languages like CF. ;)
Save yourself a world of troubles, treat CF as if it's strongly typed. CFIF
only accepts boolean conditions (your 'bitPublish' is not a boolean), and
only compare like types (with some fudge for floating point to integer and
such).
Not sure where the blame lies, but CF's type system is hosed like a building
that's on fire. Whether this specific situation that is common across
weakly-typed languages, or specific to CF, I don't know. It seems to me
that CF should coerce the int to a boolean first, because boolean is the
"more strict" type, so both of your expressions would be true.
Cheers,
barneyb
> -----Original Message-----
> From: Nando [mailto:[EMAIL PROTECTED]
> Sent: Friday, February 13, 2004 11:25 AM
> To: CF-Talk
> Subject: Strange inconsistancy evaluating True/False
>
> I just tripped over something a little strange ...
>
> <cfset bitPublish = 255>
>
> With EQ<br>
> <cfif bitPublish EQ true>
> true
> <cfelse>
> false
> </cfif>
>
> <br><br>
>
> Without EQ<br>
> <cfif bitPublish>
> true
> <cfelse>
> false
> </cfif>
>
> If you run this test code on CFMX 6.1, you'll see that the
> first case evaluates to
> false and the second evalutes to true.
>
> using a value of 1 evaluates both to true.
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]
