Why String>>adaptToNumber: rcvr andSend: selector

Do we reeeeeally want to do

5 + '4'

5 + 'hello'

?

What we've been discussing with Guille is that sometimes the 'type checking' (or actually, the conversion) is done, and sometimes not.

For example:

5 + true

true does not understand 'adaptToNumber:andSend:', so no magic conversion is done, but strings understand that message so we get things like:

5 + 'foo' = 5

This should be consistent regardless the type of the parameter.

Cheers,
Carla

On Tue, Aug 24, 2010 at 5:10 PM, Guillermo Polito <[hidden email]> wrote:

Do we reeeeeally want to do

5 + '4'

5 + 'hello'

?

Well Collection and Point understand that message too.

One step at a time, but an issue entry now so we don't forget about it.

http://code.google.com/p/pharo/issues/detail?id=2858

On Tue, Aug 24, 2010 at 5:20 PM, Carla F. Griggio <[hidden email]> wrote:

What we've been discussing with Guille is that sometimes the 'type checking' (or actually, the conversion) is done, and sometimes not.

For example:

5 + true

true does not understand 'adaptToNumber:andSend:', so no magic conversion is done, but strings understand that message so we get things like:

5 + 'foo' = 5

This should be consistent regardless the type of the parameter.

Cheers,
Carla

On Tue, Aug 24, 2010 at 5:10 PM, Guillermo Polito <[hidden email]> wrote:

Do we reeeeeally want to do

5 + '4'

5 + 'hello'

?

I feel these pieces of code aren't really explicit in what they do

5 + 'hello'

5 + (1 to: 10)

5 + (1@5)

On Tue, Aug 24, 2010 at 5:42 PM, Guillermo Polito <[hidden email]> wrote:

Well Collection and Point understand that message too.

One step at a time, but an issue entry now so we don't forget about it.

http://code.google.com/p/pharo/issues/detail?id=2858

On Tue, Aug 24, 2010 at 5:20 PM, Carla F. Griggio <[hidden email]> wrote:

What we've been discussing with Guille is that sometimes the 'type checking' (or actually, the conversion) is done, and sometimes not.

For example:

5 + true

true does not understand 'adaptToNumber:andSend:', so no magic conversion is done, but strings understand that message so we get things like:

5 + 'foo' = 5

This should be consistent regardless the type of the parameter.

Cheers,
Carla

On Tue, Aug 24, 2010 at 5:10 PM, Guillermo Polito <[hidden email]> wrote:

Do we reeeeeally want to do

5 + '4'

5 + 'hello'

?

I think 5 + 'hello' is plain wrong.

In these cases:

5 + (1 to: 10)

5 + (1@5)

the message shouldn't be conmutative. What I mean is that I don't want that to be the same as:

(1 to: 10) + 5

(1 @ 5) + 5

On Tue, Aug 24, 2010 at 5:44 PM, Guillermo Polito <[hidden email]> wrote:

I feel these pieces of code aren't really explicit in what they do

5 + 'hello'

5 + (1 to: 10)

5 + (1@5)

On Tue, Aug 24, 2010 at 5:42 PM, Guillermo Polito <[hidden email]> wrote:

Well Collection and Point understand that message too.

One step at a time, but an issue entry now so we don't forget about it.

http://code.google.com/p/pharo/issues/detail?id=2858

On Tue, Aug 24, 2010 at 5:20 PM, Carla F. Griggio <[hidden email]> wrote:

What we've been discussing with Guille is that sometimes the 'type checking' (or actually, the conversion) is done, and sometimes not.

For example:

5 + true

true does not understand 'adaptToNumber:andSend:', so no magic conversion is done, but strings understand that message so we get things like:

5 + 'foo' = 5

This should be consistent regardless the type of the parameter.

Cheers,
Carla

On Tue, Aug 24, 2010 at 5:10 PM, Guillermo Polito <[hidden email]> wrote:

Do we reeeeeally want to do

5 + '4'

5 + 'hello'

?

But... What do they do? They are not clean.

Does the collection one add the element to the collection?

Does the point one add the number to the first, the second or both components?  I remember that a point might be added to other point, multiplied by another point or a scalar :).

On Tue, Aug 24, 2010 at 5:53 PM, Carla F. Griggio <[hidden email]> wrote:

I think 5 + 'hello' is plain wrong.

In these cases:

5 + (1 to: 10)

5 + (1@5)

the message shouldn't be conmutative. What I mean is that I don't want that to be the same as:

(1 to: 10) + 5

(1 @ 5) + 5

On Tue, Aug 24, 2010 at 5:44 PM, Guillermo Polito <[hidden email]> wrote:

I feel these pieces of code aren't really explicit in what they do

5 + 'hello'

5 + (1 to: 10)

5 + (1@5)

On Tue, Aug 24, 2010 at 5:42 PM, Guillermo Polito <[hidden email]> wrote:

Well Collection and Point understand that message too.

One step at a time, but an issue entry now so we don't forget about it.

http://code.google.com/p/pharo/issues/detail?id=2858

On Tue, Aug 24, 2010 at 5:20 PM, Carla F. Griggio <[hidden email]> wrote:

What we've been discussing with Guille is that sometimes the 'type checking' (or actually, the conversion) is done, and sometimes not.

For example:

5 + true

true does not understand 'adaptToNumber:andSend:', so no magic conversion is done, but strings understand that message so we get things like:

5 + 'foo' = 5

This should be consistent regardless the type of the parameter.

Cheers,
Carla

On Tue, Aug 24, 2010 at 5:10 PM, Guillermo Polito <[hidden email]> wrote:

Do we reeeeeally want to do

5 + '4'

5 + 'hello'

?

:)

I accept code. We are in unstable so this is the time to shake the system.

On Aug 24, 2010, at 10:17 PM, Nicolas Cellier wrote:

> Noooo, we dooon't want to.
>
> Nicolas
>
> 2010/8/24 Guillermo Polito <[hidden email]>:
>> Do we reeeeeally want to do
>>
>> 5 + '4'
>>
>> 5 + 'hello'
>>
>> ?
>>
>> _______________________________________________
>> Pharo-project mailing list
>> [hidden email]
>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>>
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Wow, morphic doesn't allow me to change the point aritmethic :D.  I supossed it ¬¬.

Here is already some code:  http://code.google.com/p/pharo/issues/detail?id=2858

On Wed, Aug 25, 2010 at 4:35 AM, Stéphane Ducasse <[hidden email]> wrote:

:)

I accept code. We are in unstable so this is the time to shake the system.

On Aug 24, 2010, at 10:17 PM, Nicolas Cellier wrote:

> Noooo, we dooon't want to.
>
> Nicolas
>
> 2010/8/24 Guillermo Polito <[hidden email]>:
>> Do we reeeeeally want to do
>>
>> 5 + '4'
>>
>> 5 + 'hello'
>>
>> ?
>>
>> _______________________________________________
>> Pharo-project mailing list
>> [hidden email]
>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
>>
>
> _______________________________________________
> Pharo-project mailing list
> [hidden email]
> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project