Smalltalk › Pharo › Pharo Smalltalk Developers

RBScanner>>#initializeClassificationTable

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

14 messages Options

Benjamin Van Ryseghem (Pharo)

RBScanner>>#initializeClassificationTable

Hello guys,

Is there a reason this method use :

(Scanner isLiteralSymbol: '_')
ifTrue: [ self initializeChars: '_' to: #alphabetic ]
ifFalse: [ self initializeChars: '_' to: #special ].

and not

(Scanner allowUnderscoreAsAssignment)
ifTrue: [ self initializeChars: '_' to: #special ]
ifFalse:[ self initializeChars: '_' to: #alphabetic ]

?

Cheers,

Ben

Henrik Sperre Johansen

Re: RBScanner>>#initializeClassificationTable

On Nov 25, 2010, at 10:58 44AM, Benjamin wrote:

> Hello guys,
>
> Is there a reason this method use :
>
> (Scanner isLiteralSymbol: '_')
> ifTrue: [ self initializeChars: '_' to: #alphabetic ]
> ifFalse: [ self initializeChars: '_' to: #special ].
>
> and not
>
> (Scanner allowUnderscoreAsAssignment)
> ifTrue: [ self initializeChars: '_' to: #special ]
> ifFalse:[ self initializeChars: '_' to: #alphabetic ]
>
> ?
>
>
> Cheers,
>
> Ben

Compatibility I would guess.
In Squeak, allowUnderscoreAsAssignment is still a Preferences method.

Cheers,
Henry

Henrik Sperre Johansen

Re: RBScanner>>#initializeClassificationTable

On Nov 25, 2010, at 11:12 59AM, Henrik Johansen wrote:

>
> On Nov 25, 2010, at 10:58 44AM, Benjamin wrote:
>
>> Hello guys,
>>
>> Is there a reason this method use :
>>
>> (Scanner isLiteralSymbol: '_')
>> ifTrue: [ self initializeChars: '_' to: #alphabetic ]
>> ifFalse: [ self initializeChars: '_' to: #special ].
>>
>> and not
>>
>> (Scanner allowUnderscoreAsAssignment)
>> ifTrue: [ self initializeChars: '_' to: #special ]
>> ifFalse:[ self initializeChars: '_' to: #alphabetic ]
>>
>> ?
>>
>>
>> Cheers,
>>
>> Ben
>
> Compatibility I would guess.
> In Squeak, allowUnderscoreAsAssignment is still a Preferences method.
>
> Cheers,
> Henry
>

P.S. At the very least it could be changed to Scanner isLiteralSymbol: #_
:)

Levente Uzonyi-2

Re: RBScanner>>#initializeClassificationTable

In reply to this post by Henrik Sperre Johansen

On Thu, 25 Nov 2010, Henrik Johansen wrote:

No, it never was a method of Preferences. It's a pragma preference, but in
Squeak 4.1 it has a different name, IIRC prefAllowUnderscoreAssignment.

Levente

>
> Cheers,
> Henry
>
>

Levente Uzonyi-2

Re: RBScanner>>#initializeClassificationTable

In reply to this post by Henrik Sperre Johansen

On Thu, 25 Nov 2010, Henrik Johansen wrote:

>
> On Nov 25, 2010, at 11:12 59AM, Henrik Johansen wrote:
>
>>
>> On Nov 25, 2010, at 10:58 44AM, Benjamin wrote:
>>
>>> Hello guys,
>>>
>>> Is there a reason this method use :
>>>
>>> (Scanner isLiteralSymbol: '_')
>>> ifTrue: [ self initializeChars: '_' to: #alphabetic ]
>>> ifFalse: [ self initializeChars: '_' to: #special ].
>>>
>>> and not
>>>
>>> (Scanner allowUnderscoreAsAssignment)
>>> ifTrue: [ self initializeChars: '_' to: #special ]
>>> ifFalse:[ self initializeChars: '_' to: #alphabetic ]
>>>
>>> ?
>>>
>>>
>>> Cheers,
>>>
>>> Ben
>>
>> Compatibility I would guess.
>> In Squeak, allowUnderscoreAsAssignment is still a Preferences method.
>>
>> Cheers,
>> Henry
>>
>
> P.S. At the very least it could be changed to Scanner isLiteralSymbol: #_

I guess not. There was a bug in Squeak (and probably Pharo too), because
the compiler replaced #_ to #':=' during compilation. See
http://bugs.squeak.org/view.php?id=7571 .

Levente

> :)
>
>

Henrik Sperre Johansen

Re: RBScanner>>#initializeClassificationTable

In reply to this post by Levente Uzonyi-2

On Nov 25, 2010, at 11:41 07AM, Levente Uzonyi wrote:

On Thu, 25 Nov 2010, Henrik Johansen wrote:

On Nov 25, 2010, at 10:58 44AM, Benjamin wrote:

Hello guys,

Is there a reason this method use :

(Scanner isLiteralSymbol: '_')
ifTrue: [ self initializeChars: '_' to: #alphabetic ]
ifFalse: [ self initializeChars: '_' to: #special ].

and not

(Scanner allowUnderscoreAsAssignment)
ifTrue: [ self initializeChars: '_' to: #special ]
ifFalse:[ self initializeChars: '_' to: #alphabetic ]

?

Cheers,

Ben

Compatibility I would guess.
In Squeak, allowUnderscoreAsAssignment is still a Preferences method.

No, it never was a method of Preferences. It's a pragma preference, but in Squeak 4.1 it has a different name, IIRC prefAllowUnderscoreAssignment.

Levente

My bad, in the 4.1 image there's

Preferences >> allowUnderscoreAssignment

^ self

valueOfFlag: #allowUnderscoreAssignment

ifAbsent: [false]

(Added by Parser class>>#initialize), I didn't check if it was actually in use anymore.

Compatibility seems the be the reason anyways though, as you said the Scanner pragma preference has a different name.

On Nov 25, 2010, at 11:42 56AM, Levente Uzonyi wrote:

On Thu, 25 Nov 2010, Henrik Johansen wrote:

P.S. At the very least it could be changed to Scanner isLiteralSymbol: #_

I guess not. There was a bug in Squeak (and probably Pharo too), because the compiler replaced #_ to #':=' during compilation. See http://bugs.squeak.org/view.php?id=7571 .

Can't reproduce it in Pharo 1.0 onwards, with allowUnderscore on/off, though.

Cheers,

Henry

Levente Uzonyi-2

Re: RBScanner>>#initializeClassificationTable

On Thu, 25 Nov 2010, Henrik Johansen wrote:

> My bad, in the 4.1 image there's
> Preferences >> allowUnderscoreAssignment
> ^ self
> valueOfFlag: #allowUnderscoreAssignment
> ifAbsent: [false]
> (Added by Parser class>>#initialize), I didn't check if it was actually in use anymore.
>
> Compatibility seems the be the reason anyways though, as you said the Scanner pragma preference has a different name.

I doubt there was any effort to write compatible code, there are countless
packages that were originally written in Squeak, but now they are
incompatble with it.
In Squeak the pragma preference in Scanner was renamed recently to
allowUnderscoreAsAssignment to resolve the incompatibility.

>
>
> On Nov 25, 2010, at 11:42 56AM, Levente Uzonyi wrote:
>
>> On Thu, 25 Nov 2010, Henrik Johansen wrote:
>>
>>> P.S. At the very least it could be changed to Scanner isLiteralSymbol: #_
>>
>> I guess not. There was a bug in Squeak (and probably Pharo too), because the compiler replaced #_ to #':=' during compilation. See http://bugs.squeak.org/view.php?id=7571 .
>
> :(
> Can't reproduce it in Pharo 1.0 onwards, with allowUnderscore on/off, though.

That's fair, because you can't use underscores in selectors in Pharo.

Levente

>
> Cheers,
> Henry

Stéphane Ducasse

Re: RBScanner>>#initializeClassificationTable

In reply to this post by Henrik Sperre Johansen

>
> That's fair, because you can't use underscores in selectors in Pharo.

Strange because I was sure that we integrate it that behavior to help people calling C function.

Stef

Lukas Renggli

Re: RBScanner>>#initializeClassificationTable

>> That's fair, because you can't use underscores in selectors in Pharo.
>
> Strange because I was sure that we integrate it that behavior to help people calling C function.

The above statement is wrong. Of course underscores can be used in
selectors and variables for a long time already.

Lukas

--
Lukas Renggli
www.lukas-renggli.ch

Stéphane Ducasse

Re: RBScanner>>#initializeClassificationTable

In reply to this post by Stéphane Ducasse

On Nov 25, 2010, at 9:36 PM, Lukas Renggli wrote:

>>> That's fair, because you can't use underscores in selectors in Pharo.
>>
>> Strange because I was sure that we integrate it that behavior to help people calling C function.
>
> The above statement is wrong. Of course underscores can be used in
> selectors and variables for a long time already.

Thanks lukas. I was surprised and I was starting to look in the bugentry. So you saved me some time.
Apparently there is no setting to control it. Do you know how to configure it so that I enter a bug entry to make sure that we get a setting?

>
> Lukas
>
> --
> Lukas Renggli
> www.lukas-renggli.ch
>

Levente Uzonyi-2

Re: RBScanner>>#initializeClassificationTable

In reply to this post by Lukas Renggli

On Thu, 25 Nov 2010, Lukas Renggli wrote:

>>> That's fair, because you can't use underscores in selectors in Pharo.
>>
>> Strange because I was sure that we integrate it that behavior to help people calling C function.
>
> The above statement is wrong. Of course underscores can be used in
> selectors and variables for a long time already.

My statement is wrong (I forgot to turn off the setting when I checked),
but yours is also wrong. If underscore assignments are allowed, then you
can't use underscore selectors and vica versa.

Levente

>
> Lukas
>
> --
> Lukas Renggli
> www.lukas-renggli.ch
>
>

Lukas Renggli

Re: RBScanner>>#initializeClassificationTable

In reply to this post by Stéphane Ducasse

The setting is Compiler > Allow underscore as assignment.

Lukas

On 25 November 2010 17:42, Stéphane Ducasse <[hidden email]> wrote:

>
> On Nov 25, 2010, at 9:36 PM, Lukas Renggli wrote:
>
>>>> That's fair, because you can't use underscores in selectors in Pharo.
>>>
>>> Strange because I was sure that we integrate it that behavior to help people calling C function.
>>
>> The above statement is wrong. Of course underscores can be used in
>> selectors and variables for a long time already.
>
> Thanks lukas. I was surprised and I was starting to look in the bugentry. So you saved me some time.
> Apparently there is no setting to control it. Do you know how to configure it so that I enter a bug entry to make sure that we get a setting?
>
>
>>
>> Lukas
>>
>> --
>> Lukas Renggli
>> www.lukas-renggli.ch
>>
>
>
>

--
Lukas Renggli
www.lukas-renggli.ch

Stéphane Ducasse

Re: RBScanner>>#initializeClassificationTable

In reply to this post by Stéphane Ducasse

On Nov 25, 2010, at 11:12 PM, Lukas Renggli wrote:

> The setting is Compiler > Allow underscore as assignment.

Ok I was confused. This is more a toggle and we should change the description of the setting to convey that.

Stef

>>>>> That's fair, because you can't use underscores in selectors in Pharo.
>>>>
>>>> Strange because I was sure that we integrate it that behavior to help people calling C function.
>>>
>>> The above statement is wrong. Of course underscores can be used in
>>> selectors and variables for a long time already.
>>
>> Thanks lukas. I was surprised and I was starting to look in the bugentry. So you saved me some time.
>> Apparently there is no setting to control it. Do you know how to configure it so that I enter a bug entry to make sure that we get a setting?
>>
>>
>>>
>>> Lukas
>>>
>>> --
>>> Lukas Renggli
>>> www.lukas-renggli.ch
>>>
>>
>>
>>
>
>
>
> --
> Lukas Renggli
> www.lukas-renggli.ch
>

Lukas Renggli

Re: RBScanner>>#initializeClassificationTable

>> The setting is Compiler > Allow underscore as assignment.
>
> Ok I was confused. This is more a toggle and we should change the description of the setting to convey that.

In Pharo 1.0 we added the option (underscore assignments).

In Pharo 1.1 we disabled the option (underscore assignments) by default.

In Pharo 1.2 we hopefully remove the option (underscore assignments)
to get rid of them for all and forever.

Lukas

--
Lukas Renggli
www.lukas-renggli.ch