ANyone have a fix for Scanner isLiteralSymbol: #<=?

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

ANyone have a fix for Scanner isLiteralSymbol: #<=?

Eliot Miranda-2
Hi All,

    In trunk

        Scanner isLiteralSymbol: #<=

    answers false ?!?!

This is breaking my Cog generated code disassembly decoration so that instead of

    000a572c: movl $0x0045a458=#<=, %ecx : B9 58 A4 45 00 
    000a5731: call .+0xfff5ad5a (0x00000490=ceSend1Args) : E8 5A AD F5 FF 

I'm merely getting

    000a572c: movl $0x0045a458, %ecx : B9 58 A4 45 00 
    000a5731: call .+0xfff5ad5a (0x00000490=ceSend1Args) : E8 5A AD F5 FF 

Anyone have a fix?

best
Eliot


Reply | Threaded
Open this post in threaded view
|

Re: ANyone have a fix for Scanner isLiteralSymbol: #<=?

Juan Vuletich-4
Hi Eliot,

Eliot Miranda wrote:
> Hi All,
>
>     In trunk
>
>         Scanner isLiteralSymbol: #<=
>
> ...

It works ok in Cuis. You can see #isLiteralSymbol: there.

BTW, this was changed when I integrated the Compiler package from you,
in April this year. If you want to see what else changed then, see
change sets 0482, 0483 and 0484 in the Cuis updates.

You can get Cuis and the zip file with the updates from
http://www.jvuletich.org/Cuis/Index.html .

Cheers,
Juan Vuletich

Reply | Threaded
Open this post in threaded view
|

Re: [Pharo-project] ANyone have a fix for Scanner isLiteralSymbol: #<=?

Levente Uzonyi-2
In reply to this post by Eliot Miranda-2
On Thu, 18 Nov 2010, Eliot Miranda wrote:

> Hi All,
>
>    In trunk
>
>        Scanner isLiteralSymbol: #<=
>
>    answers false ?!?!
>
> This is breaking my Cog generated code disassembly decoration so that
> instead of
>
>    000a572c: movl $0x0045a458=#<=, %ecx : B9 58 A4 45 00
>    000a5731: call .+0xfff5ad5a (0x00000490=ceSend1Args) : E8 5A AD F5 FF
>
> I'm merely getting
>
>    000a572c: movl $0x0045a458, %ecx : B9 58 A4 45 00
>    000a5731: call .+0xfff5ad5a (0x00000490=ceSend1Args) : E8 5A AD F5 FF
>
> Anyone have a fix?

There seems to be a fix in the method, but it's commented out to be
compatible with the previous version of the method:

  type == #xBinary ifTrue: [^i = 1]. "Here we could extend to
  ^(2 to: i) allSatisfy: [:j |
  ascii := (aSymbol at: j) asciiValue.
  (TypeTable at: ascii ifAbsent: []) == #xBinary]"


Levente

>
> best
> Eliot
>