[squeak-dev] Looking for a volunteer to help with DataStream on Spur

Tobias Pape Das.Linux at gmx.de
Wed Jun 24 17:36:37 UTC 2015


Hi,

On 03.02.2015, at 19:17, Eliot Miranda <eliot.miranda at gmail.com> wrote:

> Hi All,
> 
>    Spur adds two new immediate classes, Character and SmallFloat64.  It also adds 61-bit SmallIntegers in 64-bits.  DataStream/ReferenceStream/SmartRefStream are all affected by this.  e.g.
> 
>        ReferenceStream testWith: $a
> 
> fails on 32-bit and 64-bit Spur.
> 
>    ReferenceStream testWith: SmallInteger maxVal
> 
> fails (silently, to be fixed soon) on 64-bit Spur.
> 

Do you have a 64bit spur vm and image i can play around with?

best 
	-tobias

> Is anyone kind enough and motivated enough to give this code some love and at least make it work for
> 
> 32-bit Spur => 32-bit Spur
> 64-bit Spur => 64-bit Spur
> 
> but also hopefully for
> 32-bit Spur <=> 64-bit Spur
> 32-bit V3 => 32-bit Spur
> 32-bit V3 => 64-bit Spur
> 
> ?
> 
> On Tue, Feb 3, 2015 at 8:59 AM, Stéphane Rollandin <lecteur at zogotounga.net> wrote:
> Yes.  Can you send me a pair of doits for me to test this, one I can run in v3 and one I can run in Spur?
> 
> I'm trying to set-up a faulty expression, but it is not so easy.
> 
> Meanwhile I stumbled upon a similar bug:
> 
>        $a asReferenceStream reset fileInObjectAndCode
> 
> The above succeeds in V3 and raises an error in Spur. Same reason: basicNew being sent to Character.
> 
> 
> But I'n curious, why codePoint: rather than
> 
> createFrom: aSmartRefStream size: varsOnDisk version: instVarList
> 
>    ^ self value: instVarList first
> 
> ?
> 
> Well I got used to codePoint: ... Also value: has a lot more implementors and senders and does not mean much by itself. Actually I never thought about this.





More information about the Squeak-dev mailing list