[Glass] MAStringReader>>visitNumberDescription broken?

[Dale Henrichs via Glass]

Iwan,

 From your information below, it does sound like having NumberParser use 
SqNumberParser as a default class for all of the class-side methods 
would likely solve the problem ... in Pharo3.0 the concrete class is 
called NumberParser and in GsDevKit the concrete clas is SqNumberParser 
and likely at some point in the past the Pharo folks collapsed the two 
classes, so from a compatibility perspective treating NumberParser as a 
concrete class is desired behavior in general...

With the flurry of emails in the last day, my todo list is building 
faster than I can knock items off the list so I don't know when I can 
find the time to make the changes myself ...

If you could a) verify that this approach would work and b) submit a 
pull request against https://github.com/glassdb/glass with your changes 
then we could get this issue resolved for everyone ...

Dale

On 09/09/2015 03:05 AM, Iwan Vosloo via Glass wrote:
> Hi Dale,
>
> On 08/09/2015 21:48, Dale Henrichs via Glass wrote:
>> Okay, then I need to dive into the Magritte implementation directly 
>> and try to figure things out from there ... I naively thought that 
>> the problem was the direct use of NumberParser in the code, instead 
>> of SqNumberParser - it would be relatively easy to change 
>> NumberParser to redirect instance creation to SqNumberParser, but It 
>> sounds like that isn't the source ofthe problem ...
>
> AFAIK the implementation of NumberParser>>isNumber: should not call 
> parse:onError: on a hardcoded NumberParser, but rather on self. (This 
> regardless of platform.)
>
> But then MAStringReader>>visitNumberDescription: should also rather 
> call isNumber: on SqNumberParser as you said above. (In the GemStone 
> case.)
>
>>
>> BTW are Magritte tests failing on GemStone or is this an untested 
>> chunk of code?
> There are a couple of failures if I run all tests on 
> MANumberDescriptionTest, including testFromString which fails exactly 
> because of this issue.
>
>
> I have not tried to figure out how Mariano's posting is related to this.
>
> Regards
> - Iwan
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gemtalksystems.com/mailman/private/glass/attachments/20150909/9ca1320a/attachment.html>