[squeak-dev] The Inbox: Tools-ct.937.mcz

Tobias Pape Das.Linux at gmx.de
Wed Jan 29 21:01:03 UTC 2020


> On 29.01.2020, at 21:48, Jakob Reschke <forums.jakob at resfarm.de> wrote:
> 
> Funny... not sure whether this is an intended and desirable feature though.

Reminds me of the "lobby" object in Self.

-t

> 
> <commits at source.squeak.org> schrieb am Mi., 29. Jan. 2020, 18:27:
> Christoph Thiede uploaded a new version of Tools to project The Inbox:
> http://source.squeak.org/inbox/Tools-ct.937.mcz
> 
> ==================== Summary ====================
> 
> Name: Tools-ct.937
> Author: ct
> Time: 29 January 2020, 6:26:57.711894 pm
> UUID: 34af35bc-83e2-9b40-9aa3-8f6af607ac6e
> Ancestors: Tools-mt.929
> 
> Fixes a styling issue in the Workspace that missstyled class variables from Object as invalid (that is DependentsFields). This is a bug because you can access DependentsFields indeed.
> 
> There is a subtle but important difference between the classOrMetaclass nil and the classOrMetaclass Object: The former does not respect any state for styling, but the latter respects class variables from Object (as well as, hypothetically, ProtoObject).
> 
> Please note that it would be even more accurate to specify nil class instead of Object here, but I don't know whether we really want to emphasize the nuances of UndefinedObject. It feels rather like an implementation detail for me, at least in this context. (Just imagine UndefinedObject had a UniqueInstance class variable - would we really to access it from any Workspace?) Open for opinions.
> 
> =============== Diff against Tools-mt.929 ===============
> 
> Item was changed:
>   ----- Method: Workspace>>aboutToStyle: (in category 'code pane') -----
>   aboutToStyle: aStyler
> 
>         self shouldStyle ifFalse: [ ^false ].
> +       aStyler
> +               parseAMethod: false;
> +               classOrMetaClass: Object;
> -       aStyler 
> -               classOrMetaClass: nil;
>                 workspace: self.
>         ^true!
> 
> 
> 




More information about the Squeak-dev mailing list