The pop-up is a result of a failing test in:

CodeHolder>>
validateMessageSource: sourceString forSelector: aSelector inClass: theClass
	"Check whether there is evidence that method source is invalid"

	(theClass newParser parseSelector: sourceString asString) = aSelector
		ifFalse: [self informPossiblyCorruptSource].

In my example, the source string is 'ba_z ^true'. However, the result of #parseSelector: is #ba but the selector is actually #'ba_z'. So the test fails resulting in 'self #informPossiblyCorruptSource'.

However, I can successfully doIt:
'Bar new ba_z' ==> true  
meaning that the selector is recognized in this case.

So, Parser>>#parseSelector doesn't understand the underscore, but I'm in unfamiliar territory here.

Doug



On Jun 26, 2014, at 11:08 , Eliot Miranda wrote:

> 
> 
> 
> On Thu, Jun 26, 2014 at 11:05 AM, Eliot Miranda <eliot.miranda at gmail.com> wrote:
> Hi Doug,
> 
>     this is going to be either Shout or Text>>makeSelectorBoldIn: that need changes, probably the latter.  The latter probably needs to just accept selectors containing underscores.  Shout would need to obey the preference. HTH
> 
> hmmm, this works: 'i_contain_underscores ^self' asText makeSelectorBoldIn: UndefinedObject
> 
> So can you show a stack trace?  e.g. by replacing the pop-up with a halt?
>  
> 
> 
> On Wed, Jun 25, 2014 at 11:19 PM, Douglas McPherson <djm1329 at san.rr.com> wrote:
> I want to use underscores in the selectors which define external functions, to match the external function name as much as possible. So I checked the preference "allowUnderscoreSelectors". I can define and use such selectors, however each time I browse a method whose selector has an underscore, I get the attached modal pop-up:
> 
>  <UserDialogBox.png>
> 
> This seems to happen in both 4.4 and 4.5. 
> 
> Is there something else I am supposed to do to properly enable underscores in selectors?
> 
> Thanks,
> Doug 
> 
> 
> 
> 
> 
> 
> -- 
> best,
> Eliot
> 
> 
> 
> -- 
> best,
> Eliot
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20140626/3bc7486b/attachment-0001.htm