Hi Eliot, yes, I am missing the mini-stack too, the plan was is to fix
it after the release. Tim and I had a brief talk about it the other
day.
http://lists.squeakfoundation.org/pipermail/squeak-dev/2015-July/185153.htmlI think your suggestion to check for the prefix string "Warning" could
be a quick and easy fix, and I almost did that exact same thing
(instead of cmm 7/2/2015 11:18), but after briefly exploring the API
between ToolSet and Debugger and seeing that there are already
configuration hints being passed, and so may be potential
opportunities for improvement there (i.e., passing the signaling
Exception so it can decide the debugger config). Something like that,
too much for this release.
On Tue, Jul 14, 2015 at 8:28 PM, David T. Lewis <
[hidden email]> wrote:
> On Tue, Jul 14, 2015 at 05:58:20PM -0700, Eliot Miranda wrote:
>>
>> On Tue, Jul 14, 2015 at 5:42 PM, Eliot Miranda <
[hidden email]> wrote:
>> >
>> > has anyone else noticed that now the Notifier doesn't contain a stack
>> > trace, only a copy of the title, the message text of the exception? Isn't
>> > this a regression from the recent Warning changes? Can we please have the
>> > stack trace back in cases other than Warning? I'm playing with events in
>> > the VM simulator which are breaking the UI, and its difficult to debug
>> > without a stack trace to show me where I've botched things up.
>>
>>
>> yes, Chris, the change to Debugger>>openNotifierContents:label:, cmm
>> 7/2/2015 11:18, borks the stack trace. i.e. this
>>
>> spec := self buildNotifierWith: builder label: label message: msg.
>>
>> became
>>
>> spec := self buildNotifierWith: builder label: label message: (msg ifNil:
>> [label]).
>>
>> which means no stack trace ever. I think it needs to be something like
>>
>> spec := self buildNotifierWith: builder label: label message: (msg ifNil:
>> [(label beginsWith: 'Warning:') ifTrue: [label]]).
>>
>> Is this serious enough to go for another release candidate? IMO it is, but
>> I'd like to canvas opinions.
>
> Yes I have noticed it, and yes I would like to see it fixed.
>
> No, it is not serious enough to require another release candidate. There is
> (or will be) a 4.6 update stream available to manage updates like this.
>
> The 4.6 release is almost done. Let's make it so.
>
> Dave
>
>
Locked
