on 5/14/05 1:13 PM, Björn Eiríksson at bjorn at einhugur dot com wrote:
> On 14.5.2005, at 17:04, Jonathan Johnson wrote:
>
>> Wait, you're saying it's crashing for you, not that it's returning
>> nil? What does the crash log look like? Which variation of
>> REALnewInstance are you using?
>
> Huh !!
>
> Jonathan, I have been reporting this in every version lately, yes it
> crashes and does not return nil. It fails the unit test that I sent to
> you in same way as before. Crashes where REALnewInstance is called.
>
> I believe you posted once a log from your copy where it got past that
> point, but we have never gotten a version that gets past that point.
I have been using REALnewInstance like REALnewInstance("MyAttributes"). It
is called after PluginEntry in the constructor of the control.
I know that you responded , when b4 was released, that it does not matter
when REALnewInstance is called, as long as it occurs after PluginEntry.
Björn calls REALnewInstance in the Open callback of the control, so in both
cases the timing to call it should be correct.
You also posted in April that because of a lack of plugins available for
testing, you created your own plugin indicating that REALnewInstance would
work. In that time I raised the question what might be different between
your test plugin and our plugins.
Apparently, the unit test passes your copies of RB always, whereas our
copies always crash.
Given all of the above, would it be possible that our plugin SDK is outdated
and that we need a new plugin api that does not call
CallResolver("PluginNewInstance")
but instead uses something else ?
(excluding "CreateInstance" which is used on a REALclassRef)
Alfred
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
|