Script Editor disassociates - a lossy LC IDE bug

[Ben Rubinstein]

I've seen something similar - on Mac - it may be the same in which case I have 
some hope for you.

The bug I've seen is that I make some changes, run the script, am surprised 
that my changes haven't fixed the behaviour I'm trying to correct. Twiddle the 
script, swear, same. I try setting a breakpoint on the code, it doesn't fire.

Then I set a breakpoint in another script, which calls this one; and step into 
the handler that I've been editing. At that point it seems that I am stepping 
into the version of the code that I've been editing - and my fixes take effect 
and all is well.

So, my inference is that the script editor is holding on to the latest version 
of the script; but that 'apply' is not doing its job, and - if this is the 
same as the bug you've seen - the script is also not being actually saved to 
the object. But stepping into the script in the debugger seems to break it out 
of this dissociative state - so then you can save the script, and save the 
stack, and keep your changes.

No recipe, but I've seen this a number of times, most recently yesterday.

On 06/12/2019 01:29, Paul Dupuis via use-livecode wrote:
> Windows 10 for me
> 
> 
> On 12/5/2019 8:11 PM, Tom Glod via use-livecode wrote:
>> Are you all on Mac?  This does not ring a bell.
>>
>> On Thu, Dec 5, 2019 at 8:02 PM Paul Dupuis via use-livecode <
>> use-livecode at lists.runrev.com> wrote:
>>
>>> On 12/5/2019 6:38 PM, Bob Sneidar via use-livecode wrote:
>>>>> On Dec 5, 2019, at 15:04 , Paul Dupuis via use-livecode <
>>> use-livecode at lists.runrev.com> wrote:
>>>>> On 12/5/2019 5:42 PM, Curry Kenworthy via use-livecode wrote:
>>>>>> Wondering if this bug has already been reported?
>>>>>>
>>>>>> I've seen a wonderful LC 9 IDE bug: sometimes the Script Editor window
>>> seems to disassociate from the object being edited. Thus after this bug is
>>> triggered, any typed script changes may not really be applied and saved,
>>> although they are shown as applied. Next LC session, those changes are
>>> gone; the stack code is still as it was prior to the bug being triggered
>>> and prior to editing the script.
>>>>>> An alternate form is that the script changes actually are applied, but
>>> in a bad way: first the script disassociates completely and is empty, then
>>> the empty script is saved, overwriting the previous script. Nice! (I think
>>> this is my favorite variation; if we're going to cause loss of work/code,
>>> might as well do it boldly and not beat around the bush.)
>>>>>> I suspect this happens more on Mac than Windows, but since I don't
>>> have a recipe, not sure. I've seen this happen to other people rather than
>>> on my own machine. On LC 905 for sure, and probably on 950 also.
>>>>>> If Script Editor disassociation does happen, I advise immediate but
>>> careful no-save quit or force quit, and pray that you're already in the
>>> habit of backing up beforehand. Copy and paste the current script text to
>>> an external text editor if necessary. Of course this particular bug just
>>> might be able to defeat the safety of stack backups too in some situations,
>>> but hey, that's why it's an extra-fabulous bug!
>>>>>> So, in addition to being a bit crashy and glitchy, the 9 IDE is a bit
>>> work-lossy. That could turn off both new and old users if they are the
>>> humor-deficient type to frown on losing some work. But shucks, ain't no
>>> thang! Only features features features (and new side projects) matter! Glad
>>> we all agree on that.
>>>>>> Any reports or recipes on this issue? I'll post a report if there is
>>> none. Thanks....
>>>>> I saw a similar error in earlier versions of LC9 where the script I was
>>> editing was for a different object that what was indicate the Project
>>> Browser. So the Project Browser was showing (highlighting) Object A but
>>> actually thought it was on Object B.
>>>>> It occurred fairly intermittently. I never developed a recipe and I
>>> have not see this in the most recent releases (9.0.5 and 9.5.1)
>>>> Ditto. Hasn't happened for a while though.
>>>>
>>>> Bob S
>>>>
>>>>
>>> Of course, I am not saying that the bug isn't still in the current
>>> releases, just that I have not noticed it in the current releases ;-)
>>>
>>>
>>> _______________________________________________
>>> use-livecode mailing list
>>> use-livecode at lists.runrev.com
>>> Please visit this url to subscribe, unsubscribe and manage your
>>> subscription preferences:
>>> http://lists.runrev.com/mailman/listinfo/use-livecode
>>>
>>
> 
> 
> _______________________________________________
> use-livecode mailing list
> use-livecode at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your subscription 
> preferences:
> http://lists.runrev.com/mailman/listinfo/use-livecode
>