[Cuis-dev] Question re: recent updates (4342)

[Phil B]

I tried updating straight through from 4322->4351 (i.e. ignoring the
message on 4342 saying I need to restart, since that would be the only way
to avoid hanging after 4342) but it still white screens for me on restart.
Does the world state fix need to be applied either right before or right
after 4342? (I'm OK with a manual step if fixing it in the update
changesets isn't feasible, I just need to know what the fix is and when to
apply it)

On Mon, Aug 17, 2020 at 1:17 PM Juan Vuletich <juan at jvuletich.org> wrote:

> On 8/16/2020 11:55 PM, Phil B via Cuis-dev wrote:
>
> That appears to fix the hang when applying 4342, but when I save and quit
> the image still hangs on restart (white screen).
>
>
> In the recent large changes to Morphic I introduced an unrelated bug that
> wouldn't be apparent until saving and restarting the image. I just pushed a
> fix to GitHub.
>
> I tested a bit, and I hope it is ok now.
>
> Thanks for reporting!
>
>
> On Sun, Aug 16, 2020 at 4:29 PM Juan Vuletich <juan at jvuletich.org> wrote:
>
>> Wow. I think there was a random process scheduling bug. It worked OK if
>> installing updates from another starting point (I don't even remember
>> which), but not if starting from #4322.
>>
>> I just pushed a new version of update #4342. The fix was to add a Delay
>> after spawning the new Morphic process, to give it time to be setup before
>> interrupting it. It seems to work now. Please try it.
>>
>> Thanks, Phil!
>>
>> On 8/16/2020 11:59 AM, Phil B via Cuis-dev wrote:
>>
>> This was my starting point: https://github.com/pbella/4322
>>
>> On Sun, Aug 16, 2020 at 9:52 AM Phil B <pbpublist at gmail.com> wrote:
>>
>>> Juan,
>>>
>>> This isn't trying to update a development image, it's trying to
>>> reproduce the base image in the repo.  My starting point was a 4322 base
>>> image.  So the issue is one of not being able to reproducibly apply the
>>> updates.
>>>
>>> Thanks,
>>> Phil
>>>
>>>
>>> On Sun, Aug 16, 2020, 7:35 AM Juan Vuletich <juan at jvuletich.org> wrote:
>>>
>>>> On 8/15/2020 12:24 PM, Phil B via Cuis-dev wrote:
>>>>
>>>> Nope, that didn't fix it.  When I restart the image, it hangs.  Not
>>>> sure what's different with my environment re: 4342 but it looks like it's
>>>> got problems.
>>>>
>>>> On Sat, Aug 15, 2020 at 11:10 AM Phil B <pbpublist at gmail.com> wrote:
>>>>
>>>>> Something about 4342 results in a VectorGraphics (font?) dependency
>>>>> which I didn't see a message for until I interrupted the hanging image.   I
>>>>> ended up removing the #beSilent: code and setting author initials manually
>>>>> while starting to troubleshoot which then solved the problem so that's as
>>>>> far as I took it before moving on to the next issue.
>>>>>
>>>>> On Fri, Aug 14, 2020 at 12:59 PM Phil B <pbpublist at gmail.com> wrote:
>>>>>
>>>>>> When I just tried to update my image, the image hangs while
>>>>>> processing 4342.  I see the Transcript message that says 'Please restart
>>>>>> [Install New Updates]' but the UI is frozen...
>>>>>>
>>>>>> Thanks,
>>>>>> Phil
>>>>>>
>>>>>
>>>> Hi Phil,
>>>>
>>>> For most people, the general advice is to just use a new, updated image
>>>> from the repo. In general, I think that keeping own work separate from the
>>>> image, and loading it every day, is the best workflow. It works for me.
>>>>
>>>> Still, if you want to update your personal image, use the FileList to
>>>> take a look at update #4342. See what it does. It does only what it says:
>>>> It removes an unused ivar from HandMorph. Just not applying this update
>>>> will not have serious consequences. This can be done by editing the doIt in
>>>> the changeset. In your image, HandMorph will still have an unused ivar. It
>>>> is not a big deal.
>>>>
>>>> Another alternative, of course, is to debug the problem. Extract the
>>>> doit from the changeset. Evaluate it in a Workspace. If it crashes, repeat
>>>> with smaller parts. Can you fork a process that just sets a global
>>>> variable? If so, can you make it print to Transcript. (This is an extremely
>>>> useful technique when messing with Morphic itself, that's why we have a
>>>> Transcript that runs independent of Morphic). Ok. Can you stop the Morphic
>>>> process and start a new one? Can you redefine class HandMorph with the same
>>>> definition as it has today? And so on. Part of what comes with the personal
>>>> mastery of your own image is this. Its problems are also yours!
>>>>
>>>> If all this fails, you can make your image available to me and I'll
>>>> take a look.
>>>>
>>>> Cheers,
>>>>
>>>> --
>>>> Juan Vuletichwww.cuis-smalltalk.orghttps://github.com/Cuis-Smalltalk/Cuis-Smalltalk-Devhttps://github.com/jvuletichhttps://www.linkedin.com/in/juan-vuletich-75611b3
>>>> @JuanVuletich
>>>>
>>>>
>>
>> --
>> Juan Vuletichwww.cuis-smalltalk.orghttps://github.com/Cuis-Smalltalk/Cuis-Smalltalk-Devhttps://github.com/jvuletichhttps://www.linkedin.com/in/juan-vuletich-75611b3
>> @JuanVuletich
>>
>>
>
> --
> Juan Vuletichwww.cuis-smalltalk.orghttps://github.com/Cuis-Smalltalk/Cuis-Smalltalk-Devhttps://github.com/jvuletichhttps://www.linkedin.com/in/juan-vuletich-75611b3
> @JuanVuletich
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20200817/f0c46f83/attachment.htm>