Sean Meighan

Software => Bugs in xLights/Nutcracker => Bugs (Please dont post here, soon to be deprecated) => Topic started by: Phrog30 on June 03, 2015, 07:25:54 AM

Title: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 07:25:54 AM
I created a simple sequence for testing.  If I double click in the sequencer grid, the app will crash.  After looking around, it only happens when clicking outside of a timing mark.  So, if I have no timing marks, double clicking anywhere in the grid will cause a crash.  If I have timing marks, it will only crash if I click outside (to the right) of the last mark. 


[attachment deleted by admin]
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 07:30:57 AM
And this didn't happen in the last version?
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 07:33:52 AM
I don't know. I could install an earlier version and try.
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 07:38:44 AM
That's up to you.  I can test it out either way but not till tonight.  It does make it easier when problems are posted for us to know if its a problem that just changed in the last release.

Hmm...just a quick browse of the code and we don't even define an event handler for double-click so the grid shouldn't be capable of receiving that event.  Only the waveform currently processes a double-click.
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 07:47:16 AM
Installed 4.0.35, it was fine. Installed 4.1.0, it crashed. So, the last 2 releases.
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 07:49:37 AM
I'm sorry, it does it on a single click as well. So I guess not specific to double click.
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 08:08:07 AM
Ok then it must be the new logic trying to figure out where you selected in the grid.  It's failing because there is no timing effect above your click.
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 08:10:20 AM
That's the problem when we do these rapid beta releases.  I think what's causing this crash is due to changes that were fixing another type of crash.
Title: Re: Crash 4.1.1
Post by: flyinverted on June 03, 2015, 10:13:30 AM
Here is what I see when I test

4.01.0   Left click on the sequencer with no timing mark = crash
4.01.01 Left click on the sequencer with no timing mark = no crash
4.01.01 Right click on the sequencer with no timing mark = crash

Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 10:49:44 AM
Yeah I did a checkin last night before this release that I thought fixed the crash when left clicking in the grid with no timing mark.  When you reported the right click problem last night I couldn't repeat it so I thought those other changes I checked in after that had fixed it.  Usually I find people don't give me enough details when they talk about problems.  Thinking back I created a new sequence and right clicked in the timing row which didn't crash.  Left click was causing a crash so I fixed that one.  Since you are saying right click is still crashing I can only guess that maybe you added a model and are right clicking there.
Title: Re: Crash 4.1.1
Post by: sean on June 03, 2015, 11:08:28 AM
i cant seem to duplicate this. i have no timing marks, 80 models and carol of the bells.mp3

i click and double click all over the place, no crash.

could you film your screen with your phone and post the video someplace? That usually helps us see exactly how to duplciate what you do
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 11:32:58 AM
Gil, were you able to duplicate?
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 11:38:15 AM
Sorry gotta finish my day job....can't test till tonight.
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 12:22:03 PM
No problem, just curious. I won't be able to make a video for a few days.
Title: Re: Crash 4.1.1
Post by: flyinverted on June 03, 2015, 12:44:03 PM
Here's the video. All clicks were with the left mouse button. Sean, I used your setup.

http://stevesimages.com/xl4/xl4-click-crash.mp4 (http://stevesimages.com/xl4/xl4-click-crash.mp4)

I know.. there's absolutely no valid reason to click on an empty panel because there's nothing you can do there.
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 01:10:16 PM
That must be the new user detection code.  It has a list of people who said they like Vixen and induces random crashes...lol.
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 02:46:26 PM
Here's the video. All clicks were with the left mouse button. Sean, I used your setup.

http://stevesimages.com/xl4/xl4-click-crash.mp4 (http://stevesimages.com/xl4/xl4-click-crash.mp4)

I know.. there's absolutely no valid reason to click on an empty panel because there's nothing you can do there.
No valid reason? Maybe you've been inverted for too long.
Title: Re: Crash 4.1.1
Post by: flyinverted on June 03, 2015, 03:54:42 PM
Do tell Phrog, why would you need to click on the sequencer grid before you have any timing marks or effects?
Maybe I'm missing something.
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 04:06:47 PM
It does seem like more of a test what can break action which is great if that's what your trying to do.  Maybe we should send development builds to Phrog before releases. :)
Title: Re: Crash 4.1.1
Post by: flyinverted on June 03, 2015, 04:16:44 PM
You can click on the wave-form to add timing marks, and that is on the sequencer tab, but above the grid.
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 05:01:09 PM
Do tell Phrog, why would you need to click on the sequencer grid before you have any timing marks or effects?
Maybe I'm missing something.
Simple, a new person, just creating a sequence. You really want that to be their first experience?

I tend to think about those things. You don't see the importance, I do.

James
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 05:02:22 PM
It does seem like more of a test what can break action which is great if that's what your trying to do.  Maybe we should send development builds to Phrog before releases. :)
Absolutely! I told you several days again, I can't write high level code, but I sure can break it.
Title: Re: Crash 4.1.1
Post by: gerry on June 03, 2015, 05:09:39 PM
In defence of Phrog  :), I have quite often done the same , but did not get any crashes.

I am newish to xlights and come from an LSP background , where when u create a new sequence, the default option is to to create timing marks (you can pick the frequency or deliberately say None) and the default option is to copy all your models in (you can nominate from which seq ) - most people use the previous seq.

When you open  LSP and select open a seq , it shows u a list of recently used sequences as part of the dialogue with the last one on the top .

So when the sequencer tab is then presented (very similar to the xlights view now), you are ready to go.

I am on my 3rd seq in xlights now (and have been doing odd imports ) and I mostly get caught out either not having timing marks defined or not having models defined on the sequencer tab, but know now after a few seconds what I have missed.
(mostly when I see everything greyed out)

Just a question of getting used to it, but I can see how anyone would click on the seq tab to see if it would make a diff to the greyed out icons - I have done that previously.
 
 
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 06:16:29 PM
Well I was just able to watch this video.  I see it crashed in CheckForPartialCell which means you can blame me since I just added that function last 3 days.  It uses some variables that apparently aren't always set so I'm trying to figure out all the protections to put around the call so we only call it when everything is valid.
Title: Re: Crash 4.1.1
Post by: Gilrock on June 03, 2015, 06:28:54 PM
By the way the comments to Phrog were meant in good fun.  I was laughing to myself thinking wow he's taking this beta testing seriously.
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 06:28:57 PM
Nah, no blame. It's the nature of the beast. 8)
Title: Re: Crash 4.1.1
Post by: Phrog30 on June 03, 2015, 06:29:54 PM
By the way the comments to Phrog were meant in good fun.  I was laughing to myself thinking wow he's taking this beta testing seriously.
Didn't take it wrong. I do take things serious though. That's a good and bad thing.
Title: Re: Crash 4.1.1
Post by: gerry on June 04, 2015, 12:28:28 AM
Heh Gilrock,

Even I am taking this seriously :) .
Fully plan to use what I am doing for my show this year. Best time to learn .
And if it so happens to test the Beta that's great . If no one works on it now , then the bugs will just remain dormant !

Plus if u have a new user trying to use it , more idiotic things will be done that normally would not get exercised and tested.
 
Title: Re: Crash 4.1.1
Post by: Gilrock on June 04, 2015, 07:22:14 AM
Yeah but I joked that he was taking the beta testing seriously.  What that means is you do a bunch of outrageous things a normal user wouldn't do to try to break the code.   If your sequencing for your actual show great...but that's not serious beta testing. :)
Title: Re: Crash 4.1.1
Post by: mararunr on June 04, 2015, 07:33:49 AM
Yeah but I joked that he was taking the beta testing seriously.  What that means is you do a bunch of outrageous things a normal user wouldn't do to try to break the code.   If your sequencing for your actual show great...but that's not serious beta testing. :)
That's alpha testing (trying to break it), greatly reduces beta breaks.  That's why some github issues got closed without resolution.  I believe they were called irrelevant bugs.
Title: Re: Crash 4.1.1
Post by: Gilrock on June 04, 2015, 07:47:03 AM
Alpha/Beta...did that ever matter to anyone what we were calling it? :)  According to Sean's definition we should have never been in beta anyway.  Beta was supposed to mean we think we've found and fixed all the bugs and crashes that we know about.  Well we knew about a lot of the issues on github way before anyone else found them.
Title: Re: Crash 4.1.1
Post by: gerry on June 04, 2015, 02:53:50 PM
Yeah but I joked that he was taking the beta testing seriously.  What that means is you do a bunch of outrageous things a normal user wouldn't do to try to break the code.   If your sequencing for your actual show great...but that's not serious beta testing. :)
Ahhh,  I see .  This is my first time using a beta product, mostly by the time I start using something, other early adopters have moved on to the next great thing. I thought beta testing was when those who wished to use it early , used it and then reported issues in .
But this year , am using xlights now , and have a beagle bone black with P10s on order. Must be a renaissance  :)
Title: Re: Crash 4.1.1
Post by: sean on June 04, 2015, 06:52:47 PM
here is what i see as our phases of testing

alpha: Code is not complete, bugs everywhere. You cant sequence a complete song. Alpha testers should be highly technical to help describe what they are seeing. testers augment development to get more fingers on the code. For users, they are learning a new look and feel. This code is buggy and incomplete.

beta: code is complete to teh extent that you can make a complete sequence. There are bugs, some significant. A larger test audience. We bring in the "normal" users who are not as technical. The users get the benefit of starting to learn a brand new sequencer but at the cost of "It worked before, now there are new bugs".
Developers get expansive user base to try functionality and UI usefulness.

production: No crashes with normal use. This version can be used for xmas 2015. you , the user, could freeze on this release and you would have a show.
There should be no major bug. This would be bugs blocking you creating a show.
Wide spread user base, stability very high.


for us
alpha: was jan 2015
beta: was mar 2015
prod: estimated june/july 2015. If it is not stable, we will hold on to beta testing maybe an additional month. I do believe we will be in production before expo (July 16th)