1st of June 2020, news on Vectorblade

It has been more than two weeks without news on Vectorblade (unless you watch the beta-tester forum). I’ll update to the current level of progress and also tell you that I’ll pause developing for about at least two weeks – because I am again on a two weeks holiday trip.

The mentioned “method” of trying to find the bug (see last blog entry) took some time to actually work. I implemented it alright… but the damn bug was not willing to show itself. It took a couple of days – me playing several games, sometimes bad (only getting to level fourtyish) sometimes better (getting to levels around the 150).

A couple of days of unsuccessfull playing (no crash) -it finally happened, playing a “monster” game of 474 levels… and in that level finally the crash happened! (after having played alltogether over 1000 levels in these tryouts)

It was passed midnight (having played a more than 2 hours game) and I was tired… nonetheless I just HAD to check it out.

The crash happened again after I reset the vectrex and started in “playback” mode. YES!

Than I loaded the saved game into Vide – to see what Vide had to say. The crash happened there too!

YES!

This means two things:

a) the crash could have happened on Vide all the time – I had just very bad luck and it didn’t.

b) I can REPRODUCE it – and finally get to the bottom of this (if this is the only crash-bug)

But that night I was definitely to tired but I went to bed happily (when can we ever say that – going to bed happily after the own program just crashed!)


Long story short – the cause of the crash was found quite “easily” once I could actually debug it. It was inserted not to long ago (beta 34) with an update of the collision detection algorithm (player shots – enemy). Under very seldom circumstances it could happen, that an already removed shot could be tested again for collision with an enemy. That “already” removed shot would than be removed again – which could lead to an endless loop.

It is not 100% certain that this was the only crash causing bug – but at least this one was removed quite unspectacularily… and at least I have not seen another crash.


Now I can move on and do the finishing touches to Vectorblade and finally RELEASE it.
(No, still no date, no still no preorder!)

The last couple of months I didn’t do anything new with Vectorblade – and honestly I had been a little bit “down” – for having that unfindable crash bug.

But now I can go on and hopefully finish what I started.

Following things I have added/changed the last two weeks:

  • Save/ load game in progress
    (a loaded game has highscores and such disabled)
  • load/save will be available with the flash version of VB and with the VecFever version
  • a new “bonus” – “drunken” – again taken as a reminder of one of Warblades most iconic bonuses
  • also as in Warblade you will be able to use “drunken” to get a special achievement (finishing a minestorm while being drunk)
  • Minestorm will give different bonus/results depending on the speed you finish
  • higher difficulty levels got more “variety”
    (boss fights will feature different behaviours, level aliens will use attack patterns while in intro phase …)
  • some icons changed
  • some shop prices were “fixed”
  • overall quite a lot of long outstanding bug fixes

I will also change the game over sequence, and perhaps one or two other “cinematics”. On facebook you might have seen the “candle” tryout – I thought to use that as a game over scence – as a death symbolism… but my beta tester collectively rejected the idea as “not in line with the vectorblade scenario”. I guess they are right… but it looks so nice.

After my holiday I will come back to this and try to do a space debris explosion kind of something… which will be more in line with vectorblade.

After finishing that – I’ll do one last final “release candidate” beta test round – and go on to manufacture the cards.

Gee – I will be glad when that thing is finally finished :-).

On that note… the instruction manual… Next week I will receive the third incarnation of the manual. None of it being my fault…

The printer:

  • first stapled the title page wrongly (upside down/inside out)
  • than something went wrong in the RIP – and all images really looked bad
  • I hope the next version will finally be ok.

I will see when I get home.

Cheers

Malban

Tagged on:

10 thoughts on “1st of June 2020, news on Vectorblade

  1. Phillip Eaton

    Congratulations on finding the bug! How long ago did the bug get created? Was it your game recording system that gave you the data you needed to do the root cause analysis? Are there any lessons learned?

    1. Malban Post author

      Bug creation date: somewhat August last year (2019)
      Yes – because of the recording I could find it.
      Not really – being able to record your game is great 🙂
      Cheers!

  2. David Galloway

    Amazing perseverance to locate this bug. I guess you leveled up!

    Have you reflected on how you could have prevented yourself from this bug in the first place? I’d you had religiously used git you could have binary searched your check-ins with the help of your legions of fan testers. I certainly have used that approach many times in the past. Code review is another approach.

    I’m sure the tools you’ve developed to overcome this crash will be very helpful in the future.

    Thanks for everything! And thank goodness this crash is behind you!

Leave a Reply to Graham Toal Cancel reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.