#53

Little René Patch with Turtle : https://www.instagram.com/p/BZLlmtDlNDB/?taken-by=chapelierfoumusic

4 Likes
#54

:+1: that should work

That dirty flag is pretty useful…

Have you got the PDF stuff sorted? If it’s still causing you problems I should be able to look into it this week.

#55

I’m only doing it for the beta. I want to be able to identify the build easily.

Negative. I got distracted by reptiles :wink:

Pretty sure it’s just my build tools, so maybe you want to try building from my fork and see if it works on your system?

#56

No 3D printer today, so beta5 (zip, hex) is ready. Here’s the syntax:

Most things appear to be working, but I might eat those words.

Caveats:

  • Familiarize yourself with SPEED being 100 units between cells in each axis. It permits you to set speeds proportional to your angle to facilitate predictable behaviour. Trig functions are absent on the Teletype so it’ll be hard to do that dynamically.

  • Position is tracked fractionally. This permits us to have meaningful movement in more than 8 directions. There is no access to the underlying fractional data, but it carries.

  • When you set your position, you set it to the center of the cell. Therefore, to align with the center

    @Y @Y; @X @X

  • The fence might behave unexpectedly in terms of time spent on a cell before reflection in bounce mode. This is because the turtle travels the distance of the cell twice, as the fence is on the other side of the cell. This can be changed if there’s a preferable way. I could put the fence halfway.

  • I’ve changed and read the code so much that I’ve lost faith in it. Despite using ddd to track things down and increase my confidence in every line of code, the effect has been the opposite. I encourage an enterprising soul to investigate turtle_normalize_position and turtle_step and tell me how I did.

  • It was a mistake to add this feature to the beta, for obvious reasons and by definition. I will clean up the rest of the easy ops and call it a day on code additions bar bug fixes. The next alpha test will be handled differently. :slight_smile:

  • I will continue to write tests to hunt for bugs in the existing feature set that I am familiar / involved with in the meantime. Let’s call beta6 a weekend thing at this point. I need the turtle to get his legs for a bit, so I can work out the kinks until then.

4 Likes
#57

Does it mean the turtle is moving by default ? (Edit, no way, sorry)

This also made me think of a cool feature request : what about interpolating the carried value with the neighbor values when the turtle is “between” to cells ?

#58

You’re 5 hours late!

And you’d be interpolating between up to 4 cells.

2 Likes
#59

Amazing ! Many thanks !

#60

As nobody has downloaded it yet, I was able to ninja edit in a couple of bug fixes into beta5.

  • CTRL-F1 to CTRL-F8 enable/disable scripts
  • Script number is shaded in editor when muted
  • CTRL-F9 en/disables the metro (is there already a keystroke for this?)
  • I forgot to shade the M when disabled
  • KILL now clears stack
  • KILL now disables metro
  • Updated KILL documentation
3 Likes
#61

Is it possible that my turtle was blocked on the bottom left corner ?
Also, are we supposed to set @WRAP 1 or just @WRAP ?

#62

Aaaah, maybe there is a typo, isn’t it supposed to be

@F x1 y1 x2 y2 rather than @F x1 x2 y1 y2 ?

#63

If you find a script that seems to misbehave, share it and I’ll check.

And it’s @WRAP 1.

Typo fixed, thanks!

1 Like
#64

this is great, thank you. just checking that you’re updating the CHANGELOG on your branch with these changes?

interpolation is great, for sure. but it’s not returned by default with @ correct? it’d be better to have not a mode switch, but just a separate op that returns the interp value ie @INTERP

2 Likes
#65

It would be more coherent to some of other TT syntaxes, indeed.

#66

Whoops! I’ll edit my commit messages into there.

#67

The PDF from your beta5 branch is building fine on my Arch Linux system. Haven’t tried it on OSX, but I assume it’s fine there too. Let me know if you need some more help.

A few little things that I noticed…

For this commit:

You probably wouldn’t have seen in the PDF, but you’ll need a blank line before the

## v2.0

in the change log to get it to render properly (GitHub flavoured Markdown vs Pandoc Markdown…)

For this commit:

You’ll want get rid of this change, otherwise when you merge, the Travis icon will be for your branch rather than upstreams.

#68

Glad to hear the docs build fine :slight_smile: Guess I have to actually do the reinstall of pandoc like I said I would

Thanks! Will change that.

Yes, I was planning to roll that back.

1 Like
#69

@Leverkusen here’s a 2.0.1 beta with the screen glitches fixed (at least my initial stress tests show as much).

simply updated libavr32 with the fixes from @scanner_darkly and made any changes needed to get it compiling. looks good.

if anyone else could test prior to me posting this up on github, i’d appreciate it. then @sliderule can close out his new batch of features.

teletype-2.0.1b1.zip (104.9 KB)

if someone can torture this with i2c as well, that’d be helpful.

10 Likes
#71

I am a bit short of time at the moment but just flashed 2.0.1b1 and two patches that had given me a lot of glychees over the last days seem to work fine at a shot glance when randomly skipping through the modes and scripts, which got me to a lot of unreadable lines before.

The patches/scripts also make use of i2c remote for classic meadowphysics and all four TXo oscillators.

So no hard tortures possible yet but seems to look good.

:slight_smile:

6 Likes
#72

Well, I only just finished writing a better test framework, so my next build won’t be until tomorrow, but it will only contain turtle fixes and more testing.

My weekend got bogged down in 3D modelling and printing parts for everything on the honey do list, then my server administration job had a 8 hours of unexpected work. Soooo… tomorrow night!

2 Likes
#73

i’ll get the merge together then. you shouldn’t have any conflicts pulling it into your branch.

1 Like