[Mark SiblyKeymaster]

Cool, might have to buy that! Pricey though…

Going by this (p30)…

http://sites.ieee.org/scv-cs/files/2013/03/Right-SizingPrecision1.pdf

…the key trick seems to be the addition of the ‘uncertainty’ bit and the addition of ‘exponent length’ and ‘fraction length’ fields.

I assume there are 2 length fields ‘coz he’s talking about variable lengthed vars, but with fixed length vars surely you only need one, eg: it could contain the length of the exponent and the length of the fraction would just be some constant minus that? Or is this variable length thing part of what makes it useful.

 

[Mark SiblyKeymaster]

Thanks!

As for demos, you’re probably best to start with the xmas demo (if you haven’t seen it already):

http://monkey2.monkey-x.com/2015/12/23/xmas-demo-2015

Most of the recent work has been under-the-hood stuff and website+docs.

However, I hope to have a new ‘playable’ demo out ‘soon’ which in Sibly-time means under a month, except when it doesn’t…

 

[Mark SiblyKeymaster]

Ok, found the time limit on edits setting and changed it to 999999.

Not sure if it works retroactively though – can you edit now?

Look like it’ll take me a while to get the forums ‘tuned’. Still way better than having to write them myself!

[Mark SiblyKeymaster]

Cool, didn’t realize tagging things in github created a release!

Anyway, just did a download and rebuildall.sh myself and it seemed to work.

I don’t get where this is coming from though:

> MX2CC monkey2 compiler V0.001
> sh: tmp.txt: Permission denied
> Can’t find gcc – please install gcc tools or update your bin/env.txt file

When does this appear? This is the old monkey1 version of the compiler!

If it happens when you try and compile something in ted, perhaps you have not renamed the new monkey2 folder to ‘monkey2’. It looks like Ted might still be used the old monkey2 folder.

Ted only looks for ../monkey2/, so you’ll need to rename new version to ‘monkey2’ each update (sorry).

Another thing to try after rebuildall.sh is:

../bin/mx2cc_macos makeapp ../hello-world.monkey2

 

 

 

 

[Mark SiblyKeymaster]

I guess the “Blitz” name isn’t coming back any time soon…

Pretty much – I mean, where do you go after ‘blitz max’? ‘blitz infinite’?!?

 

[Mark SiblyKeymaster]

This looks like an old-ish version – have you tried the latest from github?

Click on the ‘code’ link above.

 

 

[Mark SiblyKeymaster]

All I know about npm is that it broke the internet recently! But that page is interesting, I’ll definitely take a closer look at it.

I’ve vaguely been thinking about implementing something similar to haxelib…

http://lib.haxe.org/

…but I’m not sure how that deals with version conflicts.

I kind of like how macos frameworks do it – there’s just top-level ‘VersionA’, ‘VersionB’ etc. subdirs inside each framework. Simple but effective.

 

[Mark SiblyKeymaster]

All monkey2 modules will appear on the modules page eventually. I’ve just restricted it to monkey, libc and std for now to make life easier while I get things going.

So any monkey1 modules that ends up ‘converted’ to monkey2 will also appear on the modules page including (soon!) mojo2. Other monkey1 modules will be monkey2-ized over time as user demand dictates!

DataBuffer, Stream and the filesystem functions are already there in the monkey2 std module. These should be familiar to monkey1 users, although have been modified in a few places.

Now that I think about it, I quite like that ‘APIs reference’ overview too. Will probably do something like it for monkey2.

Hope that answers your question!

 

[Mark SiblyKeymaster]

> crazy semantics become far more doable.

Not really – the ‘front end’ (ie: parsing, type checking etc) is pretty much the same as a real compiler’s. Only the back end is simplified, and that doesn’t have to deal with syntax at all, as all that stuff is out of the way.

To support this, in addition to the above, the return statement would need to change, the compiler would have to be able to ‘unpack’ values so func1( func2() ) could work, a new ‘type’ would need to be added etc, at which point you’re effectively implementing tuples!

 

 

 

[Mark SiblyKeymaster]

To do this right now, the obvious way would be to subclass something like the Generator<T> class with a NeighborGenerator<Vector2> class. That would work, but it’s not as elegant as an enumerator method.

An interesting potential alternative would be to ‘push’ the neighbors to a generator (consumer?) eg:

Monkey

Method Neighbors( g:Generator<Vector2> )    g.Yield( this+up )    g.Yield( this+left )    ...etc... End

1 2 3 4 5 6 7 8 9

Method Neighbors( g:Generator<Vector2> )      g.Yield( this+up )      g.Yield( this+left )      ...etc...   End

Then, a global function like ‘Generate’ could be used to do the generation (not sure if it’s this simple…!) eg:

Monkey

Function Generate<T>:Generator<T>( func( g:Generator<T> ) ) Local g:=new Generator<T> func( g ) Return g End

1 2 3 4 5 6 7

Function Generate<T>:Generator<T>( func( g:Generator<T> ) )      Local g:=new Generator<T>    func( g )    Return g   End

…which could be used by client code like this…

Monkey

Function Main()    For Local v2:=Eachin Generate( v1.Neighbors )       ...    Next End

1 2 3 4 5 6 7 8 9 10

Function Main()      For Local v2:=Eachin Generate( v1.Neighbors )         ...      Next   End

This is actually kind of interesting – instead of the Neighbors method returning a ‘producer’ ala c#, we do the opposite and pass it a ‘consumer’.

If course, it’d be nice to support both approaches, and the compiler can indeed help (although it’s *really* nice if it doesn’t have to, or a minor tweak ‘enables’ something cool like this!), but I’m still getting my head around a lot of this stuff and don’t want to commit to anything yet.

[Mark SiblyKeymaster]

Wont be happening in the near future, if ever.

Support for ‘tuples’ would likely involve major changes to the compiler, and given I’ve never really used them or even felt the need for them (perhaps very infrequently…) they’re just not really in the picture for me right now.

What *may* be happening is bmx style ‘var’ parameters (or ‘out’ params in c# lingo) which would achieve something similar.

This is a way off yet (probably look into it when doing reflection) but my vague plan here is to mirror the super-flexible function pointer mechanism with a similar ‘variable pointer’ system that can be used to create pointers to fields/properties/globals. Maybe!

 

[Mark SiblyKeymaster]

> in the top left of the console there’s a little mobile icon

Well, so there is…!

 

[Mark SiblyKeymaster]

Ok, added this to the main style.css file – does it work?

 

 

[Mark SiblyKeymaster]

How do you unsubscribe from threads? I can only find “Notify me of follow-up replies via email” when posting…

Uncheck the ‘notify me of…’ checkbox.

This was defaulting to ‘checked’ until this morning, so you’ll still be subscribed to threads you posted to yesterday and will have to unsubscribe from them.

 

 

[Mark SiblyKeymaster]

> I work in WordPress all the time so if you need any pointers on anything give me a shout!

Thanks!

I had to learn about functions/bbpress.php (I think it was) this morning to turn of the default ‘notify me of replies…’ check (which turned out to be a theme issue). I wasn’t too happy about having to resort to ‘programming’, but it only involved commenting out a function so it wasn’t too bad.

A quick question: one of the reasons I choose this theme was because it worked reasonably well with bbpress, and a lot of themes don’t.  Are there any other cool themes you can recommend that also ‘just work’ with bbpress?

> CleanTalk does a great job at stopping spambots for only about $7 a year

Yep, I’ve got CleanTalk installed but haven’t forked out for the anti-spam yet. Now the sites live I’ll expect I have to soon, but it’s already done a great job blocking spam registrations. I accidentally left the site ‘live’ a while ago, came back a few hours later and there were like 10 new spam users!

[edit]Just subscribed to CleanTalk![/edit]

 

[Author]

Posts