« first day (237 days earlier)      last day (423 days later) » 

quartata
quartata
01:10
> Waking up your workspace from hibernation
@El'endiaStarman Getting back on the Pytek grind
Where did we even leave off?
El'endia Starman
El'endia Starman
I think I was working on functions last.
quartata
quartata
I thought we finished them?
except for yield
El'endia Starman
El'endia Starman
Not 100%. Take a look at TODO.txt.
quartata
quartata
Oh OK. Let's see...
I went ahead and cloned a fresh copy locally so that I don't have to put up with C9's speed; lemme know if you make a change
Sketching generators out in my head a bit
also we need some proper lists. Should we use dynamic arrays or unrolled linked lists?
Conor O'Brien
Conor O'Brien
the inspector checks back in: shortest fibonacci?
El'endia Starman
El'endia Starman
01:22
@ConorO'Brien o.O
Uh, lemme see...
Conor O'Brien
Conor O'Brien
I know your quine is very short, almost esoteric XD
quartata
quartata
Hrm no ternaries yet. Just regular if
closed form might be shorter actually
oh wait no
Conor O'Brien
Conor O'Brien
I am filled with confidence :P
quartata
quartata
@El'endiaStarman Three options for generators:
Option 1. They just store previous values in a list with no lookahead and no removal of previous items
Option 2. They keep all items around but also make sure they have at least k items ahead of their position calculated
El'endia Starman
El'endia Starman
@ConorO'Brien It should be n=4func:f(x){\if(x>1){\f(x-1)+\f(x-2)}{x}}\corePrint(\f(n)) but the parser seems to be having issues.
quartata
quartata
01:32
Option 3. They keep at most n items cached, n/2 behind n/2 ahead
With appropriate modifications if that's beyond the generator bounds
Conor O'Brien
Conor O'Brien
@El'endiaStarman I see.
El'endia Starman
El'endia Starman
I haven't tried doing a loop yet.
Conor O'Brien
Conor O'Brien
What class is pytek? (Along the lines of functional, object-oriented, etc.)
quartata
quartata
Of course if they do not have a value they generate up to that index
@ConorO'Brien procedural
@El'endiaStarman Thoughts?
El'endia Starman
El'endia Starman
func:f(x){a,b=0,1\for(x){a,b=b,a+b}} should be the loop version. Lemme see if it works.
quartata
quartata
01:37
...actually just realized that lookahead isn't a speed improvement. woops
El'endia Starman
El'endia Starman
Nope, parsing issues still. Grr.
And dangit quartata, I'm already working on the game mechanics project. I can't split my attention between that and Pytek.
quartata
quartata
I'd say Pytek is the more critical of the two; can't let Cheddar get any farther ahead ;)
Did you know they have an HTTP library
Conor O'Brien
Conor O'Brien
puts on spy hat, writes notes furiously
 
10 hours later…
ASCII-only
ASCII-only
12:10
@quartata Don't worry, Cheddar will need to be rewritten for performance, which is going to take a very long time
 
5 hours later…
quartata
quartata
16:48
@El'endiaStarman Ooh, we need to update the unit tests....
> ValueError: Function print has not been defined yet.
These must be old, we don't have \corePrint in them...
El'endia Starman
El'endia Starman
...yeah...
quartata
quartata
OK, most of them fail because we took out some of the arthimetic functions
But exponent fails with something interesting:
\corePrint(2 ** 8) -> TypeError: 'NoneType' object is not iterable
Looks like it happens when collapsing
lemme take a peek
El'endia Starman
El'endia Starman
The parser seems to be quite borked and I haven't taken a look to figure out why. Because, y'know, game mechanics taxonomy...
quartata
quartata
FizzBuzz and Collatz both seem to be working so that's something
@El'endiaStarman Just out of curiosity, why is there a SQLite database in the root?
El'endia Starman
El'endia Starman
@quartata Comes built-in in Django. Probably hitched a ride in the setting up of git.
quartata
quartata
17:10
You're not the only one with silly bugs
if it's reassuring
 
2 hours later…
quartata
quartata
19:13
@El'endiaStarman So what do I need to override to extend pIterator?
I suppose I better take a look at pRange
How are we planning to do assignment in list? Just a.3 = x;?
 
2 hours later…
El'endia Starman
El'endia Starman
21:19
@quartata I believe so, yes.
quartata
quartata
21:44
@El'endiaStarman What are you working on currently? If I add a new node will everything explode
quartata
quartata
22:03
AST_list is basically just AST_args with a different object type :PP
Huh, oh wait....
@El'endiaStarman Right now a list will actually be parsed into AST_opts, but I seem to recall there being a way to choose which object node is used for an AST node
So that there can be multiple types
How does that work? Ideally I'd like to have pOpts and pList
quartata
quartata
22:19
@El'endiaStarman Do we have multiline strings yet?
El'endia Starman
El'endia Starman
23:03
@quartata I think we had those from the start.
quartata
quartata
Yeah, I found that out
El'endia Starman
El'endia Starman
@quartata I'm not working on anything at the moment, bit the parser needs fixing first.
@quartata I forget the names, but there's a list of object types it can parse into, and the parser tries converting to each type in turn, stopping at the first success.
quartata
quartata
OK, nice.
I assume pOpts will fail if it's not after a function
So I can just put pList afterwards
@El'endiaStarman I want to add a @doc directive but I don't have multi-line directives so I'm not quite sure how to handle it
Something like this doesn't work: 
@doc "
  blahblahblah
"
Since the " is consumed as part of the directive
And then the rest is ignored
I guess what I'll do is split it up into a bunch of directives for each documentation field
quartata
quartata
23:37
@El'endiaStarman What exactly is wrong with the parser? All the code I've tried worked
El'endia Starman
El'endia Starman
23:50
@quartata Hmm, maybe there's an issue with the lack of newlines.
quartata
quartata
@El'endiaStarman Try stripping the newlines from Collatz or something?

« first day (237 days earlier)      last day (423 days later) »