bentall / cat-language Goto Github PK

The manual (http://www.cat-language.com/manual.html) refers to a construct
qv which seems to be identical to quote.  The only reference I could find
for qv in the primitive reference is one or two places that use it.  Should
this be replaced with quote?

When locale defines decimal separator as comma, not dot, the grammar is not in 
sound with parser: grammar tells to read dot (and interpreter does it), but 
then parser can not recognize floating point literal as it expects comma 
instead of dot.

Can be (and is) solved by parsing in this vay:

value = double.Parse(ToString(), 
System.Globalization.CultureInfo.InvariantCulture.NumberFormat);

Thanks!

What steps will reproduce the problem?
1. Implement the level 0 primitives
2. Implement a way of defining the level 1 primitives using level 0 primitives
3. Attempt to define whilene and the other level 1 primitives (about 17)
that depend on it, and run all of the unit tests.

What is the expected output? What do you see instead?

The expected output in all cases should be the stack containing only true
-- specifically each of the tests should pass.  Instead, you see uncons (or
some other list opeartion) complain because it ends up being called on
something that isn't a list.

Please provide any additional information below.

The reason for this problem is that the definition of whilene in the
primitives documentation online is stated as:
   Implementation   [empty not] while pop
Unfortunately, the first time this executes the (empty not), the list will
be taken off the stack (see the definition of empty).  To fix this, change
the definition to
   Implementation   [dup empty not] while pop
This will make a copy of the list on the top of the stack, then replace the
top copy with whether it is empty.  This will make everything work.  I have
done this on my scheme implementation of all of the primitives from level 0
and 1 (except for the combinators) and this is the only way to get all of
the unit tests to pass.

Final types are not checked (e.g. int and string are considered 
compatible).

The dynamic method resolution algorithm has some bugs.

What steps will reproduce the problem?
1. Enter the code for factorial online (see below for link)
   define rec_fac
      { dup 1 > [] [dec rec_fac *] if }
2. Attempt to compute the factorial of some number (specefically, 1 and
something greater than 1).

What is the expected output? What do you see instead?
Two problems occur.  In the base cases this recurses instead of doing
nothing, so the comparison is flipped (or the blocks are in the wrong
order).  Second, once that works the recursive case computes n - 1, then
the factorial of that, and then does multiply -- but there is nothing else
on the stack.  Hence, we need an additional dup.

Here is a working version of the above function as tested in my scheme
interpreter.
   (define-cat rec_fac => 
     dup 1 lteq_int [] [dup dec rec_fac mul_int] if)

Please provide any additional information below.
http://www.cat-language.com/manual.html

When running a complex program in the online interpreter, such as a bin_rec
program, or a map, etc., it would be very helpful to be able to step
through the program one step at a time. As in a debugger. And to see the
state of the stack after each step.

The partial evaluator no longer works correctly.

What steps will reproduce the problem?
define is0or1 { dup 0 == swap 1 == or }

What is the expected output?
function defined: is0or1 ( int -> bool )

What do you see instead?
exception from type inference engine

What version of the product are you using? On what operating system?
Cat 0.16, Windows XP, C# Express 2005

I want all primitives to be represented in Cat source code files, with 
tests, documentation, type, etc. 

This way a file parser can get the information out of the source code 
files, without needing to hack/read the source code. 

This may be due to macro expansion or a bug in CatGraphWindow or both.
File attached.

typed:
hilbert
expected: a loop to be drawn (once) and the turtle to take the paint pot
with him.
got: the loop flashed as it was drawn.
Gets slower like Shlemiel the Painter's algorithm described on
http://www.joelonsoftware.com/articles/fog0000000319.html

hilbert pop // (all on one line)
expected: loop should be drawn
got: nothing happened. Does the type system believe this has no effect?

hilbert // (on two separate lines)
st
expected: turtle to appear on st
got: the curve was drawn again


What version of the product are you using? On what operating system?
Cat 0.16, Windows XP, C# Express 2005

User defined functions can't be used within other user-defined functions 
after the initial open-window call. This is because the scope gets 
duplicated in the window's version of the executor stack, and new 
definitions aren't added to it. 

What steps will reproduce the problem?
1. [1 2 3] list 
2. dhead
3. drop

What is the expected output? 
stack: (1,2,3) 

What do you see instead?
stack: ()

What version of the product are you using? On what operating system?

0.18.2 on Windows XP 
Please provide any additional information below.

What steps will reproduce the problem?
define lookup : ( int -> int ) { (2,3,5,7,11) swap nth nip }

What is the expected output?
type: (int -> int)

What do you see instead?
type error in function lookup
inferred type (int -> any)
declared type (int -> int)

What version of the product are you using? On what operating system?
Cat 0.16

Please provide any additional information below.
Hopefully 0.17 support for "statically typed homogeneous lists" will fix this.

and type-vector variables which are are identified by an apostrophe and a
name starting with a ****lowercase letter**** ('A, 'Myvar).

This is found on http://www.cat-language.com/manual.html.

I believe lowercase here should be uppercase. I had to read it three times
before I picked up on this.

The CatList and CatStack class are more or less the same thing and should 
probably be merged.

On page http://www.cat-language.com/metacat.html

In the matching paragraph there is a typo 'implementatino.' should be
implementation.

What steps will reproduce the problem?
1. In Conditionals section of tutorial
(http://www.cat-language.com/tutorial.html), read the example code:
    1 1 +
    eq 2
    ["As I expected!"]
    ["I need to be repaired!"]
    if
Note that "eq" before the 2 makes no sense: at that point, there is only
one value on the stack, therefore eq (which takes two arguments) should
cause an error. Presumably, the intended code was 
    1 1 +
    2 eq
    ...
2. Enter the sample code in the interpreter: 1 1 + eq 2
What is the expected output?
true (on the stack)
 What do you see instead?
false 2 (on the stack)

What version of the product are you using? On what operating system?

Tutorial at http://www.cat-language.com/tutorial.html as of April 21, 2008.

Please provide any additional information below.

Cool language, enjoyed the Dr Dobbs article.

http://www.cat-language.com/

Joy link: http://www.latrobe.edu.au/phimvt/joy.html
«langauge» -> «language»

http://www.cat-language.com/tutorial.html

«experssion» -> «expression»
«Becaue» -> «Because»
«cosntructor» -> «constructor»
«the to write» -> ???
«to to each it» -> ???
«foreach» -> «for_each» ?

The current release of the Cat interpreter does not support type inference 
or type checking at compile-time. Any [TypeAnnotation type annotations] 
are  parsed correctly but ignored.

Some recursive programs "dup apply" etc. now are no longer legal, however 
others are. The problem is that it is not clear where the distinction is 
between legal recursive programs and non-legal recursive programs. 

There is no way to load/import/includ a module/code file other than 
directly on the command line. 

I want a logo-style turtle graphics library. Some of the framework already 
exists. For example, we have windows that operate on a single thread. 

What steps will reproduce the problem?
(1,2) [dup] map

What is the expected output?
Either: (1,1,2,2)
 or: Error, expected function which leaves stack balanced at [dup]
as argument for 'map'

What do you see instead?
Throws uncaught exception, auxilliary stack becomes unbalanced

What version of the product are you using? On what operating system?
Cat 0.16, Windows XP, C# Express 2005

Higher order list functions use an auxiliary stack. It strikes me as a 
very inefficient and possibly buggy way to deal with them. I have to fix 
it. 

What steps will reproduce the problem?
1. In cat online interpreter, type 1 3 <enter>. Note 3 is now on top of the
stack. 1 is second (if I'm understanding "top" and "second" correctly).
2. Now type lteq <enter>.
3.

What is the expected output?
false (pushed onto the stack in place of 3 and 1).
http://www.cat-language.com/primitives.html#lteq says "desc:
Pushes true if the top value is less than or equal to the second value,
false otherwise." 3 is not <= 1.
 What do you see instead?
true on top of the stack
Am I just nuts?

What version of the product are you using? On what operating system?
cat interpreter 1.3, Firefox 2.0.0.14

Please provide any additional information below.

The current parsing error reporting mechanism is poor. 
I want to highlight the line and character index. 

Goal: fix by mid-June.

The docs for bin_rec at http://www.cat-language.com/primitives.html#bin_rec
say:
('a ('a -> bool) ('a -> 'b) ('a -> 'C 'a 'a) ('C 'b 'b -> 'b) -> 'b)
  {{ desc:
      execute a binary recursion process"
The text part is not very specific. What are the 5 arguments on the input?
All I can tell from the type signature is that the 5th one ('a) is a
starting value and the 4th (('a -> bool)) is a termination condition.
It kinda looks like the 3rd is what to do in the base case; 2nd is a
splitting function and 1st is a folding function. But it would be nice to
have that in the desc if correct.
(Ah... the Dr Dobbs article lists the arguments as "input, cond, term,
unfold, fold". It would be good to have that in the desc if it is still true.)

Also, it would be nice to have bin_rec implemented in the online
interpreter, since it is featured prominently in the DDJ article.

bin_rec was 

What is the expected output? What do you see instead?


What version of the product are you using? On what operating system?


Please provide any additional information below.

What steps will reproduce the problem?
((1)(2)) flatten

What is the expected output?
Stack: (1,2)

What do you see instead?
uncaught exception, stack was not displayed

What version of the product are you using? On what operating system?
Cat 0.16, Windows XP, C# Express 2005

When defining new functions, the inferred type should be output.

"[dup] m" returns "( -> 'a 'a)" which is clearly wrong. 

Possibly the problem is the rolling up phase.

What steps will reproduce the problem?
1. Type in 1.0

What is the expected output? 
stack: 1.0

What do you see instead?
stack: 1

What steps will reproduce the problem?
1.
2.
3.

What is the expected output? What do you see instead?


Please use labels and text to provide additional information.

What steps will reproduce the problem?
1. type in help

('A -> 'B) -> ('A -> 'B)

This is an illegal type signature but is parsed successfully

Typo is found on
http://www.cat-language.com/tutorial.html

Under the expressions section.  There is a repeat of the such as 'the the', 
actual caption

higher-order languages and """the the""" theory of languages.

List comprehensions are missing and have to be added.