Some bugs?

[Jacob Howe]

Dear God Of Lispyscript,

    First of all, I would like to thank you for creating this

project. I am a lisp hack in training, and your project ticks all of

my boxes for a great language. I too read "Beating the Averages"; I

think it's a good sign this project was inspired by that essay. I'll

be using this language for creating an online game (if it all works

out ;) ), so I'm reading the documentation now and I've found a few

unclear things/errors. Correct me if I'm wrong on any of the below. If

you need help fixing them let me know (I'm just learning javascript

now so I don't know much, but wherever I can help I will)

First: a bug in the website. There's a discrepancy between "Try it!" and command line compiler.

On the web site, in the "Try it!" tab, when I try compiling

    (function (-> ($ "#xyz") (.required) (.alphanum) (.min 3) (.max 30) (.with "email")))

It gives 

function(->,$,"#xyz",.required,.alphanum,.min,3,.max,30,.with,"email") { };

which just seems wrong.

However, compiling with the command line tool lispy gives:

    // Generated by LispyScript v0.3.6

    ((((($("#xyz")).required()).alphanum()).min(3)).max(30)).with("email");

You must not be..... expanding macros? Check out the _compile function you exported in lib/ls.js I think.

And besides, what's with all of those extra parens? Shouldn't it just be:

    $("#xyz").required().alphanum().min(3).max(30).with("email");

 or so? (not knowing jquery)

Also, on the documentation page, you say the usage for the `->` macro is as follows:

    (-> (function expression) (.key1 args1 ...) (.key2 args2 ...) ...)

Looking at the source code for the macro `->`, I still don't

understand. It doesn't seem as though the arguments match up, in other

words it seems like the useage should be something like:

    (-> expression (.key1 args1 ...) (.key2 args2 ...) ...)

assuming scheme documentation style ellipses (where the preceding sexp is repeated 0 or more times) 

But I still cannot understand the macro. looking at lib/ls.js, it seems

`rest...` is syntactically significant, but this isn't documented any

which where. Same with #args-shift and it's kin.

Also, I was wondering if you used the closure compiler.

Let me know if I can help, 

   Jacob Howe

[Santosh Rajan]

My comments inline.

On Wednesday, 7 August 2013 22:32:04 UTC+5:30, Jacob Howe wrote:

    First of all, I would like to thank you for creating this

project. I am a lisp hack in training, and your project ticks all of

my boxes for a great language. I too read "Beating the Averages"; I

think it's a good sign this project was inspired by that essay. I'll

be using this language for creating an online game (if it all works

out ;) ), so I'm reading the documentation now and I've found a few

unclear things/errors. Correct me if I'm wrong on any of the below. If

you need help fixing them let me know (I'm just learning javascript

now so I don't know much, but wherever I can help I will)

Thanks for the encouraging words.

 

First: a bug in the website. There's a discrepancy between "Try it!" and command line compiler.

On the web site, in the "Try it!" tab, when I try compiling

    (function (-> ($ "#xyz") (.required) (.alphanum) (.min 3) (.max 30) (.with "email")))

It gives 

function(->,$,"#xyz",.required,.alphanum,.min,3,.max,30,.with,"email") { };

which just seems wrong.

However, compiling with the command line tool lispy gives:

    // Generated by LispyScript v0.3.6

    ((((($("#xyz")).required()).alphanum()).min(3)).max(30)).with("email");

You must not be..... expanding macros? Check out the _compile function you exported in lib/ls.js I think.

This is because the tryit page uses an old version of LispyScript. The tryit is only for trying out simple things. I need to mention this somewhere (or update it to 3.6)

 

And besides, what's with all of those extra parens? Shouldn't it just be:

    $("#xyz").required().alphanum().min(3).max(30).with("email");

 or so? (not knowing jquery)

The extra parens are there because we need whatever is in the parens to be evaluated. Unfortunately in the case of nesting it becomes to many nested parens. This is not a problem for the JavaScript compilers which are will optimize this. However I agree it does not look good.

 

Also, on the documentation page, you say the usage for the `->` macro is as follows:

    (-> (function expression) (.key1 args1 ...) (.key2 args2 ...) ...)

Looking at the source code for the macro `->`, I still don't

understand. It doesn't seem as though the arguments match up, in other

words it seems like the useage should be something like:

    (-> expression (.key1 args1 ...) (.key2 args2 ...) ...)

assuming scheme documentation style ellipses (where the preceding sexp is repeated 0 or more times) 

You are right I need to make this clear. It should read <function expression>. I will fix that in the docs.

 

But I still cannot understand the macro. looking at lib/ls.js, it seems

`rest...` is syntactically significant, but this isn't documented any

which where. Same with #args-shift and it's kin.

rest... is documented here http://lispyscript.com/docs/#macros

However the homoiconic expressions #args-shift etc are not yet documented. I need to do that.

 

Also, I was wondering if you used the closure compiler.

No I did not use the closure compiler.

 

Let me know if I can help, 

Thanks for the offer. Any help is welcome.

Thanks again.

 

   Jacob Howe

[Jacob Howe]

Alright, I'll keep on running through the documentation and I'll let you know of anything else I find. 

Jacob Howe

[Jacob Howe]

Same error with str: Should read

(str <string expression 1> <string expression 2> ...)

again assuming scheme style ellipses. 

[Jacob Howe]

(array element1 element2 ...)

should read (array element1 ...)

if you're doing scheme ellipses

Also 

(object key1 val1 key2 val2 ....)

should be something like

(object key1 val1 ...)

but that would imply that val is being repeated, so some notation

should be made up to compensate for this?

(object {key1 val1} ...)

of course people will get the idea, but consistency is good.

(new Classname arg1 arg2 ....)

should be 

(new Classname arg1 ...)

Really, a lot of things need to be updated if you're going to maintain

consistent notation for your function calls. If you send me the source

for the webpage (assuming you didn't use a static page) I'll update it

so you don't have to, then you can compare the diffs and fix my

mistakes, if you want me too.

[Santosh Rajan]

The source of docs page is here.

https://github.com/santoshrajan/lispyscript/blob/gh-pages/docs/index.html

Thanks

[Jacob Howe]

    Can't find your email anywhere and I don't think that I can attach

a file here. I don't know how to work github (yet!). Let me know how

to send you `index.html`. So, I updated `index.html`, fixing typos and

an inconsistent function notation. I have some other outstanding

problems I think I uncovered:

    First, with the emacs mode: The highlighting is off. For instance,

If I type in `doSomething` as a function, it highlights the `do` but

not the `Something`. Also, `doMonad` doesn't get highlighted. I know

this is Kris's deal, so should I let him know?

    Also, `each2d` takes a third argument, but it doesn't do anything

with it.

    The function `javascript` isn't documented.

    I've already let you know about the old interpreter on the

website, and the family of `#args-shift` functions. putting this here

for easy reference

    Also, I found out that macros must take at least one expression;

This is inconsistent with common lisp's implementation. Did you mean

to make it this way?

    Plus, this paragraph isn't finished: 

    "The bindings for a and b are made for every element of the

corresponding arrays. In the second example we applied a when

condition to the result expression to keep out results greater than

six. This when condition can only be used with the array monad and

maybe monad. Or to be more specific, monads that define mZero"

    And the withMonad function isn't documented.

    aaand there's a gotcha with html5 templates: you can't use the

object and eventsource tags.

    -> still works with normal functions; The arguments don't have to

be .key functions, they can be normal functions and -> still

compiles. Did you mean to do this?

    Not trying to load you with work, but I think it would be worth a

run through of your code to see which functions are being exposed and

document those.  Of course, not knowing javascript, I don't know if

there's an export function or whatever, so I cannot immediately tell

if a function is exported, and besides if it was I would not know what

it did, so I feel as though you're the better candidate to do that.

[Santosh Rajan]

Thanks. You can attach file here. I can see the option above the reply box. I will go through the file and make the necessary changes.

On Thursday, 8 August 2013 20:23:42 UTC+5:30, Jacob Howe wrote:

    First, with the emacs mode: The highlighting is off. For instance,

If I type in `doSomething` as a function, it highlights the `do` but

not the `Something`. Also, `doMonad` doesn't get highlighted. I know

this is Kris's deal, so should I let him know?

I am not using emacs so I wouldnt know. Maybe you should let the emacs mode author know.

 

    Also, `each2d` takes a third argument, but it doesn't do anything

with it.

 

    The function `javascript` isn't documented.

Yes it is not documented, because it is meant for internal use only.

 

    I've already let you know about the old interpreter on the

website, and the family of `#args-shift` functions. putting this here

for easy reference

    Also, I found out that macros must take at least one expression;

This is inconsistent with common lisp's implementation. Did you mean

to make it this way?

Yes it is meant to be this way. There are reasons for it.

 

    Plus, this paragraph isn't finished: 

    "The bindings for a and b are made for every element of the

corresponding arrays. In the second example we applied a when

condition to the result expression to keep out results greater than

six. This when condition can only be used with the array monad and

maybe monad. Or to be more specific, monads that define mZero"

    And the withMonad function isn't documented.

Will do this.

 

    aaand there's a gotcha with html5 templates: you can't use the

object and eventsource tags.

Yes `object` does not work because it clashes with lispyscript `object`. Need to fix that. Need to add support for eventsource.

 

    -> still works with normal functions; The arguments don't have to

be .key functions, they can be normal functions and -> still

compiles. Did you mean to do this?

Yes

 

    Not trying to load you with work, but I think it would be worth a

run through of your code to see which functions are being exposed and

document those.  Of course, not knowing javascript, I don't know if

there's an export function or whatever, so I cannot immediately tell

if a function is exported, and besides if it was I would not know what

it did, so I feel as though you're the better candidate to do that.

Yes I agree. Thanks. 

[Jacob Howe]

I think that you have permission to post files but other users don't. So, here's the file:

https://docs.google.com/file/d/0B-iXq8LOsCkydTBjM3F2cHRqa3c/edit?usp=sharing