On Sunday, August 18, 2019 at 7:10:26 AM UTC+3, Brandon Barker wrote:
>
> Hi Artyom - would you mind sharing the repo URL, assuming it is open?
>
>
Well, here it is:
https://github.com/ashalkhakov/libatsc
> On Thursday, August 8, 2019 at 2:49:43 AM UTC-4, Artyom Shalkhakov wrote:
>>
>> Hi Brand
Hi Artyom - would you mind sharing the repo URL, assuming it is open?
On Thursday, August 8, 2019 at 2:49:43 AM UTC-4, Artyom Shalkhakov wrote:
>
> Hi Brandon, Alexander,
>
> On Wednesday, August 7, 2019 at 5:31:04 AM UTC+3, Brandon Barker wrote:
>
>> I just want to say this is quite intriguing, I
Hi Brandon, Alexander,
On Wednesday, August 7, 2019 at 5:31:04 AM UTC+3, Brandon Barker wrote:
> I just want to say this is quite intriguing, I think it would be very
> appealing to have a standard library built on this idea (a modification of
> Temptory?), perhaps with some more fleshed out ex
I just want to say this is quite intriguing, I think it would be very
appealing to have a standard library built on this idea (a modification of
Temptory?), perhaps with some more fleshed out examples of IO splits before
really digging in.
On Thu, Mar 21, 2019, 8:18 PM gmhwxi wrote:
>
> One can
This looks interesting, will follow up if/when I will got something from it
:)
--
You received this message because you are subscribed to the Google Groups
"ats-lang-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to ats-lang-users+unsubscr...@goog
The problem, as I understand it, is very common in practice.
The solution you outlined in Haskell is what I call a "closed-world"
solution.
What do I mean by a "closed-world" solution? For instance, in LISP, there
are a lot of functions that are like: creating a resource, using it
and then clos
Thanks for directions to investigate!
>
--
You received this message because you are subscribed to the Google Groups
"ats-lang-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to ats-lang-users+unsubscr...@googlegroups.com.
To view this discussion o
Thanks!
There are quite a few issues mentioned in your message. Let me address them
separately.
>>The point of `IO` data type in pure language is not about precision. It's
purpose is to control where the source and distribution of side-effects in
your program.
In haskell, it is done in an opposit
Thanks for response.
понедельник, 5 августа 2019 г., 17:56:43 UTC пользователь gmhwxi написал:
>
> If I understand correctly, I think I have dealt with this issue before.
>
> When an iterator is taken out, you can stop the object from being used
> until
> the iterator is returned. For instance, t
If I understand correctly, I think I have dealt with this issue before.
When an iterator is taken out, you can stop the object from being used until
the iterator is returned. For instance, the 'minus' view in
PATSHOME/prelude/SATS/extern.sats
is precisely introduced for this purpose.
To me, a fun
Hi all, I want to popup this topic again.
And I will start with some context: recently, I had debugged pcmanfm file
manager's segmentation faults, that happen during navigation through hidden
folders. I found out, that the cause of this error is in per-directory
settings of displaying hidden fil
Thanks, Dambaev, for the excellent explanation.
On Sunday, May 5, 2019 at 1:48:48 PM UTC-4, Dambaev Alexander wrote:
>
>
>
>> Admittedly I don't really understand what RT is, but from what I
>> understand, in Haskell the expression like [print "ha"] is basically a
>> command to the top-level i
>
> So, being pure, Haskell pretends that "program is started with some
> unknown RealWorld state of value0 and each IO action modifies this state
> and thus, this value is unique for all expressions and thus. complier
> should evaluate them all instead of caching result of the first evaluation
>
> Admittedly I don't really understand what RT is, but from what I
> understand, in Haskell the expression like [print "ha"] is basically a
> command to the top-level interpreter (which is the language runtime) to
> perform an effect on the console (moreover, it will be evaluated on
> as-ne
On Tuesday, March 26, 2019 at 5:14:34 PM UTC+2, gmhwxi wrote:
>
> Boxing is removed:
>
> https://pastebin.com/JacNgK2t
>
> Some hacks are used to make it work...
>
>
This is awesome. Now we don't need to allocate every time we call a 'read'
function. :)
Thanks! I've updated my gist with this and
Boxing is removed:
https://pastebin.com/JacNgK2t
Some hacks are used to make it work...
On Tue, Mar 26, 2019 at 10:14 AM Hongwei Xi wrote:
> >>Did you mean to use vtype in {a,b:vtype} here? Or not?
>
> Typo: vtype should be vt@ype.
>
> On Tue, Mar 26, 2019 at 10:12 AM Artyom Shalkhakov <
> a
>>Did you mean to use vtype in {a,b:vtype} here? Or not?
Typo: vtype should be vt@ype.
On Tue, Mar 26, 2019 at 10:12 AM Artyom Shalkhakov <
artyom.shalkha...@gmail.com> wrote:
> On Tuesday, March 26, 2019 at 4:07:38 PM UTC+2, gmhwxi wrote:
>>
>> A "standard" solution is to use call-by-reference:
On Tuesday, March 26, 2019 at 4:07:38 PM UTC+2, gmhwxi wrote:
>
> A "standard" solution is to use call-by-reference:
>
> extern
> fun
> runCommand
> {a:vt@ype}
> (c:Command(a:vt@ype), &a? >> a): void
>
>
I think we covered this a few months back when Chris asked about it. This
finally reminded me
A "standard" solution is to use call-by-reference:
extern
fun
runCommand
{a:vt@ype}
(c:Command(a:vt@ype), &a? >> a): void
datavtype
Command(vt@ype) =
| Nop(unit)
| Read(string)
| Print(unit) of string
| Seq(unit) of (Command(unit), Command(unit))
| {a,b:vtype} Bind(b) of (Command(a), a?
On Tuesday, March 26, 2019 at 3:55:47 PM UTC+2, gmhwxi wrote:
>
> >> Now, what about boxing, can we do something with boxing?
>
> Do you mean that you want to have 'a:t@ype' instead of 'a:type'?
>
>
Yes. I think any use of 'bind' is highly discouraged if we have 'a:type'
restriction (e.g. want to
>> Now, what about boxing, can we do something with boxing?
Do you mean that you want to have 'a:t@ype' instead of 'a:type'?
On Tue, Mar 26, 2019 at 9:36 AM Artyom Shalkhakov <
artyom.shalkha...@gmail.com> wrote:
> Hi Hongwei,
>
> On Tuesday, March 26, 2019 at 3:22:41 PM UTC+2, gmhwxi wrote:
>>
On Tuesday, March 26, 2019 at 3:02:42 PM UTC+2, gmhwxi wrote:
>
> Nice!
>
> But I am very surprised that this code actually works.
> My understanding is that It works because of a bug in patsopt :)
>
>
May I suggest this bug to be classified as a feature. :)
Basically, runCommand should be impleme
Hi Hongwei,
On Tuesday, March 26, 2019 at 3:22:41 PM UTC+2, gmhwxi wrote:
>
> Here is a linear version:
>
> https://pastebin.com/sqXcRhnf
>
> Also, Command is a linear datatype (i.e., dataviewtype).
>
>
Great! Now, what about boxing, can we do something with boxing? I expect
such code, if it's re
Here is a linear version:
https://pastebin.com/sqXcRhnf
Also, Command is a linear datatype (i.e., dataviewtype).
On Tue, Mar 26, 2019 at 9:02 AM Hongwei Xi wrote:
> Nice!
>
> But I am very surprised that this code actually works.
> My understanding is that It works because of a bug in patsopt
Nice!
But I am very surprised that this code actually works.
My understanding is that It works because of a bug in patsopt :)
Basically, runCommand should be implemented as a polymorphic function
(instead of a function
template). And 'a:t0ype' should be 'a:type'.
fun runCommand: {a:type} Command
I've made a gist:
https://gist.github.com/ashalkhakov/c3577e97b20020fde31f84447fd1e056
It actually works. It illustrates the basics (sequencing, bind, input and
output). Nice. It doesn't have Haskell's "return" though, but that is
pretty simple to add (it's something that "creates" IO where the
Hi Brandon,
On Friday, March 22, 2019 at 8:49:29 PM UTC+2, Brandon Barker wrote:
>
> Hey Artyom,
>
> Thanks for the very interesting analysis and response.
>
>
Glad you found it useful!
On Fri, Mar 22, 2019 at 4:06 AM Artyom Shalkhakov > wrote:
>
>> Hi Brandon,
>>
>> This is a very lively discus
Sounds like I have to fix atspkg :p
If you want to use the monads package with npm + Makefiles, I suppose I
could upload to npm (it might even be a good idea given the permanence
of packages there).
As an aside: I suspect that what makes IO so nice in the context of
Haskell is that you get the sa
On Friday, March 22, 2019 at 2:49:29 PM UTC-4, Brandon Barker wrote:
>
> Hey Artyom,
>
> Thanks for the very interesting analysis and response.
>
> On Fri, Mar 22, 2019 at 4:06 AM Artyom Shalkhakov <
> artyom.shalkha...@gmail.com> wrote:
>
>> Hi Brandon,
>>
>> This is a very lively discussion, th
Hey Artyom,
Thanks for the very interesting analysis and response.
On Fri, Mar 22, 2019 at 4:06 AM Artyom Shalkhakov <
artyom.shalkha...@gmail.com> wrote:
> Hi Brandon,
>
> This is a very lively discussion, thanks for bringing it up.
>
> On Friday, March 22, 2019 at 5:48:07 AM UTC+2, Brandon Bar
Hi Brandon,
This is a very lively discussion, thanks for bringing it up.
On Friday, March 22, 2019 at 5:48:07 AM UTC+2, Brandon Barker wrote:
>
> And for what it's worth, here is an Idris program for haha using Effects:
>
> module Main
>
>
> import Effects
> import Effect.StdIO
>
>
> hello : Eff
Effect type, not effect time! But it is time for bed.
On Thu, Mar 21, 2019 at 11:48 PM Brandon Barker
wrote:
> And for what it's worth, here is an Idris program for haha using Effects:
>
> module Main
>
>
> import Effects
> import Effect.StdIO
>
>
> hello : Eff () [STDIO]
> hello = let ha = StdI
And for what it's worth, here is an Idris program for haha using Effects:
module Main
import Effects
import Effect.StdIO
hello : Eff () [STDIO]
hello = let ha = StdIO.putStr "ha" in ha *> ha
main : IO ()
main = run hello
It prints "ha" twice, despite being a strict language. I presume, bu
On Thu, Mar 21, 2019 at 8:18 PM gmhwxi wrote:
>
> One can definitely build a monad-based library to support IO:
>
> absvtype IO(a:vt@ype) = ptr
>
If really using monads, would it also be reasonable to do (or ideally start
at Functor and then build up to Monad):
absvtype Monad(a:vt@ype) = ptr
Monads can be used to track effects in call-by-value languages as well.
I was once told that SPJ wanted to implement a call-by-value Haskell if he
could
have started again :)
In Ur/Web, Adam Chlipala's functional call-by-value language for
web-programming,
monads are used extensively for signals,
I'm not sure what the point of an IO monad is in a strict language -
Haskell has good reason to employ them, but I don't think it's really
worth it in other cases (at least, I actually know much about
PureScript, but I don't think ATS needs it).
I think the way that ATS uses linear types for IO is
One can definitely build a monad-based library to support IO:
absvtype IO(a:vt@ype) = ptr
The problem with IO monad is that it is so broad. With linear types,
a programmer can specify a lot more precisely.
>>is that ATS doesn't (by default?) model an IO effect.
No, it doesn't if you use the de
On Thursday, March 21, 2019 at 9:30:40 AM UTC-4, Brandon Barker wrote:
>
> Hi Artyom,
>
> I'm also grappling with the issue of RT in this case as I'd so far only
> thought about it in terms of function calls, but what you and Vanessa say
> helped me to understand the issue. Though I haven't man
On Wednesday, March 20, 2019 at 8:45:24 PM UTC-5, Brandon Barker wrote:
>
> I'm a little rusty, so can't come up with many good examples.
>
> Apparently it is possible to do something like this in OCaml:
>
> implement
> main0 () = {
> val () = let
> val ha = print("ha")
> in
> (ha; ha)
I see. Then you have to blame C :)
My observation is that engineers have a tendency to not
treat things uniformly. For instance, in C, you cannot declare
a variable (or a field in a struct) to be of the type 'void'. In other
words, 'void' is treated specially as a type. This special treatment
of
Hi Artyom,
I'm also grappling with the issue of RT in this case as I'd so far only
thought about it in terms of function calls, but what you and Vanessa say
helped me to understand the issue. Though I haven't managed to get ATS to
have the same behavior as OCaml in the "let expression" above, I
On Thursday, March 21, 2019 at 8:39:10 AM UTC-4, gmhwxi wrote:
>
>
> >>ATS doesn't seem to allow a sequence expression in the "in" position of
> a let expression,
>
> It does. You can write:
>
> implement
> main0 () =
> let
> val ha = print("ha")
> in
> ha; ha
> end
>
>
I think I tried that
>>then bad things happen (beta reduction is no longer valid!)
ML is call-by-value. It only supports beta-v (that is call-by-value beta).
There are a lot of problems with supporting beta. For instance, it is in
general
very difficult to tell the memory usage of Haskell programs.
On Wed, Mar 20,
>>ATS doesn't seem to allow a sequence expression in the "in" position of a
let expression,
It does. You can write:
implement
main0 () =
let
val ha = print("ha")
in
ha; ha
end
To get two ha's, you can write:
implement
main0 () =
let
val ha = $delay(print("ha"))
in
!ha; !ha
end
If you w
Hi Brandon,
Admittedly I don't really understand what RT is, but from what I
understand, in Haskell the expression like [print "ha"] is basically a
command to the top-level interpreter (which is the language runtime) to
perform an effect on the console (moreover, it will be evaluated on
as-needed
On Wednesday, March 20, 2019 at 10:40:34 PM UTC-4, Vanessa McHale wrote:
>
> I think that might have do with laziness? If you have a side-effecting
> expression and you try to pretend it's call-by-need, then bad things happen
> (beta reduction is no longer valid!)
>
> Yes, that seems to be the
I think that might have do with laziness? If you have a side-effecting
expression and you try to pretend it's call-by-need, then bad things
happen (beta reduction is no longer valid!)
Do you have an example in OCaml? I admit I am curious as to why their
compiler would do such a thing.
On 3/20/19
I'm a little rusty, so can't come up with many good examples.
Apparently it is possible to do something like this in OCaml:
implement
main0 () = {
val () = let
val ha = print("ha")
in
(ha; ha) // How to get two ha's here?
end
}
After running the program, you would only see one "ha
48 matches
Mail list logo
