Dear All,

We are happy to announce a new edition
of the Dogelog player:

- Enhanced '$YIELD'/1:
Previously the '$YIELD'/1 instruction was hard coded
to understand an integer for sleep/1 and an atom for
native library loading. The implementation has been
change to accept a Promise object, known from
JavaScript. Emulations were made available for
Java and Python.

- Async I/O:
Based on the new '$YIELD'/1 and a buffer fast path
idea, the console reading, file property and file
reading have been asyncified. There is no change in
the Prolog API, files are still to be opened via the
ISO core open/3 and accessed via ISO core get_code/2
respectively ISO core peek_code/2. But input can now
be executed quasi-parallel by multiple tasks.

- HTTP client:
The ISO core open/4 has been moved to the library(spin).
The support of the HTTP and HTTPS protocol in open/3
and open/4, has been extended from the JavaScript
browser to the JavaScript command line (by means of
new fetch in nodejs >21). Further supported targets
are now Python and Java (by means of new HTTP client
in JDK >21).

Have Fun!

Jan Burse, http://www.xlog.ch/ , 07.03.2024

Dogelog Player is a Prolog system that is 100% written
in Prolog itself. It is avsailable for the JavaScript,
Python and Java platform. We show how the stable
ordering of Prolog variables in Dogelog Player allows
highly efficient constraint solving approaches.

We convert the undirected graph into a directed graph
and create adjacency lists. This inequality graph doesn’t
have the same delayed evaluation overhead than
(#=)/2 and dif/2. We demonstrate map coloring and
sudoku solving. Further improvements in a
subsequent article.

See also:

Map Coloring and Sudoku Solving
https://twitter.com/dogelogch/status/1766885890712883276

Map Coloring and Sudoku Solving
https://www.facebook.com/groups/dogelog

Mild Shock schrieb:
> Dear All,
>
> We are happy to announce a new edition
> of the Dogelog player:
>
> - Enhanced '$YIELD'/1:
>   Previously the '$YIELD'/1 instruction was hard coded
> to understand an integer for sleep/1 and an atom for
> native library loading. The implementation has been
> change to accept a Promise object, known from
> JavaScript. Emulations were made available for
> Java and Python.
>
> - Async I/O:
>   Based on the new '$YIELD'/1 and a buffer fast path
> idea, the console reading, file property and file
> reading have been asyncified. There is no change in
> the Prolog API, files are still to be opened via the
> ISO core open/3 and accessed via ISO core get_code/2
> respectively ISO core peek_code/2. But input can now
> be executed quasi-parallel by multiple tasks.
>
> - HTTP client:
>   The ISO core open/4 has been moved to the library(spin).
> The support of the HTTP and HTTPS protocol in open/3
> and open/4, has been extended from the JavaScript
> browser to the JavaScript command line (by means of
> new fetch in nodejs >21). Further supported targets
> are now Python and Java (by means of new HTTP client
> in JDK >21).
>
> Have Fun!
>
> Jan Burse, http://www.xlog.ch/ , 07.03.2024

Small implementation node about the new HTTP client
in Dogelog Player for JavaScript. I am using fetch()
everywhere. No need to get help from libraries such as:

> Axios is a promise-based HTTP Client for node.js
> and the browser. It is isomorphic (= it can run in
> the browser and nodejs with the same codebase).
> On the server-side it uses the native node.js http
> module, while on the client (browser) it uses XMLHttpRequests.
https://axios-http.com/docs/intro

First of all XMLHttpRequests is deprecated in the
browser, use fetch() instead. And nodejs has fetch()
since release 21.x.x as well. So they have become
isomorphic making AXIOS obsolete:

> Notable Changes - Oct 17, 2023
> The recent update to Node.js, version 21, includes
> an important change to the fetch module as well as
> WebStreams . Both modules were marked as stable
> after a recent update.
https://nodejs.org/en/blog/announcements/v21-release-announce

Mild Shock schrieb:
>
> Dogelog Player is a Prolog system that is 100% written
> in Prolog itself. It is avsailable for the JavaScript,
> Python and Java platform. We show how the stable
> ordering of Prolog variables in Dogelog Player allows
> highly efficient constraint solving approaches.
>
> We convert the undirected graph into a directed graph
> and create adjacency lists. This inequality graph doesn’t
> have the same delayed evaluation overhead than
> (#=)/2 and dif/2. We demonstrate map coloring and
> sudoku solving. Further improvements in a
> subsequent article.
>
> See also:
>
> Map Coloring and Sudoku Solving
> https://twitter.com/dogelogch/status/1766885890712883276
>
> Map Coloring and Sudoku Solving
> https://www.facebook.com/groups/dogelog
>
>
> Mild Shock schrieb:
>> Dear All,
>>
>> We are happy to announce a new edition
>> of the Dogelog player:
>>
>> - Enhanced '$YIELD'/1:
>>    Previously the '$YIELD'/1 instruction was hard coded
>> to understand an integer for sleep/1 and an atom for
>> native library loading. The implementation has been
>> change to accept a Promise object, known from
>> JavaScript. Emulations were made available for
>> Java and Python.
>>
>> - Async I/O:
>>    Based on the new '$YIELD'/1 and a buffer fast path
>> idea, the console reading, file property and file
>> reading have been asyncified. There is no change in
>> the Prolog API, files are still to be opened via the
>> ISO core open/3 and accessed via ISO core get_code/2
>> respectively ISO core peek_code/2. But input can now
>> be executed quasi-parallel by multiple tasks.
>>
>> - HTTP client:
>>    The ISO core open/4 has been moved to the library(spin).
>> The support of the HTTP and HTTPS protocol in open/3
>> and open/4, has been extended from the JavaScript
>> browser to the JavaScript command line (by means of
>> new fetch in nodejs >21). Further supported targets
>> are now Python and Java (by means of new HTTP client
>> in JDK >21).
>>
>> Have Fun!
>>
>> Jan Burse, http://www.xlog.ch/ , 07.03.2024
>

On Sun, 10 Mar 2024 23:49:38 +0100, Mild Shock wrote:

> First of all XMLHttpRequests is deprecated in the browser, use fetch()
> instead.

I don’t know why you would say that. Fetch is only for downloads
<https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API>, while
XMLHttpRequest works in both directions
<https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest>.

Do you mean POST requests?
You can do them also with fetch:

There is a POST example as the first fetch example:
https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API/Using_Fetch

So you can use fetch also for upload.
Not only for download.

Lawrence D'Oliveiro schrieb:
> On Sun, 10 Mar 2024 23:49:38 +0100, Mild Shock wrote:
>
>> First of all XMLHttpRequests is deprecated in the browser, use fetch()
>> instead.
>
> I don’t know why you would say that. Fetch is only for downloads
> <https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API>, while
> XMLHttpRequest works in both directions
> <https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest>.
>

Its straight forward:

async function postData(url = "", data = {}) {
const response = await fetch(url, {
method: "POST",
[...]
body: /* put your upload content here */

Mild Shock schrieb:
> Do you mean POST requests?
> You can do them also with fetch:
>
> There is a POST example as the first fetch example:
> https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API/Using_Fetch
>
> So you can use fetch also for upload.
> Not only for download.
>
> Lawrence D'Oliveiro schrieb:
>> On Sun, 10 Mar 2024 23:49:38 +0100, Mild Shock wrote:
>>
>>> First of all XMLHttpRequests is deprecated in the browser, use fetch()
>>> instead.
>>
>> I don’t know why you would say that. Fetch is only for downloads
>> <https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API>, while
>> XMLHttpRequest works in both directions
>> <https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest>.
>>
>

On Mon, 11 Mar 2024 09:07:44 +0100, Mild Shock wrote:

> async function postData(url = "", data = {}) {
> const response = await fetch(url, {
> method: "POST",
> [...]
> body: /* put your upload content here */

How is that better than XMLHttpRequest?

The same code runs on nodeJS.
No need to install some XHR library.

I wrote this already:

> Notable Changes - Oct 17, 2023
> The recent update to Node.js, version 21, includes
> an important change to the fetch module as well as
> WebStreams . Both modules were marked as stable
> after a recent update.
https://nodejs.org/en/blog/announcements/v21-release-announce

You get the most bang out of fetch, when
you use it with webstreams. Here is a little example:

await response.body
.pipeThrough(new TextDecoderStream())
.pipeTo(targetDiv.writable)

See also here:

Streaming element with backpressure demo
https://streams.spec.whatwg.org/demos/streaming-element-backpressure.html

Lawrence D'Oliveiro schrieb:
> On Mon, 11 Mar 2024 09:07:44 +0100, Mild Shock wrote:
>
>> async function postData(url = "", data = {}) {
>> const response = await fetch(url, {
>> method: "POST",
>> [...]
>> body: /* put your upload content here */
>
> How is that better than XMLHttpRequest?
>

On Tue, 12 Mar 2024 00:19:57 +0100, Mild Shock wrote:

> The same code runs on nodeJS.
> No need to install some XHR library.

You don’t know how to use XMLHttpRequest directly?

Natively Node.js does not provide the browser XHR API.
There is, however, a node module xmlhttprequest that does.

Lawrence D'Oliveiro schrieb:
> On Tue, 12 Mar 2024 00:19:57 +0100, Mild Shock wrote:
>
>> The same code runs on nodeJS.
>> No need to install some XHR library.
>
> You don’t know how to use XMLHttpRequest directly?
>

In 1997, Hong Kong judge Wayne Gould saw a partly
completed puzzle in a Japanese bookshop. Over six years,
he developed a computer program to produce unique
puzzles rapidly. The program tries to keep one puzzle
ahead of you, by generating the next puzzle while you
are solving the present one.

In the following we show a Prolog program, where
the random generation of a Puzzle is performed in
less than 2 seconds. The Prolog program was mainly
developed for Dogelog Player. We could test the
Prolog program also with Prolog systems such as
SWI-Prolog, Scryer Prolog and Trealla Prolog.

The smallest domain first variable ordering heuristic
allowed us to solve some hard problems below a minute.
Turning the heuristic into a static ordering before
search gave us a further boost and the baseline for
randomization. Measurement showed that solving blank
Sudokus doesn't have a large time variation.

See also:

Birthday Paradox and Sudoku Generation
https://twitter.com/dogelogch/status/1767714755827908991

Birthday Paradox and Sudoku Generation
https://www.facebook.com/groups/dogelog

On Wed, 13 Mar 2024 02:05:10 +0100, Mild Shock wrote:

> Natively Node.js does not provide the browser XHR API.

Browsers do, though.

Please go see a psychatrist, the point of
departure of my post was this advertisement,
which adresses the problem how to have

a single application code base for both browser
and node.JS, which you can also archive by using
3rd party node XHR API module. Or the 3rd party

library AXIOS is yet another alternative:

> Axios is a promise-based HTTP Client for node.js
> and the browser. It is isomorphic (= it can run in
> the browser and nodejs with the same codebase).
> On the server-side it uses the native node.js http
> module, while on the client (browser) it uses XMLHttpRequests.
https://axios-http.com/docs/intro

I then remarked that AXIOS isn't necessary anymore,
since fetch() is also isomorphic among Browser
and Node.js since release 21.x.x. With the benefit

that you wont have any 3rd party dependency anymore,
with fetch() everything will be native,
in the browser and in node.JS:

> Notable Changes - Oct 17, 2023
> The recent update to Node.js, version 21, includes
> an important change to the fetch module as well as
> WebStreams . Both modules were marked as stable
> after a recent update.
https://nodejs.org/en/blog/announcements/v21-release-announce

What don't you understand here? How stupid on
a scale of 1-1000 are you? Level 9000 stupid?

Lawrence D'Oliveiro schrieb:
> On Wed, 13 Mar 2024 02:05:10 +0100, Mild Shock wrote:
>
>> Natively Node.js does not provide the browser XHR API.
>
> Browsers do, though.
>

Maybe its a language problem, what is your native
language Lawrence D'Oliveiro <ldo@nz.invalid>.
Do you understand what AXIOS advertizes, and

what they mean when they write "same codebase"?
Any idea what "same [application] codebase
[across browser and node.JS"] could mean?

Or are you simply too lazy to read what people cite?

Mild Shock schrieb:
> Please go see a psychatrist, the point of
> departure of my post was this advertisement,
> which adresses the problem how to have
>
> a single application code base for both browser
> and node.JS, which you can also archive by using
> 3rd party node XHR API module. Or the 3rd party
>
> library AXIOS is yet another alternative:
>
> > Axios is a promise-based HTTP Client for node.js
> > and the browser. It is isomorphic (= it can run in
> > the browser and nodejs with the same codebase).
> > On the server-side it uses the native node.js http
> > module, while on the client (browser) it uses XMLHttpRequests.
> https://axios-http.com/docs/intro
>
> I then remarked that AXIOS isn't necessary anymore,
> since fetch() is also isomorphic among Browser
> and Node.js since release 21.x.x. With the benefit
>
> that you wont have any 3rd party dependency anymore,
> with fetch() everything will be native,
> in the browser and in node.JS:
>
> > Notable Changes - Oct 17, 2023
> > The recent update to Node.js, version 21, includes
> > an important change to the fetch module as well as
> > WebStreams . Both modules were marked as stable
> > after a recent update.
> https://nodejs.org/en/blog/announcements/v21-release-announce
>
> What don't you understand here? How stupid on
> a scale of 1-1000 are you? Level 9000 stupid?
>
> Lawrence D'Oliveiro schrieb:
>> On Wed, 13 Mar 2024 02:05:10 +0100, Mild Shock wrote:
>>
>>> Natively Node.js does not provide the browser XHR API.
>>
>> Browsers do, though.
>>
>

Dogelog Player is a Prolog system that features
a generational garbage collector scheme and a
change_arg/3 primitive built-in. So far the two
didn’t go together and we had to use a cludge.
With the upcoming release we provide a proper
solution based on marking write-barriers.

Our remembered set on the variable binding trail
has the benefit that it can reduce itself during
backtracking. By adopting conductor variables, i.e.
instantiated variables that are not trailed, from
formerly Jekejeke Prolog, we could recreate a
similar benefit for Dogelog Players change_arg/3.

See also:

Marking Write-Barriers for Dogelog Player
https://twitter.com/dogelogch/status/1771559896741192186

Marking Write-Barriers for Dogelog Player
https://www.facebook.com/groups/dogelog

On 23/03/2024 15:43, Mild Shock wrote:
>
> Dogelog Player is a Prolog system that features

<snip>

Tell us, in outline, about some of the data structures in the JavaScript
version of the system and how they are implemented.

John

Some of the datastructures are documented here:
https://www.novacuor.ch/doctab/rscsrv/docs/17_host/package.pdf

But you can also inspect the source code:
git clone https://www.dogelog.ch/dogelog/.git

John Harris schrieb:
> Tell us, in outline, about some of the data structures
> in the JavaScript version of the system and
> how they are implemented.

On 24/03/2024 11:23, Mild Shock wrote:
>
> Some of the datastructures are documented here:
> https://www.novacuor.ch/doctab/rscsrv/docs/17_host/package.pdf

<snip>

No they're not. Only functions are described.

John

Fucking liar, look again. Like hier:

4.1 Folder "util"
Section "meta"
4.2 Folder "albufeira"
Section "instruction"
Section "stages"
4.3 Folder "player"
Section "platform"
4.4 Folder "playerpy"
Section "platform"
4.5 Folder "playerj"
Section "platform"
Section “handlers”

A Prolog interpreter doesn't need much datastructures.

John Harris schrieb:
> On 24/03/2024 11:23, Mild Shock wrote:
>>
>> Some of the datastructures are documented here:
>> https://www.novacuor.ch/doctab/rscsrv/docs/17_host/package.pdf
>
>   <snip>
>
> No they're not. Only functions are described.
>
>   John
>
>

Also Datastructures are burried inside the functions, like here:

Term API

deref(T): (host language)
Return the dereferencing of the term T.
copy_term(T): (host language)
Return a copy of the term T.
new Variable(): (host language)
Create a Prolog variable.
new Compound(F, A): (host language)
Create a Prolog compound with functor F and arguments A.

What do you want? An UML diagram? Forget it.

Mild Shock schrieb:
> Fucking liar, look again. Like hier:
>
>
> 4.1 Folder "util"
> Section "meta"
> 4.2 Folder "albufeira"
> Section "instruction"
> Section "stages"
> 4.3 Folder "player"
> Section "platform"
> 4.4 Folder "playerpy"
> Section "platform"
> 4.5 Folder "playerj"
> Section "platform"
> Section “handlers”
>
> A Prolog interpreter doesn't need much datastructures.
>
> John Harris schrieb:
>> On 24/03/2024 11:23, Mild Shock wrote:
>>>
>>> Some of the datastructures are documented here:
>>> https://www.novacuor.ch/doctab/rscsrv/docs/17_host/package.pdf
>>
>>    <snip>
>>
>> No they're not. Only functions are described.
>>
>>    John
>>
>>
>

On 26/03/2024 19:21, Mild Shock wrote:
> Fucking liar, look again. Like hier:
>
>
> 4.1 Folder "util"
> Section "meta"

<snip>

Section "meta" has 13 lines of text. Tell us which ones describe a data
structure. And no, "cross_loading/1" is not a description.

Your JavaScript code includes some form of garbage collection. Has that
influenced the data structures ?

John

Do you have even the slightest clue of logic programming?

Dogelog Player is a Prolog system 100% written
in Prolog itself. Where do you expect the data to
be stored? For your informaton moron, this here:

cross_loading/1

Is a so called predicate indicator (PI), giving
you the formation of the predicate name and predicate
arity. So the data model in very shortened relational
notation is this here:

cross_loadeing(_)

Where:

cross_loading: The table name

_: The single column.

Hope this helps. More accademic names for data inside
Prolog are, the so called extensional database (EDB):

The set of facts is called the extensional
database or EDB of the Datalog program
https://en.wikipedia.org/wiki/Datalog

John Harris schrieb:
> On 26/03/2024 19:21, Mild Shock wrote:
>> Fucking liar, look again. Like hier:
>>
>>
>> 4.1 Folder "util"
>> Section "meta"
>
>   <snip>
>
> Section "meta" has 13 lines of text. Tell us which ones describe a data
> structure. And no, "cross_loading/1" is not a description.
>
> Your JavaScript code includes some form of garbage collection. Has that
> influenced the data structures ?
>
>   John
>
>

I mean I repeat in every post and blog I make
about Dogelog Player that it is written in 100%
Prolog itself. This means it uses Prolog for its

own data. You call these posts nauseating. I call
you a complete fucking moron with blistering
herpes furuncle on top of your neck.

Gets some brain cells idiot. Its should be also
explained in the docu what it means that Dogelog
Player is 100% written in Prolog itself.

But I do not intend make a citation now.

Just get lost and don't waste everybodies time.

Mild Shock schrieb:
> Do you have even the slightest clue of logic programming?
>
> Dogelog Player is a Prolog system 100% written
> in Prolog itself. Where do you expect the data to
> be stored? For your informaton moron, this here:
>
> cross_loading/1
>
> Is a so called predicate indicator (PI), giving
> you the formation of the predicate name and predicate
> arity. So the data model in very shortened relational
> notation is this here:
>
> cross_loadeing(_)
>
> Where:
>
> cross_loading: The table name
>
> _: The single column.
>
> Hope this helps. More accademic names for data inside
> Prolog are, the so called extensional database (EDB):
>
> The set of facts is called the extensional
> database or EDB of the Datalog program
> https://en.wikipedia.org/wiki/Datalog
>
> John Harris schrieb:
>> On 26/03/2024 19:21, Mild Shock wrote:
>>> Fucking liar, look again. Like hier:
>>>
>>>
>>> 4.1 Folder "util"
>>> Section "meta"
>>
>>    <snip>
>>
>> Section "meta" has 13 lines of text. Tell us which ones describe a
>> data structure. And no, "cross_loading/1" is not a description.
>>
>> Your JavaScript code includes some form of garbage collection. Has
>> that influenced the data structures ?
>>
>>    John
>>
>>
>

John Harris schrieb:
> Your JavaScript code includes some form of garbage
> collection. Has that influenced the data structures ?

There is a long thread about the early Garbage Collection
of Dogelog Player in comp.lang.python. Especially whether
it can be made fast, with useful input from a few
Pythonistas.

Then I post regularly about the evolution of the GC
of Dogelog Player. It was just posted a few days ago,
when you were still on your silly tangent that my

posts are nauseating. Now you are asking stuff, that
I am anyway posting about. Whats is wrong with you?
Do you have a PhD in being the most stupid person

on this planet? See here, few days ago,
to be exact 5 days ago:

Mild Shock schrieb 23.03.2024, 16:43:
>
> Dogelog Player is a Prolog system that features
> a generational garbage collector scheme and a
> change_arg/3 primitive built-in. So far the two
> didn’t go together and we had to use a cludge.
> With the upcoming release we provide a proper
> solution based on marking write-barriers.
>
> Our remembered set on the variable binding trail
> has the benefit that it can reduce itself during
> backtracking. By adopting conductor variables, i.e.
> instantiated variables that are not trailed, from
> formerly Jekejeke Prolog, we could recreate a
> similar benefit for Dogelog Players change_arg/3.
>
> See also:
>
> Marking Write-Barriers for Dogelog Player
> https://twitter.com/dogelogch/status/1771559896741192186
>
> Marking Write-Barriers for Dogelog Player
> https://www.facebook.com/groups/dogelog

Of course before reading anything about Dogelog Player
you have to start with the reference section also known
as bibliography, which is found at the end of PDF here:

Dogelog Player Host Version 1.1.5, December 17, 2023
https://www.novacuor.ch/doctab/rscsrv/docs/17_host/package.pdf

You need to first understand the below things before
you dig and put your nose into Dogelog Player:

[1] ISO (1995): Prolog, Part 1: General Core,
International Standard ISO/IEC 13211-1, First Edition, 1995-06-01
http://www.iso.org/standard/21413.html

[2] Clocksin, W. (1983): A portable Prolog compiler, Logic Programming
Workshop, Albufeira Portugal, January 1983
http://www.softwarepreservation.org/projects/prolog/lisbon/lpw83/p74-Bowen.pdf

[3] Carlson, M. et al. (1988): Garbage collection for Prolog based on
WAM. Communications of the ACM 31, 6, 719–740, June 1988
http://dl.acm.org/doi/10.1145/62959.62968

Dogelog Player implements variants of [1], [2] and [3].
SWI-Prolog also implementes [3], but it doesn't implement
the generational part of [3]. Recently I managed to also

implement the generatioal part of [3] via some coloring
approach. If I am not mistaken you find a blog about it
also announced here and somewhere on medium.com.

And most recently I removed a cludge from change_arg/3
also related to [3]. And according blog I also announced
here to be exact 5 days ago. Its also on medium.com.

Maybe you need medium.com membership to be able to read it.

Mild Shock schrieb:
> John Harris schrieb:
>> Your JavaScript code includes some form of garbage collection. Has
>> that influenced the data structures ?
>
> There is a long thread about the early Garbage Collection
> of Dogelog Player in comp.lang.python. Especially whether
> it can be made fast, with useful input from a few
> Pythonistas.
>
> Then I post regularly about the evolution of the GC
> of Dogelog Player. It was just posted a few days ago,
> when you were still on your silly tangent that my
>
> posts are nauseating. Now you are asking stuff, that
> I am anyway posting about. Whats is wrong with you?
> Do you have a PhD in being the most stupid person
>
> on this planet? See here, few days ago,
> to be exact 5 days ago:
>
> Mild Shock schrieb 23.03.2024, 16:43:
> >
> > Dogelog Player is a Prolog system that features
> > a generational garbage collector scheme and a
> > change_arg/3 primitive built-in. So far the two
> > didn’t go together and we had to use a cludge.
> > With the upcoming release we provide a proper
> > solution based on marking write-barriers.
> >
> > Our remembered set on the variable binding trail
> > has the benefit that it can reduce itself during
> > backtracking. By adopting conductor variables, i.e.
> > instantiated variables that are not trailed, from
> > formerly Jekejeke Prolog, we could recreate a
> > similar benefit for Dogelog Players change_arg/3.
> >
> > See also:
> >
> > Marking Write-Barriers for Dogelog Player
> > https://twitter.com/dogelogch/status/1771559896741192186
> >
> > Marking Write-Barriers for Dogelog Player
> > https://www.facebook.com/groups/dogelog