Smalltalk › Squeak › Squeak - Dev

HttpClient is intended to be used for...

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

16 messages Options

timrowledge

HttpClient is intended to be used for...

... what exactly? There's no class comment to help out.

So far as I can work out it seems to be partially a proxy for working out if the image is running in a web browser (can we do that these days? I was under the impression that was no longer doable with current browsers) and partially a sort of shim for some network-file-fetching-putting that would be different if actually running in a browser.

If we can't do the browser thing we probably ought to clean out the related code.

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
"bOtHeR" said Pooh, mistaking the LSD tablet for aspirin

Tobias Pape

Re: HttpClient is intended to be used for...

Hi Tim,

> On 19.01.2019, at 07:13, tim Rowledge <[hidden email]> wrote:
>
> ... what exactly? There's no class comment to help out.
>
> So far as I can work out it seems to be partially a proxy for working out if the image is running in a web browser (can we do that these days? I was under the impression that was no longer doable with current browsers) and partially a sort of shim for some network-file-fetching-putting that would be different if actually running in a browser.
>
> If we can't do the browser thing we probably ought to clean out the related code.

Sorry for the nitpick, but do you mean HTTPClient or HttpClient?
I see only the former in my trunk image but remember that there's been some package somewhere that has the latter…
Best regards
-Tobias

>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
> "bOtHeR" said Pooh, mistaking the LSD tablet for aspirin
>
>
>

Hannes Hirzel

Re: HttpClient is intended to be used for...

A good question:

Versions 3.2, 3.6, 3.8, 5.0, 5.1 and 5.2 all have HTTPClient.
e.g http://wiki.squeak.org/squeak/6569

http://wiki.squeak.org/squeak/3747 is a good place to put a summary of
this question and discussion ....

There should also be discussion about a wrapper object so that Zn
(Pharo classes) based code may be executed in Squeak. I think there is
some code with which this was started.

--Hannes

On 1/19/19, Tobias Pape <[hidden email]> wrote:

> Hi Tim,
>
>> On 19.01.2019, at 07:13, tim Rowledge <[hidden email]> wrote:
>>
>> ... what exactly? There's no class comment to help out.
>>
>> So far as I can work out it seems to be partially a proxy for working out
>> if the image is running in a web browser (can we do that these days? I was
>> under the impression that was no longer doable with current browsers) and
>> partially a sort of shim for some network-file-fetching-putting that would
>> be different if actually running in a browser.
>>
>> If we can't do the browser thing we probably ought to clean out the
>> related code.
>
> Sorry for the nitpick, but do you mean HTTPClient or HttpClient?
> I see only the former in my trunk image but remember that there's been some
> package somewhere that has the latter…
> Best regards
> -Tobias
>
>>
>> tim
>> --
>> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
>> "bOtHeR" said Pooh, mistaking the LSD tablet for aspirin
>>
>>
>>
>
>
>

Hannes Hirzel

Re: HttpClient is intended to be used for...

WebClient/ZnClient wrapper
http://wiki.squeak.org/squeak/3863

On 1/21/19, H. Hirzel <[hidden email]> wrote:

> A good question:
>
> Versions 3.2, 3.6, 3.8, 5.0, 5.1 and 5.2 all have HTTPClient.
> e.g http://wiki.squeak.org/squeak/6569
>
>
> http://wiki.squeak.org/squeak/3747 is a good place to put a summary of
> this question and discussion ....
>
> There should also be discussion about a wrapper object so that Zn
> (Pharo classes) based code may be executed in Squeak. I think there is
> some code with which this was started.
>
> --Hannes
>
>
> On 1/19/19, Tobias Pape <[hidden email]> wrote:
>> Hi Tim,
>>
>>> On 19.01.2019, at 07:13, tim Rowledge <[hidden email]> wrote:
>>>
>>> ... what exactly? There's no class comment to help out.
>>>
>>> So far as I can work out it seems to be partially a proxy for working
>>> out
>>> if the image is running in a web browser (can we do that these days? I
>>> was
>>> under the impression that was no longer doable with current browsers)
>>> and
>>> partially a sort of shim for some network-file-fetching-putting that
>>> would
>>> be different if actually running in a browser.
>>>
>>> If we can't do the browser thing we probably ought to clean out the
>>> related code.
>>
>> Sorry for the nitpick, but do you mean HTTPClient or HttpClient?
>> I see only the former in my trunk image but remember that there's been
>> some
>> package somewhere that has the latter…
>> Best regards
>> -Tobias
>>
>>>
>>> tim
>>> --
>>> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
>>> "bOtHeR" said Pooh, mistaking the LSD tablet for aspirin
>>>
>>>
>>>
>>
>>
>>
>

Hannes Hirzel

Re: HttpClient is intended to be used for...

http://wiki.squeak.org/squeak/2275
says that since 2015 WebClient should be used instead of HttpClient.

What is the background information for this?

On 1/21/19, H. Hirzel <[hidden email]> wrote:

> WebClient/ZnClient wrapper
> http://wiki.squeak.org/squeak/3863
>
> On 1/21/19, H. Hirzel <[hidden email]> wrote:
>> A good question:
>>
>> Versions 3.2, 3.6, 3.8, 5.0, 5.1 and 5.2 all have HTTPClient.
>> e.g http://wiki.squeak.org/squeak/6569
>>
>>
>> http://wiki.squeak.org/squeak/3747 is a good place to put a summary of
>> this question and discussion ....
>>
>> There should also be discussion about a wrapper object so that Zn
>> (Pharo classes) based code may be executed in Squeak. I think there is
>> some code with which this was started.
>>
>> --Hannes
>>
>>
>> On 1/19/19, Tobias Pape <[hidden email]> wrote:
>>> Hi Tim,
>>>
>>>> On 19.01.2019, at 07:13, tim Rowledge <[hidden email]> wrote:
>>>>
>>>> ... what exactly? There's no class comment to help out.
>>>>
>>>> So far as I can work out it seems to be partially a proxy for working
>>>> out
>>>> if the image is running in a web browser (can we do that these days? I
>>>> was
>>>> under the impression that was no longer doable with current browsers)
>>>> and
>>>> partially a sort of shim for some network-file-fetching-putting that
>>>> would
>>>> be different if actually running in a browser.
>>>>
>>>> If we can't do the browser thing we probably ought to clean out the
>>>> related code.
>>>
>>> Sorry for the nitpick, but do you mean HTTPClient or HttpClient?
>>> I see only the former in my trunk image but remember that there's been
>>> some
>>> package somewhere that has the latter…
>>> Best regards
>>> -Tobias
>>>
>>>>
>>>> tim
>>>> --
>>>> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
>>>> "bOtHeR" said Pooh, mistaking the LSD tablet for aspirin
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>
>

timrowledge

Re: HttpClient is intended to be used for...

> On 2019-01-21, at 1:25 AM, H. Hirzel <[hidden email]> wrote:
>
> What is the background information for this?

HTTPClient appears to be essentially dead, which means I'd like to remove it and avoid the cognitive load of having pointless stuff in the image that can trap the unwary.

So far as I can tell it was a wrapper for sockets that might be actual sockets of browser-plugin-pseudo-socket-things; since we can't run in a browser any more it has no real purpose. Unless I'm wrong about the browser thing, we could clean out some cruft and simplify our socket story a bit.

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Useful random insult:- One flower short of an arrangement.

timrowledge

Re: HttpClient is intended to be used for...

> On 2019-01-30, at 2:42 PM, tim Rowledge <[hidden email]> wrote:
>
>
>
>> On 2019-01-21, at 1:25 AM, H. Hirzel <[hidden email]> wrote:
>>
>> What is the background information for this?
>
> HTTPClient appears to be essentially dead, which means I'd like to remove it and avoid the cognitive load of having pointless stuff in the image that can trap the unwary.
>
> So far as I can tell it was a wrapper for sockets that might be actual sockets of browser-plugin-pseudo-socket-things; since we can't run in a browser any more it has no real purpose. Unless I'm wrong about the browser thing, we could clean out some cruft and simplify our socket story a bit.

I can find some connection to checking for browser plugin stuff in just three VM files;
sqMacNSPluginUILogic.c
sqUnixMozilla.c
sqWin32PluginSupport.c
Are any of these current, usable, useful?

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Klingon Code Warrior:- 4) "This machine is a piece of GAGH! I need dual G5 processors if I am to do battle with this code!"

Levente Uzonyi

Re: HttpClient is intended to be used for...

In reply to this post by timrowledge

On Wed, 30 Jan 2019, tim Rowledge wrote:

>
>
>> On 2019-01-21, at 1:25 AM, H. Hirzel <[hidden email]> wrote:
>>
>> What is the background information for this?
>
> HTTPClient appears to be essentially dead, which means I'd like to remove it and avoid the cognitive load of having pointless stuff in the image that can trap the unwary.

+1

Levente

>
> So far as I can tell it was a wrapper for sockets that might be actual sockets of browser-plugin-pseudo-socket-things; since we can't run in a browser any more it has no real purpose. Unless I'm wrong about the browser thing, we could clean out some cruft and simplify our socket story a bit.
>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
> Useful random insult:- One flower short of an arrangement.

timrowledge

Re: HttpClient is intended to be used for...

I have a first pass at doing this that I can share best as a changeset fileout, I think. It touches quite a few packages and before going any further it needs some checking by someone with a decent knowledge of the remote directory/server stuff. It hasn't broken the update loading, so at least some of it is correct ;-)

Let me know if you could be that reviewer.

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Klingon Code Warrior:- 9) "A TRUE Klingon warrior does not comment his code!"

David T. Lewis

Re: HttpClient is intended to be used for...

On Wed, Jan 30, 2019 at 06:02:22PM -0800, tim Rowledge wrote:
> I have a first pass at doing this that I can share best as a changeset fileout, I think. It touches quite a few packages and before going any further it needs some checking by someone with a decent knowledge of the remote directory/server stuff. It hasn't broken the update loading, so at least some of it is correct ;-)
>
> Let me know if you could be that reviewer.
>

I am not competent to be the reviewer, but please do post the change set,
I would like to see it.

Thankfully there is nothing in the Squeak development process that explicitly
requires actual competence :-)

Dave

timrowledge

Re: HttpClient is intended to be used for...

> On 2019-01-30, at 6:38 PM, David T. Lewis <[hidden email]> wrote:
>
> On Wed, Jan 30, 2019 at 06:02:22PM -0800, tim Rowledge wrote:
>> I have a first pass at doing this that I can share best as a changeset fileout, I think. It touches quite a few packages and before going any further it needs some checking by someone with a decent knowledge of the remote directory/server stuff. It hasn't broken the update loading, so at least some of it is correct ;-)
>>
>> Let me know if you could be that reviewer.
>>
>
> I am not competent to be the reviewer, but please do post the change set,
> I would like to see it.

Fair enough -

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
"Virtual Memory" means never knowing where your next byte is coming from.

RemoveHTTPClient.1.cs (20K) Download Attachment

jrm

Re: HttpClient is intended to be used for...

In reply to this post by timrowledge

As one of the unwary who has baked HTTPClient into my application, I went to the Squeak Wiki to see what I could find about deprecation and found the documentation about Deprecation and deprecated. The deprecated entry had three instances of self reference and no link to Deprecation, so I made some changes to make it more helpful.

- jrm

On Wed, Jan 30, 2019 at 3:42 PM tim Rowledge <[hidden email]> wrote:

> On 2019-01-21, at 1:25 AM, H. Hirzel <[hidden email]> wrote:
>
> What is the background information for this?

HTTPClient appears to be essentially dead, which means I'd like to remove it and avoid the cognitive load of having pointless stuff in the image that can trap the unwary.

So far as I can tell it was a wrapper for sockets that might be actual sockets of browser-plugin-pseudo-socket-things; since we can't run in a browser any more it has no real purpose. Unless I'm wrong about the browser thing, we could clean out some cruft and simplify our socket story a bit.

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Useful random insult:- One flower short of an arrangement.

Levente Uzonyi

Re: HttpClient is intended to be used for...

In reply to this post by timrowledge

Hi Tim,

I went through the change set and took the following notes:

After loading the change set, CodeLoader and CachingCodeLoader are
unreferenced, so those can be removed (deprecated) too.

In the first comment of Project >> #tellAFriend:, replace tellAFrien with
tellAFriend. There is another comment with "fallback for dtp servers"
which should probably be "fallback for ftp servers".

In URI class >> #initialize you wrote "at: 'http' put: #HTTPClient; <_ not
sure what to replace this with, if anything. Is URI even wnated anymore?".
Based on the users of URI, no. We also have Url which seems to be more
widely used, so I think we should get rid of URI as well.

Levente

On Fri, 1 Feb 2019, tim Rowledge wrote:

>
>
>> On 2019-01-30, at 6:38 PM, David T. Lewis <[hidden email]> wrote:
>>
>> On Wed, Jan 30, 2019 at 06:02:22PM -0800, tim Rowledge wrote:
>>> I have a first pass at doing this that I can share best as a changeset fileout, I think. It touches quite a few packages and before going any further it needs some checking by someone with a decent knowledge of the remote directory/server stuff. It hasn't broken the update loading, so at least some of it is correct ;-)
>>>
>>> Let me know if you could be that reviewer.
>>>
>>
>> I am not competent to be the reviewer, but please do post the change set,
>> I would like to see it.
>
> Fair enough -
>
>

timrowledge

Re: HttpClient is intended to be used for...

> On 2019-02-03, at 7:26 AM, Levente Uzonyi <[hidden email]> wrote:
>
> Hi Tim,
>
> I went through the change set and took the following notes:

Thanks; I feel good that only tpyos (All Hail the Doggess Tpyo!) stood out.

Deprecating still more is good. We have so much left-behind cruft scattered about..

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Useful random insult:- Calls people to ask them their phone number.

timrowledge

Re: HttpClient is intended to be used for...

In reply to this post by Levente Uzonyi

A Further thought -

> On 2019-02-03, at 7:26 AM, Levente Uzonyi <[hidden email]> wrote:
>
> I went through the change set and took the following notes:
>
> After loading the change set, CodeLoader and CachingCodeLoader are unreferenced, so those can be removed (deprecated) too.

Actually in my image it seems that CodeLoader is used in ProjectLauncher>>#startUpAfterLogin. Is it different in your image or was that simply missed?

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Strange OpCodes: EFB: Emulate Five-volt Battery mode

Levente Uzonyi

Re: HttpClient is intended to be used for...

On Mon, 4 Feb 2019, tim Rowledge wrote:

> A Further thought -
>
>> On 2019-02-03, at 7:26 AM, Levente Uzonyi <[hidden email]> wrote:
>>
>> I went through the change set and took the following notes:
>>
>> After loading the change set, CodeLoader and CachingCodeLoader are unreferenced, so those can be removed (deprecated) too.
>
> Actually in my image it seems that CodeLoader is used in ProjectLauncher>>#startUpAfterLogin. Is it different in your image or was that simply missed?

Right. My fault, I didn't realize that the method references CodeLoader
twice. The other class, CachingCodeLoader still seems to be unreferenced.

Levente

>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
> Strange OpCodes: EFB: Emulate Five-volt Battery mode