Hi Chris --

> But until we do that, and whole hog like Eliot suggested,
> what we will have are *some* domain things that String
> can do that Text can't -- a partial overlap.

It's rather easy. Once we have a CharacterCollection, we can
finally see the special cases on String. The common stuff can be
moved up to then benefit both String and Text.

> In other words, an incomplete mess for an indefinite period of time.

Disagree. This kind of refactoring does not look too difficult. One
would not rename "String" to "CharacterCollection" but add the
latter between String/Text and ArrayedCollection. Then it can
remain backwards compatible 100%. Piece by piece, we can then
start to remove the duplication and carve out the special cases
for String and Text.

> By removing all the domain stuff [...]

I think that we have a different understand of the term "domain"
here. Maybe you are worried about Magma. If so, please 
elaborate your concerns from that perspective.

> #format: was introduced to Text in 2019.

And long overdue since at least 2015. ;-P Thanks again,
Christoph (ct) for adding it! It made GUI programming much
easier. I had that one in mind for many years now.

> I don't think updates to Text will or should occur except
> when driven by specific need.

The need is there. Code duplication is a problem for
maintenance and our small community. We should work
on that. Let's also learn from VisualWorks in this regard.

Best,
Marcel
Am 14.07.2022 02:29:01 schrieb Chris Muller <asqueaker at gmail.com>:
Hi all,

There are over 300 methods on String, and 100 on Text.  Having Text be responsible for domain duty implies to blow up Text's protocol to around 4X of what it is now with copies from String or, worse, unique implementations requiring separate maintenance.  The CharacterCollection idea at least makes it feel a bit less crazy.  But until we do that, and whole hog like Eliot suggested, what we will have are *some* domain things that String can do that Text can't -- a partial overlap.  In other words, an incomplete mess for an indefinite period of time.

The only other way besides introducing the common CharacterCollection superclass would be by *emphasizing* the separation of responsibility with Text by removing most domain methods from Text, but providing a path to the underlying string for the domain accessing behaviors.  For example, both Text and String currently implement #string, so that's safe to send to either.  By removing all the domain stuff, client code would be encouraged to send #string and go through that for those operations.

But Jakob's example of a piece of code wanting to #format: Text with another Text (with embedded attributes) shows a case actually needing that, so we're in a bit of a pickle.


#format: was introduced to Text in 2019.  It's not being used with alphanumeric tokens, and that is just one of dozens of differences in capability between the two (due to the aforementioned incompleteness).  Until a final decision on how to reconcile the two is made, I don't think updates to Text will or should occur except when driven by specific need.

Best,
  Chris

On Wed, Jul 13, 2022 at 10:45 AM Thiede, Christoph <Christoph.Thiede at student.hpi.uni-potsdam.de [mailto:Christoph.Thiede at student.hpi.uni-potsdam.de]> wrote:

Hi all,

please keep in mind that Text >> #format: does more than the String equivalent as it also deals with attributes from both the receiver and the arguments. I think we should definitely support that.

Best,
Christoph
Von: Squeak-dev <squeak-dev-bounces at lists.squeakfoundation.org [mailto:squeak-dev-bounces at lists.squeakfoundation.org]> im Auftrag von Eliot Miranda <eliot.miranda at gmail.com [mailto:eliot.miranda at gmail.com]>
Gesendet: Mittwoch, 13. Juli 2022 16:50:50
An: The general-purpose Squeak developers list
Betreff: Re: [squeak-dev] String & Text
 


> On Jul 13, 2022, at 2:14 AM, Rein, Patrick <Patrick.Rein at hpi.de [mailto:Patrick.Rein at hpi.de]> wrote:
>
> To me String and Text should share all protocols related to dealing with a collection of characters. Text is a collection of characters with attached formatting information, thus Text "only" has to make sure that the attributes match whatever was done to the character collection. I regularly get slightly annoyed when I try using a String selector with a Text object and get a DNU.
>
> Some time ago, someone proposed to have an abstract CharacterCollection as the superclass of both classes. I very much like that idea, as it would allow us to make the shared protocol explicit. We might not be able to factor much out to this shared superclass but at least, we can have a clear expectation of what should work for both. This would also signify (somewhat) that the behavior should be consistent between the two. In practice I would probably expect Text behavior to be a superset of the String behavior,

+1

> but I am not really sure about that.

The VisualWorks team back before Cincom (ie the team that started ParcPlace) were very sure on this point.

>
> Just my humble oppinion though. :)
>
> Cheers,
> Patrick
>
>
> ________________________________________
> From: Squeak-dev <squeak-dev-bounces at lists.squeakfoundation.org [mailto:squeak-dev-bounces at lists.squeakfoundation.org]> on behalf of Jakob Reschke <jakres+squeak at gmail.com [mailto:jakres%2Bsqueak at gmail.com]>
> Sent: Wednesday, July 13, 2022 10:12:29 AM
> To: The general-purpose Squeak developers list
> Subject: Re: [squeak-dev] String & Text
>
> A common idiom is 'string literal' translated format: {...}
>
> While textual placeholders may make the correct placement of the placeholders in the translated phrase easier, there is also the danger that an inexperienced translator will translate the placeholder's identifier. That would obviously break the intended formatting.
>
> This is independent of the question whether Text and String should both support textual placeholders or not.
>
>
> Am Mi., 13. Juli 2022 um 10:09 Uhr schrieb Jakob Reschke <jakres+squeak at gmail.com [mailto:jakres%2Bsqueak at gmail.com]<mailto:jakres%2Bsqueak at gmail.com [mailto:jakres%252Bsqueak at gmail.com]>>:
> Hi,
>
> Maybe it would be better if they did not share so much of the protocol. So that the system guided you better when to use String and when to use Text, separating the responsibilities more.
>
> For example, I would argue that the asText send is misplaced in this example from AbstractFont>>#browseGlyphsByCategoryOf:select:label:
>
> separatorBlock := [:codePoints :category |
>    (('\{1}\\' withCRs asText
>        format: { Unicode generalCategoryLabels at: category+1 ifAbsent: ['n/a'] })
>        addAttribute: (TextFontReference toFont: TextStyle defaultFixedFont);
>        addAttribute: (PluggableTextAttribute evalBlock: [self browseGlyphsByCategoryOf: codePoints select: aBlock label: aLabelOrNil]);
>        yourself) ].
>
> asText could just as well come after the format:.
>
> Later in the same method comes an example, which is not criticized so easily, where pieces with attributes already applied are inserted via format:
>
> contents := (('Family name: {1}{6}\   Emphasis: {2}\ Point size: {3} ({4}ppi {5}px{7})\' withCRs asText format: { self familyName asText addAttribute: (PluggableTextAttribute evalBlock: [self explore]); yourself. [self emphasisString] on: Error do: [self subfamilyName]. self pointSize. self pixelsPerInch. self height. isRange ifTrue: [''] ifFalse: [' (selected code points)']. (self isTTCFont and: [(tmp := self extraGlyphScale) ~= 1]) ifFalse: [''] ifTrue: [' ', (tmp * 100) rounded asString, '%'] }) addAttribute: (TextFontReference toFont: TextStyle defaultFixedFont); yourself).
>
> Many of the occurrences of Text>>format: that I saw are used to insert links in the text template, or highlighted pieces.
>
> So I guess my opening statement is just dreaming without much practical expertise in the concrete matter. Since Text and String are already interchangeable in certain aspects, you cannot simply roll back without breaking lots of things. Under these circumstances I agree that they should behave equivalently for the protocol that they do share. Otherwise an existing pair of '...' and format: may suddenly fail if you put asText between the two later.
>
> +1 for adding symbolic placeholders to Text >> #format:, unless format: gets removed from Text or the feature gets removed from String again. ;-)
>
> Kind regards,
> Jakob
>
>
> Am Mi., 13. Juli 2022 um 08:48 Uhr schrieb Marcel Taeumel <marcel.taeumel at hpi.de [mailto:marcel.taeumel at hpi.de]<mailto:marcel.taeumel at hpi.de [mailto:marcel.taeumel at hpi.de]>>:
> Hi all --
>
> What are your thoughts on String and Text. In GUI programming, it
> is rather annoying to have to sprinkle #asText all over the code. It's
> nice to have most important protocol shared between String and Text.
>
> The recent change to only String >> #format: (see Collections-cmm.1016)
> to support symbols as placeholders indicates kind of a disagreement
> in how Text should be used in (GUI) code. Well, if some message is
> not there in Text, it is easy to find out. However, having a protocol there
> with different details feels rather challenging.
>
> I accept that not all protocols are shared between String and Text.
> I do not like inconsistencies of method implementations where the
> message (and signature) is identical.
>
> +1 for adding symbolic placeholders to Text >> #format: as well.
>
> Best,
> Marcel
>
>


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20220714/f568bc45/attachment.html>