Hi Marcel,


I am sending pictures and sounds to a server that only accepts mutually different items. However, if multiple of these items are generated using the same code, e.g. Morph new imageForm, there is no chance to identify this duplicate at the moment.

<http://www.hpi.de/>

Thanks for the hints, this sounds indeed too complex for my use case. For now, I decided to simply catch and ignore the server errors about the duplicates. It's easier to ask for forgiveness than permission. :-)

Best,
Christoph
________________________________
Von: Squeak-dev <squeak-dev-bounces at lists.squeakfoundation.org> im Auftrag von Taeumel, Marcel
Gesendet: Donnerstag, 1. Oktober 2020 14:43:38
An: squeak-dev
Betreff: Re: [squeak-dev] Overriding #= in Graphics and Sounds

Hi Christoph.

Hmmm... maybe it is not straightforward to implement? Especially whether #class (or #species) should match or not. Reminds me of SequenceableCollection >> #= and #hasEqualElements:. I keep on using #hasEqualElements: in tests because I don't care about Array-vs-OrderedCollection in several places.

[cid:aa12d832-9f5e-4b25-ba80-37b31d0c2dba]

And Form has also a non-trivial hierarchy of variations:

[cid:3ca1fe29-c01e-4821-91ee-d30a841a9712]

For example, would you ignore the #depth if it is black and white anyway?
Or would you break uses of Dictionaries that are virtually IdentityDictionaries because of their current keys and the missing #=?

It can be hard. Maybe it needs to be maintined afterwards. So, YAGNI? :-)

Best,
Marcel


Am 01.10.2020 14:23:37 schrieb Thiede, Christoph <christoph.thiede at student.hpi.uni-potsdam.de>:

Hi all,


for a current project, I need to compare form or sound objects with each other, but I found out that they do not override #=, so for example, we get:


(ColorForm extent: 1 @ 1) = (ColorForm extent: 1 @ 1) --> false


Wasn't this implemented

a) simply because it looked like YAGNI?

b) for design reasons? For example, should (Form extent: 1 @ 1) be equal to (ColorForm extent: 1 @ 1) or not?

c) for performance reasons? I don't know if there any guidelines for implement #= (except that you need to implement #hash as well), but is an equality comparison generally allowed to be an expensive operation if two objects are equal? If two large forms have a different resolution or a different bitmap hash, comparison will be fast, but if they are equal, we will need to compare every single bit. Would this be okay or a no-go?


If implementing #= as proposed is not possible, how would you think about implementing it as a keyword message on Form, e.g. #sameAs:?


The same questions apply to the AbstractSound hierarchy, too, where I'm not sure whether two sounds being played for different duration should equal or not.

For a similar discussion, see also this thread: [squeak-dev] FormInspector, or also: Text>>#= and its consequences<http://forum.world.st/FormInspector-or-also-Text-gt-gt-and-its-consequences-td5121599.html>.

Best,
Christoph
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20201001/673dd3a5/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 98235 bytes
Desc: image.png
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20201001/673dd3a5/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 81778 bytes
Desc: image.png
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20201001/673dd3a5/attachment-0003.png>