workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
|---|---|---|---|---|
elmlang
|
general
|
The key is that the server response/request objects are generated code
|
2019-03-14T23:25:07.152800
|
Rosaria
|
elmlang
|
general
|
So modifying them or assimilating them into a completely flattened model isn't really feasible.
|
2019-03-14T23:25:24.153200
|
Rosaria
|
elmlang
|
general
|
Nor desirable.
|
2019-03-14T23:25:26.153400
|
Rosaria
|
elmlang
|
general
|
ok
|
2019-03-14T23:25:31.153600
|
Hoyt
|
elmlang
|
general
|
They serialize back and forth, and encompass all the types the server understands.
|
2019-03-14T23:25:58.154200
|
Rosaria
|
elmlang
|
general
|
It's a very nice model tbh.
|
2019-03-14T23:26:00.154400
|
Rosaria
|
elmlang
|
general
|
I believe it
|
2019-03-14T23:26:05.154600
|
Hoyt
|
elmlang
|
general
|
But when you go to update these objects inside your model
|
2019-03-14T23:26:16.155000
|
Rosaria
|
elmlang
|
general
|
One of your eyes always twitches a few times
|
2019-03-14T23:26:24.155400
|
Rosaria
|
elmlang
|
general
|
:slightly_smiling_face: your = my :slightly_smiling_face:
|
2019-03-14T23:26:31.155600
|
Hoyt
|
elmlang
|
general
|
Different designs are optimized for different situations
|
2019-03-14T23:27:33.157000
|
Hoyt
|
elmlang
|
general
|
Can I ask how you handle the problem for your own stuff?
|
2019-03-14T23:28:08.157800
|
Rosaria
|
elmlang
|
general
|
You've mentioned abstractions
|
2019-03-14T23:28:15.158100
|
Rosaria
|
elmlang
|
general
|
But that's fairly abstract.
|
2019-03-14T23:28:19.158400
|
Rosaria
|
elmlang
|
general
|
I generally have a module representing working with whatever the nested value represents
|
2019-03-14T23:28:32.158900
|
Hoyt
|
elmlang
|
general
|
If I find I need to update it, then I generally find I need to do other things with it
|
2019-03-14T23:28:47.159400
|
Hoyt
|
elmlang
|
general
|
So, I usually end up with (and, of course, specific context matters)
|
2019-03-14T23:29:00.159800
|
Hoyt
|
elmlang
|
general
|
```
{ model | prop = Abstraction.doSomethingWith model.prop }
```
|
2019-03-14T23:29:20.160300
|
Hoyt
|
elmlang
|
general
|
There is usually a reason I’m updating a specific aspect of the nested structure. I try to map the update to that reason
|
2019-03-14T23:29:44.160900
|
Hoyt
|
elmlang
|
general
|
But, again, very abstract explanation and it always comes down to the specific context of the codebase
|
2019-03-14T23:30:06.161600
|
Hoyt
|
elmlang
|
general
|
Right.
|
2019-03-14T23:30:11.161800
|
Rosaria
|
elmlang
|
general
|
So, there isn’t a “one size fits all”
|
2019-03-14T23:30:18.162100
|
Hoyt
|
elmlang
|
general
|
but I will say that I don’t find myself needing `{ model.prop | attribute = newValue }`
|
2019-03-14T23:30:35.162800
|
Hoyt
|
elmlang
|
general
|
Because my module’s “doSomethingWith” receives the `prop` and updates it
|
2019-03-14T23:30:52.163400
|
Hoyt
|
elmlang
|
general
|
So, I don’t need the nested syntax
|
2019-03-14T23:30:58.163700
|
Hoyt
|
elmlang
|
general
|
I try to build in terms of the behaviours, rather than the data, so I generally find myself with these functions
|
2019-03-14T23:31:35.165000
|
Hoyt
|
elmlang
|
general
|
I’m sure there are impacts of this style in other aspects
|
2019-03-14T23:31:51.165700
|
Hoyt
|
elmlang
|
general
|
that perhaps over-complicate things in other parts of my system. ¯\_(ツ)_/¯ So many trade-offs
|
2019-03-14T23:32:07.166300
|
Hoyt
|
elmlang
|
general
|
Always is in programming.
|
2019-03-14T23:32:24.166600
|
Rosaria
|
elmlang
|
general
|
for sure
|
2019-03-14T23:32:28.166800
|
Hoyt
|
elmlang
|
general
|
Why I always try to say “there’s no best design, only better design for the context”
|
2019-03-14T23:32:44.167400
|
Hoyt
|
elmlang
|
general
|
It's generated code, so we could generate the helpers as well.
|
2019-03-14T23:33:47.168300
|
Rosaria
|
elmlang
|
general
|
It just seems like such a waste
|
2019-03-14T23:33:50.168600
|
Rosaria
|
elmlang
|
general
|
Adding potentially thousands of helper functions, most of which would never be used
|
2019-03-14T23:34:07.169100
|
Rosaria
|
elmlang
|
general
|
There are some utilities that generate the nested helpers, for sure
|
2019-03-14T23:34:11.169300
|
Hoyt
|
elmlang
|
general
|
Add compile times etc.
|
2019-03-14T23:34:14.169600
|
Rosaria
|
elmlang
|
general
|
Luckily we won’t compile in the functions that aren’t used
|
2019-03-14T23:34:24.170000
|
Hoyt
|
elmlang
|
general
|
Does 0.19 do DCE?
|
2019-03-14T23:34:45.170400
|
Rosaria
|
elmlang
|
general
|
My memory is fuzzy on that.
|
2019-03-14T23:34:57.170800
|
Rosaria
|
elmlang
|
general
|
yeah
|
2019-03-14T23:34:59.170900
|
Hoyt
|
elmlang
|
general
|
I see.
|
2019-03-14T23:35:02.171100
|
Rosaria
|
elmlang
|
general
|
(as per my understanding)
|
2019-03-14T23:35:08.171600
|
Hoyt
|
elmlang
|
general
|
yep, function and type level DCE
|
2019-03-14T23:35:35.171900
|
Earlean
|
elmlang
|
general
|
In any case, thanks <@Rosaria> for the chat. I’m a bit drunk and ready for bed. :slightly_smiling_face:
|
2019-03-14T23:36:33.172500
|
Hoyt
|
elmlang
|
general
|
I’m always intrigued by people’s thoughts on these things
|
2019-03-14T23:37:23.173000
|
Hoyt
|
elmlang
|
general
|
I appreciate your explaining where youa re coming from
|
2019-03-14T23:37:31.173400
|
Hoyt
|
elmlang
|
general
|
I'll probably gather a bunch of data and attempt to write a proposal on the Discourse. There's certainly enough data pointing to this as a problem. But unlike the verbosity of JSON decoders this one seems solvable to me.
|
2019-03-14T23:38:01.174100
|
Rosaria
|
elmlang
|
general
|
Also thanks for the counter-example <@Earlean>
|
2019-03-14T23:38:19.174600
|
Rosaria
|
elmlang
|
general
|
awesome. I’ll be interested in reading it.
|
2019-03-14T23:38:29.175000
|
Hoyt
|
elmlang
|
general
|
There is definitely value in “I design this way, because this is the facility I have”
|
2019-03-14T23:38:58.176000
|
Hoyt
|
elmlang
|
general
|
It's been something that has been discussed for a few years now and Evan doesn't seem interested in changing it
|
2019-03-14T23:39:18.176800
|
Earlean
|
elmlang
|
general
|
Or, rather, “definitely something there”
|
2019-03-14T23:39:20.176900
|
Hoyt
|
elmlang
|
general
|
I like hearing people’s thoughts on it
|
2019-03-14T23:39:34.177800
|
Hoyt
|
elmlang
|
general
|
<@Earlean> Ahh, the benevolent dictator. Only so benevolent I suppose.
|
2019-03-14T23:39:43.178300
|
Rosaria
|
elmlang
|
general
|
It often boils down to people’s design philosophies, so I like hearing them.
|
2019-03-14T23:39:52.178700
|
Hoyt
|
elmlang
|
general
|
Whether it is leading to a change or not
|
2019-03-14T23:39:59.179000
|
Hoyt
|
elmlang
|
general
|
`{myRecord.data | field = value}` (where it returns `data`) not working was originally considered a bug in the compiler
|
2019-03-14T23:40:14.179500
|
Earlean
|
elmlang
|
general
|
<@Hoyt> I like these sorts of discussions too, as a kind of language nerd myself.
|
2019-03-14T23:40:21.179700
|
Rosaria
|
elmlang
|
general
|
and was not fixed because it was considered that code doing that was worse code
|
2019-03-14T23:40:44.180200
|
Earlean
|
elmlang
|
general
|
<@Earlean> And how about code that's returning `myRecord` instead? Does that make it better? :smile:
|
2019-03-14T23:41:20.180800
|
Rosaria
|
elmlang
|
general
|
it was generally considered to be a bit of a confusing thing to do `{myrecord | ...}` updates `myrecord`, but `{myRecord.data| }` doesn't update `data`, that's weird...
|
2019-03-14T23:42:58.182500
|
Earlean
|
elmlang
|
general
|
should `{(someFunction someValue) | field = value}` work?
|
2019-03-14T23:43:33.183100
|
Earlean
|
elmlang
|
general
|
if so, what part of the value does the update apply to?
|
2019-03-14T23:43:54.183600
|
Earlean
|
elmlang
|
general
|
do we need some other different syntax for when I want to update `myRecord.data` without first assigning `myRecord.data` to a variable in a `let`?
|
2019-03-14T23:45:04.184800
|
Earlean
|
elmlang
|
general
|
if Elm supports both cases then won't people get them confused
|
2019-03-14T23:45:56.185900
|
Earlean
|
elmlang
|
general
|
just latching on to the last part of the convo, i would love to see a setter syntax just like how we currently have the getter `.` syntax
|
2019-03-14T23:45:56.186000
|
Ruthann
|
elmlang
|
general
|
<@Ruthann> I think everyone would love that :stuck_out_tongue:
|
2019-03-14T23:49:56.188300
|
Rosaria
|
elmlang
|
general
|
```
setData: a -> {b | data: a} -> b
setData a b = {b | data = a}
(setData someData myRecord)
```
|
2019-03-14T23:50:31.188900
|
Earlean
|
elmlang
|
general
|
<@Earlean> But it does update `data`. But immediately it assigns that to `myRecord.data` as it's specified there in the update statement. But it returns `myRecord` because ultimately you have modified that. And if you do not return that, you've lost part of your update.
|
2019-03-14T23:51:14.189900
|
Rosaria
|
elmlang
|
general
|
Also I would think that `{(someFunction someValue) | field = value}` is outside of the scope of this conversation
|
2019-03-14T23:52:26.191100
|
Rosaria
|
elmlang
|
general
|
<@Rosaria> yeah, but why should it do that? I passed the `data` value to it, why should it care about `myRecord`?
|
2019-03-14T23:52:34.191500
|
Earlean
|
elmlang
|
general
|
<@Earlean> You passed `myRecord's` `.data` to it. Therefore it must care.
|
2019-03-14T23:52:55.191900
|
Rosaria
|
elmlang
|
general
|
It depends how you perceive the `{ | }` syntax I suppose.
|
2019-03-14T23:53:26.192700
|
Rosaria
|
elmlang
|
general
|
```
let
data = myRecord.data
in
{data | field = value}
```
^ data is the same value there
|
2019-03-14T23:53:55.193600
|
Earlean
|
elmlang
|
general
|
To me I find it a magical beast that defies the very nature of Elm. It barely has a place in the language. It works with almost no constructs that work everywhere else in Elm.
|
2019-03-14T23:53:55.193700
|
Rosaria
|
elmlang
|
general
|
Right, but at that point you're only passing data
|
2019-03-14T23:54:07.194000
|
Rosaria
|
elmlang
|
general
|
And it would only return data
|
2019-03-14T23:54:11.194200
|
Rosaria
|
elmlang
|
general
|
And if you wanted it to behave that way, that's the code you would write.
|
2019-03-14T23:54:17.194500
|
Rosaria
|
elmlang
|
general
|
I believe that that use case is far less used (though I'd have to get the data to prove that)
|
2019-03-14T23:54:35.195200
|
Rosaria
|
elmlang
|
general
|
what if I wanted the above behaviour? would I have to do the `let..in`?
|
2019-03-14T23:54:45.195700
|
Earlean
|
elmlang
|
general
|
Yep. As you have to do now.
|
2019-03-14T23:54:56.196000
|
Rosaria
|
elmlang
|
general
|
Since no `.` is accepted anywhere in the record update syntax.
|
2019-03-14T23:55:08.196300
|
Rosaria
|
elmlang
|
general
|
so why one behaviour instead of the other?
|
2019-03-14T23:55:21.196600
|
Earlean
|
elmlang
|
general
|
> I believe that that use case is far less used (though I'd have to get the data to prove that)
|
2019-03-14T23:55:39.197000
|
Rosaria
|
elmlang
|
general
|
I'd actually like to see the data
|
2019-03-14T23:55:52.197200
|
Rosaria
|
elmlang
|
general
|
I wonder if it's possible to comb some elm repositories
|
2019-03-14T23:56:02.197500
|
Rosaria
|
elmlang
|
general
|
Looking for these exact patterns
|
2019-03-14T23:56:05.197700
|
Rosaria
|
elmlang
|
general
|
And produce meaningful statistics on it.
|
2019-03-14T23:56:13.197900
|
Rosaria
|
elmlang
|
general
|
I'd additionally say that, it's harder to do the opposite if you want it. Consider the two use cases:
```
-- Update nested data
let data = myRecord.data
newData = { data | thing = 5 }
in { myRecord | data = newData }
-- Update nested data without updating each level of nest
let
data = myRecord.data
in
{data | field = value}
```
|
2019-03-14T23:58:16.199900
|
Rosaria
|
elmlang
|
general
|
This of course becomes even more clear when you have more levels of nesting (though I still discourage deeply nested hierarchies)
|
2019-03-14T23:58:49.200400
|
Rosaria
|
elmlang
|
general
|
```
let
data = myRecord.data
in
{ myRecord | data = { data | thing = 5 } }
```
|
2019-03-15T00:00:08.202000
|
Earlean
|
elmlang
|
general
|
Ah of course. Silly me.
|
2019-03-15T00:00:47.202400
|
Rosaria
|
elmlang
|
general
|
3 levels deep and I'd agree, but at that point you're writing worse code
|
2019-03-15T00:01:23.203200
|
Earlean
|
elmlang
|
general
|
Perhaps. Perhaps not. Sometimes it might not even be something you wrote!
|
2019-03-15T00:02:07.203900
|
Rosaria
|
elmlang
|
general
|
But regardless, those are the reasons that I'd choose one return over the other.
|
2019-03-15T00:02:17.204200
|
Rosaria
|
elmlang
|
general
|
I think one has genuinely more impact and utility.
|
2019-03-15T00:02:25.204500
|
Rosaria
|
elmlang
|
general
|
Whereas the other is less frequently used and easier to deal with when it does arise
|
2019-03-15T00:02:37.204900
|
Rosaria
|
elmlang
|
general
|
And I'm happy to have someone call me out for a lack of data on that
|
2019-03-15T00:02:48.205400
|
Rosaria
|
elmlang
|
general
|
It's just my anecdotes.
|
2019-03-15T00:02:57.205700
|
Rosaria
|
elmlang
|
general
|
It's also one of those things where it's not a lot of code..but if it is then you write a function and use that function
|
2019-03-15T00:02:59.205800
|
Earlean
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.