then again it is my intention to have things like the ReSharper/Rider plugin help out a lot in sayin

then again it is my intention to have things like the ReSharper/Rider plugin help out a lot in saying "you're using unsafe, wanna use the better, more concise way?"

Kai•8/4/23, 9:12 PM

BCL has mostly decided that Span<byte> is a strict improvement to byte[] and done that everywhere.
I might be wrong / outdated here but I don't think new APIs are introduced with

byte[]

byte[]

where

Span

Span

ROSpan

ROSpan

make more sense

Kai•8/4/23, 9:12 PM

(sometimes maybe for equality with old legacy APIs)

PerkseyOP•8/4/23, 9:14 PM

yes 100%

PerkseyOP•8/4/23, 9:14 PM

AND GUESS WHAT

PPerksey then again it is my intention to have things like the ReSharper/Rider plugin hel...

Lyris the Kitori•8/4/23, 9:14 PM

in an ideal world, rider/resharper would autocomplete parameters for all the types that are implicitly convertible to

Ptr<T>

Ptr<T>

, but i dont think its smart enough to filter by that

PerkseyOP•8/4/23, 9:14 PM

ARRAY IMPLICITLY CASTS

PerkseyOP•8/4/23, 9:14 PM

JUST LIKE HERE

PerkseyOP•8/4/23, 9:14 PM

HUZZAH

Maxine•8/4/23, 9:14 PM

okay, i reckon if you understand unsafe then it is an easier bridge from byte* to Ptr<byte> than it is for someone who's a beginner and uses byte[]

LLyris the Kitori in an ideal world, rider/resharper would autocomplete parameters for all the typ...

PerkseyOP•8/4/23, 9:14 PM

you can do that in a ReSharper plugin though

PerkseyOP•8/4/23, 9:14 PM

i've seen plugins do that

PerkseyOP•8/4/23, 9:14 PM

i.e. we could put

thing.AsPtr2D

thing.AsPtr2D

at the top of the list

MMaxine okay, i reckon if you understand unsafe then it is an easier bridge from byte* t...

Maxine•8/4/23, 9:15 PM

so providing unsafe overloads seems excessive

Lyris the Kitori•8/4/23, 9:15 PM

thats kindof why i prefer the name

PtrLike<T>

PtrLike<T>

or maybe

PtrRepresentable<T>

PtrRepresentable<T>

to be even more verbose

PerkseyOP•8/4/23, 9:16 PM

yeah the name originally was PointerLike

PerkseyOP•8/4/23, 9:16 PM

inspired by numpy actually

Lyris the Kitori•8/4/23, 9:16 PM

PtrRepresentable<T>

PtrRepresentable<T>

would indicate that whatever you pass in can be represented by a pointer, but its kindof a mouthful to type?

Maxine•8/4/23, 9:16 PM

that would kill people to type

PerkseyOP•8/4/23, 9:16 PM

it's also physically difficult to type

Lyris the Kitori•8/4/23, 9:16 PM

PtrBacked<T>

PtrBacked<T>

PerkseyOP•8/4/23, 9:16 PM

Lyris the Kitori•8/4/23, 9:16 PM

a pointer is backed by a pointer, an array is on a surface level backed by a pointer

Maxine•8/4/23, 9:17 PM

i feel no matter what, a name itself isn't gonna give you enough information to grasp it, but i'm not saying this from a position of enlightenment

Lyris the Kitori•8/4/23, 9:17 PM

PtrBacked<T>

PtrBacked<T>

would probably let the "knows just enough to be dangerous" crowd intutiively realize arrays work here

PerkseyOP•8/4/23, 9:17 PM

100% agree, and if that user doesn't have the due diligence to read one documentation page then i don't want to support that user

Lyris the Kitori•8/4/23, 9:17 PM

but we would still need ample docs

LLyris the Kitori `PtrBacked<T>` would probably let the "knows just enough to be dangerous" crowd ...

PerkseyOP•8/4/23, 9:17 PM

true

Maxine•8/4/23, 9:18 PM

like this is all working under the assumption that someone can just pick up and use silk without reading pages and pages of docs

Lyris the Kitori•8/4/23, 9:18 PM

i mean thats how i started

Maxine•8/4/23, 9:18 PM

which just isn't true

Lyris the Kitori•8/4/23, 9:18 PM

i didnt read a lick of documentation

Lyris the Kitori•8/4/23, 9:18 PM

and had no unsafe experience

Lyris the Kitori•8/4/23, 9:18 PM

just learned as i went poking about with the overloads

MMaxine so providing unsafe overloads seems excessive

Aqua•8/4/23, 9:18 PM

not providing unsafe overloads causes problems though
it will drive away users like me, so granted not many, but it will mean i will most likely end up switching back to opentk or writing my own which do provide unsafe overloads
my library's internal implementation has no concept of arrays or safe or whatever
the safe overloads are merely a wrapper for unsafe ones, and it uses the advantage that silk is entirely unsafe to work well

AAqua not providing unsafe overloads causes problems though it will drive away users l...

Lyris the Kitori•8/4/23, 9:19 PM

it will be the same API,

T*

T*

implicitly converts to

Ptr<T>

Ptr<T>

Maxine•8/4/23, 9:19 PM

yeah

Maxine•8/4/23, 9:19 PM

what beyley said

Lyris the Kitori•8/4/23, 9:19 PM

it will still be unsafe, just with some different names

PerkseyOP•8/4/23, 9:19 PM

Kai•8/4/23, 9:19 PM

Wondering why. I get the usability question, but there's not really any concerns out of it being the wrong type?

Maxine•8/4/23, 9:19 PM

maybe PtrLike is fine

Lyris the Kitori•8/4/23, 9:20 PM

im not a big fan of PtrLikePtrLike, an array in C# land isnt really "like a pointer" to most people

PerkseyOP•8/4/23, 9:20 PM

it's a good thing we have C&C tomorrow haha

Aqua•8/4/23, 9:20 PM

i'm not so sure implcit conversion is a good idea
personally i think at the point you are relying on implicit conversion, your api should have a redesign
i used to use implicit conversion, realized it was causing all sorts of problems, readability included
and rewrote everything to be explicit

Kai•8/4/23, 9:20 PM

I would like to suggest PtrRpr just for more fun when showing others my code

Lyris the Kitori•8/4/23, 9:20 PM

i think PtrBackedPtrBacked is the clearest name without being hard to type

AAqua i'm not so sure implcit conversion is a good idea personally i think at the poin...

PerkseyOP•8/4/23, 9:20 PM

again team-talk

Volte•8/4/23, 9:20 PM

perksey#0000

it's worth noting what we're discussing here is not Silk.NET 3.0. This is SilkX, a completely different library. there is still room for another contributor to create a different library to propose that it gets blessed as 3.0

Original Message

Click here

Quoted by perksey#0000

PerkseyOP•8/4/23, 9:21 PM

please show us how it's done

AAqua i'm not so sure implcit conversion is a good idea personally i think at the poin...

Kai•8/4/23, 9:21 PM

I get this, but really this is just how C# works. We don't have traits or associated types or anything cool, so we have to make do with what we got

then again it is my intention to have things like the ReSharper/Rider plugin help out a lot in sayin

Similar Threads