feat: std re-organization #2425

thehowl · 2024-06-24T14:46:42Z

This is a work-in-progress PR, but it contains some key ideas of the new design

References:

This will be worked into a complete state so that we can see the full extent of the transition to the new system, then it will be broken down into a few PRs for ease of reviewing, similar to what happened with #1695.

Objectives of this effort:

Remove std in favour of chain, runtime and debug/testing.
Remove testing stdlibs.
(This has not been discussed much yet, but the current "dual system" for standard libraries, is that p/ packages that serve as "testing aids" cannot use test-context functions due to the type checking. So I'm unifying the stdlibs, but disabling test-only features on-chain).
standardise import path instead of pkgpath
This is something I wanted to add for a while for consistency with Go.
make sure internal/ rules are respected
Ie. make them work as in Go.
Change "send" into "deposit"

These are all individually up for discussion, but I think they fit nicely into a better std; so here it is.

Contributors' checklist...

Added new tests, or not needed, or not feasible
Provided an example (e.g. screenshot) to aid review or the PR is self-explanatory
Updated the official documentation or not needed
No breaking changes were made, or a BREAKING CHANGE: xxx message was included in the description
Added references to related issues and PRs
Provided any useful hints for running manual tests
Added new benchmarks to generated graphs, if any. More info here.

grepsuzette · 2024-07-12T03:56:40Z

@thehowl A bit late to the party, but it's still a draft. As said in #1475, I advocate renaming std to gno.
There's too many benefits for beginners to not discuss it, /me think.

moul · 2024-07-12T10:48:29Z

gnovm/stdlibs/chain/chain.gno

+// associated import path.
+type Realm struct {
+	address    Address
+	importPath string


What's the problem with PkgPath?

Edit: I think we should keep pkgPath over importPath. Because ultimately, the package path is an identifier which, yes, can be imported, but it's an identifier first, just as a std.Address is an identifier for an account. This one is the identifier for the realm and its assets, and the cool thing is that it's importable. It's like we would say that Go's package path can be imported, so let's not name them ImportPath. But we can also go mod download them, so let's not rename them goModDownloadPath either. I think we should be clear that the identifier is the unique identifier of a package, and then importPath is currently the same 1-1, but maybe in the future, we'll support import aliases for versioning or to support, for instance, r/morgan/foo and r/g1morganxxx/foo to be the same pkgPath but different importPath.

I am for a different name that pkgpath, because it conflicts most of the time when explaining the topics, making it very confusing. possibly URL, URI, ID, just path, etc. in the end its a unique identifier for code, and it can be used to look up code or import code.

realm path - better than realm package path

pure package path - better than pure package package path.

moul · 2024-07-12T10:51:42Z

gnovm/stdlibs/runtime/runtime.gno

+// Package runtime allows access to the GnoVM's runtime information.
+package runtime
+
+func NewReadOnlyBanker() chain.Banker {


I don't understand the benefit of having a distinct runtime package that ultimately implements helpers that will return chain types. Wouldn't it be simpler and less confusing to have a single package that defines the types and implements the helpers?

moul · 2024-07-12T10:58:05Z

I like the refactoring idea and believe that most of the renames make sense. However:

I don't see the need for separate chain and runtime packages instead of a single package.
I prefer keeping std or use os more than using chain and/or runtime or gno.
I really like the new testing and context approaches. <3
My comment about pkgPath vs importPath.

moul · 2024-07-12T13:27:16Z

gnovm/stdlibs/testing/context.gno

+}
+
+func (c Context) Run(f func()) {
+}


func CurrentFileLine() string { return ufmt.Sprintf("%s:%d", prevFrame.Filepath, prevFrame.LineNumber) }

Or better GetFrameAt(x int).FileLine().

could be useful for things like uassert.

thehowl · 2024-07-16T04:07:22Z

gnovm/stdlibs/testing/context.gno

+import (
+	"chain"
+)
+


XXX:

testing calls each test using a new Context{} (so that they're running always on a clean state)

same goes for subtests

a test may still call SetContext to just alter variables

I think we should just run a new gnomachine, it’s not only about context,but about sharing any state (global variable too).

leohhhn · 2024-09-23T10:30:19Z

gnovm/stdlibs/runtime/runtime.gno

+func NewSendBanker() chain.Banker {
+}
+
+func NewIssueBanker() chain.Banker {


Suggested change

func NewIssueBanker() chain.Banker {

func NewIssuingBanker() chain.Banker {

leohhhn · 2024-09-23T10:31:30Z

gnovm/stdlibs/runtime/runtime.gno

+func AssertOriginCall() {
+}
+
+func CallerRealm() chain.Realm {


I am of a hot take that we should completely remove the possibility of a realm being an EOA. It's a very confusing concept, especially when talking about realm package, package path, etc. Then, we could just have Caller, which can be a realm or an EOA/user.

leohhhn · 2024-09-23T10:33:52Z

gnovm/stdlibs/runtime/runtime.gno

+func CurrentRealm() chain.Realm {
+}
+
+func Deposits() chain.Coins {


We need to make a distinction between prevrealm deposits and origin deposits; rather, you only need to read the immediate previous realm deposit; I'm not sure of a usecase where origin deposit could be needed.

Related: #1887

leohhhn · 2024-09-23T10:37:18Z

gnovm/stdlibs/testing/context.gno

+// Context contains execution context variables which are used primarily by the
+// [runtime] package to give information about the current GnoVM execution.
+type Context struct {
+	IsOrigin     bool


what does this bool represent? that the caller is the origin caller?

leohhhn · 2024-09-23T10:42:17Z

gnovm/stdlibs/chain/chain.gno

+
+// NewCodeRealm creates a new realm, representing a code realm with a published
+// import path, existing on-chain.
+func NewCodeRealm(importPath string) Realm {


See comment

feat: re-organize std

ef3b264

thehowl self-assigned this Jun 24, 2024

github-actions bot added the 📦 🤖 gnovm Issues or PRs gnovm related label Jun 24, 2024

This was referenced Jun 24, 2024

[META] improved native bindings #814

Open

Issues with std.TestXYZ functions #1521

Closed

Kouteki mentioned this pull request Jun 26, 2024

Minutes: Core Staff Weekly Syncs [every Monday] gnolang/meetings#36

Open

This was referenced Jun 28, 2024

[GnoVM] Reset machine context & realm state after each Test function #1982

Open

fix(gno.land): pre-load all standard libraries in vm.Initialize #2504

Merged

michelleellen mentioned this pull request Jul 10, 2024

Minutes: Public Dev Calls [every Wednesday] gnolang/meetings#37

Open

thehowl mentioned this pull request Jul 10, 2024

[GnoVM] TestSetRealm not working properly in gno test #2371

Open

moul reviewed Jul 12, 2024

View reviewed changes

thehowl commented Jul 16, 2024

View reviewed changes

thehowl mentioned this pull request Sep 19, 2024

[Docs]: Lack of information about std.Test functions #2817

Open

leohhhn reviewed Sep 23, 2024

View reviewed changes

moul mentioned this pull request Sep 27, 2024

Gnodev to give useful info about failing replays #2867

Open

thehowl mentioned this pull request Oct 28, 2024

refactor: improve the std API #848

Closed

11 tasks

thehowl closed this Oct 31, 2024

thehowl mentioned this pull request Oct 31, 2024

META - splitting and refactoring std #1475

Open

thehowl deleted the dev/morgan/the-big-std-renaming branch December 8, 2024 16:22

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: std re-organization #2425

feat: std re-organization #2425

thehowl commented Jun 24, 2024 •

edited

Loading

grepsuzette commented Jul 12, 2024

moul Jul 12, 2024 •

edited

Loading

leohhhn Sep 23, 2024 •

edited

Loading

moul Jul 12, 2024

moul commented Jul 12, 2024 •

edited

Loading

moul Jul 12, 2024 •

edited

Loading

thehowl Jul 16, 2024

moul Jul 16, 2024

moul Oct 25, 2024

leohhhn Sep 23, 2024

leohhhn Sep 23, 2024 •

edited

Loading

leohhhn Sep 23, 2024 •

edited

Loading

leohhhn Sep 23, 2024

leohhhn Sep 23, 2024

	func NewIssueBanker() chain.Banker {
	func NewIssuingBanker() chain.Banker {

feat: std re-organization #2425

feat: std re-organization #2425

Conversation

thehowl commented Jun 24, 2024 • edited Loading

grepsuzette commented Jul 12, 2024

moul Jul 12, 2024 • edited Loading

Choose a reason for hiding this comment

leohhhn Sep 23, 2024 • edited Loading

Choose a reason for hiding this comment

moul Jul 12, 2024

Choose a reason for hiding this comment

moul commented Jul 12, 2024 • edited Loading

moul Jul 12, 2024 • edited Loading

Choose a reason for hiding this comment

thehowl Jul 16, 2024

Choose a reason for hiding this comment

moul Jul 16, 2024

Choose a reason for hiding this comment

moul Oct 25, 2024

Choose a reason for hiding this comment

leohhhn Sep 23, 2024

Choose a reason for hiding this comment

leohhhn Sep 23, 2024 • edited Loading

Choose a reason for hiding this comment

leohhhn Sep 23, 2024 • edited Loading

Choose a reason for hiding this comment

leohhhn Sep 23, 2024

Choose a reason for hiding this comment

leohhhn Sep 23, 2024

Choose a reason for hiding this comment

thehowl commented Jun 24, 2024 •

edited

Loading

moul Jul 12, 2024 •

edited

Loading

leohhhn Sep 23, 2024 •

edited

Loading

moul commented Jul 12, 2024 •

edited

Loading

moul Jul 12, 2024 •

edited

Loading

leohhhn Sep 23, 2024 •

edited

Loading

leohhhn Sep 23, 2024 •

edited

Loading