61 low level api #77

ddd-mtl · 2018-06-29T13:38:06Z

This branch is getting huge so I'll not finish all three functions here.
I did lay out some work for COMMIT and VALIDATE, but this branch is focused on GENESIS

Genesis is more complex than expected because it is called during the InitApplication Action and generates more Actions to complete. Thus Init is finished asynchronously so I added a NucleusStatus member to properly keep track of that.

Here is what a stack of Actions might look like when calling InitApplication on a app with two Zomes with each one there genesis() that commits something to the source chain.

InitApplication
	ExecuteZomeFunction - Zome0.Genesis
	ExecuteZomeFunction - Zome1.Genesis
	...
		Commit - Zome0
			ValidateEntry
			...
		Commit - Zome1
			ValidateEntry
			...
	...
	ReturnZomeFunctionResult - Zome0.Genesis
	ReturnZomeFunctionResult - Zome1.Genesis
	...
ReturnInitializationResult

I copied some of Instance's methods out of Instance because those methods cannot be called in NucleusState because there is no instance object available.

What's missing in my current implementation is handling error cases where not all of the Zomes have done their genesis() for some reason and we would want a second call to InitApplication to complete the Init and do genesis() only on the remaining Zomes.

…evel-api

Started implementing commit() low level API

Moved dispatch_action_and_wait() to instance.rs genesis WIP in nucleus InitApplication reduce()

Calling Genesis loop in init WIP Added UML-class.png Added more Error types Refactoring ExecuteZome for better Error returns

Fixed tests

Handled genesis return value Added genesis tests Refactored testing in core/src/lib.rs

Conflicts: core/src/lib.rs core/src/nucleus/mod.rs core/src/nucleus/ribosome.rs

Using cargo.lock from develop

thedavidmeister · 2018-06-30T08:50:27Z

Cargo.lock.bak

@@ -1,447 +0,0 @@
-[[package]]


don't forget to clean this up before merging

I understand that @lucksus added this file in develop 3 days ago, probably by mistake. I'm erasing it here. So I'm not sure what you are asking. Should I leave it?

oh! yes i see, good job then 👍

Oh! Ooops.. sorry. Yes, this needs deleted..

thedavidmeister · 2018-06-30T08:51:20Z

core/src/lib.rs

+    use std::fs::File;
+
+    pub fn create_wasm_from_file(fname: &str) -> Vec<u8> {
+        use std::io::prelude::*;


is this normal, sticking a use inside a fn?

thedavidmeister · 2018-06-30T08:51:36Z

core/src/lib.rs

-              r#"
+    use std::fs::File;
+
+    pub fn create_wasm_from_file(fname: &str) -> Vec<u8> {


can you put some doc comments /// on each fn?

thedavidmeister · 2018-06-30T08:52:55Z

core/src/lib.rs

+        use std::io::prelude::*;
+        let mut file = File::open(fname).unwrap();
+        let mut buf = Vec::new();
+        file.read_to_end(&mut buf).unwrap();


unwrap() panics, right? i think we're trying to avoid that for recoverable errors (e.g. the file might not exist)

Not in tests. We want tests to panic.
And anyway those unwraps are already in develop. They come from PR#44.

thedavidmeister · 2018-06-30T08:55:52Z

core/src/lib.rs

-          .unwrap();
+            nucleus::ribosome::RESULT_OFFSET
+        );
+        let wat_str = match wat {


i think unwrap_or could help here https://doc.rust-lang.org/std/option/enum.Option.html#method.unwrap_or

thedavidmeister · 2018-06-30T08:56:33Z

core/src/lib.rs

+            .canonicalize_lebs(false)
+            .write_debug_names(true)
+            .convert(wat_str)
+            .unwrap();


unwrap could be bad because of panic

thedavidmeister · 2018-06-30T08:56:55Z

core/src/lib.rs

+    }
+
+
+    // Prepare valid DNA struct with that WASM in a zome's capability


doc comment is triple slash ///

thedavidmeister · 2018-06-30T08:57:51Z

core/src/lib.rs

    }


+    // Prepare valid DNA struct helper


what's this? still needed?

Moved WAT for testing from test_utils to tests Added FromStr trait to Reserved Enums as suggested by clippy

sphinxc0re

Good job! The nucleus section is pretty huge but I guess we'll have to stick with it for now

sphinxc0re · 2018-07-02T06:54:02Z

core/src/lib.rs

        capability.code = DnaWasm { code: wasm };
-        zome.name = "test_zome".to_string();
+        zome.name = zome_name.to_string();


the to_string() call here is redundant.

sphinxc0re · 2018-07-02T06:55:07Z

core/src/lib.rs

            }
-            Err(_) => assert!(false),
+            _ => { assert!(false) },


I don't think you need the { }

sphinxc0re · 2018-07-02T06:56:48Z

core/src/nucleus/mod.rs

+        self.status.clone()
+    }
+    pub fn has_initialized(&self) -> bool {
+        self.status == NucleusStatus::Initialized
    }


There should be empty lines between the function definitions.

sphinxc0re · 2018-07-02T06:59:48Z

core/src/nucleus/mod.rs

+impl EntrySubmission {
+    pub fn new<S>(zome_name: S, type_name: S, content: S) -> Self
+        where
+          S: Into<String>,


This is good but a little too verbose. What about:

pub fn new<S: Into<String>>(zome_name: S, type_name: S, content: S) -> Self { ...

Improved some comments Trying --warn clippy

sphinxc0re · 2018-07-02T15:32:47Z

Makefile

@@ -27,7 +27,7 @@ C_BINDING_CLEAN = $(foreach dir,$(C_BINDING_DIRS),$(dir)Makefile $(dir).qmake.st
 # apply formatting / style guidelines, and build the rust project
 main:
 	cargo fmt
-	cargo clippy -- --deny clippy
+	cargo clippy -- --warn clippy


Why would you do that? We want the process to fail if it doesn't comply

I want to be able to put returns when it makes the code more understandable and clippy is preventing me from doing that, so I'm trying to make clippy less annoying. To my knowledge, there is no agreement to what the code is suppose to comply to.

zippy

only one tiny change (other than the others requested) about linking the println to #21.

All in all quite a bunch of great code using the redux infrastructure. Thanks!

zippy · 2018-07-02T16:16:53Z

core/src/lib.rs

-        assert_eq!(instance.state().nucleus().initialized(), true);
+        // Wait for Init to finish
+        while instance.state().history.len() < 2 {
+            println!("Waiting... {}", instance.state().history.len());


This println! should be converted to either a call to the app logger, or to the core debug log. For now we can just add a TODO that references #21.

zippy · 2018-07-02T16:17:14Z

core/src/lib.rs

+
+        // Wait for Init to finish
+        while instance.state().history.len() < 4 {
+            println!("Waiting... {}", instance.state().history.len());


zippy · 2018-07-02T17:04:42Z

core/src/nucleus/mod.rs

+    if has_succeeded {
+        (*nucleus_state).status = NucleusStatus::Initialized
+    } else {
+        *nucleus_state = NucleusState::new();


I'm not sure I totally understand what you were thinking here, but it seems like if the initialization didn't succeed we can't just reset the state, to uninitialized, because a client of the state object actually might want to know that the initialization failed and distinguish between these states.

My thought was that if init fails, then undo everything that was done in the InitApplication Action reducer, which basically means resetting the state. In the absence of clarity for the whole App init procedure (#86) it seemed like the safest thing to do for now. I agree about needing some callback to notify that the init failed. I am not satisfied with the wait loop on history counts.

Made clippy less annoying.

codecov-io · 2018-07-02T20:54:18Z

Codecov Report

Merging #77 into develop will decrease coverage by 2.9%.
The diff coverage is 54.05%.

@@             Coverage Diff             @@
##           develop      #77      +/-   ##
===========================================
- Coverage    82.26%   79.35%   -2.91%     
===========================================
  Files           13       13              
  Lines          451      499      +48     
===========================================
+ Hits           371      396      +25     
- Misses          80      103      +23

Impacted Files	Coverage Δ
core/src/error.rs	`56.75% <0%> (-6.88%)`	⬇️
core/src/state.rs	`85.71% <100%> (+0.52%)`	⬆️
core/src/persister.rs	`95.45% <100%> (+0.21%)`	⬆️
dna/src/zome/capabilities.rs	`65.11% <33.33%> (-22.89%)`	⬇️
core/src/nucleus/ribosome.rs	`69.35% <40%> (-7.92%)`	⬇️
core/src/instance.rs	`81.94% <79.31%> (+4.67%)`	⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d0d33b2...06657df. Read the comment docs.

Conflicts: core/src/lib.rs core/src/nucleus/mod.rs core/src/nucleus/ribosome.rs core/src/state.rs core_api/src/lib.rs dna/src/lib.rs

zippy · 2018-07-03T01:14:56Z

I do agree with @sphinxc0re that we will have to refactor reduce() in nucleus/mod.rs but also that it's fine for now. (whoops sorry, clicked on wrong button and closed the PR for a sec)

ddd-mtl · 2018-07-03T01:22:59Z

Yes, we can easily pull out functions for each Action reducer like I did for the ReturnInitializationResult reducer. Definitely the next step here.

ddd-mtl added 8 commits June 26, 2018 10:30

index on develop: 24b8cde Merge pull request #36 from holochain/19

bccb1cd

On develop: stuff for #61

acd1580

Merge commit 'acd1580b2bfbdef7c170e17c6c31ef423f7d83d7' into 61-low-l…

8665010

…evel-api

Added action & observer channel to Runtime.

5c1d742

Started implementing commit() low level API

code formatting and FIXME changed to TODO

6a059b2

Added Registered / Reserved Capabilities and functions.

a631a5f

Moved dispatch_action_and_wait() to instance.rs genesis WIP in nucleus InitApplication reduce()

Added NucleusStatus for handling async initialization

f6d2b25

Calling Genesis loop in init WIP Added UML-class.png Added more Error types Refactoring ExecuteZome for better Error returns

Handled Missing Capability case

f03247d

Fixed tests

ddd-mtl added the review label Jun 29, 2018

ddd-mtl requested review from lucksus, zippy and thedavidmeister June 29, 2018 13:38

Removed printfs

e0e3e71

Handled genesis return value Added genesis tests Refactored testing in core/src/lib.rs

ddd-mtl requested a review from sphinxc0re June 29, 2018 19:22

ddd-mtl added 4 commits June 29, 2018 15:35

Cleaner comments and fixed commented asserts in tests

0a7cd13

Merge branch 'develop' (#62)

91521e9

Conflicts: core/src/lib.rs core/src/nucleus/mod.rs core/src/nucleus/ribosome.rs

Merge remote-tracking branch 'origin/develop' into 61-low-level-api

d3dbcc7

Fixed merge with entry_types update

9479eac

Using cargo.lock from develop

ddd-mtl changed the title ~~WIP - 61 low level api~~ 61 low level api Jun 30, 2018

thedavidmeister suggested changes Jun 30, 2018

View reviewed changes

Refactored code in instance

704e305

Moved WAT for testing from test_utils to tests Added FromStr trait to Reserved Enums as suggested by clippy

sphinxc0re suggested changes Jul 2, 2018

View reviewed changes

Minor refactors based on Julian's review.

0665a7f

Improved some comments Trying --warn clippy

sphinxc0re suggested changes Jul 2, 2018

View reviewed changes

zippy suggested changes Jul 2, 2018

View reviewed changes

Minor refactors / comments based zippy's review and clippy.

9f8c4fe

Made clippy less annoying.

Merge branch 'develop'

06657df

Conflicts: core/src/lib.rs core/src/nucleus/mod.rs core/src/nucleus/ribosome.rs core/src/state.rs core_api/src/lib.rs dna/src/lib.rs

zippy approved these changes Jul 3, 2018

View reviewed changes

zippy closed this Jul 3, 2018

zippy reopened this Jul 3, 2018

zippy added review and removed review labels Jul 3, 2018

ddd-mtl merged commit d69e81f into develop Jul 3, 2018

ddd-mtl removed the review label Jul 3, 2018

sphinxc0re deleted the 61-low-level-api branch July 5, 2018 15:26

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

61 low level api #77

61 low level api #77

ddd-mtl commented Jun 29, 2018 •

edited

Loading

thedavidmeister Jun 30, 2018

ddd-mtl Jun 30, 2018

thedavidmeister Jun 30, 2018

lucksus Jul 2, 2018

thedavidmeister Jun 30, 2018

thedavidmeister Jun 30, 2018

thedavidmeister Jun 30, 2018

ddd-mtl Jun 30, 2018

thedavidmeister Jun 30, 2018

thedavidmeister Jun 30, 2018

thedavidmeister Jun 30, 2018

thedavidmeister Jun 30, 2018

sphinxc0re left a comment

sphinxc0re Jul 2, 2018

sphinxc0re Jul 2, 2018

sphinxc0re Jul 2, 2018

sphinxc0re Jul 2, 2018

sphinxc0re Jul 2, 2018

ddd-mtl Jul 2, 2018

zippy left a comment

zippy Jul 2, 2018

zippy Jul 2, 2018

zippy Jul 2, 2018

ddd-mtl Jul 2, 2018

codecov-io commented Jul 2, 2018 •

edited

Loading

zippy commented Jul 3, 2018 •

edited

Loading

ddd-mtl commented Jul 3, 2018

		}


		// Prepare valid DNA struct with that WASM in a zome's capability

61 low level api #77

61 low level api #77

Conversation

ddd-mtl commented Jun 29, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

sphinxc0re left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

zippy left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codecov-io commented Jul 2, 2018 • edited Loading

Codecov Report

zippy commented Jul 3, 2018 • edited Loading

ddd-mtl commented Jul 3, 2018

ddd-mtl commented Jun 29, 2018 •

edited

Loading

codecov-io commented Jul 2, 2018 •

edited

Loading

zippy commented Jul 3, 2018 •

edited

Loading