Added `serde1` feature to Serialize/Deserialize `WeightedIndex` by CGMossa · Pull Request #974 · rust-random/rand

CGMossa · 2020-05-01T14:53:10Z

It required adding serde as an optional dependency.
Also added a test using bincode like in the other crate.

Referering to this: #967

I hope this is up to standard, otherwise I'll take all of your feedback and ensure it is as correct
as possible.

It required adding serde as an optional dependency. Also added a test using bincode like in the other crate.

dhardy · 2020-05-01T17:28:57Z

Thanks @CGMossa. Hmm, would it be possible to do the reformatting first in its own commit? Also please add a note to the changelog.

https://rust-random.github.io/book/contributing.html

CGMossa · 2020-05-01T17:51:53Z

I have the intellij auto use rustfmt. Which is not admittedly cool, but I thought that would be fine. I'll add a line in the changelog. I thought you'd suggest more things to add the derive attribute to. I'd be happy to add it everywhere needed. But I don't know which places.

…

On Fri, May 1, 2020, 19:29 Diggory Hardy ***@***.***> wrote: Thanks @CGMossa <https://github.com/CGMossa>. Hmm, would it be possible to do the reformatting *first* in its own commit? Also please add a note to the changelog. https://rust-random.github.io/book/contributing.html — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#974 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIDVSCCVXXWLSJQCZ62MLLRPMBGPANCNFSM4MXET6CQ> .

dhardy · 2020-05-01T18:31:15Z

I thought you'd suggest more things to add the derive attribute to. I'd be
happy to add it everywhere needed. But I don't know which places.

It crossed my mind that this might be a good idea, but I didn't look into where yet. Probably would only be applicable to distributions.

This reverts commit 9c823fa.

This time without the rustfmt invocation.

CGMossa · 2020-05-02T16:36:31Z

These are the changes you've asked for. I will experiment with making all the distributions serde complaint.

…

On Fri, May 1, 2020, 20:31 Diggory Hardy ***@***.***> wrote: I thought you'd suggest more things to add the derive attribute to. I'd be happy to add it everywhere needed. But I don't know which places. It crossed my mind that this might be a good idea, but I didn't look into where yet. Probably would only be applicable to distributions. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#974 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIDVSFNXMAM2WC6EOIEWR3RPMIQDANCNFSM4MXET6CQ> .

vks · 2020-05-02T16:42:59Z

It looks like the minimal supported serde version is incorrectly specified.

CGMossa · 2020-05-03T18:10:48Z

I wrote it exactly like it is written in rand_core and rand_pcg. Which version should I write instead?
Should I use this

serde = "1.0.106"

or

serde = "1.0.*"

?

Hopefully this will do it.

CGMossa · 2020-05-03T18:24:47Z

I see that tha CI stuff is resulting in an error. I am unable to see the same error here. I've added a version number to one of the Cargo.toml and I'll wait and see. Even if I run the command that is making the error cargo test --features=serde1,log I don't get an error. But I have serde 1.0.105 installed and the latest is 1.0.106 and maybe the thing on the here has an earlier version?

CGMossa · 2020-05-03T18:32:51Z

This worked. I've no idea why, since the other crates in the rand-workspace also indicated serde with "1". Maybe they fixed a bug? Although, not this works.

dhardy · 2020-05-04T07:32:08Z

We test with the minimal version of Serde which crates say they support — but this must be unified across all dependencies, so will be the highest minimum version. The idea is to find the real minimum version, but I think most people only care that it's high enough to work. That said, 1.0.105 is less than two months old. If an older version suffices that would be preferable. You can test locally via cargo generate-lockfile -Z minimal-versions

dhardy · 2020-05-04T07:40:09Z

For completeness's sake (if you don't mind), we should also consider: Bernoulli, Uniform, UniformFloat, UniformDuration, StepRng, IndexVec (maybe).
The following structs have zero size (thus don't need serialising), but doing so might ease deriving of other types: OpenClosed01, Open01, Standard, Alphanumeric.

CGMossa · 2020-05-04T07:49:11Z

Yep, I'll take a look later today. Thanks for the follow up!

…

On Mon, 4 May 2020 at 09:40, Diggory Hardy ***@***.***> wrote: For completeness's sake (if you don't mind), we should also consider: Bernoulli, Uniform, UniformFloat, UniformDuration, StepRng, IndexVec (maybe). The following structs have zero size (thus don't need serialising), but doing so might ease deriving of other types: OpenClosed01, Open01, Standard, Alphanumeric. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#974 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIDVSGSA6JM5OM46NQ4D33RPZWONANCNFSM4MXET6CQ> .

CGMossa · 2020-05-04T10:06:00Z

Minimal supported serde version is 1.0.103. Release note says:

Support deserializing untagged unit variants from formats that treat unit as None (#1668)

Test command:

rustup run nightly cargo generate-lockfile -Z minimal-versions & cargo test --features=serde1,log

#TODO: Test of equality between ser/de is needed.

#TODO: Unable to test equality in mode between de-serialized and original UniformDuration

dhardy · 2020-05-04T10:29:38Z

1.0.103

This version is almost six months old. Since we're developing against the master branch (will be a breaking release), this is fine IMO.

… `Open01`, and `Standard`

CGMossa · 2020-05-04T10:44:44Z

I've added the serialize and deserialize annotations. I also wanted to add tests using bincode, but some of these types have no fields I can test equality against, and PartialEq is not implemented for them so.. I don't know.

I've left some skeleton code, awaiting further feedback.

(this is like my biggest attempt at contributing yet! Thanks for the PR-mentoring!!)

CGMossa · 2020-05-14T00:45:13Z

Am I missing more things here?

dhardy · 2020-05-14T15:56:12Z

Sorry, I'll try to review in the next couple of days.

dhardy

Sorry, I haven't been paying this project much attention lately!

Your PR looks good. Those unfinished bits shouldn't be hard (I gave a couple hints).

CGMossa · 2020-05-15T13:40:44Z

I've added the tests based on your suggestions.

Do I resolve them or do I wait for you to click that they are resolved?

dhardy

Looks good to me. (There are some over-long lines, and you didn't need to match cases which can't occur in the tests, but it's still good enough.)

I'll wait a couple of days in case @vks wishes to review, then merge.

vks

Looks good!

vks · 2020-05-15T15:25:32Z

This should probably be squashed before merging. I think this is technically a breaking change.

CGMossa · 2020-05-15T17:51:39Z

I don't know how to squash?

vks · 2020-05-15T20:25:55Z

~~I think GitHub offers this option to the maintainers, but it seems like it is disabled for this repository.~~

Edit: I just enabled it, you don't have to do anything, we can let GitHub do it.

If you want to do that yourself, you can use git rebase -i to specify which commits should be squashed.

dhardy · 2020-05-16T09:49:43Z

master already contains breaking changes, so no problem there.

CGMossa · 2020-05-17T15:35:37Z

I've added the missing spaces. Thanks for enabling that feature, as I've got no training in using rebase or squashing. I'd appreciate a link to learn about this.

dhardy · 2020-05-18T08:21:47Z

Thanks for the PR!

git rebase is quite useful. Also git add -p. You may want to test them out on a dummy repo!

Implements rust-random#974. As in rust-random#975, but defining equality such that the user is not exposed to the fact that one logical state may have different representations in an implementation-specific way.

Implements #974. As in #975, but defining equality such that the user is not exposed to the fact that one logical state may have different representations in an implementation-specific way.

Added serde1 feature to WeightedIndex

9c823fa

It required adding serde as an optional dependency. Also added a test using bincode like in the other crate.

CGMossa added 4 commits May 2, 2020 00:27

Revert "Added serde1 feature to WeightedIndex"

f6871d0

This reverts commit 9c823fa.

Added Serialize/Deserialize to WeightedIndex and UniformInt

50bc078

This time without the rustfmt invocation.

Updated changelog

2bf89f6

Unreleased changes

9ddd57e

Cahnged the minimal req. serde version

1ec330f

Hopefully this will do it.

CGMossa added 4 commits May 4, 2020 12:12

Minimal supported serde version

8bb81d2

Ser/de Bernoulli

0e97b60

Added ser/de to Uniform

11ade61

#TODO: Test of equality between ser/de is needed.

Added ser/de for UniformDuration

5b23d21

#TODO: Unable to test equality in mode between de-serialized and original UniformDuration

CGMossa added 2 commits May 4, 2020 12:42

Added ser/deto IndexVec, StepRng, Alphanumeric, OpenClosed01,…

b393d75

… `Open01`, and `Standard`

Update float.rs

fda3639

dhardy reviewed May 15, 2020

View reviewed changes

Comment thread src/seq/index.rs Outdated

Comment thread src/distributions/uniform.rs Outdated

Added tests for Uniform, UniformDurationSampler, and IndexVec

fd72a6e

dhardy approved these changes May 15, 2020

View reviewed changes

vks reviewed May 15, 2020

View reviewed changes

Comment thread Cargo.toml

vks reviewed May 15, 2020

View reviewed changes

Comment thread src/seq/index.rs Outdated

vks approved these changes May 15, 2020

View reviewed changes

vks changed the title ~~Added serde1 feature to Serialize/Deseriaslize WeightedIndex~~ Added serde1 feature to Serialize/Deserialize WeightedIndex May 15, 2020

Added spaces

dc3cda6

dhardy merged commit 7ede440 into rust-random:master May 18, 2020

kazcw mentioned this pull request May 19, 2020

impl PartialEq, Eq for rngs #979

Merged

vks mentioned this pull request Aug 1, 2020

Serialize/Deserialize WeightedIndex<T> is not enabled with feature "serde1" #967

Closed

CAD97 mentioned this pull request Dec 21, 2020

Update rand requirement from 0.7 to 0.8 bevyengine/bevy#1114

Merged

sync-by-unito Bot mentioned this pull request Dec 21, 2020

Bump rand from 0.7.3 to 0.8.0 ProvableHQ/leo#510

Closed

This was referenced Mar 9, 2021

Update rand requirement from 0.7 to 0.8 hacspec/hacspec#80

Merged

chore(deps): update rand requirement from 0.7 to 0.8 transparencies/zip#2

Open

Bump rand from 0.7.3 to 0.8.3 Lionjudge9061-corp/mobilecoin#2

Open

dependabot Bot mentioned this pull request Mar 16, 2021

Bump rand from 0.7.3 to 0.8.3 ZeusWPI/zauth#65

Closed

Uh oh!

Conversation

CGMossa commented May 1, 2020

Uh oh!

dhardy commented May 1, 2020

Uh oh!

CGMossa commented May 1, 2020 via email

Uh oh!

dhardy commented May 1, 2020

Uh oh!

CGMossa commented May 2, 2020 via email

Uh oh!

vks commented May 2, 2020

Uh oh!

CGMossa commented May 3, 2020

Uh oh!

CGMossa commented May 3, 2020

Uh oh!

CGMossa commented May 3, 2020

Uh oh!

dhardy commented May 4, 2020

Uh oh!

dhardy commented May 4, 2020

Uh oh!

CGMossa commented May 4, 2020 via email

Uh oh!

CGMossa commented May 4, 2020

Uh oh!

dhardy commented May 4, 2020

Uh oh!

CGMossa commented May 4, 2020

Uh oh!

CGMossa commented May 14, 2020

Uh oh!

dhardy commented May 14, 2020

Uh oh!

dhardy left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

CGMossa commented May 15, 2020

Uh oh!

dhardy left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

vks left a comment

Choose a reason for hiding this comment

Uh oh!

vks commented May 15, 2020

Uh oh!

CGMossa commented May 15, 2020

Uh oh!

vks commented May 15, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dhardy commented May 16, 2020

Uh oh!

CGMossa commented May 17, 2020

Uh oh!

dhardy commented May 18, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

vks commented May 15, 2020 •

edited

Loading