Add support for generic timeline #525

Fusionnée
Plume_migration_agent a fusionné 58 révision(s) à partir de timeline vers master il y a 5 ans
Propriétaire

fix #450
Query example : lang in [fr, en] and license in [cc] or followed or title contains "Plume"

todolist

  • design a query language for timeline definition
  • make a parser for it
  • make the parser return meaning-full syntax error
  • add tests for parser
  • make a query executor for it
  • make the query executor return meaning-full runtime error (list not found...)
  • add a keyword to show/hide boosts/likes
  • add tests for query executor
  • define timeline in sql model
  • implement timeline in rust
  • add tests for timeline
  • define list in sql model
  • implement lists in rust
  • add tests for lists
  • create migration for pgsql
  • create migration for sqlite
  • modify circleci config to run migration before compilling
  • make a gui to manage timelines for user
  • make a gui to manage timelines for instance
  • make a gui to manage lists for user&instance
  • replace old timeline system with this one
  • add tests for front end??

Abstract syntax for query language

S -> A or S | A
A -> B and A | B
B -> ( S ) | C
C -> not D | D
D -> E in L | F contains W | G
E -> blog | author | license | tags | lang
F -> title | subtitle | content
G -> followed H | has_cover | local | all
H -> include I H | exclude I H | ε
I -> boosts | boost | likes | like
L -> [ M ] | W
M -> W, M | W
W -> <char+>
fix #450 Query example : `lang in [fr, en] and license in [cc] or followed or title contains "Plume"` #### todolist - [x] design a query language for timeline definition - [x] make a parser for it - [x] make the parser return meaning-full syntax error - [x] add tests for parser - [x] make a query executor for it - [x] make the query executor return meaning-full runtime error (list not found...) - [x] add a keyword to show/hide boosts/likes - [x] add tests for query executor - [x] define timeline in sql model - [x] implement timeline in rust - [x] add tests for timeline - [x] define list in sql model - [x] implement lists in rust - [x] add tests for lists - [x] create migration for pgsql - [x] create migration for sqlite - [x] modify circleci config to run migration before compilling - [x] ~~make a gui to manage timelines for user~~ - [x] ~~make a gui to manage timelines for instance~~ - [x] ~~make a gui to manage lists for user&instance~~ - [x] replace old timeline system with this one - [x] ~~add tests for front end??~~ #### Abstract syntax for query language ``` S -> A or S | A A -> B and A | B B -> ( S ) | C C -> not D | D D -> E in L | F contains W | G E -> blog | author | license | tags | lang F -> title | subtitle | content G -> followed H | has_cover | local | all H -> include I H | exclude I H | ε I -> boosts | boost | likes | like L -> [ M ] | W M -> W, M | W W -> <char+> ```
codecov[bot] a commenté il y a 5 ans (Migré de github.com)

Codecov Report

Merging #525 into master will increase coverage by 5.48%.
The diff coverage is 69.63%.

@@            Coverage Diff             @@
##           master     #525      +/-   ##
==========================================
+ Coverage   34.35%   39.83%   +5.48%     
==========================================
  Files          68       72       +4     
  Lines        8104     9368    +1264     
  Branches     1923     2219     +296     
==========================================
+ Hits         2784     3732     +948     
- Misses       4538     4580      +42     
- Partials      782     1056     +274
# [Codecov](https://codecov.io/gh/Plume-org/Plume/pull/525?src=pr&el=h1) Report > Merging [#525](https://codecov.io/gh/Plume-org/Plume/pull/525?src=pr&el=desc) into [master](https://codecov.io/gh/Plume-org/Plume/commit/b36e12b65460f25f1d6aeb7823bc1154d928b1fd?src=pr&el=desc) will **increase** coverage by `5.48%`. > The diff coverage is `69.63%`. ```diff @@ Coverage Diff @@ ## master #525 +/- ## ========================================== + Coverage 34.35% 39.83% +5.48% ========================================== Files 68 72 +4 Lines 8104 9368 +1264 Branches 1923 2219 +296 ========================================== + Hits 2784 3732 +948 - Misses 4538 4580 +42 - Partials 782 1056 +274 ```
Éditeur
Propriétaire

blocked by #555
interface for managing timelines and lists will be made in a further pr

blocked by #555 interface for managing timelines and lists will be made in a further pr
elegaanz (Migré de github.com) révisé il y a 5 ans
@ -0,0 +5,4 @@
use schema::{posts, timeline, timeline_definition};
use std::ops::Deref;
use {Connection, Error, PlumeRocket, Result};
elegaanz (Migré de github.com) a commenté il y a 5 ans

What is this field for? Is it important to always have a defined order for timelines? And how should it be determined: by the user, or automatically?

What is this field for? Is it important to always have a defined order for timelines? And how should it be determined: by the user, or automatically?
trinity-1686a révisé il y a 5 ans
@ -0,0 +5,4 @@
use schema::{posts, timeline, timeline_definition};
use std::ops::Deref;
use {Connection, Error, PlumeRocket, Result};
Éditeur
Propriétaire

hum, I don't know, my through was that the first one would be on top, the second under and so on, but it clearly limit what can be done, I'll revert that

hum, I don't know, my through was that the first one would be on top, the second under and so on, but it clearly limit what can be done, I'll revert that
igalic (Migré de github.com) révisé il y a 5 ans
igalic (Migré de github.com) laisser un commentaire

*snort

*snort
@ -0,0 +268,4 @@
let tl2_u1 = Timeline::new_for_user(
conn,
users[0].id,
"another timeline".to_owned(),
igalic (Migré de github.com) a commenté il y a 5 ans

haah

haah
trinity-1686a révisé il y a 5 ans
@ -0,0 +268,4 @@
let tl2_u1 = Timeline::new_for_user(
conn,
users[0].id,
"another timeline".to_owned(),
Éditeur
Propriétaire

I'm not good at this whole "giving name to lists" thing

I'm not good at this whole "giving name to lists" thing
igalic (Migré de github.com) révisé il y a 5 ans
@ -0,0 +268,4 @@
let tl2_u1 = Timeline::new_for_user(
conn,
users[0].id,
"another timeline".to_owned(),
igalic (Migré de github.com) a commenté il y a 5 ans

i disagree, i think you're very good 😸

i disagree, i think you're very good 😸
elegaanz a commenté il y a 5 ans (Migré de github.com)

I think this PR is now complete. If you are fine with my commits, I think we can merge it?

I think this PR is now complete. If you are fine with my commits, I think we can merge it?
Éditeur
Propriétaire

there seems to be some failure when running migrations

there seems to be some failure when running migrations
trinity-1686a révisé il y a 5 ans
trinity-1686a laisser un commentaire
Éditeur
Propriétaire

at first glance this looks good. I'll run some tests when I get the time, to make sure all is ok

at first glance this looks good. I'll run some tests when I get the time, to make sure all is ok
@ -0,0 +23,4 @@
)))
}
// TODO
Éditeur
Propriétaire

will it be done in this pr, or is this just basic support and actual customization will be added later?

will it be done in this pr, or is this just basic support and actual customization will be added later?
elegaanz (Migré de github.com) révisé il y a 5 ans
@ -0,0 +23,4 @@
)))
}
// TODO
elegaanz (Migré de github.com) a commenté il y a 5 ans

I will add it later I think, this PR is big enough.

I will add it later I think, this PR is big enough.
trinity-1686a révisé il y a 5 ans
trinity-1686a laisser un commentaire
Éditeur
Propriétaire

I think you are not calling the query parser when handling local like/reshare

I think you are not calling the query parser when handling local like/reshare
Éditeur
Propriétaire

I might be reading this wrong, but this enum (Kind) is made to tell if this is a new post, or a like/reshare of an "old" one, and by whom. This is in the part that handle likes, so I think this should be Kind::Like(&actor). There is probably the same issue with reshare in other places

I might be reading this wrong, but this enum (Kind) is made to tell if this is a new post, or a like/reshare of an "old" one, and by whom. This is in the part that handle likes, so I think this should be `Kind::Like(&actor)`. There is probably the same issue with reshare in other places
Éditeur
Propriétaire

see previous comment

see previous comment
elegaanz (Migré de github.com) révisé il y a 5 ans
elegaanz (Migré de github.com) a commenté il y a 5 ans

Indeed, thanks. I probably copied it from somewhere else and forgot to change it (I should really stop copy/pasting code…)

Indeed, thanks. I probably copied it from somewhere else and forgot to change it (I should really stop copy/pasting code…)
elegaanz a commenté il y a 5 ans (Migré de github.com)

Thanks for the review, everything is fixed now.

But… I think posts are correctly removed from timelines when deleted, but not when un-liked or un-boosted (but this can probably be fixed later?).

Thanks for the review, everything is fixed now. But… I think posts are correctly removed from timelines when deleted, but not when un-liked or un-boosted (but this can probably be fixed later?).
Éditeur
Propriétaire

they are remove because of cascading remove from sql. The easiest way to have the same comportment might be to add a reference to the like/reshare in 2 optional (and useless) columns, that cascade remove too

they are remove because of cascading remove from sql. The easiest way to have the same comportment might be to add a reference to the like/reshare in 2 optional (and useless) columns, that cascade remove too
trinity-1686a révisé il y a 5 ans
trinity-1686a laisser un commentaire
Éditeur
Propriétaire

I can't approve this myself as it's my pr, so if someone could review it that would be nice
we should also do a little todolist of things yet missing.
What come to mind is :

  • boost/like deletion doesn't remove from tl
  • new tl are not automatically filled (make a magic sql query that does this?)
  • related to the above, modifying the query of a tl should clear the tl and rebuild it
  • having an actual way to modify a tl, else they are not really custom
  • I believe query parsing errors are not translated
I can't approve this myself as it's my pr, so if someone could review it that would be nice we should also do a little todolist of things yet missing. What come to mind is : - boost/like deletion doesn't remove from tl - new tl are not automatically filled (make a magic sql query that does this?) - related to the above, modifying the query of a tl should clear the tl and rebuild it - having an actual way to modify a tl, else they are not really custom - I believe query parsing errors are not translated
@ -0,0 +1,51 @@
#![allow(dead_code)]
Éditeur
Propriétaire

?

?
elegaanz (Migré de github.com) révisé il y a 5 ans
@ -0,0 +1,51 @@
#![allow(dead_code)]
elegaanz (Migré de github.com) a commenté il y a 5 ans

It is because of the unimplemented routes, to avoid a few compilation warnings.

It is because of the unimplemented routes, to avoid a few compilation warnings.
elegaanz a commenté il y a 5 ans (Migré de github.com)

I can't approve this myself as it's my pr, so if someone could review it that would be nice

I seems OK for me (excepted for the things you listed), but I worked on it too, maybe you want to review it, @igalic ? (don't feel forced to do it, I think we can safely merge it if someone else approves the few commits I pushed).

And for the issue you listed, maybe it could be handled in another PR ? The basics are here, and this branch is already pretty big…

> I can't approve this myself as it's my pr, so if someone could review it that would be nice I seems OK for me (excepted for the things you listed), but I worked on it too, maybe you want to review it, @igalic ? (don't feel forced to do it, I think we can safely merge it if someone else approves the few commits I pushed). And for the issue you listed, maybe it could be handled in another PR ? The basics are here, and this branch is already pretty big…
igalic a commenté il y a 5 ans (Migré de github.com)

I'll see what i can do tonight.

I'll see what i can do tonight.
trinity-1686a révisé il y a 5 ans
@ -0,0 +1,51 @@
#![allow(dead_code)]
Éditeur
Propriétaire

There is not warning about them, it's actually... dead code badoum tss

(no warning because functions are public, so the are maybe just used in a crate that depends on Plume, aka none)

There is not warning about them, it's actually... dead code *badoum tss* (no warning because functions are public, so the are maybe just used in a crate that depends on Plume, aka none)
Éditeur
Propriétaire

What I listed was just to have a small todolist of things to add later, not now. This should end in an issue to not get forgotten

What I listed was just to have a small todolist of things to add later, not now. This should end in an issue to not get forgotten
igalic (Migré de github.com) révisé il y a 5 ans
igalic (Migré de github.com) laisser un commentaire

weak 👀 very tired, not finished this review sorry

weak 👀 very tired, not finished this review sorry
igalic (Migré de github.com) a commenté il y a 5 ans

why is this mutable?
can't we do:

for i in i.. {
    let users = super::users::User::get_local_page(conn, (i * 20, (i + 1) * 20))
    if users.is_err() {
        break;
    }

    let users = users.ok();
    if users.is_empty() {
        return;
    }

    for u in users {
        super::timeline::Timeline::new_for_user(conn, u.id, "Your feed".into(), format!("followed or author in [ {} ]", u.fqn)).expect("User feed creation error");
    }
}
why is this mutable? can't we do: ```rust for i in i.. { let users = super::users::User::get_local_page(conn, (i * 20, (i + 1) * 20)) if users.is_err() { break; } let users = users.ok(); if users.is_empty() { return; } for u in users { super::timeline::Timeline::new_for_user(conn, u.id, "Your feed".into(), format!("followed or author in [ {} ]", u.fqn)).expect("User feed creation error"); } } ```
igalic (Migré de github.com) a commenté il y a 5 ans

same as above.

same as above.
igalic (Migré de github.com) a commenté il y a 5 ans

👍

👍
igalic (Migré de github.com) a commenté il y a 5 ans

Invalid without the e

`Invalid` without the `e`
igalic (Migré de github.com) a commenté il y a 5 ans

i dunno, this feels like an error…

i dunno, this feels like an error…
@ -0,0 +23,4 @@
0 => Ok(ListType::User),
1 => Ok(ListType::Blog),
2 => Ok(ListType::Word),
3 => Ok(ListType::Prefix),
igalic (Migré de github.com) a commenté il y a 5 ans

it would be nice if we had the same enum in SQL… but i digress.

it would be nice if we had the same enum in SQL… but i digress.
trinity-1686a révisé il y a 5 ans
Éditeur
Propriétaire

there are some errors in your code (i..instead of 0.., ok() instead of unwrap...) but overall using for seems more logic than loop

there are some errors in your code (`i..`instead of `0..`, ok() instead of unwrap...) but overall using for seems more logic than loop
igalic (Migré de github.com) révisé il y a 5 ans
igalic (Migré de github.com) laisser un commentaire

slightly less tired, more comments, would love to hear your feedback

slightly less tired, more comments, would love to hear your feedback
igalic (Migré de github.com) a commenté il y a 5 ans

👍

👍
@ -0,0 +1,4 @@
-- This file should undo anything in `up.sql`
DELETE FROM timeline_definition WHERE name = 'Your feed';
DELETE FROM timeline_definition WHERE name = 'Local feed' AND query = 'local';
DELETE FROM timeline_definition WHERE name = 'Federared feed' AND query = 'all';
igalic (Migré de github.com) a commenté il y a 5 ans

how does this work with translations?

how does this work with translations?
igalic (Migré de github.com) a commenté il y a 5 ans

i'm confused… why can't we just say .first(&*c.conn)?;

i'm confused… why can't we just say `.first(&*c.conn)?;`
igalic (Migré de github.com) a commenté il y a 5 ans

mumbles something about macros

*mumbles something about macros*
igalic (Migré de github.com) a commenté il y a 5 ans

what does this do?

what does this do?
igalic (Migré de github.com) a commenté il y a 5 ans

if we're searching by name, why do we need the user_id (i understand it's Optional, but still)

if we're searching by name, why do we need the user_id (i understand it's `Option`al, but still)
@ -0,0 +284,4 @@
.select(list_elems::word)
.load::<Option<String>>(conn)
.map_err(Error::from)
.map(|r| r.into_iter().filter_map(|o| o).collect::<Vec<String>>())
igalic (Migré de github.com) a commenté il y a 5 ans

i'm confused about the filter_map() in these functions.
can you explain what it does?

i'm confused about the `filter_map()` in these functions. can you explain what it does?
igalic (Migré de github.com) a commenté il y a 5 ans

is this still needed, or was this just for debugging?

is this still needed, or was this just for debugging?
@ -0,0 +53,4 @@
.map_err(Error::from)
} else {
timeline_definition::table
.filter(timeline_definition::user_id.is_null())
igalic (Migré de github.com) a commenté il y a 5 ans

this is the only difference between the two expressions. I find it odd that there's no sensible, generic way to solve this.

this is the only difference between the two expressions. I find it odd that there's no sensible, generic way to solve this.
@ -0,0 +68,4 @@
.map_err(Error::from)
} else {
timeline_definition::table
.filter(timeline_definition::user_id.is_null())
igalic (Migré de github.com) a commenté il y a 5 ans

again, this duplication begs to be solved generically

again, this duplication begs to be solved generically
igalic (Migré de github.com) a commenté il y a 5 ans

if you turn the logic in this if around, we get one indentation level less.

if you turn the logic in this `if` around, we get one indentation level less.
igalic (Migré de github.com) a commenté il y a 5 ans

turn logic in if around to break out on top, and get one indentation level less.

turn logic in `if` around to break out on top, and get one indentation level less.
@ -0,0 +121,4 @@
*$state = Some(Token::Word($i,1,&""));
vec![]
},
_ => unreachable!(),
igalic (Migré de github.com) a commenté il y a 5 ans

how informative is this unreachable!() actually gonna be, if we do ever hit it?

how informative is this `unreachable!()` actually gonna be, if we do ever hit it?
@ -0,0 +151,4 @@
/// Private internals of TimelineQuery
#[derive(Debug, Clone, PartialEq)]
enum TQ<'a> {
igalic (Migré de github.com) a commenté il y a 5 ans

what does TQ mean?

what does `TQ` mean?
@ -0,0 +490,4 @@
fn parse_d<'a, 'b>(mut stream: &'b [Token<'a>]) -> QueryResult<(&'b [Token<'a>], Arg<'a>)> {
match stream.get(0).map(Token::get_text)? {
s @ "blog" | s @ "author" | s @ "license" | s @ "tags" | s @ "lang" => {
igalic (Migré de github.com) a commenté il y a 5 ans

what does this mean?

what does this mean?
igalic (Migré de github.com) a commenté il y a 5 ans

yay for refs

yay for refs
@ -0,0 +27,4 @@
#[get("/timeline/new")]
pub fn new() -> Result<Ructe, ErrorPage> {
unimplemented!()
igalic (Migré de github.com) a commenté il y a 5 ans

what are these supposed to do ?

what are these supposed to do ?
@ -111,0 +113,4 @@
"Your feed" => i18n!(cat, "Your feed"),
"Local feed" => i18n!(cat, "Local feed"),
"Federated feed" => i18n!(cat, "Federated feed"),
n => n.to_string(),
igalic (Migré de github.com) a commenté il y a 5 ans

neato…
also, i just realized what this does, thanks to this awful example: https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=ba47f87eb9d03239e40287e8c405e4ae

neato… also, i just realized what this does, thanks to this awful example: https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=ba47f87eb9d03239e40287e8c405e4ae
igalic (Migré de github.com) a commenté il y a 5 ans

i'm quite happy with how the repetition of these ⬆️ in the templates you deleted is gone

i'm quite happy with how the repetition of these ⬆️ in the templates you deleted is gone
igalic (Migré de github.com) a commenté il y a 5 ans

i think we'll need dir="auto" here

i think we'll need `dir="auto"` here
@ -0,0 +9,4 @@
@:base(ctx, tl.name.clone(), {}, {}, {
<section class="flex wrap" dir="auto">
<h1 class="grow">@i18n_timeline_name(ctx.1, &tl.name)</h1>
igalic (Migré de github.com) a commenté il y a 5 ans

dir="auto"?

`dir="auto"`?
igalic (Migré de github.com) révisé il y a 5 ans
@ -0,0 +1,4 @@
-- This file should undo anything in `up.sql`
DELETE FROM timeline_definition WHERE name = 'Your feed';
DELETE FROM timeline_definition WHERE name = 'Local feed' AND query = 'local';
DELETE FROM timeline_definition WHERE name = 'Federared feed' AND query = 'all';
igalic (Migré de github.com) a commenté il y a 5 ans

i think i found the answer to this somewhere below in the code!

i think i found the answer to this somewhere below in the code!
trinity-1686a révisé il y a 5 ans
Éditeur
Propriétaire

First return a Result<Blog, _>, optional convert it to a Result<Option<Blog>,...> so that if the error is not found, we can fetch the blog instead of not knowing it ever, but if the error is anything else, it's still an error

First return a `Result<Blog, _>`, [optional](https://docs.diesel.rs/diesel/result/trait.OptionalExtension.html) convert it to a `Result<Option<Blog>,...>` so that if the error is not found, we can fetch the blog instead of not knowing it ever, but if the error is anything else, it's still an error
trinity-1686a révisé il y a 5 ans
@ -0,0 +284,4 @@
.select(list_elems::word)
.load::<Option<String>>(conn)
.map_err(Error::from)
.map(|r| r.into_iter().filter_map(|o| o).collect::<Vec<String>>())
Éditeur
Propriétaire

It map each Option<String> to it's inner String, and filter Nones away

It map each `Option<String>` to it's inner `String`, and filter `None`s away
trinity-1686a révisé il y a 5 ans
Éditeur
Propriétaire

this gives this module more visibility from the outer code, while keeping it private from other crates. It was probably necessary at some point, but rn it could be plain private

this gives this module more visibility from the outer code, while keeping it private from other crates. It was probably necessary at some point, but rn it could be plain private
trinity-1686a révisé il y a 5 ans
@ -0,0 +27,4 @@
#[get("/timeline/new")]
pub fn new() -> Result<Ructe, ErrorPage> {
unimplemented!()
Éditeur
Propriétaire

allow creation/edition/deletion of custom timelines

allow creation/edition/deletion of custom timelines
trinity-1686a révisé il y a 5 ans
@ -0,0 +121,4 @@
*$state = Some(Token::Word($i,1,&""));
vec![]
},
_ => unreachable!(),
Éditeur
Propriétaire
thread 'main' panicked at 'internal error: entered unreachable code', src/main.rs:2:5
note: Run with `RUST_BACKTRACE=1` environment variable to display a backtrace.

I could probably rewrite that without unreachable! if it's needed (it's truly unreachable atm)

``` thread 'main' panicked at 'internal error: entered unreachable code', src/main.rs:2:5 note: Run with `RUST_BACKTRACE=1` environment variable to display a backtrace. ``` I could probably rewrite that without unreachable! if it's needed (it's truly unreachable atm)
trinity-1686a révisé il y a 5 ans
@ -0,0 +151,4 @@
/// Private internals of TimelineQuery
#[derive(Debug, Clone, PartialEq)]
enum TQ<'a> {
Éditeur
Propriétaire

TimelineQuery, but more hidden (it's private while TimelineQuery is public)

TimelineQuery, but more hidden (it's private while `TimelineQuery` is public)
trinity-1686a révisé il y a 5 ans
@ -0,0 +490,4 @@
fn parse_d<'a, 'b>(mut stream: &'b [Token<'a>]) -> QueryResult<(&'b [Token<'a>], Arg<'a>)> {
match stream.get(0).map(Token::get_text)? {
s @ "blog" | s @ "author" | s @ "license" | s @ "tags" | s @ "lang" => {
Éditeur
Propriétaire

if it's equal to "blog", "author", "license"... store the value in s and match the given arm

if it's equal to "blog", "author", "license"... store the value in s and match the given arm
trinity-1686a révisé il y a 5 ans
Éditeur
Propriétaire

I have no idea, @BaptisteGelez ?

I have no idea, @BaptisteGelez ?
trinity-1686a révisé il y a 5 ans
Éditeur
Propriétaire

I's a find by (list) name, in context of the given user (or instance if None)

I's a find by (list) name, in context of the given user (or instance if None)
trinity-1686a révisé il y a 5 ans
@ -0,0 +53,4 @@
.map_err(Error::from)
} else {
timeline_definition::table
.filter(timeline_definition::user_id.is_null())
Éditeur
Propriétaire

Welcome to the party!
Diesel consider that anything eq None unless we are that explicit

Welcome to the party! Diesel consider that anything eq None unless we are that explicit
igalic (Migré de github.com) révisé il y a 5 ans
igalic (Migré de github.com) a commenté il y a 5 ans

that needs either documentation or different names, because this way it's just way too confusing

that needs either documentation or different names, because this way it's just way too confusing
igalic (Migré de github.com) révisé il y a 5 ans
igalic (Migré de github.com) a commenté il y a 5 ans

thank you for the explanation!

thank you for the explanation!
igalic (Migré de github.com) révisé il y a 5 ans
@ -0,0 +151,4 @@
/// Private internals of TimelineQuery
#[derive(Debug, Clone, PartialEq)]
enum TQ<'a> {
igalic (Migré de github.com) a commenté il y a 5 ans

perhaps add a comment such as:

/// These are the private internals of TImeQuery

or something along the lines

perhaps add a comment such as: ```rust /// These are the private internals of TImeQuery ``` or something along the lines
igalic (Migré de github.com) révisé il y a 5 ans
@ -0,0 +490,4 @@
fn parse_d<'a, 'b>(mut stream: &'b [Token<'a>]) -> QueryResult<(&'b [Token<'a>], Arg<'a>)> {
match stream.get(0).map(Token::get_text)? {
s @ "blog" | s @ "author" | s @ "license" | s @ "tags" | s @ "lang" => {
igalic (Migré de github.com) a commenté il y a 5 ans

i have never seen that before! where is that documented?

i have never seen that before! where is that documented?
igalic (Migré de github.com) révisé il y a 5 ans
@ -0,0 +27,4 @@
#[get("/timeline/new")]
pub fn new() -> Result<Ructe, ErrorPage> {
unimplemented!()
igalic (Migré de github.com) a commenté il y a 5 ans

I the user interface? for which we have no design concept yet?
or via APIs?

I the user interface? for which we have no design concept yet? or via APIs?
trinity-1686a révisé il y a 5 ans
@ -0,0 +27,4 @@
#[get("/timeline/new")]
pub fn new() -> Result<Ructe, ErrorPage> {
unimplemented!()
Éditeur
Propriétaire

for the user interface. APIs are in their own directory

for the user interface. APIs are in their own directory
trinity-1686a révisé il y a 5 ans
@ -0,0 +151,4 @@
/// Private internals of TimelineQuery
#[derive(Debug, Clone, PartialEq)]
enum TQ<'a> {
Éditeur
Propriétaire

I think we should #![warn(missing_docs)] and progressively document everything in general

I think we should `#![warn(missing_docs)]` and progressively document everything in general
trinity-1686a révisé il y a 5 ans
@ -0,0 +490,4 @@
fn parse_d<'a, 'b>(mut stream: &'b [Token<'a>]) -> QueryResult<(&'b [Token<'a>], Arg<'a>)> {
match stream.get(0).map(Token::get_text)? {
s @ "blog" | s @ "author" | s @ "license" | s @ "tags" | s @ "lang" => {
Éditeur
Propriétaire

https://doc.rust-lang.org/1.5.0/book/patterns.html#bindings
It's not useful very often, but when it is, it's very nice to have

https://doc.rust-lang.org/1.5.0/book/patterns.html#bindings It's not useful very often, but when it is, it's very nice to have
trinity-1686a révisé il y a 5 ans
@ -0,0 +53,4 @@
.map_err(Error::from)
} else {
timeline_definition::table
.filter(timeline_definition::user_id.is_null())
Éditeur
Propriétaire

I've tried for a few hours to add a eq_optional via some trait, but the best I got was "only" 4 compiler error, of which the shortest was 4 lines long. I'm giving up on this

I've tried for a few hours to add a `eq_optional` via some trait, but the best I got was "only" 4 compiler error, of which the shortest was 4 lines long. I'm giving up on this
igalic (Migré de github.com) révisé il y a 5 ans
@ -0,0 +53,4 @@
.map_err(Error::from)
} else {
timeline_definition::table
.filter(timeline_definition::user_id.is_null())
igalic (Migré de github.com) a commenté il y a 5 ans

kudos for trying.

kudos for trying.
Éditeur
Propriétaire

and I failed my merge >_<

and I failed my merge >_<
igalic (Migré de github.com) révisé il y a 5 ans
igalic (Migré de github.com) laisser un commentaire

👍

👍
@ -0,0 +119,4 @@
}
})
{
Err(err)?;
igalic (Migré de github.com) a commenté il y a 5 ans

That is 21 lines of error handling

That is 21 lines of error handling
@ -0,0 +157,4 @@
}
})
{
Err(err)?;
igalic (Migré de github.com) a commenté il y a 5 ans

that's another 21 lines of error handling

that's another 21 lines of error handling
igalic (Migré de github.com) a commenté il y a 5 ans

there should be a better way to do that …
(I'm not saying you have to find that now)

there should be a better way to do that … (I'm not saying you have to find that *now*)
@ -0,0 +724,4 @@
/*
#[test]
fn test_matches_lists_saved() {
let r = &rockets();
igalic (Migré de github.com) a commenté il y a 5 ans

why is this one commented out?

why is this one commented out?
@ -0,0 +805,4 @@
conn,
users[0].id,
"Linux subtitle".to_owned(),
"subtitle contains Stallman".to_owned(),
igalic (Migré de github.com) a commenté il y a 5 ans

(i'd rather it didn't… lolsob…)

(i'd rather it didn't… lolsob…)
trinity-1686a révisé il y a 5 ans
@ -0,0 +724,4 @@
/*
#[test]
fn test_matches_lists_saved() {
let r = &rockets();
Éditeur
Propriétaire

I have no idea

I have no idea
trinity-1686a révisé il y a 5 ans
@ -0,0 +805,4 @@
conn,
users[0].id,
"Linux subtitle".to_owned(),
"subtitle contains Stallman".to_owned(),
Éditeur
Propriétaire

shall I edit that?

shall I edit that?
igalic (Migré de github.com) révisé il y a 5 ans
@ -0,0 +805,4 @@
conn,
users[0].id,
"Linux subtitle".to_owned(),
"subtitle contains Stallman".to_owned(),
igalic (Migré de github.com) a commenté il y a 5 ans

We can do a cleanup PR and just remove all problematic techmen from our test examples

We can do a cleanup PR and just remove all problematic techmen from our test examples
elegaanz (Migré de github.com) ces changements ont été approuvés il y a 5 ans
elegaanz (Migré de github.com) laisser un commentaire

LGTM, this great PR have been waiting for too long!

LGTM, this great PR have been waiting for too long!

Relecteurs

La demande d'ajout a été fusionnée en 006b44f580.
Vous pouvez également voir les instructions en ligne de commande.

Étape 1:

Depuis le dépôt de votre projet, sélectionnez une nouvelle branche et testez les modifications.
git checkout -b timeline master
git pull origin timeline

Étape 2:

Fusionner les modifications et mettre à jour sur Forgejo.
git checkout master
git merge --no-ff timeline
git push origin master
Connectez-vous pour rejoindre cette conversation.
Aucune évaluation
Aucun jalon
Pas d'assignataires
2 participants
Notifications
Échéance
La date d’échéance est invalide ou hors plage. Veuillez utiliser le format 'aaaa-mm-dd'.

Aucune échéance n'a été définie.

Dépendances

No dependencies set.

Reference: Plume/Plume#525
Chargement…
Il n'existe pas encore de contenu.