 | Diesel | 734 Members |
| A safe, extensible ORM and Query Builder for Rust | 10 Servers |
| Diesel | 734 Members |
| A safe, extensible ORM and Query Builder for Rust | 10 Servers |
| Timestamp | Message | |
|---|---|---|
| 19 Feb 2019 | ||
| 23:54:54 |  John Halbert (Gitter) | So, sometime mismatching on number of actual vs expected binds, or the prepared statement either exists or does not unexpectedly. |
| 23:57:57 |  Georg Semmler (Gitter) | Hmm, that should not happen. Could you provide some minimal example how to reproduce that? (Maybe as issue?) |
| 23:59:00 | John Halbert (Gitter) | Okay, yes, I can do that. Will open an issue. Just wanted to check first that there wasn't a simple solution. |
| 20 Feb 2019 | ||
| 00:37:32 | John Halbert (Gitter) | Actually, I'm seeing now this has to do with connection pool being provided by PgBouncer using transaction pooling, which disallows PREPARE/DEALLOCATE. |
| 02:34:12 |  Farzeen (Gitter) | @r-arias Here's what I went with: https://github.com/happycoder97/DSA-Testcases-Repo-Backend/blob/master/src/database/impl_crud.rs |
| 02:41:13 | Farzeen (Gitter) | My current design: I wanted the rest of my code to be entirely independent of what persistence mechanism I was using. So I created a repository module with the required structs and some traits. https://github.com/happycoder97/DSA-Testcases-Repo-Backend/blob/master/src/repository.rs But Rust doesn't allow to add #[derive(..)] to a struct after it has been defined. So I put all my struct inside a macro model_struct! { .. }. It is the responsibility of the implementor of repository to add the necessary derives to each struct inside model_struct! . Here is the implementation of model_struct! for diesel: https://github.com/happycoder97/DSA-Testcases-Repo-Backend/blob/master/src/database/model_struct.rs |
| 02:53:42 | Farzeen (Gitter) | PS: I would like to avoid typing table name inside model_struct! {...} in repository.rs, but I can't find a way to do it. |
| 06:28:11 |  Hal Gentz (Gitter) | I think I've found a bug |
| 06:29:13 | Hal Gentz (Gitter) | https://i.imgur.com/kX0LMfa.png |
| 06:29:39 | Hal Gentz (Gitter) | https://i.imgur.com/q3dj4qB.png |
| 06:30:22 | Hal Gentz (Gitter) | Maybe I'm wrong, but 7 is not greater than 255 ;) |
| 06:30:54 | Hal Gentz (Gitter) | I'll have to investigate more tomorrow to figure out why this is happening... |
| 06:32:13 | Hal Gentz (Gitter) | Interestingly, 7 != 7 according to diesel... |
| 06:37:05 | Hal Gentz (Gitter) | For those interested, diesel is generating this sql when using eq |
| 06:37:09 | Hal Gentz (Gitter) | [src/account_actions.rs:314] diesel::debug_query::<crate::DbBack, _>(&e).to_string() = "SELECT `ddb_session_roots`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_user_id`, `ddb_session_token_pairs`.`ddb_scrypt_log_n` FROM (`ddb_session_token_pairs` INNER JOIN `ddb_session_roots` ON `ddb_session_token_pairs`.`ddb_root_id` = `ddb_session_roots`.`ddb_id`) WHERE `ddb_session_token_pairs`.`ddb_cookie_token` = ? AND `ddb_session_token_pairs`.`ddb_csrf_token` = ? AND `ddb_session_token_pairs`.`ddb_scrypt_log_n` = ? AND `ddb_session_roots`.`ddb_token_head_id` = `ddb_session_token_pairs`.`ddb_id` AND `ddb_session_roots`.`ddb_end_time` IS NULL AND `ddb_session_roots`.`ddb_start_time` >= ? AND `ddb_session_roots`.`ddb_ip` = ? AND `ddb_session_roots`.`ddb_is_ipv4` = ? -- binds: [[183, 164, 170, 206, 240, 134, 84, 218, 4, 114, 140, 129, 248, 112, 148, 236, 52, 94, 57, 12, 228, 157, 54, 91, 73, 110, 34, 141, 180, 171, 249, 160, 219, 18, 15, 204, 84, 174, 77, 217, 234, 110, 245, 20, 32, 29, 102, 40, 94, 117, 160, 128, 231, 130, 89, 41, 220, 39, 243, 173, 44, 57, 85, 254], [185, 158, 189, 141, 19, 85, 34, 20, 40, 116, 183, 169, 140, 99, 42, 74, 253, 83, 196, 237, 208, 104, 185, 192, 254, 105, 192, 92, 231, 99, 195, 174, 98, 59, 163, 124, 45, 46, 47, 118, 172, 165, 36, 183, 216, 149, 126, 230, 55, 137, 91, 109, 208, 55, 114, 170, 110, 197, 66, 73, 22, 144, 113, 6], [7], 1550643684, [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1], [0]]" |
| 06:37:24 | Hal Gentz (Gitter) | (edited) ```[src/account_actions.rs:314] diesel::debug_query::<crate::DbBack, ... => ``` [src/account_actions.rs:314] diesel::debug_query::<crate::DbBack, ... |
| 06:48:29 | Hal Gentz (Gitter) | Hehehe |
| 06:48:43 | Hal Gentz (Gitter) | This also yields nothing according to diesel: [src/account_actions.rs:315] diesel::debug_query::<crate::DbBack, _>(&e).to_string() = "SELECTddb_session_token_pairs.ddb_id,ddb_session_token_pairs.ddb_id,ddb_session_token_pairs.ddb_idFROMddb_session_token_pairsWHEREddb_session_token_pairs.ddb_scrypt_log_n= ? -- binds: [[7]]" |
| 06:48:52 | Hal Gentz (Gitter) | (edited) ... diesel: `[src/account_actions.rs:315] diesel::debug_query::<crate::DbBack, _>(&e).to_string() = "SELECT `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id` FROM `ddb_session_token_pairs` WHERE `ddb_session_token_pairs`.`ddb_scrypt_log_n` = ? -- binds: [[7]]"` => ... diesel: ```[src/account_actions.rs:315] diesel::debug_query::<crate::DbBack, _>(&e).to_string() = "SELECT `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id` FROM `ddb_session_token_pairs` WHERE `ddb_session_token_pairs`.`ddb_scrypt_log_n` = ? -- binds: [[7]]"``` |
| 06:49:45 | Hal Gentz (Gitter) | Php admin says otherwise: https://i.imgur.com/KYWQRgH.png |
| 09:01:48 |  r-arias (Gitter) | @happycoder97 thanks for the response! I'll have a look! |
| 09:02:14 | Farzeen (Gitter) | Feel free to DM me if you have any further questions :) |
| 09:03:20 | r-arias (Gitter) | great, thanks :) |
| 11:28:25 | Georg Semmler (Gitter) | @johnhalbert Why not just use a connection pool inside your applications? (That's provided by r2d2) |
| 11:38:21 | Georg Semmler (Gitter) | @ZeGentzy As far as I can see the sql you tested in php admin does not match the sql produced by diesel. Therefore I would guess that is's just one of the other filters that prevents loading an result. Not sure what's going on the the other problem, but without any further information about the schema, the generated query and the involved data it's not possible to help you there. |
| 16:16:28 | John Halbert (Gitter) | @weiznich the application won't scale using r2d2 alone. We're already using that. r2d2 has to maintain its own pool, but as new instances of these services come online, they eventually exhaust the connections allowed to the db. PgBouncer was chosen as it allows a reasonable number of connections to be allowed on the database (500 or so), while still allowing the application to scale out without exhausting those connections. |
| 23:58:15 | Hal Gentz (Gitter) | @weiznich It produced SELECT `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id` FROM `ddb_session_token_pairs` WHERE `ddb_session_token_pairs`.`ddb_scrypt_log_n` = ? and I tested SELECT `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id`, `ddb_session_token_pairs`.`ddb_id` FROM `ddb_session_token_pairs` WHERE `ddb_session_token_pairs`.`ddb_scrypt_log_n` = 7 |
| 23:58:32 | Hal Gentz (Gitter) | Only difference is I got rid of the bind |
| 23:58:46 | Hal Gentz (Gitter) | Anyways, let me come up with a minimal example. |
| 21 Feb 2019 | ||
| 01:53:37 |  Dustin Knopoff (Gitter) | I feel like I'm missing something really obvious, but I want to chain optional filters. That is, I have a struct Filters with fields that are optional. They all map to a column in a table. So for something like a Post, I would and a given filter Filter { count: Some(10), title: Some("Cookies"), author: None, /* etc */ } I'd like to chain a posts.filter(title.like(filter.title.unwrap()). It's a bit more complicated than that but that's the essential piece I can't figure out. I've tried using the into_boxed() method which does help with a couple things but it seems that any filters I apply are ignored. Thanks for the help! |