lemmy/src/lib.rs

432 lines
15 KiB
Rust
Raw Normal View History

pub mod api_routes_v3;
pub mod api_routes_v4;
pub mod code_migrations;
pub mod prometheus_metrics;
pub mod scheduled_tasks;
pub mod session_middleware;
2022-05-10 12:06:32 +00:00
use crate::{code_migrations::run_advanced_migrations, session_middleware::SessionMiddleware};
use activitypub_federation::config::{FederationConfig, FederationMiddleware};
use actix_cors::Cors;
Fix #3366: Wrap plain-text error responses from the API in JSON (#3559) * Fix #3366: API does return plain HTML errors * Fix Clippy errors * Improve api response times by doing send_activity asynchronously (#3493) * do send_activity after http response * move to util function * format * fix prometheus * make synchronous federation configurable * cargo fmt * empty * empty --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> * Updating `login.rs` with generic `incorrect_login` response. (#3549) * Adding v0.18.1 and v0.18.0 release notes. (#3530) * Update RELEASES.md (#3556) added instruction to find the location of your docker directory (especially useful for those who used ansible since they never had to setup docker manually) * Use async email sender (#3554) * Upgrade all dependencies (#3526) * Upgrade all dependencies * as base64 * Adding phiresky to codeowners. (#3576) * Error enum fixed (#3487) * Create error type enum * Replace magic string slices with LemmyErrorTypes * Remove unused enum * Add rename snake case to error enum * Rename functions * clippy * Fix merge errors * Serialize in PascalCase instead of snake_case * Revert src/lib * Add serialization tests * Update translations * Fix compilation error in test * Fix another compilation error * Add code for generating typescript types * Various fixes to avoid breaking api * impl From<LemmyErrorType> for LemmyError * with_lemmy_type * trigger ci --------- Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> * Only update site_aggregates for local site (#3516) * Fix #3501 - Fix aggregation counts for elements removed and deleted (#3543) Two bugs were found and fixed: - previously elements removal and deletion were counted as two separate disappearances - removing comments did not affect post aggregations * Use LemmyErrorType also make error_type compulsory * Add missing import for jsonify_plain_text_errors * Fix formatting * Trying to make woodpecker run again --------- Co-authored-by: phiresky <phireskyde+git@gmail.com> Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: rosenjcb <rosenjcb@gmail.com> Co-authored-by: nixoye <12674582+nixoye@users.noreply.github.com> Co-authored-by: dullbananas <dull.bananas0@gmail.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> Co-authored-by: Sander Saarend <sander@saarend.com> Co-authored-by: Piotr Juszczyk <74842304+pijuszczyk@users.noreply.github.com>
2023-07-10 20:44:14 +00:00
use actix_web::{
dev::{ServerHandle, ServiceResponse},
middleware::{self, Condition, ErrorHandlerResponse, ErrorHandlers},
Fix #3366: Wrap plain-text error responses from the API in JSON (#3559) * Fix #3366: API does return plain HTML errors * Fix Clippy errors * Improve api response times by doing send_activity asynchronously (#3493) * do send_activity after http response * move to util function * format * fix prometheus * make synchronous federation configurable * cargo fmt * empty * empty --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> * Updating `login.rs` with generic `incorrect_login` response. (#3549) * Adding v0.18.1 and v0.18.0 release notes. (#3530) * Update RELEASES.md (#3556) added instruction to find the location of your docker directory (especially useful for those who used ansible since they never had to setup docker manually) * Use async email sender (#3554) * Upgrade all dependencies (#3526) * Upgrade all dependencies * as base64 * Adding phiresky to codeowners. (#3576) * Error enum fixed (#3487) * Create error type enum * Replace magic string slices with LemmyErrorTypes * Remove unused enum * Add rename snake case to error enum * Rename functions * clippy * Fix merge errors * Serialize in PascalCase instead of snake_case * Revert src/lib * Add serialization tests * Update translations * Fix compilation error in test * Fix another compilation error * Add code for generating typescript types * Various fixes to avoid breaking api * impl From<LemmyErrorType> for LemmyError * with_lemmy_type * trigger ci --------- Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> * Only update site_aggregates for local site (#3516) * Fix #3501 - Fix aggregation counts for elements removed and deleted (#3543) Two bugs were found and fixed: - previously elements removal and deletion were counted as two separate disappearances - removing comments did not affect post aggregations * Use LemmyErrorType also make error_type compulsory * Add missing import for jsonify_plain_text_errors * Fix formatting * Trying to make woodpecker run again --------- Co-authored-by: phiresky <phireskyde+git@gmail.com> Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: rosenjcb <rosenjcb@gmail.com> Co-authored-by: nixoye <12674582+nixoye@users.noreply.github.com> Co-authored-by: dullbananas <dull.bananas0@gmail.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> Co-authored-by: Sander Saarend <sander@saarend.com> Co-authored-by: Piotr Juszczyk <74842304+pijuszczyk@users.noreply.github.com>
2023-07-10 20:44:14 +00:00
web::Data,
App,
HttpResponse,
Fix #3366: Wrap plain-text error responses from the API in JSON (#3559) * Fix #3366: API does return plain HTML errors * Fix Clippy errors * Improve api response times by doing send_activity asynchronously (#3493) * do send_activity after http response * move to util function * format * fix prometheus * make synchronous federation configurable * cargo fmt * empty * empty --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> * Updating `login.rs` with generic `incorrect_login` response. (#3549) * Adding v0.18.1 and v0.18.0 release notes. (#3530) * Update RELEASES.md (#3556) added instruction to find the location of your docker directory (especially useful for those who used ansible since they never had to setup docker manually) * Use async email sender (#3554) * Upgrade all dependencies (#3526) * Upgrade all dependencies * as base64 * Adding phiresky to codeowners. (#3576) * Error enum fixed (#3487) * Create error type enum * Replace magic string slices with LemmyErrorTypes * Remove unused enum * Add rename snake case to error enum * Rename functions * clippy * Fix merge errors * Serialize in PascalCase instead of snake_case * Revert src/lib * Add serialization tests * Update translations * Fix compilation error in test * Fix another compilation error * Add code for generating typescript types * Various fixes to avoid breaking api * impl From<LemmyErrorType> for LemmyError * with_lemmy_type * trigger ci --------- Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> * Only update site_aggregates for local site (#3516) * Fix #3501 - Fix aggregation counts for elements removed and deleted (#3543) Two bugs were found and fixed: - previously elements removal and deletion were counted as two separate disappearances - removing comments did not affect post aggregations * Use LemmyErrorType also make error_type compulsory * Add missing import for jsonify_plain_text_errors * Fix formatting * Trying to make woodpecker run again --------- Co-authored-by: phiresky <phireskyde+git@gmail.com> Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: rosenjcb <rosenjcb@gmail.com> Co-authored-by: nixoye <12674582+nixoye@users.noreply.github.com> Co-authored-by: dullbananas <dull.bananas0@gmail.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> Co-authored-by: Sander Saarend <sander@saarend.com> Co-authored-by: Piotr Juszczyk <74842304+pijuszczyk@users.noreply.github.com>
2023-07-10 20:44:14 +00:00
HttpServer,
};
use actix_web_prom::PrometheusMetricsBuilder;
Add custom migration runner, forbid some `diesel migration` commands, fix old migrations (#4673) * Update schema.rs * rename * stuff * finish new implementation of schema_setup::run (not including revert, test, etc.) * fmt * refactor * fix sql * migriation run command * use trigger on migrations table * add Options with disable_migrations field for test * rename to enable_forbid_diesel_cli_trigger * fix * fix merge * diff_checker (partial) * Revert "diff_checker (partial)" This reverts commit 6709882e148ceff48d6f097895af18d13fbdf4bf. * Revert "Revert "diff_checker (partial)"" This reverts commit d4bdda5d11216f4acd13b9585d2392ca8c252a73. * diff check * improve schema diff * timestamp replacement * ignore column order * remove fedi_name default * stuff * improve diff * stuff * attempt parallel pg_dump * attempt 2 * Revert "attempt 2" This reverts commit a909d2d6438d0b53382f762cf916ee92286a9965. * Revert "attempt parallel pg_dump" This reverts commit 592a12795428000ab15b572f89cbd9c6e25fb76c. * improve diff check * finish fixing migrations * stuff * use advisory lock * stuff * Update lib.rs * fmt * fmt * clippy * Update diff_check.rs * Update .woodpecker.yml * Update lib.rs * Update lib.rs * Update lib.rs * Update .woodpecker.yml * Update .woodpecker.yml * Update lib.rs * re-run ci * fmt * fmt * Update .woodpecker.yml * Update .woodpecker.yml * create separate database in ci * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * try to fix env var * Update diff_check.rs * Remove condition that's not needed anymore * clippy * exclude views and fast tables * revert some migration changes * fix * fmt * re-attempt checking character after skipped trigger name, and make code less confusing * fmt * fix * rerun ci * rerun ci * fix strip_prefix order * fix weird big Cargo.lock change by running `git checkout upstream/main Cargo.lock` then letting it auto update again * fix * remove installation commands that were removed in main branch * Revert "remove installation commands that were removed in main branch" This reverts commit fd65234a760b8b1bb42bc5b59b0ec972b08ded13. * move create_database_user woodpecker step to make diff less weird * fix clippy * Make diff check work just like before * Move new migrations to the end * Revert changes to old migrations * don't assume that migrations are already sorted * retry CI * fix merge * find migrations dir in debug mode using CARGO_MANIFEST_DIR variable instead of current working directory * always use embedded migrations * improve doc comments for migration subcommand * clippy fix * move cfg(test) attribute to diff_check.rs * copy `o` variable instead of calling `o` function * use chrono::TimeDelta Display implementation to show migration duration
2025-01-08 10:07:32 +00:00
use clap::{Parser, Subcommand};
use lemmy_api_common::{
context::LemmyContext,
lemmy_db_views::structs::SiteView,
request::client_builder,
send_activity::{ActivityChannel, MATCH_OUTGOING_ACTIVITIES},
utils::{
check_private_instance_and_federation_enabled,
local_site_rate_limit_to_rate_limit_config,
},
};
use lemmy_apub::{
activities::{handle_outgoing_activities, match_outgoing_activities},
objects::instance::ApubSite,
VerifyUrlData,
FEDERATION_HTTP_FETCH_LIMIT,
};
Add custom migration runner, forbid some `diesel migration` commands, fix old migrations (#4673) * Update schema.rs * rename * stuff * finish new implementation of schema_setup::run (not including revert, test, etc.) * fmt * refactor * fix sql * migriation run command * use trigger on migrations table * add Options with disable_migrations field for test * rename to enable_forbid_diesel_cli_trigger * fix * fix merge * diff_checker (partial) * Revert "diff_checker (partial)" This reverts commit 6709882e148ceff48d6f097895af18d13fbdf4bf. * Revert "Revert "diff_checker (partial)"" This reverts commit d4bdda5d11216f4acd13b9585d2392ca8c252a73. * diff check * improve schema diff * timestamp replacement * ignore column order * remove fedi_name default * stuff * improve diff * stuff * attempt parallel pg_dump * attempt 2 * Revert "attempt 2" This reverts commit a909d2d6438d0b53382f762cf916ee92286a9965. * Revert "attempt parallel pg_dump" This reverts commit 592a12795428000ab15b572f89cbd9c6e25fb76c. * improve diff check * finish fixing migrations * stuff * use advisory lock * stuff * Update lib.rs * fmt * fmt * clippy * Update diff_check.rs * Update .woodpecker.yml * Update lib.rs * Update lib.rs * Update lib.rs * Update .woodpecker.yml * Update .woodpecker.yml * Update lib.rs * re-run ci * fmt * fmt * Update .woodpecker.yml * Update .woodpecker.yml * create separate database in ci * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * try to fix env var * Update diff_check.rs * Remove condition that's not needed anymore * clippy * exclude views and fast tables * revert some migration changes * fix * fmt * re-attempt checking character after skipped trigger name, and make code less confusing * fmt * fix * rerun ci * rerun ci * fix strip_prefix order * fix weird big Cargo.lock change by running `git checkout upstream/main Cargo.lock` then letting it auto update again * fix * remove installation commands that were removed in main branch * Revert "remove installation commands that were removed in main branch" This reverts commit fd65234a760b8b1bb42bc5b59b0ec972b08ded13. * move create_database_user woodpecker step to make diff less weird * fix clippy * Make diff check work just like before * Move new migrations to the end * Revert changes to old migrations * don't assume that migrations are already sorted * retry CI * fix merge * find migrations dir in debug mode using CARGO_MANIFEST_DIR variable instead of current working directory * always use embedded migrations * improve doc comments for migration subcommand * clippy fix * move cfg(test) attribute to diff_check.rs * copy `o` variable instead of calling `o` function * use chrono::TimeDelta Display implementation to show migration duration
2025-01-08 10:07:32 +00:00
use lemmy_db_schema::{schema_setup, source::secret::Secret, utils::build_db_pool};
use lemmy_federate::{Opts, SendManager};
use lemmy_routes::{feeds, images, nodeinfo, webfinger};
use lemmy_utils::{
error::{LemmyErrorType, LemmyResult},
rate_limit::RateLimitCell,
Fix #3366: Wrap plain-text error responses from the API in JSON (#3559) * Fix #3366: API does return plain HTML errors * Fix Clippy errors * Improve api response times by doing send_activity asynchronously (#3493) * do send_activity after http response * move to util function * format * fix prometheus * make synchronous federation configurable * cargo fmt * empty * empty --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> * Updating `login.rs` with generic `incorrect_login` response. (#3549) * Adding v0.18.1 and v0.18.0 release notes. (#3530) * Update RELEASES.md (#3556) added instruction to find the location of your docker directory (especially useful for those who used ansible since they never had to setup docker manually) * Use async email sender (#3554) * Upgrade all dependencies (#3526) * Upgrade all dependencies * as base64 * Adding phiresky to codeowners. (#3576) * Error enum fixed (#3487) * Create error type enum * Replace magic string slices with LemmyErrorTypes * Remove unused enum * Add rename snake case to error enum * Rename functions * clippy * Fix merge errors * Serialize in PascalCase instead of snake_case * Revert src/lib * Add serialization tests * Update translations * Fix compilation error in test * Fix another compilation error * Add code for generating typescript types * Various fixes to avoid breaking api * impl From<LemmyErrorType> for LemmyError * with_lemmy_type * trigger ci --------- Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> * Only update site_aggregates for local site (#3516) * Fix #3501 - Fix aggregation counts for elements removed and deleted (#3543) Two bugs were found and fixed: - previously elements removal and deletion were counted as two separate disappearances - removing comments did not affect post aggregations * Use LemmyErrorType also make error_type compulsory * Add missing import for jsonify_plain_text_errors * Fix formatting * Trying to make woodpecker run again --------- Co-authored-by: phiresky <phireskyde+git@gmail.com> Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: rosenjcb <rosenjcb@gmail.com> Co-authored-by: nixoye <12674582+nixoye@users.noreply.github.com> Co-authored-by: dullbananas <dull.bananas0@gmail.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> Co-authored-by: Sander Saarend <sander@saarend.com> Co-authored-by: Piotr Juszczyk <74842304+pijuszczyk@users.noreply.github.com>
2023-07-10 20:44:14 +00:00
response::jsonify_plain_text_errors,
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
settings::{structs::Settings, SETTINGS},
VERSION,
};
use prometheus::default_registry;
use prometheus_metrics::serve_prometheus;
Rewrite images to use local proxy (#4035) * Add markdown rule to add rel=nofollow for all links * Add markdown image rule to add local image proxy (fixes #1036) * comments * rewrite markdown image links working * add comment * perform markdown image processing in api/apub receivers * clippy * add db table to validate proxied links * rewrite link fields for avatar, banner etc * sql fmt * proxy links received over federation * add config option * undo post.url rewriting, move http route definition * add tests * proxy images through pictrs * testing * cleanup request.rs file * more cleanup (fixes #2611) * include url content type when sending post over apub (fixes #2611) * store post url content type in db * should be media_type * get rid of cache_remote_thumbnails setting, instead automatically take thumbnail from federation data if available. * fix tests * add setting disable_external_link_previews * federate post url as image depending on mime type * change setting again * machete * invert * support custom emoji * clippy * update defaults * add image proxy test, fix test * fix test * clippy * revert accidental changes * address review * clippy * Markdown link rule-dess (#4356) * Extracting opengraph_data to its own type. * A few additions for markdown-link-rule. --------- Co-authored-by: Nutomic <me@nutomic.com> * fix setting * use enum for image proxy setting * fix test configs * add config backwards compat * clippy * machete --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com>
2024-01-25 14:22:11 +00:00
use reqwest_middleware::ClientBuilder;
use reqwest_tracing::TracingMiddleware;
use serde_json::json;
use std::{ops::Deref, time::Duration};
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
use tokio::signal::unix::SignalKind;
use tracing_actix_web::{DefaultRootSpanBuilder, TracingLogger};
2021-11-23 12:16:47 +00:00
/// Timeout for HTTP requests while sending activities. A longer timeout provides better
/// compatibility with other ActivityPub software that might allocate more time for synchronous
/// processing of incoming activities. This timeout should be slightly longer than the time we
/// expect a remote server to wait before aborting processing on its own to account for delays from
/// establishing the HTTP connection and sending the request itself.
const ACTIVITY_SENDING_TIMEOUT: Duration = Duration::from_secs(125);
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
#[derive(Parser, Debug)]
#[command(
version,
about = "A link aggregator for the fediverse",
long_about = "A link aggregator for the fediverse.\n\nThis is the Lemmy backend API server. This will connect to a PostgreSQL database, run any pending migrations and start accepting API requests."
)]
// TODO: Instead of defining individual env vars, only specify prefix once supported by clap.
// https://github.com/clap-rs/clap/issues/3221
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
pub struct CmdArgs {
/// Don't run scheduled tasks.
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
///
/// If you are running multiple Lemmy server processes, you probably want to disable scheduled
/// tasks on all but one of the processes, to avoid running the tasks more often than intended.
#[arg(long, default_value_t = false, env = "LEMMY_DISABLE_SCHEDULED_TASKS")]
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
disable_scheduled_tasks: bool,
/// Disables the HTTP server.
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
///
/// This can be used to run a Lemmy server process that only performs scheduled tasks or activity
/// sending.
#[arg(long, default_value_t = false, env = "LEMMY_DISABLE_HTTP_SERVER")]
disable_http_server: bool,
/// Disable sending outgoing ActivityPub messages.
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
///
/// Only pass this for horizontally scaled setups.
/// See https://join-lemmy.org/docs/administration/horizontal_scaling.html for details.
#[arg(long, default_value_t = false, env = "LEMMY_DISABLE_ACTIVITY_SENDING")]
disable_activity_sending: bool,
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
/// The index of this outgoing federation process.
///
/// Defaults to 1/1. If you want to split the federation workload onto n servers, run each server
/// 1≤i≤n with these args: --federate-process-index i --federate-process-count n
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
///
/// Make you have exactly one server with each `i` running, otherwise federation will randomly
/// send duplicates or nothing.
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
///
/// See https://join-lemmy.org/docs/administration/horizontal_scaling.html for more detail.
#[arg(long, default_value_t = 1, env = "LEMMY_FEDERATE_PROCESS_INDEX")]
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
federate_process_index: i32,
/// How many outgoing federation processes you are starting in total.
///
/// If set, make sure to set --federate-process-index differently for each.
#[arg(long, default_value_t = 1, env = "LEMMY_FEDERATE_PROCESS_COUNT")]
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
federate_process_count: i32,
Add custom migration runner, forbid some `diesel migration` commands, fix old migrations (#4673) * Update schema.rs * rename * stuff * finish new implementation of schema_setup::run (not including revert, test, etc.) * fmt * refactor * fix sql * migriation run command * use trigger on migrations table * add Options with disable_migrations field for test * rename to enable_forbid_diesel_cli_trigger * fix * fix merge * diff_checker (partial) * Revert "diff_checker (partial)" This reverts commit 6709882e148ceff48d6f097895af18d13fbdf4bf. * Revert "Revert "diff_checker (partial)"" This reverts commit d4bdda5d11216f4acd13b9585d2392ca8c252a73. * diff check * improve schema diff * timestamp replacement * ignore column order * remove fedi_name default * stuff * improve diff * stuff * attempt parallel pg_dump * attempt 2 * Revert "attempt 2" This reverts commit a909d2d6438d0b53382f762cf916ee92286a9965. * Revert "attempt parallel pg_dump" This reverts commit 592a12795428000ab15b572f89cbd9c6e25fb76c. * improve diff check * finish fixing migrations * stuff * use advisory lock * stuff * Update lib.rs * fmt * fmt * clippy * Update diff_check.rs * Update .woodpecker.yml * Update lib.rs * Update lib.rs * Update lib.rs * Update .woodpecker.yml * Update .woodpecker.yml * Update lib.rs * re-run ci * fmt * fmt * Update .woodpecker.yml * Update .woodpecker.yml * create separate database in ci * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * try to fix env var * Update diff_check.rs * Remove condition that's not needed anymore * clippy * exclude views and fast tables * revert some migration changes * fix * fmt * re-attempt checking character after skipped trigger name, and make code less confusing * fmt * fix * rerun ci * rerun ci * fix strip_prefix order * fix weird big Cargo.lock change by running `git checkout upstream/main Cargo.lock` then letting it auto update again * fix * remove installation commands that were removed in main branch * Revert "remove installation commands that were removed in main branch" This reverts commit fd65234a760b8b1bb42bc5b59b0ec972b08ded13. * move create_database_user woodpecker step to make diff less weird * fix clippy * Make diff check work just like before * Move new migrations to the end * Revert changes to old migrations * don't assume that migrations are already sorted * retry CI * fix merge * find migrations dir in debug mode using CARGO_MANIFEST_DIR variable instead of current working directory * always use embedded migrations * improve doc comments for migration subcommand * clippy fix * move cfg(test) attribute to diff_check.rs * copy `o` variable instead of calling `o` function * use chrono::TimeDelta Display implementation to show migration duration
2025-01-08 10:07:32 +00:00
#[command(subcommand)]
subcommand: Option<CmdSubcommand>,
}
#[derive(Subcommand, Debug)]
enum CmdSubcommand {
/// Do something with migrations, then exit.
Migration {
#[command(subcommand)]
subcommand: MigrationSubcommand,
/// Stop after there's no remaining migrations.
#[arg(long, default_value_t = false)]
all: bool,
/// Stop after the given number of migrations.
#[arg(long, default_value_t = 1)]
number: u64,
},
}
#[derive(Subcommand, Debug)]
enum MigrationSubcommand {
/// Run up.sql for pending migrations, oldest to newest.
Run,
/// Run down.sql for non-pending migrations, newest to oldest.
Revert,
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
}
/// Placing the main function in lib.rs allows other crates to import it and embed Lemmy
pub async fn start_lemmy_server(args: CmdArgs) -> LemmyResult<()> {
// Print version number to log
println!("Starting Lemmy v{VERSION}");
Add custom migration runner, forbid some `diesel migration` commands, fix old migrations (#4673) * Update schema.rs * rename * stuff * finish new implementation of schema_setup::run (not including revert, test, etc.) * fmt * refactor * fix sql * migriation run command * use trigger on migrations table * add Options with disable_migrations field for test * rename to enable_forbid_diesel_cli_trigger * fix * fix merge * diff_checker (partial) * Revert "diff_checker (partial)" This reverts commit 6709882e148ceff48d6f097895af18d13fbdf4bf. * Revert "Revert "diff_checker (partial)"" This reverts commit d4bdda5d11216f4acd13b9585d2392ca8c252a73. * diff check * improve schema diff * timestamp replacement * ignore column order * remove fedi_name default * stuff * improve diff * stuff * attempt parallel pg_dump * attempt 2 * Revert "attempt 2" This reverts commit a909d2d6438d0b53382f762cf916ee92286a9965. * Revert "attempt parallel pg_dump" This reverts commit 592a12795428000ab15b572f89cbd9c6e25fb76c. * improve diff check * finish fixing migrations * stuff * use advisory lock * stuff * Update lib.rs * fmt * fmt * clippy * Update diff_check.rs * Update .woodpecker.yml * Update lib.rs * Update lib.rs * Update lib.rs * Update .woodpecker.yml * Update .woodpecker.yml * Update lib.rs * re-run ci * fmt * fmt * Update .woodpecker.yml * Update .woodpecker.yml * create separate database in ci * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * try to fix env var * Update diff_check.rs * Remove condition that's not needed anymore * clippy * exclude views and fast tables * revert some migration changes * fix * fmt * re-attempt checking character after skipped trigger name, and make code less confusing * fmt * fix * rerun ci * rerun ci * fix strip_prefix order * fix weird big Cargo.lock change by running `git checkout upstream/main Cargo.lock` then letting it auto update again * fix * remove installation commands that were removed in main branch * Revert "remove installation commands that were removed in main branch" This reverts commit fd65234a760b8b1bb42bc5b59b0ec972b08ded13. * move create_database_user woodpecker step to make diff less weird * fix clippy * Make diff check work just like before * Move new migrations to the end * Revert changes to old migrations * don't assume that migrations are already sorted * retry CI * fix merge * find migrations dir in debug mode using CARGO_MANIFEST_DIR variable instead of current working directory * always use embedded migrations * improve doc comments for migration subcommand * clippy fix * move cfg(test) attribute to diff_check.rs * copy `o` variable instead of calling `o` function * use chrono::TimeDelta Display implementation to show migration duration
2025-01-08 10:07:32 +00:00
if let Some(CmdSubcommand::Migration {
subcommand,
all,
number,
}) = args.subcommand
{
let mut options = match subcommand {
MigrationSubcommand::Run => schema_setup::Options::default().run(),
MigrationSubcommand::Revert => schema_setup::Options::default().revert(),
};
if !all {
options = options.limit(number);
}
schema_setup::run(options)?;
return Ok(());
}
// return error 503 while running db migrations and startup tasks
let mut startup_server_handle = None;
if !args.disable_http_server {
startup_server_handle = Some(create_startup_server()?);
}
// Set up the connection pool
let pool = build_db_pool()?;
// Run the Code-required migrations
run_advanced_migrations(&mut (&pool).into(), &SETTINGS).await?;
// Initialize the secrets
Replace `clippy::unwrap_used` in tests (#5064) * Add LemmyResult to session_middleware tests * Add LemmyResult to inboxes tests * Add LemmyResult to slurs tests * Add LemmyResult to markdown tests * Add LemmyResult to rate_limiter tests * Add LemmyResult to error tests * Add LemmyResult to api_common utils tests * Add LemmyResult to request tests * Add LemmyResult to claims tests * Propagate registration_applications errors * Remove clippy::unwrap_used from community tests * Add LemmyResult to community_view tests * Add LemmyResult to db_schema post tests * Add LemmyResult to site_aggregates tests * Add LemmyResult to private_message tests * Add LemmyResult to activity tests * Add LemmyResult to federation_allowlist tests * Add LemmyResult to comment_aggregates tests * Add LemmyResult to post_report tests * Add LemmyResult to moderator tests * Add LemmyResult to community_aggregates tests * Add LemmyResult to person_aggregates tests * Add LemmyResult to language tests * Add LemmyResult to post_aggregates tests * Add LemmyResult to db_schema comment tests * Add LemmyResult to actor_language tests * Add LemmyResult to vote_view tests * Add LemmyResult to registration_application_view tests * Add LemmyResult to private_message_view tests * Add LemmyResult to private_message_report_view tests * Add LemmyResult to post_report_view tests * Add LemmyResult to comment_report_view tests * Add LemmyResult to sitemap tests * Replace .expect() with .unwrap() * Format code * Remove clippy::unwrap_used from activity tests * Add diesel result in db_schema tests * Format code * Map to_bytes() error to LemmyErrorType * Remove clippy::unwrap_used from error tests * Removing a few more unwraps, and cleaning up language code. * Replace map_err with unwrap_or_default * Replace ok_or with and_then --------- Co-authored-by: Dessalines <tyhou13@gmx.com>
2024-10-02 12:50:21 +00:00
let secret = Secret::init(&mut (&pool).into()).await?;
// Make sure the local site is set up.
let site_view = SiteView::read_local(&mut (&pool).into()).await?;
let local_site = site_view.local_site;
let federation_enabled = local_site.federation_enabled;
if federation_enabled {
println!("Federation enabled, host is {}", &SETTINGS.hostname);
}
check_private_instance_and_federation_enabled(&local_site)?;
// Set up the rate limiter
let rate_limit_config =
local_site_rate_limit_to_rate_limit_config(&site_view.local_site_rate_limit);
Refactor rate limiter and improve rate limit bucket cleanup (#3937) * Update rate_limiter.rs * Update mod.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update mod.rs * Update scheduled_tasks.rs * Shrink `RateLimitBucket` * Update rate_limiter.rs * Update mod.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update mod.rs * Update rate_limiter.rs * fmt * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * rerun ci * Update rate_limiter.rs * Undo changes to fields * Manually undo changes to RateLimitBucket fields * fmt * Bucket cleanup loop in rate_limit/mod.rs * Remove rate limit bucket cleanup from scheduled_tasks.rs * Remove ; * Remove UNINITIALIZED_TOKEN_AMOUNT * Update rate_limiter.rs * fmt * Update rate_limiter.rs * fmt * Update rate_limiter.rs * fmt * Update rate_limiter.rs * stuff * MapLevel trait * fix merge * Prevent negative numbers in buckets * Clean up MapLevel::check * MapLevel::remove_full_buckets * stuff * Use remove_full_buckets to avoid allocations * stuff * remove tx * Remove RateLimitConfig * Rename settings_updated_channel to rate_limit_cell * Remove global rate limit cell * impl Default for RateLimitCell * bucket_configs doc comment to explain EnumMap * improve test_rate_limiter * rename default to with_test_config --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com>
2023-10-19 13:31:51 +00:00
let rate_limit_cell = RateLimitCell::new(rate_limit_config);
println!(
"Starting HTTP server at {}:{}",
SETTINGS.bind, SETTINGS.port
);
let client = ClientBuilder::new(client_builder(&SETTINGS).build()?)
.with(TracingMiddleware::default())
.build();
let context = LemmyContext::create(
pool.clone(),
client.clone(),
secret.clone(),
rate_limit_cell.clone(),
);
if let Some(prometheus) = SETTINGS.prometheus.clone() {
serve_prometheus(prometheus, context.clone())?;
}
let mut federation_config_builder = FederationConfig::builder();
federation_config_builder
.domain(SETTINGS.hostname.clone())
.app_data(context.clone())
.client(client.clone())
.http_fetch_limit(FEDERATION_HTTP_FETCH_LIMIT)
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
.debug(cfg!(debug_assertions))
.http_signature_compat(true)
.url_verifier(Box::new(VerifyUrlData(context.inner_pool().clone())));
if local_site.federation_signed_fetch {
let site: ApubSite = site_view.site.into();
federation_config_builder.signed_fetch_actor(&site);
}
let federation_config = federation_config_builder.build().await?;
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
MATCH_OUTGOING_ACTIVITIES
.set(Box::new(move |d, c| {
Box::pin(match_outgoing_activities(d, c))
}))
.map_err(|_e| LemmyErrorType::Unknown("couldnt set function pointer".into()))?;
let request_data = federation_config.to_request_data();
let outgoing_activities_task = tokio::task::spawn(handle_outgoing_activities(
request_data.reset_request_count(),
));
Add custom migration runner, forbid some `diesel migration` commands, fix old migrations (#4673) * Update schema.rs * rename * stuff * finish new implementation of schema_setup::run (not including revert, test, etc.) * fmt * refactor * fix sql * migriation run command * use trigger on migrations table * add Options with disable_migrations field for test * rename to enable_forbid_diesel_cli_trigger * fix * fix merge * diff_checker (partial) * Revert "diff_checker (partial)" This reverts commit 6709882e148ceff48d6f097895af18d13fbdf4bf. * Revert "Revert "diff_checker (partial)"" This reverts commit d4bdda5d11216f4acd13b9585d2392ca8c252a73. * diff check * improve schema diff * timestamp replacement * ignore column order * remove fedi_name default * stuff * improve diff * stuff * attempt parallel pg_dump * attempt 2 * Revert "attempt 2" This reverts commit a909d2d6438d0b53382f762cf916ee92286a9965. * Revert "attempt parallel pg_dump" This reverts commit 592a12795428000ab15b572f89cbd9c6e25fb76c. * improve diff check * finish fixing migrations * stuff * use advisory lock * stuff * Update lib.rs * fmt * fmt * clippy * Update diff_check.rs * Update .woodpecker.yml * Update lib.rs * Update lib.rs * Update lib.rs * Update .woodpecker.yml * Update .woodpecker.yml * Update lib.rs * re-run ci * fmt * fmt * Update .woodpecker.yml * Update .woodpecker.yml * create separate database in ci * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * try to fix env var * Update diff_check.rs * Remove condition that's not needed anymore * clippy * exclude views and fast tables * revert some migration changes * fix * fmt * re-attempt checking character after skipped trigger name, and make code less confusing * fmt * fix * rerun ci * rerun ci * fix strip_prefix order * fix weird big Cargo.lock change by running `git checkout upstream/main Cargo.lock` then letting it auto update again * fix * remove installation commands that were removed in main branch * Revert "remove installation commands that were removed in main branch" This reverts commit fd65234a760b8b1bb42bc5b59b0ec972b08ded13. * move create_database_user woodpecker step to make diff less weird * fix clippy * Make diff check work just like before * Move new migrations to the end * Revert changes to old migrations * don't assume that migrations are already sorted * retry CI * fix merge * find migrations dir in debug mode using CARGO_MANIFEST_DIR variable instead of current working directory * always use embedded migrations * improve doc comments for migration subcommand * clippy fix * move cfg(test) attribute to diff_check.rs * copy `o` variable instead of calling `o` function * use chrono::TimeDelta Display implementation to show migration duration
2025-01-08 10:07:32 +00:00
if !args.disable_scheduled_tasks {
// Schedules various cleanup tasks for the DB
Add custom migration runner, forbid some `diesel migration` commands, fix old migrations (#4673) * Update schema.rs * rename * stuff * finish new implementation of schema_setup::run (not including revert, test, etc.) * fmt * refactor * fix sql * migriation run command * use trigger on migrations table * add Options with disable_migrations field for test * rename to enable_forbid_diesel_cli_trigger * fix * fix merge * diff_checker (partial) * Revert "diff_checker (partial)" This reverts commit 6709882e148ceff48d6f097895af18d13fbdf4bf. * Revert "Revert "diff_checker (partial)"" This reverts commit d4bdda5d11216f4acd13b9585d2392ca8c252a73. * diff check * improve schema diff * timestamp replacement * ignore column order * remove fedi_name default * stuff * improve diff * stuff * attempt parallel pg_dump * attempt 2 * Revert "attempt 2" This reverts commit a909d2d6438d0b53382f762cf916ee92286a9965. * Revert "attempt parallel pg_dump" This reverts commit 592a12795428000ab15b572f89cbd9c6e25fb76c. * improve diff check * finish fixing migrations * stuff * use advisory lock * stuff * Update lib.rs * fmt * fmt * clippy * Update diff_check.rs * Update .woodpecker.yml * Update lib.rs * Update lib.rs * Update lib.rs * Update .woodpecker.yml * Update .woodpecker.yml * Update lib.rs * re-run ci * fmt * fmt * Update .woodpecker.yml * Update .woodpecker.yml * create separate database in ci * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * try to fix env var * Update diff_check.rs * Remove condition that's not needed anymore * clippy * exclude views and fast tables * revert some migration changes * fix * fmt * re-attempt checking character after skipped trigger name, and make code less confusing * fmt * fix * rerun ci * rerun ci * fix strip_prefix order * fix weird big Cargo.lock change by running `git checkout upstream/main Cargo.lock` then letting it auto update again * fix * remove installation commands that were removed in main branch * Revert "remove installation commands that were removed in main branch" This reverts commit fd65234a760b8b1bb42bc5b59b0ec972b08ded13. * move create_database_user woodpecker step to make diff less weird * fix clippy * Make diff check work just like before * Move new migrations to the end * Revert changes to old migrations * don't assume that migrations are already sorted * retry CI * fix merge * find migrations dir in debug mode using CARGO_MANIFEST_DIR variable instead of current working directory * always use embedded migrations * improve doc comments for migration subcommand * clippy fix * move cfg(test) attribute to diff_check.rs * copy `o` variable instead of calling `o` function * use chrono::TimeDelta Display implementation to show migration duration
2025-01-08 10:07:32 +00:00
let _scheduled_tasks =
tokio::task::spawn(scheduled_tasks::setup(request_data.reset_request_count()));
}
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
let server = if !args.disable_http_server {
if let Some(startup_server_handle) = startup_server_handle {
startup_server_handle.stop(true).await;
}
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
Some(create_http_server(
federation_config.clone(),
SETTINGS.clone(),
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
federation_enabled,
)?)
} else {
None
};
// This FederationConfig instance is exclusively used to send activities, so we can safely
// increase the timeout without affecting timeouts for resolving objects anywhere.
let federation_sender_config = if !args.disable_activity_sending {
let mut federation_sender_config = federation_config_builder.clone();
federation_sender_config.request_timeout(ACTIVITY_SENDING_TIMEOUT);
Some(federation_sender_config.build().await?)
} else {
None
};
let federate = federation_sender_config.map(|cfg| {
SendManager::run(
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
Opts {
process_index: args.federate_process_index,
process_count: args.federate_process_count,
},
cfg,
federation: parallel sending per instance (#4623) * federation: parallel sending * federation: some comments * lint and set force_write true when a request fails * inbox_urls return vec * split inbox functions into separate file * cleanup * extract sending task code to separate file * move federation concurrent config to config file * off by one issue * improve msg * fix both permanent stopping of federation queues and multiple creation of the same federation queues * fix after merge * lint fix * Update crates/federate/src/send.rs Co-authored-by: dullbananas <dull.bananas0@gmail.com> * comment about reverse ordering * remove crashable, comment * comment * move comment * run federation tests twice * fix test run * prettier * fix config default * upgrade rust to 1.78 to fix diesel cli * fix clippy * delay * add debug to make localhost urls not valid in ap crate, add some debug logs * federation tests: ensure server stop after test and random activity id * ci fix * add test to federate 100 events * fix send 100 test * different data every time so activities are distinguishable * allow out of order receives in test * lint * comment about https://github.com/LemmyNet/lemmy/pull/4623#discussion_r1565437391 * move sender for clarity, add comment * move more things to members * update test todo comment, use same env var as worker test but default to 1 * remove else below continue * some more cleanup * handle todo about smooth exit * add federate inboxes collector tests * lint * actor max length * don't reset fail count if activity skipped * fix some comments * reuse vars * format * Update .woodpecker.yml * fix recheck time * fix inboxes tests under fast mode * format * make i32 and ugly casts * clippy --------- Co-authored-by: dullbananas <dull.bananas0@gmail.com>
2024-07-21 15:50:50 +00:00
SETTINGS.federation.clone(),
)
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
});
let mut interrupt = tokio::signal::unix::signal(SignalKind::interrupt())?;
let mut terminate = tokio::signal::unix::signal(SignalKind::terminate())?;
Add custom migration runner, forbid some `diesel migration` commands, fix old migrations (#4673) * Update schema.rs * rename * stuff * finish new implementation of schema_setup::run (not including revert, test, etc.) * fmt * refactor * fix sql * migriation run command * use trigger on migrations table * add Options with disable_migrations field for test * rename to enable_forbid_diesel_cli_trigger * fix * fix merge * diff_checker (partial) * Revert "diff_checker (partial)" This reverts commit 6709882e148ceff48d6f097895af18d13fbdf4bf. * Revert "Revert "diff_checker (partial)"" This reverts commit d4bdda5d11216f4acd13b9585d2392ca8c252a73. * diff check * improve schema diff * timestamp replacement * ignore column order * remove fedi_name default * stuff * improve diff * stuff * attempt parallel pg_dump * attempt 2 * Revert "attempt 2" This reverts commit a909d2d6438d0b53382f762cf916ee92286a9965. * Revert "attempt parallel pg_dump" This reverts commit 592a12795428000ab15b572f89cbd9c6e25fb76c. * improve diff check * finish fixing migrations * stuff * use advisory lock * stuff * Update lib.rs * fmt * fmt * clippy * Update diff_check.rs * Update .woodpecker.yml * Update lib.rs * Update lib.rs * Update lib.rs * Update .woodpecker.yml * Update .woodpecker.yml * Update lib.rs * re-run ci * fmt * fmt * Update .woodpecker.yml * Update .woodpecker.yml * create separate database in ci * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * try to fix env var * Update diff_check.rs * Remove condition that's not needed anymore * clippy * exclude views and fast tables * revert some migration changes * fix * fmt * re-attempt checking character after skipped trigger name, and make code less confusing * fmt * fix * rerun ci * rerun ci * fix strip_prefix order * fix weird big Cargo.lock change by running `git checkout upstream/main Cargo.lock` then letting it auto update again * fix * remove installation commands that were removed in main branch * Revert "remove installation commands that were removed in main branch" This reverts commit fd65234a760b8b1bb42bc5b59b0ec972b08ded13. * move create_database_user woodpecker step to make diff less weird * fix clippy * Make diff check work just like before * Move new migrations to the end * Revert changes to old migrations * don't assume that migrations are already sorted * retry CI * fix merge * find migrations dir in debug mode using CARGO_MANIFEST_DIR variable instead of current working directory * always use embedded migrations * improve doc comments for migration subcommand * clippy fix * move cfg(test) attribute to diff_check.rs * copy `o` variable instead of calling `o` function * use chrono::TimeDelta Display implementation to show migration duration
2025-01-08 10:07:32 +00:00
tokio::select! {
_ = tokio::signal::ctrl_c() => {
tracing::warn!("Received ctrl-c, shutting down gracefully...");
}
_ = interrupt.recv() => {
tracing::warn!("Received interrupt, shutting down gracefully...");
}
_ = terminate.recv() => {
tracing::warn!("Received terminate, shutting down gracefully...");
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
}
}
if let Some(server) = server {
server.stop(true).await;
}
if let Some(federate) = federate {
federate.cancel().await?;
}
// Wait for outgoing apub sends to complete
ActivityChannel::close(outgoing_activities_task).await?;
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
Ok(())
}
/// Creates temporary HTTP server which returns status 503 for all requests.
fn create_startup_server() -> LemmyResult<ServerHandle> {
let startup_server = HttpServer::new(move || {
App::new().wrap(ErrorHandlers::new().default_handler(move |req| {
let (req, _) = req.into_parts();
let response =
HttpResponse::ServiceUnavailable().json(json!({"error": "Lemmy is currently starting"}));
let service_response = ServiceResponse::new(req, response);
Ok(ErrorHandlerResponse::Response(
service_response.map_into_right_body(),
))
}))
})
.bind((SETTINGS.bind, SETTINGS.port))?
.run();
let startup_server_handle = startup_server.handle();
tokio::task::spawn(startup_server);
Ok(startup_server_handle)
}
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
fn create_http_server(
federation_config: FederationConfig<LemmyContext>,
settings: Settings,
federation_enabled: bool,
) -> LemmyResult<ServerHandle> {
// this must come before the HttpServer creation
// creates a middleware that populates http metrics for each path, method, and status code
let prom_api_metrics = PrometheusMetricsBuilder::new("lemmy_api")
.registry(default_registry().clone())
.build()
.map_err(|e| LemmyErrorType::Unknown(format!("Should always be buildable: {e}")))?;
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
let context: LemmyContext = federation_config.deref().clone();
Refactor rate limiter and improve rate limit bucket cleanup (#3937) * Update rate_limiter.rs * Update mod.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update mod.rs * Update scheduled_tasks.rs * Shrink `RateLimitBucket` * Update rate_limiter.rs * Update mod.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update mod.rs * Update rate_limiter.rs * fmt * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * Update rate_limiter.rs * rerun ci * Update rate_limiter.rs * Undo changes to fields * Manually undo changes to RateLimitBucket fields * fmt * Bucket cleanup loop in rate_limit/mod.rs * Remove rate limit bucket cleanup from scheduled_tasks.rs * Remove ; * Remove UNINITIALIZED_TOKEN_AMOUNT * Update rate_limiter.rs * fmt * Update rate_limiter.rs * fmt * Update rate_limiter.rs * fmt * Update rate_limiter.rs * stuff * MapLevel trait * fix merge * Prevent negative numbers in buckets * Clean up MapLevel::check * MapLevel::remove_full_buckets * stuff * Use remove_full_buckets to avoid allocations * stuff * remove tx * Remove RateLimitConfig * Rename settings_updated_channel to rate_limit_cell * Remove global rate limit cell * impl Default for RateLimitCell * bucket_configs doc comment to explain EnumMap * improve test_rate_limiter * rename default to with_test_config --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com>
2023-10-19 13:31:51 +00:00
let rate_limit_cell = federation_config.rate_limit_cell().clone();
Rewrite images to use local proxy (#4035) * Add markdown rule to add rel=nofollow for all links * Add markdown image rule to add local image proxy (fixes #1036) * comments * rewrite markdown image links working * add comment * perform markdown image processing in api/apub receivers * clippy * add db table to validate proxied links * rewrite link fields for avatar, banner etc * sql fmt * proxy links received over federation * add config option * undo post.url rewriting, move http route definition * add tests * proxy images through pictrs * testing * cleanup request.rs file * more cleanup (fixes #2611) * include url content type when sending post over apub (fixes #2611) * store post url content type in db * should be media_type * get rid of cache_remote_thumbnails setting, instead automatically take thumbnail from federation data if available. * fix tests * add setting disable_external_link_previews * federate post url as image depending on mime type * change setting again * machete * invert * support custom emoji * clippy * update defaults * add image proxy test, fix test * fix test * clippy * revert accidental changes * address review * clippy * Markdown link rule-dess (#4356) * Extracting opengraph_data to its own type. * A few additions for markdown-link-rule. --------- Co-authored-by: Nutomic <me@nutomic.com> * fix setting * use enum for image proxy setting * fix test configs * add config backwards compat * clippy * machete --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com>
2024-01-25 14:22:11 +00:00
// Pictrs cannot use proxy
let pictrs_client = ClientBuilder::new(client_builder(&SETTINGS).no_proxy().build()?)
.with(TracingMiddleware::default())
.build();
// Create Http server
let bind = (settings.bind, settings.port);
let server = HttpServer::new(move || {
let cors_config = cors_config(&settings);
let app = App::new()
.wrap(middleware::Logger::new(
// This is the default log format save for the usage of %{r}a over %a to guarantee to
// record the client's (forwarded) IP and not the last peer address, since the latter is
// frequently just a reverse proxy
"%{r}a '%r' %s %b '%{Referer}i' '%{User-Agent}i' %T",
))
2023-06-26 10:54:41 +00:00
.wrap(middleware::Compress::default())
.wrap(cors_config)
.wrap(TracingLogger::<DefaultRootSpanBuilder>::new())
Fix #3366: Wrap plain-text error responses from the API in JSON (#3559) * Fix #3366: API does return plain HTML errors * Fix Clippy errors * Improve api response times by doing send_activity asynchronously (#3493) * do send_activity after http response * move to util function * format * fix prometheus * make synchronous federation configurable * cargo fmt * empty * empty --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> * Updating `login.rs` with generic `incorrect_login` response. (#3549) * Adding v0.18.1 and v0.18.0 release notes. (#3530) * Update RELEASES.md (#3556) added instruction to find the location of your docker directory (especially useful for those who used ansible since they never had to setup docker manually) * Use async email sender (#3554) * Upgrade all dependencies (#3526) * Upgrade all dependencies * as base64 * Adding phiresky to codeowners. (#3576) * Error enum fixed (#3487) * Create error type enum * Replace magic string slices with LemmyErrorTypes * Remove unused enum * Add rename snake case to error enum * Rename functions * clippy * Fix merge errors * Serialize in PascalCase instead of snake_case * Revert src/lib * Add serialization tests * Update translations * Fix compilation error in test * Fix another compilation error * Add code for generating typescript types * Various fixes to avoid breaking api * impl From<LemmyErrorType> for LemmyError * with_lemmy_type * trigger ci --------- Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> * Only update site_aggregates for local site (#3516) * Fix #3501 - Fix aggregation counts for elements removed and deleted (#3543) Two bugs were found and fixed: - previously elements removal and deletion were counted as two separate disappearances - removing comments did not affect post aggregations * Use LemmyErrorType also make error_type compulsory * Add missing import for jsonify_plain_text_errors * Fix formatting * Trying to make woodpecker run again --------- Co-authored-by: phiresky <phireskyde+git@gmail.com> Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: rosenjcb <rosenjcb@gmail.com> Co-authored-by: nixoye <12674582+nixoye@users.noreply.github.com> Co-authored-by: dullbananas <dull.bananas0@gmail.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: SleeplessOne1917 <abias1122@gmail.com> Co-authored-by: Sander Saarend <sander@saarend.com> Co-authored-by: Piotr Juszczyk <74842304+pijuszczyk@users.noreply.github.com>
2023-07-10 20:44:14 +00:00
.wrap(ErrorHandlers::new().default_handler(jsonify_plain_text_errors))
.app_data(Data::new(context.clone()))
.app_data(Data::new(rate_limit_cell.clone()))
.wrap(FederationMiddleware::new(federation_config.clone()))
.wrap(SessionMiddleware::new(context.clone()))
.wrap(Condition::new(
SETTINGS.prometheus.is_some(),
prom_api_metrics.clone(),
));
// The routes
app
.configure(|cfg| api_routes_v3::config(cfg, &rate_limit_cell))
.configure(|cfg| api_routes_v4::config(cfg, &rate_limit_cell))
.configure(|cfg| {
if federation_enabled {
lemmy_apub::http::routes::config(cfg);
webfinger::config(cfg);
}
})
.configure(feeds::config)
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
.configure(|cfg| images::config(cfg, pictrs_client.clone(), &rate_limit_cell))
.configure(nodeinfo::config)
})
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
.disable_signals()
.bind(bind)?
Persistent, performant, reliable federation queue (#3605) * persistent activity queue * fixes * fixes * make federation workers function callable from outside * log federation instances * dead instance detection not needed here * taplo fmt * split federate bin/lib * minor fix * better logging * log * create struct to hold cancellable task for readability * use boxfuture for readability * reset submodule * fix * fix lint * swap * remove json column, use separate array columns instead * some review comments * make worker a struct for readability * minor readability * add local filter to community follower view * remove separate lemmy_federate entry point * fix remaining duration * address review comments mostly * fix lint * upgrade actitypub-fed to simpler interface * fix sql format * increase delays a bit * fixes after merge * remove selectable * fix instance selectable * add comment * start federation based on latest id at the time * rename federate process args * dead instances in one query * filter follow+report activities by local * remove synchronous federation remove activity sender queue * lint * fix federation tests by waiting for results to change * fix fed test * fix comment report * wait some more * Apply suggestions from code review Co-authored-by: SorteKanin <sortekanin@gmail.com> * fix most remaining tests * wait until private messages * fix community tests * fix community tests * move arg parse * use instance_id instead of domain in federation_queue_state table --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: SorteKanin <sortekanin@gmail.com>
2023-09-09 16:25:03 +00:00
.run();
let handle = server.handle();
tokio::task::spawn(server);
Ok(handle)
}
fn cors_config(settings: &Settings) -> Cors {
let self_origin = settings.get_protocol_and_hostname();
let cors_origin_setting = settings.cors_origin();
// A default setting for either wildcard, or None
let cors_default = Cors::default()
.allow_any_origin()
.allow_any_method()
.allow_any_header()
.expose_any_header()
.max_age(3600);
match (cors_origin_setting.clone(), cfg!(debug_assertions)) {
(Some(origin), false) => {
// Need to call send_wildcard() explicitly, passing this into allowed_origin() results in
// error
if origin == "*" {
cors_default
} else {
Cors::default()
.allowed_origin(&origin)
.allowed_origin(&self_origin)
.allow_any_method()
.allow_any_header()
.expose_any_header()
.max_age(3600)
}
}
_ => cors_default,
}
}
#[cfg(test)]
pub mod tests {
use activitypub_federation::config::Data;
use lemmy_api_common::context::LemmyContext;
use std::env::set_current_dir;
pub async fn test_context() -> Data<LemmyContext> {
// hack, necessary so that config file can be loaded from hardcoded, relative path.
// Ignore errors as this gets called once for every test (so changing dir again would fail).
set_current_dir("crates/utils").ok();
LemmyContext::init_test_context().await
}
}