reivilibre
/
quickpeep
1
0
Fork 0
Code Issues 11 Pull Requests Projects Releases 11 Wiki Activity

Add more error context

This commit is contained in:
Olivier 'reivilibre' 2022-11-26 22:59:14 +00:00
parent 08f4b7aeaa
commit 438beed86a
1 changed files with 13 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
quickpeep_raker/src/raking/task.rs
View File

@ -1,5 +1,5 @@
use crate::config::RerakeTimings; use crate::config::RerakeTimings;
use crate::raking::references::references_from_urlrakes; use crate::raking::references::{clean_url, references_from_urlrakes};
use crate::raking::{ use crate::raking::{
get_robots_txt_for, robots_txt_url_for, PermanentFailure, PermanentFailureReason, RakeIntent, get_robots_txt_for, robots_txt_url_for, PermanentFailure, PermanentFailureReason, RakeIntent,
RakeOutcome, Raker, RedirectReason, RobotsTxt, TemporaryFailure, TemporaryFailureReason, RakeOutcome, Raker, RedirectReason, RobotsTxt, TemporaryFailure, TemporaryFailureReason,
@ -425,7 +425,7 @@ impl TaskContext {
RakeOutcome::Redirect { reason, new_url } => { RakeOutcome::Redirect { reason, new_url } => {
let refs = RakedReferrerEntry { let refs = RakedReferrerEntry {
references: [RakedReference { references: [RakedReference {
target: new_url.to_string(), target: clean_url(&new_url).to_string(),
kind: match reason { kind: match reason {
RedirectReason::Redirected { .. } => ReferenceKind::Redirect, RedirectReason::Redirected { .. } => ReferenceKind::Redirect,
RedirectReason::NotCanonical { .. } => ReferenceKind::CanonicalUrl, RedirectReason::NotCanonical { .. } => ReferenceKind::CanonicalUrl,
@ -584,12 +584,20 @@ impl EventProcessor<'_> {
last_visited_days: datestamp, last_visited_days: datestamp,
}, },
rerake_on, rerake_on,
)?; )
.context("failed to mark URL as visited")?;
// track all the referred-to URLs! // track all the referred-to URLs!
for reference in refs.references { for reference in refs.references {
let ref_url = Url::parse(&reference.target)?; let ref_url = Url::parse(&reference.target).with_context(|| {
let domain = get_reduced_domain(&ref_url)?; format!(
"failed to parse target URL of reference: {:?}",
reference.target
)
})?;
let domain = get_reduced_domain(&ref_url).with_context(|| {
format!("failed to reduce domain: {:?}", reference.target)
})?;
// First check if this URL is an allowed URL (hence should be enqueued) // First check if this URL is an allowed URL (hence should be enqueued)
let allowed = txn let allowed = txn