Do not try to redownload the message in case of any error

Since switch to async we don't have spurious "database is busy"
errors anymore. Since an error is irrecoverable in most cases,
we can skip the message. The cost of this is we may
accidentally skip a correct message if I/O fails, but
the advantage is that we are guaranteed to never confuse
irrecoverable error with recoverable one and get stuck in
infinite loop redownloading the same message over and over.

src/authres.rs

@@ -13,8 +13,6 @@ use once_cell::sync::Lazy;
 use crate::config::Config;
 use crate::context::Context;
 use crate::headerdef::HeaderDef;
-use crate::mimeparser;
-use crate::mimeparser::ParserErrorExt;
 use crate::tools::time;
 use crate::tools::EmailAddress;
 
@@ -32,23 +30,19 @@ pub(crate) async fn handle_authres(
     mail: &ParsedMail<'_>,
     from: &str,
     message_time: i64,
-) -> mimeparser::ParserResult<DkimResults> {
+) -> Result<DkimResults> {
     let from_domain = match EmailAddress::new(from) {
         Ok(email) => email.domain,
         Err(e) => {
             // This email is invalid, but don't return an error, we still want to
             // add a stub to the database so that it's not downloaded again
-            return Err(anyhow::format_err!("invalid email {}: {:#}", from, e)).map_err_malformed();
+            return Err(anyhow::format_err!("invalid email {}: {:#}", from, e));
         }
     };
 
     let authres = parse_authres_headers(&mail.get_headers(), &from_domain);
-    update_authservid_candidates(context, &authres)
-        .await
-        .map_err_sql()?;
-    compute_dkim_results(context, authres, &from_domain, message_time)
-        .await
-        .map_err_sql()
+    update_authservid_candidates(context, &authres).await?;
+    compute_dkim_results(context, authres, &from_domain, message_time).await
 }
 
 #[derive(Default, Debug)]