From 6c61aed7a1b0069c6b539a77da2e2296c3eb157f Mon Sep 17 00:00:00 2001
From: Alexandr Mansurov <alex@eghuro.cz>
Date: Wed, 21 Jan 2026 22:49:04 +0100
Subject: [PATCH] Parse error breakdown, polish warnings

---
 src/files.rs  |  3 ---
 src/main.rs   | 24 +++++++++++++++++-------
 src/parser.rs |  7 -------
 3 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/src/files.rs b/src/files.rs
index f8557a2..8f447ec 100644
--- a/src/files.rs
+++ b/src/files.rs
@@ -47,7 +47,6 @@ impl LogFileDiscovery {
             return Ok(Some(LogFile {
                 path: gz_path,
                 compressed: true,
-                date,
             }));
         }
 
@@ -57,7 +56,6 @@ impl LogFileDiscovery {
             return Ok(Some(LogFile {
                 path: plain_path,
                 compressed: false,
-                date,
             }));
         }
 
@@ -70,7 +68,6 @@ impl LogFileDiscovery {
 pub struct LogFile {
     pub path: PathBuf,
     pub compressed: bool,
-    pub date: NaiveDate,
 }
 
 impl LogFile {
diff --git a/src/main.rs b/src/main.rs
index cb5b2b9..0082851 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -1,6 +1,7 @@
 use anyhow::{anyhow, Result};
 use chrono::NaiveDate;
 use clap::Parser;
+use std::collections::HashMap;
 use std::io::BufRead;
 use std::path::PathBuf;
 
@@ -119,7 +120,7 @@ fn process_reader(
     let mut signature_batch: Vec<SignatureEntry> = Vec::with_capacity(batch_size);
     let mut total_lines = 0u64;
     let mut parsed_lines = 0u64;
-    let mut error_lines = 0u64;
+    let mut error_counts: HashMap<String, u64> = HashMap::new();
 
     for line_result in reader.lines() {
         let line = line_result?;
@@ -136,18 +137,16 @@ fn process_reader(
                     }
                 }
                 Err(e) => {
-                    error_lines += 1;
-                    if error_lines <= 10 {
-                        eprintln!("Parse error: {}", e);
-                    }
+                    *error_counts.entry(e.to_string()).or_insert(0) += 1;
                 }
             }
         }
 
         if total_lines % 100_000 == 0 {
+            let total_errors: u64 = error_counts.values().sum();
             eprintln!(
                 "Progress: {} lines read, {} parsed, {} errors",
-                total_lines, parsed_lines, error_lines
+                total_lines, parsed_lines, total_errors
             );
         }
     }
@@ -157,11 +156,22 @@ fn process_reader(
         flush_signature_batch(db, &mut signature_batch)?;
     }
 
+    let total_errors: u64 = error_counts.values().sum();
     eprintln!(
         "File complete: {} lines read, {} parsed, {} errors",
-        total_lines, parsed_lines, error_lines
+        total_lines, parsed_lines, total_errors
     );
 
+    // Print error summary
+    if !error_counts.is_empty() {
+        eprintln!("\nParse errors breakdown:");
+        let mut errors: Vec<_> = error_counts.into_iter().collect();
+        errors.sort_by(|a, b| b.1.cmp(&a.1)); // Sort by count descending
+        for (error, count) in errors {
+            eprintln!("  {} ({}x)", error, count);
+        }
+    }
+
     Ok(())
 }
 
diff --git a/src/parser.rs b/src/parser.rs
index 5fdb0f8..ad15d01 100644
--- a/src/parser.rs
+++ b/src/parser.rs
@@ -25,9 +25,6 @@ pub struct SignatureEntry {
 /// Trait for parsing different message types from logs.
 /// Implement this trait to add support for new message formats.
 pub trait MessageParser: Send + Sync {
-    /// Returns the message type identifier (e.g., "signature")
-    fn message_type(&self) -> &'static str;
-
     /// Attempts to parse a log line. Returns None if this parser doesn't handle this message type.
     fn parse(&self, line: &str) -> Option<Result<ParsedMessage>>;
 }
@@ -49,10 +46,6 @@ static SIGNATURE_RE: LazyLock<Regex> =
 pub struct SignatureParser;
 
 impl MessageParser for SignatureParser {
-    fn message_type(&self) -> &'static str {
-        "signature"
-    }
-
     fn parse(&self, line: &str) -> Option<Result<ParsedMessage>> {
         // Check if this line contains a signature message
         if !line.contains("msg=\"signature:") {