UP

src/StellaOps.Feedser.Source.Distro.Debian/Internal/DebianListParser.cs

@@ -0,0 +1,107 @@
+using System;
+using System.Collections.Generic;
+using System.Globalization;
+using System.Text.RegularExpressions;
+
+namespace StellaOps.Feedser.Source.Distro.Debian.Internal;
+
+internal static class DebianListParser
+{
+    private static readonly Regex HeaderRegex = new("^\\[(?<date>[^\\]]+)\\]\\s+(?<id>DSA-\\d{4,}-\\d+)\\s+(?<title>.+)$", RegexOptions.Compiled);
+    private static readonly Regex CveRegex = new("CVE-\\d{4}-\\d{3,7}", RegexOptions.IgnoreCase | RegexOptions.Compiled);
+
+    public static IReadOnlyList<DebianListEntry> Parse(string? content)
+    {
+        if (string.IsNullOrWhiteSpace(content))
+        {
+            return Array.Empty<DebianListEntry>();
+        }
+
+        var entries = new List<DebianListEntry>();
+        var currentCves = new HashSet<string>(StringComparer.OrdinalIgnoreCase);
+        DateTimeOffset currentDate = default;
+        string? currentId = null;
+        string? currentTitle = null;
+        string? currentPackage = null;
+
+        foreach (var rawLine in content.Split('\n'))
+        {
+            var line = rawLine.TrimEnd('\r');
+            if (string.IsNullOrWhiteSpace(line))
+            {
+                continue;
+            }
+
+            if (line[0] == '[')
+            {
+                if (currentId is not null && currentTitle is not null && currentPackage is not null)
+                {
+                    entries.Add(new DebianListEntry(
+                        currentId,
+                        currentDate,
+                        currentTitle,
+                        currentPackage,
+                        currentCves.Count == 0 ? Array.Empty<string>() : new List<string>(currentCves)));
+                }
+
+                currentCves.Clear();
+                currentId = null;
+                currentTitle = null;
+                currentPackage = null;
+
+                var match = HeaderRegex.Match(line);
+                if (!match.Success)
+                {
+                    continue;
+                }
+
+                if (!DateTimeOffset.TryParseExact(
+                        match.Groups["date"].Value,
+                        new[] { "dd MMM yyyy", "d MMM yyyy" },
+                        CultureInfo.InvariantCulture,
+                        DateTimeStyles.AssumeUniversal | DateTimeStyles.AdjustToUniversal,
+                        out currentDate))
+                {
+                    continue;
+                }
+
+                currentId = match.Groups["id"].Value.Trim();
+                currentTitle = match.Groups["title"].Value.Trim();
+
+                var separatorIndex = currentTitle.IndexOf(" - ", StringComparison.Ordinal);
+                currentPackage = separatorIndex > 0
+                    ? currentTitle[..separatorIndex].Trim()
+                    : currentTitle.Split(' ', StringSplitOptions.RemoveEmptyEntries | StringSplitOptions.TrimEntries).FirstOrDefault();
+                if (string.IsNullOrWhiteSpace(currentPackage))
+                {
+                    currentPackage = currentId;
+                }
+
+                continue;
+            }
+
+            if (line[0] == '{')
+            {
+                foreach (Match match in CveRegex.Matches(line))
+                {
+                    if (match.Success && !string.IsNullOrWhiteSpace(match.Value))
+                    {
+                        currentCves.Add(match.Value.ToUpperInvariant());
+                    }
+                }
+            }
+        }
+
+        if (currentId is not null && currentTitle is not null && currentPackage is not null)
+        {
+            entries.Add(new DebianListEntry(
+                currentId,
+                currentDate,
+                currentTitle,
+                currentPackage,
+                currentCves.Count == 0 ? Array.Empty<string>() : new List<string>(currentCves)));
+        }
+
+        return entries;
+    }
+}