Fix

BUTR · Oct 2, 2023 · ae228f7 · ae228f7
1 parent 6f3e20d
commit ae228f7
Show file tree

Hide file tree

Showing 3 changed files with 14 additions and 13 deletions.
diff --git a/src/BUTR.Site.NexusMods.Client/Models/DemoUser.cs b/src/BUTR.Site.NexusMods.Client/Models/DemoUser.cs
@@ -3,8 +3,6 @@
 using BUTR.Site.NexusMods.ServerClient;
 using BUTR.Site.NexusMods.Shared.Helpers;
 
-using HtmlAgilityPack;
-
 using System;
 using System.Collections.Generic;
 using System.Collections.Immutable;
@@ -50,9 +48,7 @@ static string GetException(ExceptionModel? exception, bool inner = false) => exc
             var response = await client.SendAsync(request);
             var content = await response.Content.ReadAsStringAsync();
 
-            var document = new HtmlDocument();
-            document.LoadHtml(content.Replace("<filename unknown>", "NULL"));
-            return (id, CrashReportParser.ParseLegacyHtml(10, document, content));
+            return (id, CrashReportParser.ParseLegacyHtml(10, content));
         }
 
         if (_crashReports is null)

diff --git a/src/BUTR.Site.NexusMods.Server/BUTR.Site.NexusMods.Server.csproj b/src/BUTR.Site.NexusMods.Server/BUTR.Site.NexusMods.Server.csproj
@@ -18,7 +18,6 @@
     <PackageReference Include="butr.crashreport.bannerlord.parser" Version="13.0.0.14" />
     <PackageReference Include="Community.Microsoft.Extensions.Caching.PostgreSql" Version="4.0.4" />
     <PackageReference Include="DynamicExpressions.NET" Version="1.1.0" />
-    <PackageReference Include="HtmlAgilityPack" Version="1.11.53" />
     <PackageReference Include="ICSharpCode.Decompiler" Version="8.1.1.7464" />
     <PackageReference Include="Microsoft.CodeAnalysis.CSharp" Version="4.5.0" />
     <PackageReference Include="Microsoft.EntityFrameworkCore.Design" Version="8.0.0-rc.1.23419.6" PrivateAssets="all" IncludeAssets="runtime; build; native; contentfiles; analyzers; buildtransitive" />

diff --git a/src/BUTR.Site.NexusMods.Server/Jobs/CrashReportProcessorJob.cs b/src/BUTR.Site.NexusMods.Server/Jobs/CrashReportProcessorJob.cs
@@ -58,7 +58,7 @@ public CrashReportProcessorJob(ILogger<CrashReportProcessorJob> logger, IService
         _serviceScopeFactory = serviceScopeFactory ?? throw new ArgumentNullException(nameof(serviceScopeFactory));
     }
 
-    private static async Task DownloadCrashReportsAsync(TenantId tenant, IAsyncEnumerable<CrashReportFileMetadata> requests, CrashReporterClient client, ChannelWriter<HttpResultEntry> httpResultChannel, CancellationToken ct)
+    private static async Task DownloadCrashReportsAsync(TenantId tenant, IAsyncEnumerable<CrashReportFileMetadata> requests, ILogger logger, CrashReporterClient client, ChannelWriter<HttpResultEntry> httpResultChannel, CancellationToken ct)
     {
         var options = new ParallelOptions { CancellationToken = ct, MaxDegreeOfParallelism = ParallelCount };
         await Parallel.ForEachAsync(requests, options, async (entry, ct2) =>
@@ -70,9 +70,15 @@ await Parallel.ForEachAsync(requests, options, async (entry, ct2) =>
             {
                 var content = await client.GetCrashReportAsync(fileId, ct2);
 
-                var document = new HtmlDocument();
-                document.LoadHtml(content.Replace("<filename unknown>", "NULL"));
-                model = CrashReportParser.ParseLegacyHtml(version, document, content);
+                try
+                {
+                    model = CrashReportParser.ParseLegacyHtml(version, content);
+                }
+                catch (Exception e)
+                {
+                    logger.LogError(e, "Failed to parse {FileId}", fileId);
+                    throw;
+                }
             }
             else
             {
@@ -253,7 +259,7 @@ static string GetException(ExceptionModel? exception, bool inner = false) => exc
         // Disposing the DBContext will save the data
     }
 
-    private static async Task HandleFileIdDatesAsync(TenantId tenant, IServiceProvider serviceProvider, CrashReporterClient client, IAsyncEnumerable<CrashReportFileMetadata> requests, CancellationToken ct)
+    private static async Task HandleFileIdDatesAsync(TenantId tenant, IServiceProvider serviceProvider, ILogger logger, CrashReporterClient client, IAsyncEnumerable<CrashReportFileMetadata> requests, CancellationToken ct)
     {
         var options = serviceProvider.GetRequiredService<IOptions<CrashReporterOptions>>().Value;
         var httpResultChannel = Channel.CreateBounded<HttpResultEntry>(ParallelCount);
@@ -263,7 +269,7 @@ private static async Task HandleFileIdDatesAsync(TenantId tenant, IServiceProvid
 
         await Task.WhenAll(new Task[]
         {
-            DownloadCrashReportsAsync(tenant, requests, client, httpResultChannel, ct),
+            DownloadCrashReportsAsync(tenant, requests, logger, client, httpResultChannel, ct),
             FilterCrashReportsAsync(tenant, serviceProvider, httpResultChannel, databaseResultChannel, linkedCrashReportsChannel, ignoredCrashReportsChannel, ct),
             WriteIgnoredToDatabaseAsync(tenant, serviceProvider, ignoredCrashReportsChannel, ct),
             WriteLinkedToDatabaseAsync(tenant, serviceProvider, linkedCrashReportsChannel, ct),
@@ -301,7 +307,7 @@ public async Task Execute(IJobExecutionContext context)
                 fileIds.ExceptWith(await dbContextRead.CrashReportToFileIds.Where(x => fileIds.Contains(x.FileId)).Select(x => x.FileId).ToListAsync(ct));
                 if (fileIds.Count == 0) continue;
 
-                await HandleFileIdDatesAsync(tenant, scope.ServiceProvider, client, client.GetCrashReportMetadataAsync(fileIds, ct).OfType<CrashReportFileMetadata>(), ct);
+                await HandleFileIdDatesAsync(tenant, scope.ServiceProvider, _logger, client, client.GetCrashReportMetadataAsync(fileIds, ct).OfType<CrashReportFileMetadata>(), ct);
                 processed += fileIds.Count;
             }
         }