Fix mojibake in signature: detect real encoding instead of assuming UTF-8

Outlook writes signature .htm files in the system ANSI code page
(windows-1251 on Russian Windows) without a BOM, but File.ReadAllText
defaults to UTF-8 for BOM-less files - mangling Cyrillic into
replacement characters. ReadSignatureHtml now checks BOM, then a
<meta charset> tag, then falls back to Encoding.Default (the actual
system code page) before ever assuming UTF-8.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

outlook-bulk-sender/BulkArchiveMailer.AddIn/SignatureReader.cs

@@ -1,6 +1,8 @@
 using System;
 using System.Collections.Generic;
 using System.IO;
+using System.Text;
+using System.Text.RegularExpressions;
 using Microsoft.Win32;
 using Outlook = Microsoft.Office.Interop.Outlook;
 
@@ -38,7 +40,7 @@ namespace BulkArchiveMailer.AddIn
                 }
 
                 var path = ResolveSignaturePath(signatureName);
-                return File.Exists(path) ? File.ReadAllText(path) : null;
+                return File.Exists(path) ? ReadSignatureHtml(path) : null;
             }
             catch
             {
@@ -119,5 +121,141 @@ namespace BulkArchiveMailer.AddIn
 
             return null;
         }
+
+        /// <summary>
+        /// Reads a signature .htm file with the encoding it was actually
+        /// written in, instead of assuming UTF-8 the way
+        /// <see cref="File.ReadAllText(string)"/> does for BOM-less files.
+        ///
+        /// Outlook writes signature .htm files in the system ANSI code
+        /// page (windows-1251 on a Russian Windows) without a BOM and
+        /// usually without a reliable &lt;meta charset&gt; tag, so blindly
+        /// reading as UTF-8 mangles Cyrillic into replacement characters.
+        ///
+        /// Detection order:
+        /// 1. BOM (UTF-8 / UTF-16 LE / UTF-16 BE / UTF-32 LE / UTF-32 BE) - trusted if present.
+        /// 2. &lt;meta charset=...&gt; or &lt;meta http-equiv="Content-Type" content="...charset=..."&gt;
+        ///    found in the first ~1KB, if the named encoding is recognized.
+        /// 3. Encoding.Default (the system ANSI code page) - matches
+        ///    Outlook's legacy signature file format.
+        ///
+        /// Never throws: any failure along the way falls back to the
+        /// previous behavior, <see cref="File.ReadAllText(string)"/>.
+        /// </summary>
+        internal static string ReadSignatureHtml(string path)
+        {
+            try
+            {
+                var bytes = File.ReadAllBytes(path);
+
+                var bomEncoding = DetectBomEncoding(bytes, out var preambleLength);
+                if (bomEncoding != null)
+                {
+                    return bomEncoding.GetString(bytes, preambleLength, bytes.Length - preambleLength);
+                }
+
+                var metaEncoding = TryDetectMetaCharsetEncoding(bytes);
+                if (metaEncoding != null)
+                {
+                    return metaEncoding.GetString(bytes, 0, bytes.Length);
+                }
+
+                return Encoding.Default.GetString(bytes, 0, bytes.Length);
+            }
+            catch
+            {
+                try
+                {
+                    return File.ReadAllText(path);
+                }
+                catch
+                {
+                    return null;
+                }
+            }
+        }
+
+        private static Encoding DetectBomEncoding(byte[] bytes, out int preambleLength)
+        {
+            // UTF-32 BOMs must be checked before UTF-16 ones: the UTF-32 LE
+            // BOM (FF FE 00 00) starts with the same two bytes as the
+            // UTF-16 LE BOM (FF FE).
+            if (bytes.Length >= 4 && bytes[0] == 0x00 && bytes[1] == 0x00 && bytes[2] == 0xFE && bytes[3] == 0xFF)
+            {
+                preambleLength = 4;
+                return new UTF32Encoding(bigEndian: true, byteOrderMark: true);
+            }
+
+            if (bytes.Length >= 4 && bytes[0] == 0xFF && bytes[1] == 0xFE && bytes[2] == 0x00 && bytes[3] == 0x00)
+            {
+                preambleLength = 4;
+                return new UTF32Encoding(bigEndian: false, byteOrderMark: true);
+            }
+
+            if (bytes.Length >= 3 && bytes[0] == 0xEF && bytes[1] == 0xBB && bytes[2] == 0xBF)
+            {
+                preambleLength = 3;
+                return new UTF8Encoding(encoderShouldEmitUTF8Identifier: true);
+            }
+
+            if (bytes.Length >= 2 && bytes[0] == 0xFE && bytes[1] == 0xFF)
+            {
+                preambleLength = 2;
+                return new UnicodeEncoding(bigEndian: true, byteOrderMark: true);
+            }
+
+            if (bytes.Length >= 2 && bytes[0] == 0xFF && bytes[1] == 0xFE)
+            {
+                preambleLength = 2;
+                return new UnicodeEncoding(bigEndian: false, byteOrderMark: true);
+            }
+
+            preambleLength = 0;
+            return null;
+        }
+
+        private static readonly Regex MetaCharsetRegex = new Regex(
+            @"<meta\b[^>]*charset\s*=\s*[""']?\s*([A-Za-z0-9_\-]+)",
+            RegexOptions.IgnoreCase | RegexOptions.Compiled);
+
+        private static Encoding TryDetectMetaCharsetEncoding(byte[] bytes)
+        {
+            try
+            {
+                // <meta charset=...> tags are always plain ASCII, so
+                // Latin1 (single byte, 1:1 with the raw bytes) is a safe,
+                // lossless way to look for them regardless of the file's
+                // real encoding.
+                var headLength = Math.Min(bytes.Length, 1024);
+                var head = Encoding.GetEncoding("ISO-8859-1").GetString(bytes, 0, headLength);
+
+                var match = MetaCharsetRegex.Match(head);
+                if (!match.Success)
+                {
+                    return null;
+                }
+
+                var charsetName = match.Groups[1].Value.Trim();
+                if (string.IsNullOrEmpty(charsetName))
+                {
+                    return null;
+                }
+
+                try
+                {
+                    return Encoding.GetEncoding(charsetName);
+                }
+                catch
+                {
+                    // Unrecognized charset name - let the caller fall back
+                    // to Encoding.Default instead.
+                    return null;
+                }
+            }
+            catch
+            {
+                return null;
+            }
+        }
     }
 }

outlook-bulk-sender/BulkArchiveMailer.AddIn/UI/PreviewForm.Designer.cs

@@ -30,6 +30,7 @@ namespace BulkArchiveMailer.AddIn.UI
         private System.Windows.Forms.Button btnCheckAllReady;
         private System.Windows.Forms.Button btnUncheckAll;
         private System.Windows.Forms.Button btnRescan;
+        private System.Windows.Forms.Button btnSettings;
 
         private System.Windows.Forms.ProgressBar progressSend;
         private System.Windows.Forms.Label lblProgressStatus;
@@ -56,6 +57,7 @@ namespace BulkArchiveMailer.AddIn.UI
             this.btnCheckAllReady = new System.Windows.Forms.Button();
             this.btnUncheckAll = new System.Windows.Forms.Button();
             this.btnRescan = new System.Windows.Forms.Button();
+            this.btnSettings = new System.Windows.Forms.Button();
 
             this.progressSend = new System.Windows.Forms.ProgressBar();
             this.lblProgressStatus = new System.Windows.Forms.Label();
@@ -70,7 +72,7 @@ namespace BulkArchiveMailer.AddIn.UI
             // lblPasswordReminder
             //
             this.lblPasswordReminder.Location = new System.Drawing.Point(12, 10);
-            this.lblPasswordReminder.Size = new System.Drawing.Size(900, 20);
+            this.lblPasswordReminder.Size = new System.Drawing.Size(1018, 20);
             this.lblPasswordReminder.Text = "Напоминание: пароль от архивов сообщается получателям отдельно (вручную, не через это письмо и не через этот инструмент).";
             this.lblPasswordReminder.ForeColor = System.Drawing.Color.DarkRed;
             this.lblPasswordReminder.Font = new System.Drawing.Font(this.lblPasswordReminder.Font, System.Drawing.FontStyle.Bold);
@@ -80,7 +82,7 @@ namespace BulkArchiveMailer.AddIn.UI
             // lblSignatureWarning
             //
             this.lblSignatureWarning.Location = new System.Drawing.Point(12, 32);
-            this.lblSignatureWarning.Size = new System.Drawing.Size(900, 20);
+            this.lblSignatureWarning.Size = new System.Drawing.Size(1018, 20);
             this.lblSignatureWarning.Text = "Подпись не найдена — письма будут отправлены без подписи. Настройте .htm-файл вручную в Настройках.";
             this.lblSignatureWarning.ForeColor = System.Drawing.Color.DarkOrange;
             this.lblSignatureWarning.Visible = false;
@@ -90,7 +92,7 @@ namespace BulkArchiveMailer.AddIn.UI
             // grid
             //
             this.grid.Location = new System.Drawing.Point(12, 58);
-            this.grid.Size = new System.Drawing.Size(900, 420);
+            this.grid.Size = new System.Drawing.Size(1018, 420);
             this.grid.Anchor = ((System.Windows.Forms.AnchorStyles)((((System.Windows.Forms.AnchorStyles.Top | System.Windows.Forms.AnchorStyles.Bottom)
                         | System.Windows.Forms.AnchorStyles.Left)
                         | System.Windows.Forms.AnchorStyles.Right)));
@@ -215,24 +217,34 @@ namespace BulkArchiveMailer.AddIn.UI
             this.btnUncheckAll.UseVisualStyleBackColor = true;
             this.btnUncheckAll.Click += new System.EventHandler(this.btnUncheckAll_Click);
 
+            //
+            // btnSettings
+            //
+            this.btnSettings.Location = new System.Drawing.Point(416, 540);
+            this.btnSettings.Size = new System.Drawing.Size(110, 27);
+            this.btnSettings.Text = "Настройки...";
+            this.btnSettings.Anchor = ((System.Windows.Forms.AnchorStyles)((System.Windows.Forms.AnchorStyles.Bottom | System.Windows.Forms.AnchorStyles.Left)));
+            this.btnSettings.UseVisualStyleBackColor = true;
+            this.btnSettings.Click += new System.EventHandler(this.btnSettings_Click);
+
             //
             // progressSend
             //
-            this.progressSend.Location = new System.Drawing.Point(430, 543);
+            this.progressSend.Location = new System.Drawing.Point(548, 543);
             this.progressSend.Size = new System.Drawing.Size(300, 20);
             this.progressSend.Anchor = ((System.Windows.Forms.AnchorStyles)((System.Windows.Forms.AnchorStyles.Bottom | System.Windows.Forms.AnchorStyles.Left)));
 
             //
             // lblProgressStatus
             //
-            this.lblProgressStatus.Location = new System.Drawing.Point(430, 565);
+            this.lblProgressStatus.Location = new System.Drawing.Point(548, 565);
             this.lblProgressStatus.Size = new System.Drawing.Size(482, 20);
             this.lblProgressStatus.Anchor = ((System.Windows.Forms.AnchorStyles)((System.Windows.Forms.AnchorStyles.Bottom | System.Windows.Forms.AnchorStyles.Left)));
 
             //
             // btnClose
             //
-            this.btnClose.Location = new System.Drawing.Point(720, 540);
+            this.btnClose.Location = new System.Drawing.Point(838, 540);
             this.btnClose.Size = new System.Drawing.Size(90, 30);
             this.btnClose.Text = "Закрыть";
             this.btnClose.Anchor = ((System.Windows.Forms.AnchorStyles)((System.Windows.Forms.AnchorStyles.Bottom | System.Windows.Forms.AnchorStyles.Right)));
@@ -243,7 +255,7 @@ namespace BulkArchiveMailer.AddIn.UI
             //
             // btnSend
             //
-            this.btnSend.Location = new System.Drawing.Point(816, 540);
+            this.btnSend.Location = new System.Drawing.Point(934, 540);
             this.btnSend.Size = new System.Drawing.Size(96, 30);
             this.btnSend.Text = "Отправить";
             this.btnSend.Anchor = ((System.Windows.Forms.AnchorStyles)((System.Windows.Forms.AnchorStyles.Bottom | System.Windows.Forms.AnchorStyles.Right)));
@@ -256,8 +268,8 @@ namespace BulkArchiveMailer.AddIn.UI
             //
             this.AutoScaleDimensions = new System.Drawing.SizeF(7F, 15F);
             this.AutoScaleMode = System.Windows.Forms.AutoScaleMode.Font;
-            this.ClientSize = new System.Drawing.Size(924, 582);
+            this.ClientSize = new System.Drawing.Size(1042, 582);
-            this.MinimumSize = new System.Drawing.Size(760, 420);
+            this.MinimumSize = new System.Drawing.Size(878, 420);
             this.Controls.Add(this.lblPasswordReminder);
             this.Controls.Add(this.lblSignatureWarning);
             this.Controls.Add(this.grid);
@@ -266,6 +278,7 @@ namespace BulkArchiveMailer.AddIn.UI
             this.Controls.Add(this.btnRescan);
             this.Controls.Add(this.btnCheckAllReady);
             this.Controls.Add(this.btnUncheckAll);
+            this.Controls.Add(this.btnSettings);
             this.Controls.Add(this.progressSend);
             this.Controls.Add(this.lblProgressStatus);
             this.Controls.Add(this.btnClose);

outlook-bulk-sender/BulkArchiveMailer.AddIn/UI/PreviewForm.cs

@@ -58,7 +58,7 @@ namespace BulkArchiveMailer.AddIn.UI
             {
                 if (!string.IsNullOrWhiteSpace(settings.SignatureHtmlPath))
                 {
-                    return File.Exists(settings.SignatureHtmlPath) ? File.ReadAllText(settings.SignatureHtmlPath) : null;
+                    return File.Exists(settings.SignatureHtmlPath) ? SignatureReader.ReadSignatureHtml(settings.SignatureHtmlPath) : null;
                 }
 
                 return SignatureReader.TryAutoDetectSignatureHtml(outlookApplication);
@@ -128,6 +128,17 @@ namespace BulkArchiveMailer.AddIn.UI
 
         private void UpdateSummaryLabel()
         {
+            if (rows == null)
+            {
+                // Can be invoked before RunScan() ever ran - most notably,
+                // WinForms fires DataGridView.CellValueChanged as a side
+                // effect of DataGridViewColumn.HeaderText assignments in
+                // InitializeComponent(), i.e. before the constructor (and
+                // therefore RunScan()) has had a chance to populate `rows`.
+                // Nothing to summarize yet in that case.
+                return;
+            }
+
             var ready = rows.Count(r => r.Job.Status == SendStatus.Ready);
             var already = rows.Count(r => r.Job.Status == SendStatus.AlreadySent);
             var errors = rows.Count(r => r.Job.Status == SendStatus.Error);
@@ -161,6 +172,34 @@ namespace BulkArchiveMailer.AddIn.UI
             RunScan();
         }
 
+        private void btnSettings_Click(object sender, EventArgs e)
+        {
+            using (var settingsForm = new SettingsForm(outlookApplication))
+            {
+                settingsForm.ShowDialog(this);
+            }
+
+            // Regardless of DialogResult (OK or Cancel), reload settings and
+            // rescan so a changed archive folder, domain, or templates are
+            // reflected immediately without a separate "Пересканировать"
+            // click. A closed-without-saving dialog just reloads the same
+            // settings again, which is harmless.
+            try
+            {
+                settings = AppSettings.Load();
+            }
+            catch (Exception ex)
+            {
+                MessageBox.Show(this, "Не удалось прочитать настройки: " + ex.ToString(), "BulkArchiveMailer", MessageBoxButtons.OK, MessageBoxIcon.Error);
+                return;
+            }
+
+            signatureHtml = ResolveSignatureHtml();
+            lblSignatureWarning.Visible = string.IsNullOrEmpty(signatureHtml);
+
+            RunScan();
+        }
+
         private void btnCheckAllReady_Click(object sender, EventArgs e)
         {
             foreach (var row in rows)