Merge PR #280: Add support for explicit extraction file name charset

* fixed #274 and #278 encoding issue
* Added russian name sample
* Added workaround for irreversible text samples
   Some samples become broken when they're writen to byte array with their encoder
* Renamed local variable which had the same name, that static field has
* Magic codepage value for detecting auto-codepage. see #278

Co-authored-by: EvilBeaver <[email protected]>

Author: EvilBeaver

src/ICSharpCode.SharpZipLib/Zip/ZipStrings.cs

@@ -13,8 +13,8 @@ static ZipStrings()
 		{
 			try
 			{
-				var codePage = Encoding.GetEncoding(0).CodePage;
-				SystemDefaultCodePage = (codePage == 1 || codePage == 2 || codePage == 3 || codePage == 42) ? FallbackCodePage : codePage;
+				var platformCodepage = Encoding.GetEncoding(0).CodePage;
+				SystemDefaultCodePage = (platformCodepage == 1 || platformCodepage == 2 || platformCodepage == 3 || platformCodepage == 42) ? FallbackCodePage : platformCodepage;
 			}
 			catch
 			{
@@ -29,8 +29,12 @@ static ZipStrings()
 		/// In practice, most zip apps use OEM or system encoding (typically cp437 on Windows). 
 		/// Let's be good citizens and default to UTF-8 http://utf8everywhere.org/
 		/// </remarks>
-		private static int codePage = Encoding.UTF8.CodePage;
+		private static int codePage = AutomaticCodePage;
 
+		/// Automatically select codepage while opening archive
+		/// see https://github.com/icsharpcode/SharpZipLib/pull/280#issuecomment-433608324
+		/// 
+		private const int AutomaticCodePage = -1;
 
 		/// <summary>
 		/// Encoding used for string conversion. Setting this to 65001 (UTF-8) will
@@ -40,7 +44,7 @@ public static int CodePage
 		{
 			get
 			{
-				return codePage;
+				return codePage == AutomaticCodePage? Encoding.UTF8.CodePage:codePage;
 			}
 			set
 			{
@@ -125,7 +129,14 @@ public static string ConvertToString(byte[] data)
 		private static Encoding EncodingFromFlag(int flags)
 			=> ((flags & (int)GeneralBitFlags.UnicodeText) != 0)
 				? Encoding.UTF8
-				: Encoding.GetEncoding(SystemDefaultCodePage);
+				: Encoding.GetEncoding(
+					// if CodePage wasn't set manually and no utf flag present
+					// then we must use SystemDefault (old behavior)
+					// otherwise, CodePage should be preferred over SystemDefault
+					// see https://github.com/icsharpcode/SharpZipLib/issues/274
+					codePage == AutomaticCodePage? 
+						SystemDefaultCodePage:
+						codePage);
 
 		/// <summary>
 		/// Convert a byte array to a string  using <see cref="CodePage"/>

test/ICSharpCode.SharpZipLib.Tests/TestSupport/StringTesting.cs

@@ -12,6 +12,7 @@ static StringTesting()
 			AddLanguage("Greek", "Ϗΰ.txt", "windows-1253");
 			AddLanguage("Nordic", "Åæ.txt", "windows-1252");
 			AddLanguage("Arabic", "ڀڅ.txt", "windows-1256");
+			AddLanguage("Russian", "Прйвёт.txt", "windows-1251");
 		}
 
 		private static void AddLanguage(string language, string filename, string encoding)

test/ICSharpCode.SharpZipLib.Tests/Zip/FastZipHandling.cs

@@ -1,4 +1,4 @@
-using System;
+using System;
 using System.Collections.Generic;
 using System.IO;
 using System.Linq;
@@ -224,6 +224,20 @@ public void NonUnicodeText()
 				foreach((string language, string filename, string encoding) in StringTesting.GetTestSamples())
 				{
 					Console.WriteLine($"{language} filename \"{filename}\" using \"{encoding}\":");
+
+					// TODO: samples of this test must be reversible
+					// Some samples can't be restored back with their encoding.
+					// test wasn't failing only because SystemDefaultCodepage is 65001 on Net.Core and
+					// old behaviour actually was using Unicode instead of user's passed codepage
+					var encoder = Encoding.GetEncoding(encoding);
+					var bytes = encoder.GetBytes(filename);
+					var restoredString = encoder.GetString(bytes);
+					if(string.CompareOrdinal(filename, restoredString) != 0)
+					{
+						Console.WriteLine($"Sample for language {language} with value of {filename} is skipped, because it's irreversable");
+						continue;
+					}
+
 					ZipStrings.CodePage = Encoding.GetEncoding(encoding).CodePage;
 					TestFileNames(filename);
 				}