c# - XPS 编写器因 .Net 4/字体渲染错误而失败?

标签 c# .net wpf xps

我们有一个 WPF 应用程序,允许用户根据需要输入文本并设置文本样式。然后将他们的文本转换为 XPS,然后使用 ABCPdf 将其插入到 PDF 中。

我们最近切换到 .Net 4,现在 XPS 生成有时会生成格式错误的 XPS。

只有在用户使用重叠两个字符的字体(例如 Sevillana、available on Google Fonts)时,才会生成格式错误的 XPS。在 .Net 4 中,字符重叠,而在 .Net 3.5 中,它们不重叠。来自 another SO question我猜这是因为 .Net 4.0 改变了它的字体渲染引擎。

.Net 3.5 :
DotNet 3.5

.Net 4.0:(注意“d”和“A”是如何重叠的)
DotNet 4.0

看起来 .Net 4 没有考虑撇号的间距,甚至否定了它。如果我删除撇号,间距会变宽。我创建了一个简单的测试项目来演示这个问题:https://github.com/tbroust-trepia/wpf-4-font-rendering

我们将 FlowDocument 保存到 XPS 的方法与 creating an XPS Document from a FlowDocument and attach it on the fly 中的方法几乎相同。 (我怀疑原始开发人员只是复制/粘贴了该代码);它只是将 XPS 保存到文件而不是流中。

The MS tool IsXPS表示特定节点无效。事实上,在Documents\1\Pages\1.fpage中,Indices的值为;,-16;,84;;;;,30。我可以看出“0.-16”并不是真正的数字。

所以,我有很多问题:

  • 为什么会这样?我不明白这些值是如何转换成这些指数的。我猜转换引擎试图将其设置为“-0.16”但搞砸了?
  • 我可以在保存此索引之前修改/检查它的值吗?
  • 如果我不能,我之后如何检查?我找到了一种 XPS validator here但我不明白如何从那里检查索引值。我可以手动完成,读取和解析 XML,但我真的确定这不是一个好主意。

编辑

I have opened a bug at Microsoft ,但我对那里期望不高。我想我会尝试修改生成的 XPS。它很脏,但它可能有用。

最佳答案

在等待 MS 为所欲为的过程中,我找到了修复损坏的 XPS 的方法。

Warning: dirty hack ahead

public class XpsFile
{
    /// <summary>
    /// Regex to validate XPS "indices" property (gotten from ABCPDF)
    /// </summary>
    private static readonly string IndicesRegex = @"(((\(([1-9][0-9]*)(:([1-9][0-9]*))?\))?([0-9]+))?(,(\+?(([0-9]+(\.[0-9]+)?)|(\.[0-9]+))((e|E)(\-|\+)?[0-9]+)?)?(,((\-|\+)?(([0-9]+(\.[0-9]+)?)|(\.[0-9]+))((e|E)(\-|\+)?[0-9]+)?)?(,((\-|\+)?(([0-9]+(\.[0-9]+)?)|(\.[0-9]+))((e|E)(\-|\+)?[0-9]+)?))?)?)?)(;((\(([1-9][0-9]*)(:([1-9][0-9]*))?\))?([0-9]+))?(,(\+?(([0-9]+(\.[0-9]+)?)|(\.[0-9]+))((e|E)(\-|\+)?[0-9]+)?)?(,((\-|\+)?(([0-9]+(\.[0-9]+)?)|(\.[0-9]+))((e|E)(\-|\+)?[0-9]+)?)?(,((\-|\+)?(([0-9]+(\.[0-9]+)?)|(\.[0-9]+))((e|E)(\-|\+)?[0-9]+)?))?)?)?)*";

    /// <summary>
    /// Fixes the XPS problems it encounters and knows about
    /// </summary>
    public static void FixXps(string filePath)
    {
        // first we'll load the XPS file
        using (var currentPackage = Package.Open(filePath, FileMode.Open))
        {
            var pageUri = new Uri("/Documents/1/Pages/1.fpage", UriKind.Relative);

            // check that the file we'll modify exists
            if (!currentPackage.PartExists(pageUri))
            {
                throw new Exception(string.Format("Unable to find first page in XPS {0} - unable to fix this XPS !", filePath));
            }

            // assume the broken part is in the first page
            var firstPage = currentPackage.GetPart(pageUri);
            var relationships = firstPage.GetRelationships();
            var pageContent = XDocument.Load(firstPage.GetStream());

            // then we'll look up each glyph and check if their "Indices" property is valid
            XNamespace ns = pageContent.Root.GetDefaultNamespace();
            var glyphs = (from g in pageContent.Descendants(ns + "Glyphs")
                          where g.Attribute("Indices") != null
                          select g).ToList();
            for (var i = 0; i < glyphs.Count(); i ++)
            {
                glyphs[i] = FixGlyph(glyphs[i]);
            }

            // remove the current (corrupted) file from the package
            currentPackage.DeletePart(pageUri);

            // add the new (shiny) file to the package
            var newPage = currentPackage.CreatePart(pageUri, "application/vnd.ms-package.xps-fixedpage+xml", CompressionOption.NotCompressed);
            using (var ms = new MemoryStream())
            {
                // we need to remove XML declaration, so we need to use the XmlWriter
                var settings = new XmlWriterSettings();
                settings.Indent = false;
                settings.NewLineChars = string.Empty;
                settings.NewLineHandling = NewLineHandling.Replace;
                settings.OmitXmlDeclaration = true;
                using (var xw = XmlWriter.Create(ms, settings))
                {
                    pageContent.WriteTo(xw);
                }

                ms.Seek(0, SeekOrigin.Begin);
                CopyStream(newPage.GetStream(), ms);
            }

            // now we need to re-create the relationships between the Page file and the fonts
            foreach (var relation in relationships)
            {
                newPage.CreateRelationship(relation.TargetUri, relation.TargetMode, relation.RelationshipType, relation.Id);
            }
        }
    }

    /// <summary>
    /// Tries to load the XPS, and returns false if it fails
    /// </summary>
    public static bool IsValidXps(string filePath)
    {
        try
        {
            using (var xpsOld = new XpsDocument(filePath, FileAccess.Read))
            {
                var unused = xpsOld.GetFixedDocumentSequence();
            }

            return true;
        }
        catch (System.Windows.Markup.XamlParseException)
        {
            return false;
        }
    }

    /// <summary>
    /// Writes the whole content of a stream into another
    /// </summary>
    /// <remarks>
    /// http://stackoverflow.com/a/18885954/2354542
    /// </remarks>
    private static void CopyStream(Stream target, Stream source)
    {
        const int bufSize = 0x1000;
        byte[] buf = new byte[bufSize];
        int bytesRead = 0;
        while ((bytesRead = source.Read(buf, 0, bufSize)) > 0)
        {
            target.Write(buf, 0, bytesRead);
        }
    }

    /// <summary>
    /// Fixes the glyph, if necessary
    /// </summary>
    private static XElement FixGlyph(XElement g)
    {
        var matchAttribute = Regex.Match(g.Attribute("Indices").Value, IndicesRegex);
        if (!matchAttribute.Success)
        {
            return g;
        }

        var hasProblem = false;
        foreach (var token in matchAttribute.Value.Split(";".ToCharArray()))
        {
            if (token == ",")
            {
                hasProblem = true;
                break;
            }
        }

        if (hasProblem)
        {
            // the Indices attribute is not well-formed: let's try to fix the one(s) that are wrong
            var fixedTokens = new List<string>();
            foreach (var token in g.Attribute("Indices").Value.Split(";".ToCharArray()))
            {
                var newToken = token;
                var matchToken = Regex.Match(token, @",(-\d+)");
                if (matchToken.Success) // negative number, yay ! it's not allowed :-(
                {
                    newToken = ",0"; // it should be zero, I believe
                }

                fixedTokens.Add(newToken);
            }

            g.Attribute("Indices").Value = string.Join(";", fixedTokens);
        }

        return g;
    }
}

关于c# - XPS 编写器因 .Net 4/字体渲染错误而失败?,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/23246254/

相关文章:

c# - 使用变量赋值与重复方法

C# Swagger 生成的客户端如何验证和使用自动生成的代码

c# - 如何创建一个结合了CheckedListBox和详细ListView的显示?

c# - 如何在 IIS 中授予对 Web API 应用程序的写入权限?

.net - 使用 ServiceLocator Bootstrapper 在 MVVM 应用程序中注册事件处理程序的正确位置?

c# - .NET从其他线程在主线程(UI)中执行函数

c# - WPF - CanExecute 不适用于 DataGrid 的 ContextMenu

c# - WPF MultiDataTrigger AND 条件

c# - 为什么 client.GetUserDialogsAsync() 抛出错误无法读取 TLSharp 中的数据包长度

c# - RelayCommand CanExecute 行为