Rebase on .NET Standard 2.0

[danielgindi]

I know that iText 5 is EOL, but for people who are trying to migrate to newer technologies, it becomes very complex do it all at once.
Allowing itextsharp to compile for .NET Standard removes this obstacle from the way.

[andreycha]

Hi @danielgindi . I've tried your packages from Nuget.org (https://www.nuget.org/packages/Portable.itextsharp and https://www.nuget.org/packages/Portable.itextsharp.xmlworker) and found out that converting HTML to PDF using XmlWorkerHelper does not work. Calling Document.Close() throws "System.IOException: The document has no pages." (using this sample from SO). Have you tried it out and could you maybe advise anything?

[danielgindi]

Hi @danielgindi . I've tried your packages from Nuget.org (https://www.nuget.org/packages/Portable.itextsharp and https://www.nuget.org/packages/Portable.itextsharp.xmlworker) and found out that converting HTML to PDF using XmlWorkerHelper does not work. Calling Document.Close() throws "System.IOException: The document has no pages." (using this sample from SO). Have you tried it out and could you maybe advise anything?

I did not try the helper specifically, but found out that when using in .NET Core env, I needed to register the used codepages with the system encoding, as it only had just about utf8 out of the box, so iText could not load fonts. It might be related.

[andreycha]

Could you please share relevant code sample? If you meant using System.Text.Encoding.CodePages package and adding Encoding.RegisterProvider(CodePagesEncodingProvider.Instance); then unfortunately it doesn't help...

[danielgindi]

Could you please share relevant code sample? If you meant using System.Text.Encoding.CodePages package and adding Encoding.RegisterProvider(CodePagesEncodingProvider.Instance); then unfortunately it doesn't help...

Exactly what I meant...
Maybe try to debug it? You can download the fork and compile locally easily.
Then you could perhaps find what's the source of the issue. I can say that manual pdf generation works perfectly.

[andreycha]

So the issue was in DefaultTagProcessorFactory.Load(): there is no overload of Activator.CreateInstance() in .NET Standard which accepts class name. The call was throwing an exception, unable to create tag processors which resulted into an empty document. The call should look like Activator.CreateInstance(Type.GetType(className)). Now HTML to PDF conversion works.

I've double checked other calls of Activator.CreateInstance(), they look good.

Feel free to include this fix into your PR/republish packages. Thanks for quick response and help!

[danielgindi]

@andreycha Fixing, rebasing on 5.5.13.3, and releasing :-)