Porting to .net standard

[Markeli]

We are using your library in our project and recently migrated it to .net core. Are you planning to migrate library to .net standard?

[grishin]

I did a simple project type change here https://github.com/grishin/SPSS-.NET-Reader/tree/netstandard/SpssLib . All tests are passing. Didn't use it in production though.

[Mikael-Gustavsson]

Any plans on publishing the .net standard port on nuget.org?

[tonykaralis]

I would also be interested to know if this will be officially migrated. @grishin do you foresee any issues using this in production?

[Markeli]

We are already have used SPSS in production the last few weeks. First problem is that in original library default encoding was not UTF-8. But it easy to fix with explicit setting encoding via SPPSOptions. And second problem is working with fractional numbers. We are looking for the solution now.

[tonykaralis]

@Markeli Do you have issues with the default encoding when reading or writing files?
I have copied the files over to a .NET Standard project in an attempt to port it over. All tests pass except for the TestReadFile which is throwing the following:

: System.NotSupportedException : No data is available for encoding 1252. 
For information on defining a custom encoding, see the documentation for the Encoding.RegisterProvider method.

Have you faced similar issue?

[Markeli]

@tonykaralis, no, we didn't face. Our customer have some scripts for sav file analisys and they developed for Windows-1251 encoding. Did you try to isntall packages with aditional encodings (System.Text.Encoding.CodePages)?

[tonykaralis]

I downloaded the package from nuget but still no difference. I must have messed something up when porting as running @grishin fork which is .NET Standard 2.0 works fine. I am on .NET Standard 2.2. This might be an issue.

[tonykaralis]

Doing a bit more research I found the following article Core and encoding 1252. This is what i thought to be the issue after your suggestion @Markeli . Adding the following loc:

    Encoding.RegisterProvider(CodePagesEncodingProvider.Instance);

to the SpssReader constructor didnt work, neither did adding it to the .NET Core test project.

It now throws a new error

SpssLib.FileParser.SpssFileFormatException : Encoding not recognized: CP1252

This leads me to believe something else is lacking in order to make this work in a .NET Core environment.

When migrating did you have to change the underlying code of this library to get it to work?
I cant seem to get that 1 critical test to pass.

[tonykaralis]

Hi All, Sorry for the many messages.
Managed to get it working and compatible with .NET Core by making a few simple changes. It does break compatibility with .NET Framework but thats outside of the scope of this issue.
Here is a working version for anyone looking in the future .NET Core compatibility.

[Mikael-Gustavsson]

It is possible to keep compatibility with .net core AND .NET framework, I had it working on my computer before leaving for a week long vacation.

[tonykaralis]

I have added the compatibility by specifying an additional target framework. Builds and all tests pass on both platforms.

[grishin]

I apologize for the late reply. Our management team canceled the project where we expected to use this library. So I had no chance to use it in real production environment.

[Mikael-Gustavsson]

@tonykaralis SpssLib, being a .net standard library, does not need multitargeting. I suggest removing the "net471" target from SpssLib and adding <RestoreProjectStyle>PackageReference</RestoreProjectStyle>
to Test.SpssLib.csproj. The tests should pass once that is done.
See dotnet/standard#481

[tonykaralis]

I was not aware of that. Thanks for pointing that out. Will amend as suggested.

[mikhail-vazhenin]

Hi!
Are you planing to upload netstandard version to nuget?