Hi,

on a quick look reading the content-type seems to be built into the parsing
logic currently. So this would likely require more changes than simply
removing a null-check.

You can have a go at
org.apache.poi.openxml4j.opc.ZipPackage.EntryTriple#register, but I'd doubt
it suffices to only touch the code there.

You can also create a issue at https://bz.apache.org/bugzilla/ so we can
collect information, but to be honest I'd not expect much progress unless
you can invest a bit yourself first.

Some sample files seem to be available at
https://github.com/microsoft/powerbi-desktop-samples/tree/main

Dominik.

On Fri, Oct 13, 2023 at 12:50 AM Olivier Schmitt
<schm...@metaintegration.com.invalid> wrote:

> Hello,
>
> The Microsoft PowerBI PBIX file format is an OPC package file type.
> It contains [Content_Types].xml which declares the content types of parts.
> Unfortunately, the content types are declared as empty strings.
> Apache POI code throws an exception, when trying to open the file.
> (org.apache.poi.openxml4j.opc.OPCPackage)
> Is it possible to support empty string content type for parts?
>
> The DotNet source code supports empty contentTypes:
>
> https://github.com/dotnet/runtime/blob/main/src/libraries/System.IO.Packaging/src/System/IO/Packaging/ZipPackage.cs
>
> Thank you.
> Olivier
>

Reply via email to