Questions about Windows 10 Iot Core ARM

[DF4D3110]

As shown in the picture, when I tried to extract something from Windows 10 Iot Core Build 17763, MobilePackageGen didn't recognize the contents, and it didn't work after converting to a vhdx file (of course, I used the drive letter directly to get an error) If you package the data partition separately from the mainos partition as a wim file, it can recognize and extract the file normally (but the efiesp partition does not recognize and extract to the file) The problem is present in the IoT Core of Arm32, but the FFU (9924 to 18941) of amd64 and x86 is fine, and I did not test arm64 Hopefully this issue will help improve MobilePackageGen

[BlueRain-debug]

Cannot reproduce on the latest MobilePackageGen rev with 99d4451 reverted. Please provide more details.

[DF4D3110]

无法在最新的 MobilePackageGen Rev 上复制,99d4451 已恢复。请提供更多细节。

This is the structure I run using the compilation under https://github.com/BlueRain-debug/MobilePackageGen/actions/runs/15394128919, of course, FFU comes from betaarchive (e.g. https://www.betaarchive.com/database/view_release.php?uuid=fcdf7809-c18a-48fb-8b1c-06a98cb95f5a), and if it's a physical machine dump, then it's normal

[BlueRain-debug]

无法在最新的 MobilePackageGen Rev 上复制,99d4451 已恢复。请提供更多细节。

This is the structure I run using the compilation under https://github.com/BlueRain-debug/MobilePackageGen/actions/runs/15394128919, of course, FFU comes from betaarchive (e.g. https://www.betaarchive.com/database/view_release.php?uuid=fcdf7809-c18a-48fb-8b1c-06a98cb95f5a), and if it's a physical machine dump, then it's normal

You need to revert the 99d4451 commit to restore support for 512b sector (the lack of this support can result in the Bad GPT error in your third screenshot).

[BlueRain-debug]

无法在最新的 MobilePackageGen Rev 上复制,99d4451 已恢复。请提供更多细节。

This is the structure I run using the compilation under https://github.com/BlueRain-debug/MobilePackageGen/actions/runs/15394128919, of course, FFU comes from betaarchive (e.g. https://www.betaarchive.com/database/view_release.php?uuid=fcdf7809-c18a-48fb-8b1c-06a98cb95f5a), and if it's a physical machine dump, then it's normal

still can't reproduce using the 14922 ffus you linked.

[gus33000]

Keeping this issue by the way as the hack currently in place is well a, hack and we need a better fix indeed. But as @BlueRain-debug pointed out you have to revert that commit. Watch out however for ffus using a mbr layout, we don't currently support those at all