[INTERNAL] List of rufus-4.16 enhancements and fixes

This topic is flagged as INTERNAL and is therefore NOT open for comments.

Medium Term

• What the fudge is Microsoft doing with UEFI:NTFS in Hyper-V?
• In UEFI:NTFS (with Secure Boot enabled), do we want to take a peek at the SVN of the bootmgr bootloader and the system's SVN to warn users if the media they are trying to boot is going to fail security validation?
• [18ae93bf37f94764d77216cff4168e0549a14c4d] Ah are you frigging kidding me! Microsoft updated their DBX a mere hours after the release of 4.15...
• [18ae93bf37f94764d77216cff4168e0549a14c4d] We should attempt to detect when a DBX update has been saved but the files have now been deleted by the user.
• [b4a62a4e01a4e1a5b0ed536d01a0b5e3263da1ba] Also if we can't download the DBX, as is currently the case, we appear to delete the existing DBX.
• Et tu, Gentoo
• [2ea79910738184c4633c26d82fd706ac7a2ec062] Somewhere between 4.4 and 4.7 we screwed up ISO → ESP as it now uses the full disk size instead of creating a small ESP. Also Debian 13 netinst no longer detects the mmc as cdrom on Pi 4. → The Debian netinst issue has always existed when installing to mmc
• There's a GRUB 2.16 RC
• Pick the latest libcdio changes

The ever-bounced lot

• on 24H2, USB Device Tree Viewer v4.4+ can show the actual USB 3 5/10/20G speed, and we sure want to be able to do the same in Rufus. How do they do it, knowing that Microsoft have not provided IOCTL_USB_GET_NODE_CONNECTION_INFORMATION_EX_V3?
• We get way too many device list refreshes after creating a blank UEFI:NTFS drive.
• We're going to have to take matters in our own hands when it comes to the buggy AMI NTFS driver and forcefully prevent the non UEFI:NTFS partition to boot when md5sum validation is applied...
• Windows installer corrupts the backup GPT of the installer drive?!? What the hell?!? → Seen when installing Win11 and shutting down the machine after the installer bails out on the 59fd550c468f13ab086076061f26f9ce15a93d51 issue

Long Term No Rush

• Do we want to push the GDIE (grub_debug_is_enabled) to mainline GRUB, since Fedora are NOT bothering with that? We could also push for an autolookup of grub.cfg under the prefixed dir (which we could set to root) instead of having to contend with /boot/grub, /boot/grub2 and whatnot.
• Codename "Washington"
• There's an f3probe on Linux to detect fake drives. Could we pick up on it for Rufus?
• Ultimately, we probably want to replace all mentions of ISO Mode with FST Mode...
• Do we have an issue unloading old UEFI:NTFS drivers (and have Google secretly been using our NTFS drivers in their Chromebooks)? → Needs more data... which we don't have!
• Add a .esp image copy mode?
• Is there any way we can make the regular MSVC exe behave the same as the App Store one? Even as it's the same executable binary, the App Store version is a lot less forgiving about exceptions and at any rate, you'd expect the standard MSVC binary to report stuff like access violations, especially when debugging in VS, yet it doesn't...
• The case of tails and its early write-error unless you zero the drive first... And now the case of Windows complaining that it needs a restart when you write a UEFI bootloader.
• Should we try to remove volume drive letter while saving to VHD just to be safe? We do lock the physical drive, but Windows being Windows, we may still have some bad surprises there...
• Async/double buffering improvements for ISO mode
• Use VDS for GetDrivePartitionData() and use this opportunity to call on IVdsVolumeMF_GetFileSystemProperties() to get the file systems. Or we might switch to using WMI as Microsoft advises and create our own C++ wmi library interface. (Except WMI won't for Win7 users...)
• Also we should use VDS to query data for label sanitization
• Add support for .json provided custom commands to apply after the image has been created. e.g. something like:

{
  "commands": [
    {
      "image": "en_windows_7_ultimate_with_sp1_x86_dvd_619077.iso",
      "source_dir": "C:/my_dir/",
      "custom_var1": "SOME_VAR",
      "command": "my_command.exe some_flags %custom_var1% some_more_flags"
    }
  ]
}

• We don't detect Ubuntu ARM64 as Linux images where persistence can be enabled. We may want to change that.
• UI should be disabled during port cycle until we detect the cycled device again (with timeout in case the user yanked the device)
• Label changed by the user should not revert to ISO label once the drive has been created.