V5.2's Version mismatch.

[isgdre]

Version 5.2.0 r118431 (Qt5.6.2) has guest additions that report 5.2.1.118430.
What's really odd is 5.2.0 < 5.2.1 but 118431 > 118430
Does anyone know if this is ok?

[socratis]

There seems to be an issue with the Windows VirtualBox distribution and the included Guest Additions CD.

[michaln]

isgdre: Could you please share with us which VirtualBox package exactly you installed?

The mismatch shouldn't be a problem, and the seemingly backward going revision numbers are just an artifact of some build process quirks.

[isgdre]

The file name is VirtualBox-5.2.0-118431-Win.exe which I got from the big 5.2 download button on the www.virtualbox.org home page

I didn't download the guest additions separately.

[socratis]

I can confirm that it's the Windows package that has the problem, with the Linux guest additions (so far). If you're on a Win host and try the GAs, you get a 5.2.1 version. It doesn't happen on OSX as a host.

[klaus]

At least the Windows package is affected, and I don't yet understand why our normal logic for forcing consistent builds has failed for 5.2.0.

[klaus]

Staring at the last changes in this area finally solved the riddle. A recent change completely sabotaged the logic which forced a consistent build, and no one noticed because it sabotaged a condition check without a warning.

[isgdre]

Ok, So I understand it's not big deal and should work properly. Just a version labeling problem in the build and likely will be fixed for the next version.

Thanks guys.

[klaus]

Correct - the issue is purely cosmetic, shipping guest additions which have a strange version number. The actual code is the correct one (in some sense, these guest additions were not working with Linux, for that you'll need to fetch the test build, labeled development).