[vbox-dev] Drag-n-Drop is still unstable; Call for 5.0.1 BETA
al4321 at gmail.com
Wed Jul 22 16:25:50 UTC 2015
My fork, if I do, will be a development fork, focusing on some
features, not necessarily a production / permanent fork. It will be
rabased on new versions of Oracle's VirtualBox periodically -and- my
code patches will be available for Oracle for integration upstream. So
there will be collaboration between my fork and Oracle's VirtualBox.
Right now I do not have much time for forking or development, but
things could change. In the future, you could treat my fork as a
Reasons are many:
-I need few people to test my code / changes
-major regressions inside stable releases (breaking NVIDIA/3D on Linux
hosts in 4.3.x point release) - the patch(diff) between 4.3.14 and
4.3.16 sources was 20 MB (!!!) of code. Insanity. My view is that new
major release x.y.0 should be stable, and few changes need to be made
for point releases, only fixing critical bugs, that have little chance
-difficulty in including patches into VirtualBox; slow review. (which
needs to be decentralized among community members)
-kicking me from Oracle's forum / Moderator status - reduces motivation.
-inability to edit / close bugs in VirtualBox bugzilla. In years past
I verified dozens of bugs in community bugzilla, but being unable to
close bugs, I only left a comment there.
-no build system for 3rd party builds.
Unless at least some of the problems above could be addressed, a
pressure (for fork) remains. I call Oracle to give veteran Open-Source
community members more power over the project. i.e. having a build
system, that veteran community members could use to spin custom
versions of VirtualBox with custom patches, -but- label it with
"Unofficial" tag in GUI, and restricted to a specific subforum (like
"VirtualBox OSE" or make a new forum "VirtualBox spinoffs", where it
will be restricted to Volunteers and above). Build system will allow
to test new ideas & code in VirtualBox forum, will free my time from
creating a build environment, so I could be more interested to stay.
Case-in-point, the DnD-related unsolved bugs are:
1. DnD: Windows 8.1 Host/Windows Guests (5.0.0 RC3)
2. Drag-n-Drop: Guest-to-Host fails above 200 MB (Windows)
3. Drag-n-Drop: Linux/GNOME guest-to-host fail
4. Drag-n-Drop: KDE4 Linux guests still broken
5. Drag-n-Drop: VM deadlock on network share copy
6. Drag-n-Drop: Doesn't work on Mac OS X hosts.
On Wed, Jul 22, 2015 at 6:32 PM, Stéphane Charette
<stephanecharette at gmail.com> wrote:
> Alexey, your messages are getting more hostile. Take a break, get involved
> in other projects, and/or find another way to be more constructive.
> Personally, I upgraded to VB 5.0.0r101573 a few days after it was available,
> and it is working great for me in a production environment. (Ubuntu 64 bit
> hosts, Win32, Win64, and various Linux guests.)
> I recall there were some messages a short while ago on the list about
> wanting to fork things. Is this part of the frustration? Or is this just a
> English-as-a-2nd-language issue, and I'm reading too much between the lines?
> On Wed, Jul 22, 2015 at 5:21 AM, Alexey Eromenko <al4321 at gmail.com> wrote:
>> Since v5.0 was released, many known Drag-n-Drop issues remained
>> unfixed. Both for Linux and Windows guests.
>> It is fully documented in the BETA forums.
>> This precluded me from completing the full Guest Matrix Test, to which
>> I committed (I did only partial test).
>> For this reason, I call to release 5.0.1 BETA, and I will help to
>> finish this task - the testing part.
>> Best wishes, and congrats with v5.0 release !
>> -Alexey Eromenko "Technologov"
>> vbox-dev mailing list
>> vbox-dev at virtualbox.org
-Alexey Eromenko "Technologov"
More information about the vbox-dev