[vbox-dev] VBox 4.3 Beta - SDK Download

Maxime Dor maxime.dor at altherian.org
Wed Oct 2 18:54:45 UTC 2013


Hi Devs,

A bit more feedback, following on my previous already long e-mail :


The method IConsole->restoreSnapshot() requires a ISnapshot object, when
all the others require only the name or the UUID of a given snapshot. Any
particular reaosn for this difference?


If I want to find out if a VM currently has snapshot(s), I can use these
two ways :
- use getCurrentSnapshot() and check for null
- use findSnapshot() and pass a null value, and check for an exception to
be thrown, or not.
Are both acceptable or is one more recommended? Would it make sense to
provide a "hasSnapshot()" method aswell, so you hide the way to determine
if a VM is using snapshots or not?


When we are using methods that return a IProgress event, we basically have
to keep "pooling" for the state, using IProgress->waitForCompletion()
method.
Wouldn't it make more sense to send events when there is an update in the
progress? There is already an event manager in Virtualbox, so that could be
used to provide a more reactive link to a running task?


Finally, when modifying the settings of a VM, or controlling the VM, we
need to aquire a lock. There are 2 types of locks, and some settings only
require "Shared" while others definitly require "Write".
The appropriate lock can easily by known with a bit of common sense, but I
was wondering how to find out that information programmatically? I can't
find anything in the API or in the docs that would look like a table with
the different settings or actions, and the required lock type. Maybe I am
being very picky here, but since this is all about feedback, I thought I
would mention it.


All for now, more to come I am sure.

Best regards,
Max



On 26 September 2013 21:29, Maxime Dor <maxime.dor at altherian.org> wrote:

> Hi Klaus,
>
> So far so good, I can't find anything wrong in my tests. They are quite
> basic for now, but I'll keep on testing and see if something comes up.
> The medium variant list instead of an integer/long is definitly a very
> nice change. Same for the USB controllers list. Quite neat now!
>
> Taking on your API feedback offer, I do have few questions/remarks on the
> API in general, and for the Java part :
>
> If I want to get a IMedium object of a known medium, I have two
> possiblities :
> 1. I loop through 3 different arrays to find a specific value/UUID
> 2. I have to use the IVirtualbox->openMedium() method, but then I must
> know the medium type (floppy, cd/dvd, hard disk)
>
> Few questions about this :
> - Shouldn't there be a findMedium() method, just like for a VM, using
> either the path or the UUID? Else we have to build this "cache" ourselves.
> - Wouldn't be a more appropriate to "auto-detect" the medium type and give
> the options to simply open a file given a specific path? But I guess that's
> more a matter of opinioin which one is "best". But I would like to see both
> methods maybe.
> - Why not a single array?
>
>
> Still about mediums, I see that I cannot create a medium under the size of
> 2MB. I must pass a value no less than 2048000 to
> IMedium->createBaseStorage(size,variant).
> No mention of this anywhere in the API, and I get a cryptic error (invalid
> parameters) when I try. In retrospec, it is the proper error, but not
> really when you discover this limitation.
> I guess 2MB is a "physical" limitation (minimum sectors in a disk?) or is
> a Virtualbox one?
>
>
> I would like to provide a "Guest reboot" option, that would send a command
> to reboot the guest, and check when the VM has reseted, but there is no
> such event in the API.
>  I would like to send the reboot command to the guest (via the guest
> properties per exemple, or other) and either wait for this event, or a
> timeout and force the reboot by powering off the VM then restarting it.
> It would very much make sense to send a machine event telling us the VM
> has rebooted. It should be possible since you actually get a line in the VM
> log about the VM reseting and the BIOS kicking in.
> That would be a nice addition I think.
>
>
> Specific to the Java API, I am interesting into gettnig the error code for
> any VBoxException to rethrow a specific custom exception, but no such
> method is present.
> The only thing I can do is to parse the text message for the error number,
> if it is included.
> This is not very "clean" and could break if you decide to change the
> format of error messages. Is there a way to get this error code from the
> VBoxException somewhere, or not possible as of now?
>
>
> More questions to come later I am sure. Thank you for your time already!
>
> Best regards,
> Max
>
>
> On 21 September 2013 22:27, Maxime Dor <maxime.dor at altherian.org> wrote:
>
>> Oops, a bit too trigger happy hehe...
>> I got it, thank you!
>>
>> -----Original Message-----
>> From: Frank Mehnert [mailto:frank.mehnert at oracle.com]
>> Sent: 21 September 2013 07:58
>> To: vbox-dev at virtualbox.org
>> Subject: Re: [vbox-dev] VBox 4.3 Beta - SDK Download
>>
>> Maxime,
>>
>> the SDK upload was just not yet completed. Please have a look at the
>> download directory now.
>>
>> Kind regards,
>>
>> Frank
>>
>> On Saturday 21 September 2013 02:29:15 Maxime Dor wrote:
>> > Hello Klaus;
>> >
>> > You already changed all the points I had "an issue" with in the 4.2
>> > API, so once I use the 4.3 API further, I  can let you know my
>> > impressions and any ideas.
>> >
>> > Could you give me a link to the SDK for BETA3 as well please? None was
>> > uploaded. Thank you!
>> >
>> > On 18 September 2013 20:30, Klaus Espenlaub
>> <klaus.espenlaub at oracle.com>wrote:
>> > > Hi Max,
>> > >
>> > > On 16.09.2013 23:49, Maxime Dor wrote:
>> > >> Hi Klaus,
>> > >>
>> > >> Just wanted to let you know that so far, everything is very easy
>> > >> and that except for the actual API changes, which are very limited
>> > >> from my PoV, all the code is pretty much plug & play.
>> > >> I'll run all my unit tests once I have adapted the changes you made
>> > >> and see how it goes, but I don't fear.
>> > >>
>> > >> I like when it's easy like this :) And I like the nice little
>> > >> tweaks you've made - definitly makes more sense.
>> > >
>> > > We try not to break things just to break them... and let us know if
>> > > you have any ideas for improving the API usability. Any suggestions
>> > > will be reviewed, and we have a list of changes we want to do in 4.4
>> > > already (and we can't do them now because of lack of developer time).
>> > >
>> > > There will be a few more very minor changes coming in the next beta,
>> > > which you'll probably call boring as I can't imagine that any code
>> > > is using the affected methods (mostly not yet implemented stuff
>> > > which we might be able to find time to implement in some 4.3.x).
>> > >
>> > > Klaus
>> > >
>> > >> On 12 September 2013 17:46, Klaus Espenlaub
>> > >> <klaus.espenlaub at oracle.com <mailto:klaus.espenlaub@**oracle.com
>> > >> <klaus.espenlaub at oracle.com>>>
>> > >>
>> > >> wrote:
>> > >>     Hi Max,
>> > >>
>> > >>     On 12.09.2013 16:58, Maxime Dor wrote:
>> > >>      > Hello,
>> > >>      >
>> > >>      > Would it possible to get a SDK zip like with 4.2.x build so I
>> can
>> > >>      > already test my code against 4.3 and already prepare what
>> needs
>> > >>      > to
>> > >>
>> > >> be
>> > >>
>> > >>      > changed?
>> > >>
>> > >>     Was an oversight... 4.3.0_BETA1 had it, but it's not yet there
>> for
>> > >>     4.3.0_BETA2 (and there were API changes, some more to come). Will
>> be
>> > >>     fixed soon, and appear in the expected place on
>> > >>
>> http://download.virtualbox.**org/virtualbox/4.3.0_BETA2/<http://down
>> > >>     load.virtualbox.org/virtualbox/4.3.0_BETA2/>
>> > >>
>> > >>     Klaus
>> > >>
>> > >>      > Thank you.
>> > >>      >
>> > >>      > Best regards,
>> > >>      > Max
>>
>> --
>> Dr.-Ing. Frank Mehnert | Software Development Director, VirtualBox ORACLE
>> Deutschland B.V. & Co. KG | Werkstr. 24 | 71384 Weinstadt, Germany
>>
>> Hauptverwaltung: Riesstr. 25, D-80992 München
>> Registergericht: Amtsgericht München, HRA 95603
>> Geschäftsführer: Jürgen Kunz
>>
>> Komplementärin: ORACLE Deutschland Verwaltung B.V.
>> Hertogswetering 163/167, 3543 AS Utrecht, Niederlande Handelsregister der
>> Handelskammer Midden-Niederlande, Nr. 30143697
>> Geschäftsführer: Alexander van der Ven, Astrid Kepper, Val Maher
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.virtualbox.org/pipermail/vbox-dev/attachments/20131002/18357f3a/attachment.html 


More information about the vbox-dev mailing list