Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Rafael Weingärtner
> > So, your official position is that declaring a variable as constant is > typically pointless because some future programmer can always change the > variable to not be constant That is it! I do that a lot.. specially if I cannot find a reason to declare a variable final. That is why quality co

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Tutkowski, Mike
So, your official position is that declaring a variable as constant is typically pointless because some future programmer can always change the variable to not be constant. :) In this case, to appease both parties (you and Wido), I’ll leave it final, but add a comment to explain why it’s final.

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Rafael Weingärtner
No setters help to make an object immutable, but in Java we have reflection, and the only way to really avoid changing a property is using the final modifier. However, even when using final, it is possible to do so by manipulating the byte code of the class that describes the object and is loaded t

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Tutkowski, Mike
“For instance, when we have to design/create a library and we want to make sure an object is immutable.” According to your argument, though, you don’t need final here either: just make sure to never provide setters or change those values (and document). > On Jan 20, 2018, at 9:37 AM, Tutkowski,

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Tutkowski, Mike
Personally, I’m OK with it either way here. If @wido reads what you wrote and asks me to change it back to the way I wrote it initially, I’m happy to do so. I believe he sees the explicitly declared constant here not only as a protection against ourselves, but to prevent a future programmer from

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Rafael Weingärtner
Never is a strong word. In any case, let it be if you believe it is going to provide benefits… I believe `final` modifier should be used in certain situation when it is truly required. For instance, when we have to design/create a library and we want to make sure an object is immutable. Then, we

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Tutkowski, Mike
Perhaps your argument against the final keyword is that it should never be used? If so, you and @wido can debate that and let me know which way you’d like this bit of code to end up. > On Jan 20, 2018, at 9:05 AM, Tutkowski, Mike > wrote: > > It’s pretty much the reason why the final variable

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Tutkowski, Mike
It’s pretty much the reason why the final variable exists in Java: to make a variable a constant so no one accidentally changes it. @wido just wanted the code to protect against the variable being changed...that’s all. > On Jan 20, 2018, at 8:25 AM, Rafael Weingärtner > wrote: > > Changed by

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Rafael Weingärtner
Changed by a future programmer that may develop some changes in your code? Because, if you do not code any other assignment to that variable, it will never happen. On Sat, Jan 20, 2018 at 1:16 PM, Tutkowski, Mike wrote: > @wido just wanted to make sure the Boolean variable didn’t accidentally >

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Tutkowski, Mike
@wido just wanted to make sure the Boolean variable didn’t accidentally get changed later since it’s critical it keep that value. > On Jan 20, 2018, at 8:14 AM, Tutkowski, Mike > wrote: > > I made it final per @wido’s comment (in #2415). How’s about you guys hash it > out and get back to me.

Re: [GitHub] rafaelweingartner commented on a change in pull request #2416: CLOUDSTACK-10244: KVM online storage migration fails

2018-01-20 Thread Tutkowski, Mike
I made it final per @wido’s comment (in #2415). How’s about you guys hash it out and get back to me. :) > On Jan 20, 2018, at 8:06 AM, GitBox wrote: > > rafaelweingartner commented on a change in pull request #2416: > CLOUDSTACK-10244: KVM online storage migration fails > URL: https://github.c

RE: HA issues

2018-01-20 Thread Paul Angus
No problem, To be honest host-ha was developed *because* vm-ha was not reliable under a number of conditions, including a host failure. paul.an...@shapeblue.com  www.shapeblue.com 53 Chandos Place, Covent Garden, London WC2N 4HSUK @shapeblue -Original Message- From: Nux! [mailto: