[libvirt] [test-API PATCH] streamAPI: Create stream objects that are already encapsulated

Peter Krempa pkrempa at redhat.com
Mon Mar 26 09:01:07 UTC 2012


On 03/22/2012 02:51 PM, Guannan Ren wrote:
> On 03/22/2012 08:54 PM, Peter Krempa wrote:
>> With this patch, the newStream() method of the ConnectAPI class returns
>> an already encapsulated StreamAPI object .
>> *lib/connectAPI.py: - modify newStream method to return StreamAPI
>> object
>> *lib/streamAPI.py: - modify constructor to take virStream objects
>> ---
>> lib/connectAPI.py | 4 +++-
>> lib/streamAPI.py | 9 ++-------
>> 2 files changed, 5 insertions(+), 8 deletions(-)
>>
>> diff --git a/lib/connectAPI.py b/lib/connectAPI.py
>> index f52467f..2723fc8 100644
>> --- a/lib/connectAPI.py
>> +++ b/lib/connectAPI.py
>> @@ -24,6 +24,7 @@ import os
>> import re
>>
>> import libvirt
>> +from lib import streamAPI
>
> The initial idea is to have a separate file for each of classes(domain,
> storage, etc)
> in libvirt.py at least in the form of files. The direction you are
> heading here is to
> combine them together again.
> Testcase is the right place where these objects of encapsulated class
> should meet
> each other.
> I prefer the way of what you did in openConsole API. We just pass the
> instance of
> ConnectAPI to initiate SteamAPI class for getting the object of SteamAPI
> for use
> we should delete the newStream method in connectAPI.py.
> Right now, the virConnect object from libvirt.py is exposed directly in
> existing testcases,
> that's not good way compared to what you did in openConsole. I will
> clean them next.
> What do you think?

Ok, I didn't know you want to keep those classes separate. The only 
small issue with this is, that ConnectAPI.newStream() produces 
unencapsulated virStream object, but you can get that anyways using the 
StreamAPI.getStream() method.

I'll post a v2 of the next patch that adds flags to add them to the 
correct file.

Peter

>
> Guannan Ren
>
> --
> libvir-list mailing list
> libvir-list at redhat.com
> https://www.redhat.com/mailman/listinfo/libvir-list




More information about the libvir-list mailing list