Thank you Karen,

My comments are inline below.

Joe


Karen Tung wrote:
> Hi Joe,
> 
> install-finish:
> 
> I think there's no need to keep an extra variable and update it with the 
> same value of "1" in the for loop for
> every failure.  Before you call sys.exit(), you can just check 
> "errcount", if it is not zero, you return 1.
> 

Good idea. I will do this.


> perform_slim_install.c:
> 
> - line 1418: why check for OM_FAILURE, instead of != ICT_SUCCESS, so 
> that it is consistent with all the other ICT calls?
> - The run_install_finish_script() function, why not return OM_SUCCESS 
> instead of ICT_SUCCESS since this is all
> related to ICT.


OM_xyz is used by run_install_finish_script() because, for ICT Phase I, 
it is defined in liborchestrator which uses the OM prefix.

In an effort to minimize changes I dont think moving 
run_install_finish_script() from liborchestrator to libict, at this 
phase in the project, is a good idea. For ICT Phase II this 
functionality will be reworked.


Thanks again!

Joe



> 
> --Karen
> 
> 
> Joseph J VLcek wrote:
>> Hello,
>>
>> Can two people please do a code review for a fix for bug:
>>
>> 4871 ICT failures are not reported as a failed installation
>> http://defect.opensolaris.org/bz/show_bug.cgi?id=4871
>>
>>
>> The webrev is available at:
>> http://cr.opensolaris.org/~joev/bug4871/
>>
>>
>> * The modules affected and tested:
>>
>> liborchestrator
>> install-finish
>>
>> * Testing done for GUI Install
>>
>> I booted a 101 live Image and applied the updated library using 
>> LD_LIBRARY_PATH, I copied the updated install-finish script onto /sbin
>>
>> I fabricated a failure in ict.py with a hard coded error to have one 
>> of the ICT return a failure.
>>
>> I used mount -F lofs so the ict.py containing the error would be used.
>>
>> I then repeated the test using the version of ict.py we deliver to 
>> ensure the installation succeeded.
>>
>> * Results:
>>
>> When using the version of ict.py with a hard coded error the 
>> installation reported errors.
>>
>> When using the version of ict.py we deliver, without a hard coded 
>> error, the installation succeeded.
>>
>>
>> * Testing done for AI
>>
>> No AI testing was performed.
>>
>>
>> Thank you,
>> Joe
>>
>> _______________________________________________
>> caiman-discuss mailing list
>> caiman-discuss at opensolaris.org
>> http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
>>   
> 


Reply via email to