[rfc][icedtea-web] rememberable dialogues general solution - impl part 1 - safe types returned from getUserResponse

Jiri Vanek jvanek at redhat.com
Mon Jun 8 10:18:40 UTC 2015


On 06/05/2015 05:36 PM, Andrew Azores wrote:
> Hi! Generally liking where this patch is going.
>
> Jiri Vanek <jvanek at ...> writes:
>
>>
>> hi!
>>
>> So this is first part.
>> Tho most nasty and error prone.
>>
>> It is changing type of securityMEssage.response from Object to regular
> interface, and is providing
>> useful implementations of this interface.
>>
>> The legacy object was integer at first, and was used to preselect button,
> and tro return id of
>> pressed button. LAter was missuesd and returned yes/no/always/never even
> later sanbox was added and
>> last drop was complex result about shortcuts.
>
> I much prefer your new approach, obviously much better than ints and Object[]s.
>
>>
>> In this patch, there are few wrappers for types of yAnNS, which contains
> also remember logic. This
>> will be fixed in another changeset, and those wrappers will be gone.
>
> Does this mean the Yes, No, YesNo, YesNoCancel, etc. classes? And "Garbage"?

Garabge for sure:)

But not so YEs, No, YesNo - they are what dfialog can return, and what is returning. What you would like to replace them with?
If dialog returns more htne just yes/nos/sandbox/any_button then they can be rused in same way as AccessWarningPaneComplexReturn do.

Thoughts?
>
>>
>> Approach of this patch i slightly different, less verbose, with a lot of
> shared code and different
>> typesafety. I hope it is generally better. If somebody found tech-review's
> solution better, or
>> generally have better ideas, pelase let me know.
>>
>> J.
>>
>> Attachment (makeTypesOfSecurityDialogMessageReturnValueSafe.patch):
> text/x-patch, 84 KiB
>
>
> Thanks,
>
> Andrew Azores
>



More information about the distro-pkg-dev mailing list