8 Feb
2019
8 Feb
'19
3:11 p.m.
On 8 Feb 2019, at 14:00, Marius Dumitru Florea
<mariusdumitru.florea(a)xwiki.com> wrote:
On Fri, Feb 8, 2019 at 1:51 PM Vincent Massol <vincent(a)massol.net> wrote:
(e.g.
Why is it complex? For me it’s about putting references to pages in a Set. Whether the
update will come from a link or an xobject can be seen as a detail.
We could simply mention the number of pages (i.e. references) that will be updated as you
suggested and provide a LT for that (should be easy).
We could also decide to not show anything for simple users.
On 8 Feb 2019, at 12:45, Marius Dumitru Florea
<
mariusdumitru.florea(a)xwiki.com> wrote:
On Fri, Feb 8, 2019 at 10:25 AM Vincent Massol <vincent(a)massol.net>
wrote:
>
>
>> On 8 Feb 2019, at 09:20, Marius Dumitru Florea <
> mariusdumitru.florea(a)xwiki.com> wrote:
>>
>> On Fri, Feb 8, 2019 at 9:52 AM Vincent Massol <vincent(a)massol.net>
> wrote:
>>
>>> Hi Marius/All,
>>>
>>> See below
>>>
>>>> On 31 Jan 2019, at 11:29, Vincent Massol <vincent(a)massol.net>
wrote:
>>>>
>>>> Hi Marius/all,
>>>>
>>>>> On 30 Jan 2019, at 15:45, Marius Dumitru Florea <
>>> mariusdumitru.florea(a)xwiki.com> wrote:
>>>>>
>>>>> Hi devs,
>>>>>
>>>>> I'm working on https://jira.xwiki.org/browse/XWIKI-1660 (I need
it
> for
>>>>> https://jira.xwiki.org/browse/XWIKI-13352) and I'd like to change
the
>>> page
>>>>> rename job (from refactoring module) to update the existing objects
>>> when a
>>>>> class is renamed *if the "Update links" options is
checked*.
>>>>>
>>>>> Of course, we could add a new option (e.g. "Update
objects") but:
>>>>>
>>>>> * it complicates the rename UI (too many options)
>>>>> * I think most of the users understand the current "Update
links"
>>> option as
>>>>> "update the places where this page is *used*". I don't
think it
makes
>>> sense
>>>>> to have separate options (at least at the UI level) for things like
>>> "Update
>>>>> macro calls" or "Update image includes".
>>>>> * I don't see why you would want to update the back-links but not
the
>>>>> objects (or the other way
around).
>>>>
>>>> Sounds good to me in general.
>>>>
>>>>> If we agree on using a single option ("Update links") then
the next
>>>>> questions are:
>>>>>
>>>>> * Is there a better name? I think "Update links" is a good
name for
>>> simple
>>>>> users so I would keep it. Another option is "Update
references" but
it
>>> has
>>>>> a special meaning for XWiki developers.
>>>>
>>>> Maybe "Update other pages” with a hint saying “Ensure that other
pages
>>> using the renamed pages continue to
work after the rename”.
>>>>
>>>> ?
>>>>
>>>>>
>>>>> * Should we update the hint for the "Update links" option?
I think
we
>>>>> should, but only for advanced
users, since they should be aware of
the
>>>>> implications of renaming a
class. Simple users are not aware of the
>>>>> existence of objects, most probably, so I wouldn't complicate
their
>>> lives.
>>>>
>>>> Would be nicer to find a single message that work for everyone but I
>>> agree it’s not easy if we wish to provide details.
>>>>
>>>> I feel a nicer option would be to NOT show “Update other pages” for
>>> simple users since that should always be checked. Only offer the
> ability to
>>> uncheck it for advanced users and this solves the hint issue too :)
>>>
>>>
>>
>>> Nobody replied to this proposal but I really find it the best by far
and
>>> it solves your other questions too
while making the UI simpler
globally.
>>
>
> The only issue I see with this option is that by hiding the "Update
Links"
> the simple users might not be aware of the side effects of the rename
> operation: the fact that other pages will have to be updated. Seeing
that a
> page you want to rename is referenced in many places can make you think
> twice about the rename.
We could keep that info, it could be useful
indeed.
I can keep the message but then I'll probably need to display different
messages for simple and advanced users. Moreover, ideally the message
should be updated whenever the Preserve Children checkbox is clicked to indicate that there are more pages to update
if the child pages are
preserved).
By messages I meant to indicate (as information)
the number of links
leading to the renamed pages.
Sure, but it's not just links. There's also xobjects of a class that is
among those pages being renamed. There are two options:
* show a single number (e.g. "There are *10 other pages* that are going to
be updated because they are referencing the pages that are being renamed")
. The issue here is de-duplication: if you simply sum up the backlinks +
xobjects + etc. then you can have pages counted multiple times... Moreover,
we would need to provide a link to a view showing these other pages (as we
have for backlinks right now), and having a unified backlinks + xobjects +
etc. is complex. * show multiple numbers (i.e. "There are 4 pages
that have links to the
pages being renamed. There are 7 pages with xobjects defined by classes
that are going to be renamed. etc.”)
If you think that’s interesting info (and it’s probably the case), we could display this
one for advanced users, as additional info.
Thanks
-Vincent
>
> For me this is the same info whether you’re simple or advanced, no? OTOH
> the checkbox for advanced users could provide additional info as hint.
>
> Or we just don’t display this message at all for simple users. I wouldn’t
> mind either. Makes it simpler in practice :)
>
> Thanks
> -Vincent
>
>
>>
>>
>>>
>>> I was referring to hiding the option (the checkbox). This makes the UI
>>> simpler to use for simple user, which is the direction we want to go
> and I
>>> cannot find tons of reasons why simple users would want to not fix
> links…
>>> Actually in the past all issues that were raised were the opposite,
> users
>>> who didn’t check the box, and then we made it checked by default.
>>>
>>>> Now, the current hint for "Update Links" doesn't indicate
all the side
>>>> effects. For instance it indicates the number of back-links to the page
>>>> you're trying to rename but it *doesn't include back-links to
child
>>> pages*
>>>> (when child pages are preserved). So what I said above it not really
> true
>>>> ATM either.
>>>
>>> Yes, it’s actually worse in a sense :) Right now it makes it seem as if
>>> it’ll work perfectly well…
>>>
>>> Thanks
>>> -Vincent
>>>
>>>>
>>>> Thanks,
>>>> Marius
>>>>
>>>>
>>>>>
>>>>> WDYT?
>>>>>
>>>>> Thanks
>>>>> -Vincent
>>>>>
>>>>> [snip]