Ivan Nejgebauer wrote:
> Dave Page wrote:
>
>> That breaks the principle of least surprise because it introduces
>> inconsistent behaviour.
>>
>> Perhaps a better approach would be to open the data view only if
>> 'double-click-for-properties' is turned off?
>
>
> Andreas Pflug wrote:
>
>> - the default action might differ from user to user. For sure, an
>> unconditional "View Data" on a table is a bad idea. I already hear
>> those guys yelling double clicking on a 100 million row table...
>
>
> This version of the patch tries to address those issues in the following
> ways:
You left out the most important part of my comment. I won't accept
double click actions other than "properties" until we have a visual
feedback in each object's context menu which marks the default action
(this is not only a prerequisite, I'd like to see it in any case). If a
user who's familiar with pgAdmin3 is put in front of an arbitrary
pgAdmin3 installation, he should know how it behaves without analyzing
option settings prior to usage.
>
> 2. In the same tab, a parameter labeled "Maximum rows in data view" is
> introduced.
A limited rowset is too undeterministic to edit on. Use a restricting
query, i.e. filtered view data.
> The patch is against the beta1 source tree;
As a consequence of your first posting, which pointed out some
inconsistent behaviour, I reworked doubleclick (and other) actions in
the properties listview to match the treeview actions; see
CHANGELOG.txt. Thus your patch will probably generate some non-trivial
conflicts.
Any patch posted should be diff'ed against cvs HEAD.
Regards,
Andreas
